Research and analysis

Project ADDER: Impact evaluation

Published 12 February 2025

Applies to England and Wales

Authors

Verian authors Rowan Foster, Holly Captainino, Daniel McGrady and Peter Matthews, in partnership with London Economics, the Institute for Criminal Policy Research and Revolving Doors.

Acknowledgements

Verian is a world-leading independent research, evidence, evaluation and communications agency, providing services to government and the public realm. Verian was commissioned by the Home Office as the lead evaluation organisation for Project ADDER and worked with consortium partners to prepare this report.

Our thanks go to our consortium partners, Dr Gavan Conlon and Jack Booth at London Economics, Lauren Bennett at Revolving Doors and Professor Mike Hough and Dr Bina Bhardwa at the Institute for Crime & Justice Policy Research (ICPR) at Birkbeck, University of London, for their collaboration and continued support with subject matter expertise, guidance and advice throughout the design, fieldwork and analysis stages of the evaluation.

We are also grateful to the Drugs, Exploitation and Abuse Analysis Unit within Home Office Analysis and Insight team, for their regular guidance and feedback: in particular, Molly Watts, Lauren Comber, Alison Green, Katherine Weaver and Anna Reed. In addition, we received valuable support from our colleagues at the Office for Health Improvement and Disparities (OHID): in particular, Dylan Kerr.

Our thanks also go to the anonymous peer reviewers who provided us with comments during the reporting stage and, finally, to all the ADDER project stakeholders and service users who gave their time generously to the evaluation process.

Glossary

Buprenorphine: A medicine used to treat dependence on opioid (narcotic) drugs such as heroin and morphine, often referred to by the brand names ‘Buvidal’ and ‘Sublocade’. Depot buprenorphine is administered (after titration) either weekly or monthly via sub-cutaneous injection and marks a significant shift in opioid substitution therapy, which relies on daily doses of either methadone or buprenorphine.

County lines: Gangs and organised criminal networks involved in exporting illegal drugs to one or more importing areas using dedicated mobile phone lines or other forms of ‘deal lines’.

Cuckooing: Cuckooing is when drug gangs take over the home of a vulnerable person through violence and intimidation and use it as their base for selling/manufacturing drugs.^{[footnote 1]}

Delivery partners: Organisations, services and professionals involved in the delivery of Project ADDER. For example, treatment services, police forces, and the National Crime Agency (NCA).

Diversion: Initiatives that use the criminal justice system to divert people who use Class A drugs away from offending (and criminalisation) into health interventions, through programmes such as drug testing on arrest (DToA), community resolutions and Out of Court Disposals (OOCDs).

Drug testing on arrest (DToA): Police are able to test suspects in custody for Class A drugs, namely opiates and cocaine, to better understand whether it is suitable to refer an individual for support, and also to understand the role of drug misuse in certain crime types. Testing is currently used for a range of ‘trigger offences’ that are defined in law.^{[footnote 2]}

Individual Placement Support (IPS): A service that is designed to support service users to find paid employment, typically consisting of intensive individual support in relation to searching for a job and throughout the job placement.^{[footnote 3]}

Middle market targets: Criminal networks and individuals operating within the sphere of activity that lies between bulk importation traffickers/wholesalers and retail-level dealers.^{[footnote 4]}

Naloxone: Naloxone is the emergency antidote for overdoses caused by heroin and other opiates or opioids (such as methadone, morphine and fentanyl).

Organised crime group (OCG): An OCG is defined as a group which has as its purpose, or one of its purposes, the carrying on of criminal activities, and which consists of 3 or more people who agree to act together to further that purpose. ^{[footnote 5]}

Out of Court Disposal (OOCD): OOCDs are a set of cautions or orders that the police are able to give when individuals are caught for low-level offences, and which prevent the individual from being formally prosecuted or having a criminal record. OOCDs include, but are not limited to, referrals for individuals to engage with a support service, and some types of caution.

Outreach: Outreach refers to activities that target and seek to engage a specific group for treatment or support services.

Recovery support: Ongoing interventions that include intensive employment, tenancy, and peer support, which occur after structured treatment support.

Service user: Refers to an individual who is a beneficiary of, or who is in receipt of, an ADDER-funded activity or service. In a few cases the term ‘people using drugs’ is used in this report to refer to individuals who are eligible for ADDER-funded activities due to drug use.

Treatment: Within Project ADDER, drug treatment encompasses a broad spectrum of interventions/activities, including Tier 1 to 2, that is, harm reduction approaches (for example, needle exchanges), GP/primary care interventions and signposting, but the programme was primarily focused on getting people using drugs into Tier 3 to 4 structured treatment, that is, activities designed to reduce harm for service users, often involving pharmacological and psychological treatment.

Trauma-informed practice: The aim of trauma-informed practice is to raise staff understanding about the wide-ranging impact of trauma and to prevent the re-traumatisation of clients in service settings that are meant to support and assist healing.

Whole systems approach (WSA): This involves responding to a complex issue by working across multiple stakeholders and ways of working, to deliver a single system response. WSA aims to bring about sustainable, long-term change. For ADDER, this refers to working with enforcement, treatment and recovery partners to tackle drug misuse.

1. Executive summary

1.1 Policy context and introduction to Project ADDER

Project ADDER (addiction, diversion, disruption, enforcement and recovery) aims to promote an intensive whole systems approach (WSA) to tackling drug use and its consequences. The funding from Project ADDER was used to implement interventions across treatment and recovery, enforcement, and diversion, with the overall aim of reducing the prevalence of drug use, drug-related offending, and drug-related deaths. The project initially focused on opiates and crack cocaine but over time local areas were able to expand their focus to other drugs (primarily Class A), co-use of alcohol, and preventative interventions.

Five local areas (Blackpool, Hastings, Middlesbrough, Norwich and Swansea Bay) were initially invited to bid for ADDER funding due to their high levels of drug-related deaths and opiate and crack cocaine use. Implementation on a phased basis in the different areas began from January 2021, with projects originally scheduled to run until March 2023. A further 6 local areas were invited to bid at a later stage – referred to as ‘Place-based Accelerators’. These areas were Bristol, Hackney, Tower Hamlets, Merseyside (itself comprising the sub-regions of Liverpool, Knowsley and Wirral), Newcastle, and Wakefield. Phased implementation for these areas began from April 2021. Through the 10-year Drug Strategy published by the Conservative government in 2021, From harm to hope: A 10-year drugs plan to cut crime and save lives, funding was extended by the Conservative government for all areas until March 2025.

Based on available evidence, a range of interventions were proposed by the Home Office, classified into 3 broad groups: treatment and recovery, diversion, and enforcement. Each of the local areas were invited to prepare tailored project proposals according to these interventions which reflected their local contexts. As a result, the interventions delivered by each area differed significantly. In delivering these interventions, local projects sought to adopt a WSA, whereby enforcement, diversion, treatment and recovery partners function as a coherent ‘system’ to meet the needs of service users holistically.

1.2 Evaluating Project ADDER

The Home Office has commissioned Verian to conduct an independent evaluation of Project ADDER, working in partnership with London Economics, Revolving Doors, and the Institute for Crime & Justice Policy Research (ICPR) at Birkbeck, University of London. The objectives of the evaluation are to measure delivery against the stated project’s aims, investigate experiences of delivery and provide recommendations on designing or rolling out a similar programme, and provide recommendations on continuing to evaluate the effects of Project ADDER.

This report presents the findings from the impact evaluation of Project ADDER, and shares case study examples of impact from local area projects. For detailed descriptions and local-level examples of different interventions and WSA, please refer to the Project ADDER Evaluation: Practitioner Report. The impact evaluation has used a theory-based approach, known as ‘contribution analysis’, to assess the extent to which Project ADDER has contributed to key outcomes identified in the project’s Theory of Change (ToC; see Annex A).

2. Key findings

2.1 Enforcement

Project ADDER funding was used to boost the number of officers working on relevant operations or in relevant teams, meaning that it was possible to do more and to achieve more through increased resources. The funding was also used to enhance intelligence gathering and sharing: for example, through developing drug market profiles and establishing information sharing between enforcement teams.

The evaluation found that enforcement activities are likely to have contributed towards an increase in arrests of high-harm individuals ^{[footnote 6]} involved in drug supply, and seizures of drugs, assets and cash. The programme also sought to reduce the drug supply within local areas by disrupting the activities of organised crime groups (OCGs). The evaluation found evidence that disruptions were achieved, but that achieving an observable reduction in drug supply would require intensive and consistent activity over a longer period of time. Rates of drug trafficking offences and charges were also analysed as a measure of drug supply, but no consistent impact of the programme was observed. It was considered that there may be some unintended consequences from Project ADDER, such as crime (including dealing of drugs) being displaced to other areas, but we observed little evidence to support this.

Overall, the evaluation found that Project ADDER may have indicatively contributed towards achieving a reduction in drug supply but that further evidence would be needed to determine whether this is the case.

In the longer term, Project ADDER aims to achieve an overall reduction in drug-related offending. The wider literature, including research from North America, suggests that we are unlikely to observe reductions in the number of people using drugs and in levels of drug-related crime over the duration of a 2 to 3 year programme. Evidence from similar diversionary programmes has shown mixed success in terms of crime reduction. However, the evidence for the impact of treatment programmes on reducing crime is somewhat more promising.

Table 1 summarises the overall evaluation conclusions for each of the enforcement outcomes and impacts in the ToC (combining evidence from multiple sources, for example, management information (MI), stakeholder interviews). Table 2 summarises the results from the quasi-experimental impact analysis across the relevant enforcement measures.

Table 1: Summary of enforcement impact findings

ToC outcome/impact	Assessment of evidence	Rationale
Increase in arrests of high-harm individuals involved in drug supply and an increase in seizures of drugs, assets and cash	Project ADDER activities to increase drug testing capabilities and enforcement activity are likely to have contributed towards an increase in arrests of high-harm individuals involved in drug supply, and seizures of drugs, assets and cash.	Stakeholders reported that their main focus was on the quality of arrests and seizures, rather than the quantity. They felt confident that Project ADDER activities had led to an increase in arrests of high-harm individuals and seizures of drugs, assets and cash, which might not have been reflected in the overall numbers of arrests and seizures.
Reduction in drug supply to and within local areas	Project ADDER may be contributing towards a reduction in drug supply to and within local areas, but further evidence would be needed to determine whether this is the case. This evidence could include analysis over a longer time period, and inclusion of alternative metrics of drug supply (for example, purity and price).	Stakeholders reported that intensified local enforcement activity may have led to reductions in drug supply. Quasi-experimental design (QED) analysis of police recorded crime (PRC) data shows some evidence of an increase in drug trafficking offences in ADDER areas in year 1 and in drug trafficking charges in Accelerators areas in year 2, as a proxy for drug supply in the local areas. However, there was significant variation at a local level, and other projects and initiatives were operating at a similar time to Project ADDER. Evidence is also lacking as to whether increased arrests do actually disrupt the supply system.
Reduction in drug-related offending	No observable effect of Project ADDER on drug-related offending.	There was no observable impact of Project ADDER on reported drug possession offences, acquisitive crime or serious violent crime. Although there was some evidence of a reduction in sharps hospital admissions in Accelerators areas in year 2, there was also an increase in reported neighbourhood crime in those areas.
Reduction in anti-social behaviour (ASB) and re-offending	There is not enough suitable evidence to determine whether or not Project ADDER interventions had an effect on ASB or re-offending. Analysis of re-offending data over a longer period would be useful for longer-term evaluation in regard to this outcome.	Stakeholders did not feel it was possible to directly attribute reductions in ASB and re-offending to communication-related interventions, and they especially noted that these outcomes are regularly impacted by wider influences. There was no suitable quantitative data available on ASB, and no available data on re-offending to allow for QED impact analysis or descriptive trend analysis.

Table 2: Summary of QED impact analysis findings for enforcement measures

ToC outcome/impact	Measure	Evidence of impact: ADDER areas	Evidence of impact: Accelerators areas
Reduction in drug supply to and within local areas	Drug trafficking offences	Increase in drug trafficking offences in year 1, equivalent to 167 additional offences overall. No evidence of an impact in year 2.	No evidence of impact in years 1 or 2.
	Drug trafficking offences resulting in a ‘charged/summonsed’ outcome	No evidence of impact in years 1 and 2.	Increase in drug trafficking charges compared to the counterfactual, corresponding to 473 and 257 additional charges in years 1 and 2, respectively.
Reduction in drug-related offending	Drug possession offences	No evidence of impact in years 1 or 2.	No evidence of impact in years 1 or 2.
	Acquisitive crime offences	No evidence of impact in years 1 or 2.	No evidence of impact in years 1 or 2.
	Violent crime offences	No evidence of impact in years 1 or 2.	No evidence of impact in years 1 or 2.
	Neighbourhood crime offences	No evidence of impact in years 1 or 2.	Increase in recorded offences in year 1, equivalent to 3,895 additional offences.
	Patients admitted to hospital due to assault with a sharp object	No evidence of impact in years 1 or 2.	Reduction in hospital admissions due to assault with a sharp object in year 2.

2.2 Diversion

Under Project ADDER, funding was available for interventions that aimed to divert service users away from the criminal justice system (CJS) and into treatment or support. In most areas, diversionary interventions comprised increasing drug testing on arrest (DToA) and greater use of out of court disposals (OOCDs). Importantly, this was also combined with more integrated and improved care pathways, which meant that those diverted from the CJS were able to access better-quality support services and interventions.

The evaluation found that diversionary activities are likely to have contributed towards an increase in the number of people who use drugs who were referred into treatment and support services. Stakeholders reported that DToA, in-reach to prison and custody, and continuity of care following prison were important drivers for achieving this outcome. Data shows a significant increase in referrals into treatment from criminal justice pathways, and some evidence of more drug possession offences ending in OOCDs, namely community resolutions. Impacts could perhaps have been even greater if DToAs were not limited to cocaine and opiates, concerns about the suitability of OOCDs for heroin or crack users, and judiciaries in some areas being directed to use Drug Rehabilitation Requirements (DRRs) instead.

Table 3 summarises the overall evaluation conclusions for the diversion outcomes in the ToC (combining evidence from multiple sources). Table 4 summarises the results from the QED impact analysis across the relevant diversion measures.

Table 3: Summary of diversion impact findings

ToC outcome/impact	Assessment of evidence	Rationale
Referral into treatment from the CJS	Project ADDER activities related to DToA, OOCDs, and improving integrated care are likely to have contributed towards an increase in the number of people using drugs referred into support services from the CJS.	Stakeholders reported that activities had led to an increase in referrals of individuals from the CJS into treatment and support. Quantitative data was positive but somewhat mixed across areas and strands. MI data shows that DToA and the use of OOCDs either increased or remained stable during the implementation period. Impact analysis demonstrated an overall impact of Project ADDER on referrals into treatment from the CJS across Project ADDER. There was a greater proportion of drug possession offences resulting in community resolutions in Accelerators areas in year 2 but there was no such trend for ADDER areas.

Table 4: Summary of QED impact analysis findings for diversion measures

ToC outcome/impact	Measure	Evidence of impact: ADDER areas	Evidence of impact: Accelerators areas
Referral into treatment from the CJS	People in treatment reporting a CJS referral as source of entry.	Increase in referrals into treatment from CJS in year 2.	Increase in referrals into treatment from CJS in year 2.
	Drug possession offences that resulted in a community resolution.	Weak evidence of impact in year 1 and no evidence in year 2.	Increase in the proportion of drug possession offences resulting in community resolutions in year 2.

2.3 Treatment

Project ADDER funded local areas to deliver enhanced treatment, outreach, and recovery provision (including housing and employment support). Local areas were free to tailor this provision to their locality, enabling some areas to develop new services specifically for Project ADDER and other areas to increase or enhance existing provision aimed at all people who use drugs. This led to large variation between areas. However, interventions typically included activities related to recovery support, harm reduction programmes, pharmacological and psychological treatment, enhancing treatment capacity, and integrating and improving care pathways.

These activities are showing promising signs that the programme may be contributing towards an increase in the number of people using non-opiate drugs (primarily powder cocaine and cannabis) entering treatment and wider support services. However, the same affect was not observed for people using opiates, which is important to note given the focus of the programme. This may be because opiate users typically exhibit more entrenched drug use and may require more time and preparatory work before they can be referred into treatment or support, compared to non-opiate users. They may also be more likely to be referred into tier 2 (for example, provision of information and advice, harm reduction interventions) or informal forms of treatment (for example, preparing entrenched users to take up support), entries to which are not recorded in evaluation data.

Where individuals were receiving treatment, the qualitative data demonstrates real impacts on individual lives, including improved engagement with support, improved mental and physical health, improved personal relationships, and coming out of a cycle of exploitation or criminality. However, quantitative data does not show improvements in treatment completion (nor reduced drop-out), or self-reported education, employment and housing need, and the wider literature notes historical challenges in overcoming attrition along the treatment journey.

In line with this, the evaluation found less to suggest that Project ADDER has been able to achieve the outcomes that lie further along the casual pathway towards impact. It may be that these outcomes need longer to materialise: for example, completion of drug treatment for opiate users typically takes 6 years, and the broader literature suggests that effects on the prevalence of drug use and drug-related death are likely to take an extended period of time to be realised (especially in regard to seeing widespread use of treatments, such as naloxone, being realised).

It was suggested by stakeholders that we might see some unintended consequences from Project ADDER, such as displacement of crime (including dealing of drugs) to other areas, and the creation of a ‘postcode lottery’ of treatment, but we observed little evidence of this.

Table 5 summarises the overall evaluation conclusions for the treatment outcomes and impacts in the ToC (combining evidence from multiple sources). Table 6 summarises the results from the quasi-experimental impact analysis across the relevant treatment measures.

Table 5: Summary of treatment impact findings

ToC outcome/impact	Assessment of evidence	Rationale
Referral into treatment and support services	Indicatively, Project ADDER may be contributing towards an increase in the number of non-opiate users referred into treatment and wider support services. There was no observable effect of Project ADDER on the number of opiate users referred into treatment.	Stakeholders reported that referrals into treatment had increased, both from CJS pathways and from targeted outreach activities. While there was no evidence of an increase in the number of non-opiate users entering into treatment in ADDER or Accelerators areas, the trend across Project ADDER looks promising. In ADDER areas there was a reduction in the number of opiate users entering into treatment, but there was no evidence of a change in Accelerators areas.
Being supported to sustain a life that is no longer dependent on drugs	Indicatively, Project ADDER may be contributing towards more service users being supported to sustain a life that is no longer dependent on drugs, but further evidence would be needed to determine whether this is the case.	Stakeholder interviews indicated that Project ADDER interventions and activities had improved the intensity and variety of support for service users. Service users’ accounts indicate the real difference that treatment and wider support had made to their lives. However, QED analysis did not find an impact of Project ADDER on successfully completing treatment, failure to complete treatment due to drop-out or death, or a range of wider outcomes, such as being in paid/unpaid work or education, or being homeless.
Prevalence of drug use	There was no observable effect of Project ADDER on a reduction in the prevalence of drug use.	As shown by the QED analysis, there has been a significant reduction in hospital admissions for poisoning due to drug misuse in ADDER areas. There is no evidence across Project ADDER for an effect of the programme on drug-related mental and behavioural disorders. However, these are proxy measures for the prevalence of drug use. There is no evidence of an impact on self-reported drug use in National Drug Treatment Monitoring System (NDTMS) data, which is a more direct measure. Stakeholders reported, anecdotally, that some service users of Project ADDER interventions had reduced or eliminated their drug use; however, this does not provide evidence of a reduction in the general population.
Drug-related deaths	There was no observable effect of Project ADDER on drug-related deaths in year 1.	The impact analysis showed that there is no evidence so far of the impact on Office for National Statistics (ONS) recorded drug-related deaths in year 1. Due to data lags in records of drug-related deaths, it was not possible to conclusively measure impact for year 2. Stakeholders were unable to comment on the extent to which Project ADDER may have prevented deaths due to drug misuse, particularly at any scale, although the roll-out of naloxone may lead to impacts in the longer term.

Table 6: Summary of QED impact analysis findings for treatment measures

ToC outcome/impact	Measure	Evidence of impact: ADDER areas	Evidence of impact: Accelerators areas
Referral into treatment and support services	Opiate patients entering treatment	Reduction in the number of opiate patients entering treatment, equivalent to 247 fewer patients in year 2.	No evidence of impact in years 1 or 2.
	Non-opiate patients entering treatment	No evidence of impact in years 1 or 2.	No evidence of impact in years 1 or 2.
Being supported to sustain a life that is no longer dependent on drugs	Failure to complete treatment	No evidence of impact in years 1 or 2.	No evidence of impact in years 1 or 2.
	Successful completion of treatment	No evidence of impact in years 1 or 2.	No evidence of impact in years 1 or 2.
	Deaths in treatment	No evidence of impact in years 1 or 2.	No evidence of impact in years 1 or 2.
	Self-reported measures of being in work, being in education, and housing need	No evidence of impact in years 1 or 2.	No evidence of impact in years 1 or 2.
Prevalence of drug use	Hospital admissions due to drug poisoning	Significant reduction in the number of hospital admissions for drug admissions, equivalent to 331 fewer admissions.	No evidence of impact in year 1. It was not possible to conclusively measure impact for year 2.
	Hospital admissions for drug-related mental and behavioural disorders	No evidence of impact in years 1 or 2.	No evidence of impact in years 1 or 2.
	Self-reported drug use	No evidence of impact in years 1 or 2.	No evidence of impact in years 1 or 2.
Drug-related deaths	Drug-related deaths	No evidence of impact in years 1 or 2.	No evidence of impact in years 1 or 2.

2.4 Reflections and recommendations

Whilst it is inevitably challenging to generalise about a programme that was designed to be tailored and flexible in order to meet local area needs, a programme-level perspective is important in order to fill evidence gaps, and to identify suitable recommendations for future policymaking and evaluation at a national level.

Policy recommendations include the following: understanding the importance of rebuilding the professional workforce, as acknowledged within the Conservative government’s 10-year Drugs Strategy; mainstreaming the WSA, integrating enforcement, treatment and recovery efforts; ensuring flexibility in central funding to meet local needs; and pursuing continued improvements in treatment capacity and quality.

Given that Project ADDER has been extended until March 2025, there is value in continuing to evaluate the programme’s impact, including for the following purposes: to allow further time for monitoring impacts on key outcomes such as drug use and drug deaths; to identify alternative/additional metrics for key outcomes (for example, re-offending data); and to map the various programmes in the drugs policy space to fully integrate alternative explanations for any impacts or outcomes observed.

3. Introduction and methodology

3.1 Introduction to Project ADDER

In July 2021, Professor Dame Carol Black published her final report as part of an independent review of drugs in the UK.^{[footnote 7]} This report made recommendations for refreshed action across government to address the societal costs of drugs, and the detrimental impacts that funding reductions and budget constraints, in both the enforcement and the treatment sector, have had on drug-related harms. While this is difficult to quantify, the review estimated the social and economic cost of illicit drugs in England at around £19.3 billion per year, with Class A heroin and crack cocaine causing the most harms and associated crimes (that is, county lines, acquisitive and violent crime). These recommendations were accepted by the Conservative government in 2021^{[footnote 8]} and are addressed in the Drugs Strategy, From harm to hope: A 10-year drugs plan to cut crime and save lives, also published by the Conservative government in 2021.^{[footnote 9]}

Project ADDER was initially a 3-year pilot programme, operating between November 2020 and March 2023, and has since been extended to March 2025. It is led by the Home Office and the Office for Health Improvement and Disparities (OHID)^{[footnote 10]} within the Department of Health and Social Care. Further support was received from the Department for Levelling Up, Housing and Communities, the Department for Work and Pensions, the Ministry of Justice (MoJ) and the Welsh Government. The programme aims to promote an intensive WSA to tackling drug use – and in doing so, to reduce the prevalence of drug use, drug-related offending, and drug-related deaths. In total, £43.5 million has been invested in the programme at the time of writing.

Five local areas (Blackpool, Hastings, Middlesbrough, Norwich and Swansea Bay) were initially invited to bid for ADDER funding on the basis of their high levels of drug-related deaths, and opiate and crack cocaine use. Implementation on a phased basis in the different areas began in January 2021, with projects scheduled to run until March 2023. A further 6 local areas were invited to bid for the ‘Place-Based Accelerators’ programme at a later date (Bristol, Hackney, Tower Hamlets, Merseyside (itself comprising the sub-regions of Liverpool, Knowsley and Wirral), Newcastle, and Wakefield). Phased implementation in the Accelerators areas began in April 2021 and ran until March 2023.

The Place-based Accelerators programme closely mirrors the original ADDER programme – the main difference being that locations are more complex. For example, these areas may have overlapping local authority boundaries, meaning that the landscape within the defined local Accelerators area can differ significantly (for example, by different drug treatment providers), while other elements might be shared (for example, courts, police force).

Given the similarities between Project ADDER and the Place-based Accelerators programme, the wording ‘Project ADDER areas’ is used throughout this report to refer to both ADDER and Accelerators areas. Where a distinction is made between them, this will be made explicit, by using either the term ‘ADDER areas’ or ‘Accelerators areas’.

A range of evidence-based interventions were proposed by the Home Office and classified into 3 broad groups: treatment and recovery, diversion, and enforcement. Each of the local areas was then invited to prepare project proposals which reflected their local contexts, population characteristics, models of existing service provision, existing partnership structures and preferred delivery models. Local areas were able to choose a selection of activities from this ‘menu of interventions’ but were also given the discretion to propose alternatives in the design stage.

Overall, delivery plans sought to:

• enhance outreach services so that they would become more assertive, and address the lack of proactive/early intervention services
• change eligibility criteria and increase the flexibility of available support to make it easier for people using drugs to access treatment and support services
• increase the intensity of treatment offers, including improving the quality of support services and reducing caseloads to address complex needs and disengagement among service users
• boost recovery communities and so enhance recovery support through and after treatment
• utilise the WSA and embed strong partnerships to improve care pathways, provide holistic support and curtail siloed support
• purchase software and technology to enable more effective intelligence and information sharing between partners
• free up specialist resources for more targeted disruption activity, and provide additional specialist resources to address gaps
• facilitate change in management structures for greater strategic oversight

However, several pre-existing challenges were identified which final delivery plans were less able to address, including:

• the inability to support specialist, dedicated alcohol-only treatment services, and to support diversion from CJS pathways to support services for those using drugs other than crack cocaine and opiates
• a lesser focus on interventions related to non-Class A drugs (such as benzodiazepines and cannabis), which were also considered to be substances requiring treatment support or leading to engagement with the CJS by stakeholders
• a limited ability to influence housing supply/access to stable accommodation, with gaps in housing supply and tenancy support services
• an inability to support users outside the ADDER/Accelerators borders, or to address crime that is displaced to neighbouring areas
• avoiding duplication or overlap with other activities associated with the National Drugs Strategy which was launched in December 2021,^{[footnote 11]} after the design and early implementation of Project ADDER had taken place

4. Evaluation approach and objectives

4.1 Objectives

The key evaluation objectives are to:

• measure delivery against the stated project’s aims
• investigate experiences of delivery and provide recommendations on designing or rolling out a similar programme
• provide recommendations on continuing to evaluate the effects of Project ADDER

4.2 Approach

Verian developed a holistic evaluation design that combines various approaches, all underpinned by a flexible and iterative way of working together across the evaluation consortium partners, the Home Office, OHID, and all the local authority stakeholders. The overall design and specific methods align with the established Magenta Book and Green Book approaches^{[footnote 12]} and draws on a variety of data sources to inform the process and impact evaluation strands.

The evaluation is made up of 4 distinct but overlapping work streams, including:

• scoping and immersion: to produce a detailed evaluation strategy, a comprehensive programme-level ToC model (available in Annex A) and project-level ToCs
• process evaluation: to explore how Project ADDER has been implemented across all areas, to assess the extent to which delivery progressed as planned, and to identify facilitators and barriers to successful implementation
• impact evaluation: to measure whether Project ADDER interventions have produced or contributed to the desired outcomes, to understand how outcomes have occurred, and to explore any unintended outcomes; a literature review (Annex N) on the impacts of similar interventions was also conducted to contextualise the findings
• economic evaluation: to understand the costs and benefits of Project ADDER at the project and programme level, and to assess whether the programme is representing value for money; the results of the economic evaluation are not being published at this stage; this is due to the findings from the economic evaluation only reflecting the early stage of this programme where upfront investment was required and which were impacted by delays in fully mobilising due to COVID

An overview of the methods used for each workstream is provided in Annex C. This report covers the findings from the impact evaluation stream.

The report is structured by intervention type: enforcement, diversion and treatment. This was done in order to provide clarity on the theory behind different interventions, although it is noted that the use of a WSA meant that a number of interventions could be applicable to multiple sections. As a result, some of the interventions considered under ‘treatment’ may be relevant for enforcement or diversion outcomes, and vice versa.

The Project ADDER Evaluation: Practitioner Report covers the scoping phase and process evaluation results and methodology is available separately.

4.3 Impact evaluation approach and contribution analysis

The content in this report relates to insights from the impact evaluation. The primary method used for the impact evaluation was contribution analysis.

Contribution analysis is a theory-based method that seeks to understand the extent to which a programme has ‘contributed to’ the expected outcomes and impacts.^{[footnote 13]} It is an iterative process that is designed to test the underlying hypotheses in the ToC, and is typically used to understand the contribution of programmes conducted in a complex policy environment to the desired outcomes, or where quasi-experimental methods alone are otherwise not suitable. Contribution analysis ultimately allows an evaluator to determine, in a robust and evidenced way, how confident they can be that the theory behind a programme is manifesting as intended.

Contribution analysis evaluations are often tailored to fit the nature of the programme, the available evidence, and the context in question. For the evaluation of Project ADDER, Verian produced a customised contribution analysis design, broadly inspired by White’s 6 principles,^{[footnote 14]} and the Global Evaluation Initiative’s 6 steps for producing a credible contribution story.^{[footnote 15]}

In particular, the following key questions were considered to build the contribution story for Project ADDER:

Hypothesis: what was the theory or intention behind the programme? How were programme activities expected to lead to key outcomes and impacts? These questions were evidenced using:

• programme documentation and bid reviews
• workshops conducted with the Home Office to develop a programme-level ToC
• interviews and workshops conducted with local project strategic leads to develop a local-level ToC
• revision of the ToC between year 1 and 2 of the evaluation
• outcome mapping the end of year 1 and year 2 to identify key causal pathways and hypotheses in the ToC

Fidelity: were the programme activities delivered as intended? What barriers did they face? These questions were evidenced by:

• immersion interviews with local project strategic leads at the outset of the evaluation
• in-depth interviews with local project strategic leads and managerial / operational stakeholders

Evidence and limitations: did the expected outcome or impact from conducting the activities occur? How strong is the available evidence for this, and what limitations are there?

• in-depth interviews with local project strategic leads and managerial / operational stakeholders
• analysis of MI data collected by local projects and submitted to the Home Office for year 1 and 2 of the impact evaluation
• quasi-experimental analysis of outcomes for year 1 and 2 of the impact evaluation, using either difference-in-difference or generalised synthetic counterfactuals, depending on the specific outcome; data was provided by the Home Office, OHID and the ONS
• in-depth interviews with service users of the programme, conducted in year 2 of the process evaluation

Plausibility and alternative explanations: how plausible is it that the activities led to the outcomes or impacts? Are there any other explanations for why the impact or outcome did or did not occur?

• contribution analysis workshops conducted at the end of the year 1 evaluation with local project strategic leads and the Home Office
• contribution analysis in-depth interviews with strategic leads, conducted at the end of the year 2 evaluation
• contribution analysis workshops conducted with Home Office at the end of the year 2 evaluation
• literature review on evidence of the effectiveness of similar programmes

Detailed technical notes for each form of evidence are available in Annexes D to M.

4.3.1 Data timeframes

Table 7 summarises the outcome data sources and the month which outcomes were reported to.

Table 7: Data sources and timeframes

Data source	Outcomes reported up to
MI	March 2023
PRC offences and outcomes	December 2022
NDTMS treatment entry, referral source, and outcomes	March 2023
Welsh National Database for Substance Misuse (WNDSM)	December 2022
NHS Digital/NHS Wales hospitalisations	December 2022
ONS drug death registrations	March 2022

4.3.2 Interpreting contribution analysis results

Verian established a set of 5 possible conclusions that could be made for each hypothesis. These conclusions are based on a synthesis of all of the evidence (QED analysis, case study interviews and MI data) and different sources of evidence are given more prominence depending on the nature of the outcome being measured.

The contribution analysis conclusions provide an overview of the levels of confidence regarding whether Project ADDER was contributing towards the outcomes and impacts in the manner outlined in the ToC.^{[footnote 16]} These conclusions are summarised in Table 8.

Table 8: Possible contribution analysis scoring

Score	Conclusion
0	Not enough evidence to determine whether Project ADDER contributed towards the outcome
0	There has been no observable effect of Project ADDER on the outcome
1	Project ADDER may be contributing towards the outcome, but further evidence would be needed to determine if this is the case
2	Project ADDER is likely to have contributed towards the outcome, but there are some results or explanations that weaken the strength of the conclusion
3	Project ADDER has contributed towards the achievement of the outcome

Analysis was conducted first at a local level using a rubric that mapped the key contribution analysis questions against the 5 possible conclusions. This analysis was then scaled up to provide a programme-level view.

Confidence in the hypothesis was therefore increased where:

• programme activities were conducted and delivered as intended
• evidence triangulation across quantitative and qualitative methods suggested that an expected outcome occurred, and the available evidence does not have major limitations
• the causal pathway between activities and outcome is considered to be plausible by practitioners of the programme and policy experts
• there are few or no other explanations in the landscape or environment for the outcome occurring other than the programme in question

Conversely, confidence in the hypothesis was reduced where:

• programme activities were not conducted as intended, or encountered considerable challenges
• where the balance of evidence does not suggest an outcome occurred, where there is not enough evidence, or where the evidence that is available has significant limitations
• where practitioners of the programme and policy experts had enduring questions about the plausibility of the theory
• where there are many, or convincing, alternative explanations in the landscape or environment for an outcome occurring

Ultimately, no elements of Project ADDER are considered to have contributed measurably and definitively towards key outcomes. The reasons for this limited confidence are primarily the following:

• there was no observable effect of Project ADDER on the outcome: this may be because the programme is not leading to the intended effect, or because the outcome in question is expected to take a longer period of time to manifest than could be assessed within the scope of the evaluation
• the complexity of the policy environment, such as difficulties in differentiating the impact of Project ADDER from the impact of the roll-out of concurrent programmes operating in Project ADDER areas, such as projects from the National Drug Strategy and various other policies or initiatives
• evidential quality and availability, such as missing, incomplete, inconsistent, or out-of-date data on key project outcomes; general challenges also exist in accurately measuring some outcomes related to crime and justice: for example, the prevalence of drug use, drug supply, or OCG disruption; in some cases (for example, drug supply), proxy measures have been used where no direct measure is available; some degree of sample selection bias may have been present for qualitative insights

For each hypothesis, Verian has provided a detailed breakdown of the rationale for the conclusion drawn according to the key questions. This can be found in each relevant section in the main body of this report. A breakdown of detailed limitations for each form of evidence can also be found in Annexes D to M.

4.3.3 Case study evidence

Throughout this report, case studies have been included to give some additional insight into the programme-level findings, by giving an overview of the activities conducted at a local level. These case studies are primarily based on the perspectives of the stakeholders involved in implementation. For this reason, this case study evidence may not directly reflect the overall programme-level findings or conclusions for a particular hypothesis.

4.3.4 MI data

Local area enforcement teams shared monthly MI data submissions with the Home Office, which were subsequently shared with Verian. The data is local records of key outputs and outcomes relevant to the evaluation of Project ADDER, such as locally recorded arrests and seizures figures. The data typically runs from April 2021 to March 2023 in ADDER areas and from July 2021 to March 2023 in Accelerators areas, reflecting the delivery period of Project ADDER interventions. The data provides some description of the journey travelled over the course of delivery for each measure, per area.

However, there are some important limitations to using the MI data for evaluation. Firstly, local areas independently inputted and submitted data entries from their own internal records and databases, which means there were differences in how measures were interpreted and calculated, limiting comparability between areas. Secondly, there is no data from the pre-ADDER period, which limits the ability to attribute any changes in the data to Project ADDER. Finally, no equivalent MI data was collected from non-ADDER areas, which precludes any counterfactual impact analysis.

The MI data is used in this evaluation to draw comparisons with the evidence from stakeholder interviews, and to illustrate whether key measures appear to be broadly in line with expectations outlined in the ToC. Noting the limitations above, the MI data is not used to draw quantitative conclusions on the impact of Project ADDER.

4.3.5 Interpreting quasi-experimental results

For the evaluation of Project ADDER, a QED was integrated within the contribution analysis. A ‘combined’ approach provided a framework for integrating the findings from the QED with findings from the other strands of the evaluation. This approach recognises the complexity of the policy environment in which Project ADDER took place.

The primary quasi-experimental method used for the Project ADDER evaluation was generalised synthetic controls (GSC). This was used to estimate the ‘counterfactual’ – or what Verian predicts would have happened in the absence of Project ADDER. The models estimate this counterfactual using data from areas where Project ADDER was not implemented. These potential comparison areas were weighted so that the historic trend for an outcome across the comparison areas approximately matches the real trend observed for Project ADDER areas. The impact of Project ADDER is then estimated as the difference between the observed outcome and this estimated counterfactual outcome. In other words, this analysis compares what was observed in Project ADDER areas against an estimate of what would be expected without the programme. GSC controls for local project size.

The outcomes have been standardised to the population size to make comparison across areas more transparent. If a comparison of absolute estimates across areas was made, the difference in outcomes (for example, number of offences) might just be representing the difference in population size. For example, we expect that the overall level of offences recorded is higher in Bristol (with a population of 471,000) than in Hastings (with a population of 91,000). The same principle applies when comparing aggregated estimates for all ADDER and all Accelerators areas. Since the population of Accelerators areas is about 3 times the population of ADDER areas, we expect the level of offences to be higher in Accelerators areas. By showing the impact estimates in terms of per 10,000 or per 100,000 people, it is possible compare the extent of the impact across Project ADDER.

For some outcomes, a GSC model was not appropriate for use (see Annex F for further details). In these cases, difference-in-differences (DiD) analysis was conducted instead. This involves comparing the change in Project ADDER areas against the change in other areas. The underlying assumption is that the change seen in areas where Project ADDER was not taking place is the same as the change that would have happened in Project ADDER areas without the programme.

Within the contribution analysis framework, conclusions will rarely be based on the findings from QED analysis alone and different evidence sources will be weighted differently during evidence synthesis, depending on the nature of the outcome. However, for consistency, in describing impact evaluation estimates in the presentation of findings, the term ‘evidence of impact’ describes estimates which are significant with 95% confidence (a p-value equal to or less than 0.05). The description ‘weak evidence of impact’ describes estimates which are significant at the 90% confidence level (a p-value equal to or less than 0.1).

The QED has some limitations that are important to consider when interpreting the results. In particular:

• the QED models may not have completely removed the effects of external factors, such as co-current programmes or other influences at a local level; while the models attempt to account for these other factors to isolate the effects of Project ADDER, it is still possible that the findings are affected by these external influences; the contribution analysis sought to identify these alternative explanations, and confidence in the hypothesis was reduced where other factors could plausibly explain outcomes
• following the implementation of Project ADDER, funding for similar interventions was extended to non-Project ADDER areas via the National Drugs Strategy; some areas may also have located additional sources of funding in response to the report by Carol Black; this reduces the extent to which Project ADDER outcomes might be expected to differ from outcomes in other areas; in other words, it is possible that other areas may have been doing similar activities under separate funding streams, and this could make it harder to identify the effects of Project ADDER
• in some cases, there were errors, gaps or inconsistencies in the data sources being used for the QED, which could influence the results; the contribution analysis sought to identify these, and confidence in the hypothesis was reduced where limitations were present

A full breakdown of the impact evaluation methodology, including details on the QED models and the limitations of specific sources, can be found in Annexes F to M.

5. Enforcement

5.1 Enforcement summary

• Project ADDER activities related to increasing drug testing capabilities and enforcement activity are likely to have contributed to an increase in arrests of high-harm individuals involved in drug supply, and to seizures of drugs, assets and cash
• Project ADDER may be contributing towards a reduction in drug supply to and within local areas, but further evidence would be needed to determine whether this is the case
• there is not enough evidence to determine whether Project ADDER contributed towards reductions in ASB and re-offending
• there was no observable effect of Project ADDER on drug-related offending

5.2 Arrests and seizures

Project ADDER activities related to increasing drug-testing capabilities and enforcement activity are likely to have contributed towards an increase in arrests of high-harm individuals involved in drug supply, and seizures of drugs, assets and cash.

5.2.1 Hypothesis and data sources

The programme hypothesised that Project ADDER would lead to an increase in arrests of high-harm individuals involved in drug supply, and that it would also lead to an increase in seizures of drugs, assets and cash. These outcomes were expected to result from enhanced drug testing capability and enforcement activities. This includes the procurement of drug testing equipment, greater intelligence gathering and sharing, an increase in targeted local enforcement and drug warranty, more reassurance policing, enhanced activity on financial investigations, and more safeguarding of vulnerable individuals.

Although local projects experienced challenges in implementation – for example, handling the nature of shift work, competing priorities, and the time needed to set up information gathering and sharing processes – these activities largely took place as intended.^{[footnote 17]} Therefore, according to the underlying programme theory, it is reasonable to expect to see evidence that Project ADDER activities contributed towards an increase in arrests and seizures.

To investigate and test this hypothesis, the following evidence was used:

• descriptive analysis of MI data on arrests and seizures of cash, assets, and drugs over the course of the programme
• qualitative insights from year 1 and year 2 stakeholder and contribution analysis interviews

5.2.2 Evidence of outcomes

Local project stakeholders across most areas reported that Project ADDER increased intelligence-based enforcement activities, which they believed had led to more arrests of high-harm individuals and seizures. They reported that increased capacity for enforcement – for example, in funding additional officers or teams for targeted operations – had allowed increased action that would not have occurred otherwise. Enforcement activity was bolstered by increased capacity to build and share intelligence: for example, via the development of drug market profiles or setting up data sharing agreements and practices between teams. It was reported that this made enforcement more proactive, targeted and efficient. A continuation of these activities over time was also thought to have embedded these practices and developed the skills and experience of team members. As a result, stakeholders confidently reported that arrests of high-harm individuals and seizures of drugs, assets and cash had increased compared to before implementation of Project ADDER activities.

However, stakeholders also highlighted that enforcement activity was focused on targeting arrests of a small number of high-harm individuals, rather than larger numbers of low-level dealers. Therefore, they expected that the increases in arrests and seizures that were occurring through Project ADDER activities might not result in a large increase in overall arrests figures. Stakeholders also predicted that overall arrests might appear to be lower than expected, or to decrease, given that there may be periods of lower numbers of arrests after key high-harm individuals have been disrupted.

The MI data reflects this more nuanced picture of arrest patterns during programme implementation. Across both ADDER and Accelerators areas, there were decreases in the average number of overall arrests reported during programme implementation. There was a 37% decrease in ADDER areas, from 954 arrests in July to September 2021 to 603 arrests in January to March 2023, and there was a smaller 1% decrease in Accelerators areas, from 2,533 arrests to 2,520 arrests in the same timeframe.^{[footnote 18]}

In relation to drug seizures, the data for ADDER areas was broadly stable per quarter over the course of the programme, with an average of 2,460 drug seizures reported per quarter since the period April to June 2021. However, Accelerators areas saw an increase in drug seizures occurring over the course of the programme: a 20% increase since the period July to September 2021, largely driven by Merseyside.

Across Project ADDER, cash seizures increased during the implementation period, with an average of £125,000 seized every quarter in ADDER areas, and an average of £1.1 million seized each quarter in Accelerators areas.^{[footnote 19]}

Table 9: Average arrests, cash seizures and drug seizures per quarter, MI data, July to September 2021 to January to March 2023

Average arrests, cash seizures and drug seizures per quarter in MI data	ADDER areas	Accelerators areas
Arrests	796	2,485
Cash seizures	£125,000	£1,100,000
Drugs seizures	2,460	5,028

It is important to consider that there were also several other reasons, beyond Project ADDER, that may explain why local areas may have witnessed an increase in seizures. These alternative explanations include the police uplift programme,^{[footnote 20]} the introduction of the National Drugs Strategy,^{[footnote 21]} and overlapping local funding streams – for example, Operation Orochi, Project Medusa, and Project Evolve – which were also occurring at this time within some local project areas.^{[footnote 22]}

Case study: arrests and seizures in Wakefield

The Wakefield Enforcement team delivered a range of activities as part of its delivery plan: this included both overt and covert operations to gather and subsequently share intelligence with the Neighbourhood Impact Team (NIT) to inform their enforcement activity. Funding was also allocated for the use of digital investigative tools (such as Grey Key), and for an analyst to process information from police and partners in support of operations.

Wakefield reported the highest proportion of drug of warrants issued among the Accelerators areas. Additionally, as at May 2023, enforcement stakeholders reported that they had seized around £12 million of drugs and £200,000 in cash thanks to the formation of the NIT. Enforcement respondents noted that in some cases weapons had also been seized during these activities.

Targeted local enforcement through the creation of the NIT was considered the critical reason for the increase in arrests and seizures. The team allowed for dedicated resources to be utilised and targeted towards drug-related crime across Wakefield. It allowed the police to follow up on leads and to conduct enforcement activity proactively.

“The main drive of the NIT is a focus on the intelligence to drive that forwards so we’re not waiting. So, they’re able to process that information really quickly and react and respond to it. If you didn’t have that team, you’d be waiting for a neighbourhood team for a particular area to get the resources to deal with it and take them off normal everyday demand for what they’re doing in their area.”

Strategic, enforcement

Stakeholders felt that the presence of the NIT became increasingly effective over time as staff became more experienced and knowledgeable about drug-related activity in Wakefield, and as relationships were built with partners and local communities.

“Once you lift the stone and have resources to handle it this is what happens.”

Strategic, enforcement

5.2.3 Conclusion and future reflections

Overall, Project ADDER activities to increase enforcement activity are likely to have contributed towards an increase in arrests of high-harm individuals involved in drug supply, and an increase in seizures of drugs, assets and cash. Whilst definitions of ‘high-harm’ varied by area, this term was usually used to refer to individuals involved in the higher ranks of drug trafficking organisations, or those who commit extreme violence.

Local project areas experienced some barriers to implementing interventions related to increasing drug testing capabilities and enforcement activity, but despite these, activities were broadly conducted as intended. Stakeholder interviews indicate that interventions did lead to a greater number of arrests of high-harm individuals involved in drug supply, and a greater number of seizures of drugs, assets and cash. This was largely due to increased capacity for enforcement operations, and associated intelligence gathering, sharing and analysis. These mechanisms were independently reported across most areas, and there was strong corroboration between stakeholders that here had been an impact on arrests of high-harm individuals.

MI revealed that overall numbers of arrests and seizures typically remained consistent or decreased as the programme continued. Local projects often reported that their main focus was on the quality of arrests and seizures, such as focusing on high-harm individuals or drugs. Stakeholders therefore did not expect to see increases in the overall number of arrests and seizures data over the course of the programme.

However, there are a number of factors or circumstances that reduce confidence in the hypothesis that Project ADDER definitively contributed towards increased arrests of high-harm individuals and seizures. Primarily, there was no suitable area-level data available for QED analysis to determine whether the numbers of arrests and seizures in ADDER areas were significantly different from those in non-ADDER areas. Furthermore, there were also several non-ADDER-related operations active at the time of implementation, which could have affected arrests and seizures.

For future evaluations, it is recommended that data is recorded on the quality of arrests, to differentiate between arrests of ‘high-harm’ individuals and overall arrests in the local area. This would provide evidence to fully test the hypothesis that improving the targeting and quality of arrests increases the number of arrests of high-harm individuals, but may decrease arrests overall.

5.3 Drug supply

Indicatively, Project ADDER may be contributing towards a reduction in drug supply to and within local areas, but further evidence would be needed to determine whether this is the case

5.3.1 Hypothesis and fidelity

The programme hypothesised that Project ADDER activities would lead to a reduction in drug supply to and within local areas. This outcome was expected to result from increased arrests of high-harm individuals involved in drug supply and increased seizures of drugs, assets and cash, to which Project ADDER is likely to have contributed (see ‘Arrests and seizures’).

To investigate and test this hypothesis, the following evidence was used:

• descriptive analysis of MI data on OCG disruptions^{[footnote 23]}
• qualitative insights from year 1 and year 2 stakeholder and contribution analysis interviews
• QED analysis on PRC drug trafficking offences and charges data

There is no direct measure of drug supply. However, in consultation with stakeholders, it was expected that an increase in OCG disruptions should reasonably indicate a decrease in drug supply within the local area, and that disruptions could be a suitable proxy indicator. Similarly, stakeholders emphasised that an increase in drug trafficking offences and charges could also indicate disruption to drug supply, given increases in enforcement activity under ADDER. Therefore, in the early stages of the programme, stakeholders expected to see an increase in drug trafficking offences and charges, with a reduction occurring in the long term.

5.3.2 Evidence of outcomes

Stakeholders reported significant increases in disruptive activity; that activity was more targeted, intelligence-driven and efficient; and that OCGs that were not previously ‘on the map’ were now being targeted. However, some stakeholders reported that it was difficult to say whether activities constituted sufficient disruption to affect overall drug supply in the long term, given that county lines and OCGs can be quick to remobilise, particularly in areas with high demand. Some stakeholders also mentioned that disruption is most impactful when it is applied consistently and intensely over a long period of time.

The MI data supports the stakeholders’ view that there were increases in disruptive activity against OCGs during the implementation period, as shown in Table 10. OCG disruptions^{[footnote 24]} increased by 90% in ADDER areas from April to June 2021, though the majority were minor disruptions^{[footnote 25]} (73%). In Accelerators areas (excluding Bristol and Knowsley), OCG disruptions tripled from the period July to September 2021 to the period January to March 2023, with 83% being minor disruptions.

Table 10: % of major, moderate and minor OCG disruptions according to MI data

% of major, moderate and minor OCG disruptions	ADDER areas	Accelerators areas
Major OCG disruptions	6%	1%
Moderate OCG disruptions	21%	16%
Minor OCG disruptions	73%	83%
Change in total OCG disruptions since the start of implementation	+90%	+299%

Analysis of PRC data on drug trafficking offences in ADDER areas reveals a more complicated picture (Annex I, Table 1.1). The data demonstrates that there was a significant increase in drug trafficking offences recorded in ADDER areas in year 1, equivalent to 2 additional offences per 10,000 people compared to the counterfactual (or 167 additional offences overall). There was no evidence of an impact in year 2 in ADDER areas. There was no strong evidence of an impact on drug trafficking offences in Accelerators areas in years 1 or 2, compared to the counterfactual.

It is important to note that there were significant variations in the patterns of drug trafficking offences by local area. For example, the year 1 estimate is largely driven by a big increase in offences in Blackpool (Annex L, Table 1.1). However, in Norwich, there was evidence of a decrease in drug trafficking offences compared to the counterfactual in years 1 and 2, which could be related to other activities in the area targeting individuals involved in drug trafficking, which had been active for some time prior to Project ADDER.

Figure 1: Number of drug trafficking offences per 10,000 people (chart on left shows results for ADDER sites; chart on right shows results for Accelerator sites)

Notes:

1. PRC data published quarterly by the Home Office (April to June 2015 to October to December 2022).

Analysis of drug trafficking charges indicates a similarly mixed picture on the status of drug supply (Annex I, Table 1.6). There was no strong evidence that the increase in drug trafficking offences in ADDER areas in year 1 resulted in an increase in charges in years 1 or 2. However, in Accelerators areas, there was evidence of a significant increase in drug trafficking charges compared to the counterfactual, equivalent to 18 additional charges per 100,000 people in year 1 and 10 additional charges per 100,000 people in year 2 (corresponding to 473 and 257 additional charges, respectively, overall).

There was no definitive evidence to explain why there was an increase in drug trafficking charges in Accelerators areas without a corresponding increase in offences. Some potential explanations include the following: drug trafficking offences are usually only recorded as crimes when an arrest takes place; policing of street markets may have become easier during lockdowns because street dealers were more conspicuous, and because the Accelerators delivery period had a longer period of overlap with UK lockdowns due to COVID-19; relatedly, as a result of COVID-19, the reduced demand for a wide range of police services may have freed up police capacity for proactive work; there may have been greater displacement in Accelerators areas, which led to fewer drug trafficking offences; and Accelerators areas are more aligned to police borders than ADDER areas, which could have led to more effective working across the force and more efficient processes for making charges. At the area level, there was evidence of an increase compared to the counterfactual in Hackney (years 1 and 2), Knowsley (years 1 and 2), Tower Hamlets (years 1 and 2) and the Wirral (year 2) (Annex L, Table 1.5).

Figure 2: Number of drug trafficking offences resulting in a ‘charged/summonsed’ outcome per 100,000 people (chart on left shows results for ADDER sites; chart on right shows results for Accelerator sites)

Notes:

1. PRC and outcomes data published quarterly by the Home Office (April to June 2016 to October to December 2022).

Case study: drug supply in Middlesbrough

Project ADDER funded Operation Arrow with the aim of tackling large-scale OCGs, as well as reacting to the signs of exploitation by these criminal operations. Police deployed advanced covert tactics in the field, such as the use of concealed vehicles, undercover policing, and liaising with relevant partner agencies to gather information on the OCGs. The intelligence analyst funded under ADDER also bolstered the police effort through conducting effective analysis of the information and intelligence uncovered through Operation Arrow. This strategic approach enabled the police to build evidence to disrupt the OCG and its operations by arresting high-harm individuals and by seizing drugs and assets. The enforcement stakeholders reported that they believed supply was disrupted significantly due to this operation.

“[Operation Arrow] helped us identify the top-level people in the OCG and bring down their network through arrests, along with the recovery of drugs, money and 16 more children in the network.”

Strategic, enforcement

However, stakeholders reported that investigations and convictions may take more time before they reach fruition. The enforcement stakeholders also commented on the fact that there is some risk of OCGs re-emerging through their existing supply networks, even where disruption has been successful and sustained so far.

Case study: drug supply in Newcastle

In Newcastle, stakeholders reported that establishing an ‘analyst hub’ between Northumbria Police and public health had supported partnership work, including enabling enforcement activity to be more intelligence-led and targeted, as well as enabling ‘early alert’ systems for suspected drug deaths, which have since been extended force-wide under the Combatting Drugs Partnership. The analyst hub also included the development of the drug market profile partnership, which is a comprehensive public health approach to understanding drugs in the city.

The intelligence gained from the drug market profile was designed to support the mapping and disruption of OCGs, and subsequent arrests of higher-harm individuals. The MI data indicates that it was successful in doing so: the data shows that the number of OCG disruptions increased from 12 in January to March of the year 2022, to 40 in October to December of the same year.

Stakeholders reported that this data-driven approach improved the quality of enforcement activity. For example, stakeholders reported an increase in the quantity of drugs seized and in the number of arrests of higher-harm individuals. Supporting this view, the MI data shows an increase in the total number of arrests. However, it shows no increase over the course of the programme in the number of drug seizures or in the value of cash or assets seized.

Enforcement stakeholders were of the view that the overall drug supply had not been quantitatively disrupted, because as soon as one drug supply line is disrupted, the vacuum is quickly filled by another. However, they did perceive the intelligence-led Accelerators activities to have effectively disrupted the supply of particular types of drugs, or to particular population groups.

“I think the quality of seizures and the amounts has improved because it’s been more data and intelligence-led … previous campaigns targeted vulnerable cohorts, whereas we have focused on the mid to higher level of possession with intent to supply, so we have focused on the quality of arrests rather than quantity and the lower-hanging fruit.”

Strategic, enforcement

“We’ve made massive progress; we’ve taken down quite a number of serious and organised crime groups. But the reality is that as soon as you do that, in that vacuum comes another serious and organised crime group.”

Strategic, enforcement

5.3.3 Conclusion and future reflections

Overall, Project ADDER may be contributing towards a reduction in drug supply to and within local areas, but further evidence would be needed to determine whether this is the case.

Stakeholder interviews and MI data both support the view that targeted and intensified local enforcement activity, aided by the procurement of drug testing equipment, have increased OCG disruptions.

Analysis of the PRC data reveals a more mixed picture, with some evidence of an increase, compared to the counterfactual, in drug trafficking offences in ADDER areas in year 1, and in drug trafficking charges in Accelerators areas in year 2. This could be evidence of increased and effective enforcement activities as a result of Project ADDER. However, there was significant variation at a local level, and there were other projects and initiatives operating at a similar time to Project ADDER, which may have also contributed to these outcomes. For example, Operation Orochi and Project Medusa were part of the wider County Lines Programme (concentrated in London, Manchester, Merseyside and west London) which has been in operation since 2019. These programmes, and related increases in funding within some areas, may have impacted enforcement activity and drug supply in Accelerators areas, and may in part have driven the increase in drug trafficking charges. Similarly, Project Evolve, which targeted OCGs in Merseyside, may have also impacted drug supply in that area.

Moreover, drug trafficking, drug charges, and OCG disruptions are all indirect proxies for overall drug supply to and within local areas. Even where there is evidence of effective enforcement activity because of Project ADDER, this does not necessarily equate to an overall reduction in drug supply, especially given the possibility that OCGs and county lines could remobilise quickly after enforcement.

As a result, further evidence would be needed to determine whether Project ADDER has led to a reduction in drug supply. This would enable greater confidence on some existing measures, such as a higher proportion of major OCG disruptions, which might be expected to drive a long-term impact on drug supply. Efforts in this regard could include expanding the range of metrics used to measure levels of drug supply, such as analysis of wastewater, or price and purity metrics, and could include further monitoring over time to assess whether the expected increase in charges materialises.

5.4 Drug-related offending

There was no observable effect of Project ADDER on drug-related offending.

5.4.1 Hypothesis and fidelity

The programme hypothesised that Project ADDER activities would lead to an overall reduction in drug-related offending. This outcome was expected to result from a reduction in drug supply to and within local areas, and a reduction in ASB and re-offending. The latter part of this hypothesis is based on the assumption that since ASB is a common category of drug-related offending, reductions in ASB-related crimes would lead to reductions in drug-related offending overall; similarly, if individuals involved in drug-related offending desist from re-offending, this would also lead to fewer overall drug-related crimes being committed.

Although Project ADDER activities are likely to have contributed towards an increase in arrests of high-harm individuals, there is less evidence to suggest that this has gone on to affect drug supply. There is also not enough evidence to determine whether Project ADDER has contributed towards reducing ASB or re-offending. This reduces the extent to which Project ADDER might be expected to lead to an observable reduction in drug-related offending.

To investigate and test this hypothesis, the following evidence was used:

• qualitative insights from year 1 and year 2 stakeholder and contribution analysis interviews
• QED analysis of PRC data on acquisitive crime offences, drug possession, neighbourhood crime offences, serious violent crime^{[footnote 26]} and homicide
• QED analysis of NHS data on hospitalisations due to assault with a sharp object

There is no direct or consistent measure of drug-related crime. The motivations and/or factors leading to crimes are not measured in PRC data. However, wider evidence indicates that there is a link between drug use and some crime types. Therefore, to measure this outcome, a range of different types of crime were selected which are known to be associated with drug use or to be otherwise drug-related, as proxy measures for drug-related crimes (see Annex H for a full list of crimes included under acquisitive crime, neighbourhood crime, serious violence and homicides).^{[footnote 27]}

5.4.1 Evidence of outcomes

QED analysis of PRC data on offences (see Annex I, Table 1.1) indicates that there is no evidence of an impact of Project ADDER, as compared to the counterfactual, on drug possession offences, acquisitive crime or serious violent crime in either ADDER or Accelerators areas in years 1 or 2.

Notably, however, some local projects saw increases in these offences compared with the counterfactual (see Annex L, Table 1.1 for ADDER areas, and Annex M, Table 1.1 for Accelerators areas). There were increases in the following: drug possession offences in Middlesbrough in year 1; acquisitive crime offences in Blackpool (year 2) and Middlesbrough (years 1 and 2); and serious violent crime in Blackpool (years 1 and 2) and Liverpool (year 1). Only Norwich saw consistent reductions in offences for acquisitive crime (year 1) and neighbourhood crime (year 2).

With regards to neighbourhood crime,^{[footnote 28]} there is evidence of an increase in recorded offences in Accelerators areas in year 1, equivalent to 15 additional offences per 10,000 people, or 3,895 additional offences overall. These increases were particularly evident in Newcastle upon Tyne (year 2) and Tower Hamlets (year 1). There was also an increase in neighbourhood crime offences in Middlesbrough (years 1 and 2); however, there is no evidence of an impact in ADDER areas overall. Some stakeholders reported that an increase in neighbourhood crime offences reported could be related to an increase in reassurance policing, enforcement activity or engagement with the public under ADDER, rather than an actual increase in crime.

Figure 3: Number of neighbourhood crime offences per 10,000 people

Notes:

1. PRC data published quarterly by the Home Office. Data from April to June 2015 to October to December 2022.

As illustrated in Figure 5 there is weak evidence of a reduction in homicides in ADDER areas in year 1 of implementation, but not in year 2, nor in Accelerators areas during the reference period (see Annex J, Table 1.1). However, this is a relatively rare crime and therefore it is sensitive to small changes.

Figure 4: Number of homicides per 10,000 people

There is also some evidence of a reduction in hospital admissions due to assault with a sharp object (a key metric for serious violence) in Accelerators areas in year 2, as illustrated in Figure 6 (see Annex J, Table 1.2).

Figure 5: Number of patients admitted to hospital due to assault with a sharp object per 100,000 people

Notes:

1. Data collected in the Hospital Episode Statistics (HES) (England) and the Patient Episode Database for Wales (PEDW). 6 data points (from 2017 to 2022).

Generally speaking, stakeholders were often not able to say whether or not Project ADDER had directly reduced drug-related offending in their area. They commented that many external factors determine offence rates, and they pointed to issues with demarcating crimes as drug-related. Furthermore, they suggested that any changes to drug-related crime in national statistics might not be observable for some time, although many reiterated that they were optimistic this would happen in the future.

Some stakeholders did note, anecdotally, that offending among Project ADDER service users seemed to have reduced. This was corroborated by some service users themselves, a small number of whom reported that they had not re-offended since being referred to treatment or were no longer engaged in cycles of criminality.

Case study: drug-related offending in Norwich

Project ADDER funded additional posts on the Norwich county lines team, which uses data and technology to identify supply lines into the county and to target police activity. The increased capacity facilitated an improved intelligence picture and better evidence gathering (such as mobile phone data that they could attribute to the drug line controller), leading to arrests of more high-harm individuals. This finding is supported by stakeholder interviews, MI data showing an upward trend in OCG disruptions and county lines closed, and impact data demonstrating significant decreases in drug trafficking offences in years 1 and 2, suggesting a reduction in drug supply occurred.

“What I can say for certain is that within the last 12 months there is a recognisable reduction in the levels of supply of crack and heroin across Norfolk, particularly in Norwich … that is seen both through our intelligence reporting of the number of drug lines operating through the city, but also anecdotally and qualitatively.”

Managerial, Enforcement

Project ADDER also provided a support fund to help service users cover small costs, such as groceries, toiletries, and emergency accommodation. Impact data in this area shows a significant decrease in both acquisitive crime in year 1 and neighbourhood crime in year 2. Stakeholders suggested that this support fund may have contributed to reducing shoplifting among service users.

“We have a support fund and quite often we’ve had service users come in and they’ve not eaten for a few days and sometimes we will go and get them some very basic groceries. I guess if we weren’t doing that and providing them with something, potentially they could go and shoplift.”

Managerial, Treatment

This area has not seen a reduction in drug possession offences: stakeholders reported that this may be due to a lesser focus on drug possession and a greater emphasis on tackling drug trafficking. The data for serious violent crime offences has been volatile and has not reduced overall. However, the enforcement strategic lead reported that less fighting is taking place between dealers as the market is no longer oversaturated (supply previously outstripped demand so dealers would fight with each other over their share of the market). They also said that reports of cuckooing have decreased as drug dealing is not taking place to the same extent as it was before because there are fewer high-harm individuals targeting vulnerable people.

5.4.2 Conclusion and future reflections

Overall, there was no observable effect of Project ADDER on drug-related offending.

There was little evidence that Project ADDER had reduced drug supply, ASB or re-offending. This reduces the extent to which Project ADDER might be expected to have contributed towards a reduction in drug-related offending, as outlined in the ToC.

There was no observable impact of Project ADDER, compared to counterfactual analysis, on various types of crime: drug possession offences, acquisitive crime or serious violent crime. Although there was some evidence of a reduction in sharp hospital admissions compared to the counterfactual in Accelerators areas in year 2, there was also an increase in reported neighbourhood crime, which stakeholders suspected was related to greater levels of enforcement and reassurance policing.

Stakeholders were often unable to comment on whether Project ADDER was having an impact on drug-related offending, although many were optimistic that the programme would contribute towards a reduction going forwards. In line with this suggestion, programmes that have similar characteristics to Project ADDER appear to have sometimes reduced numbers of criminally involved people using drugs, and thus the volume of drug-related crime. Evidence from similar diversionary programmes has shown mixed success in terms of crime reduction. However, the evidence for the impact of treatment programmes in reducing crime is somewhat more promising. Future crime reduction might therefore depend on the extent to which impact is demonstrated on the treatment side (see Annex N).

The nature of drug dependence as a relapsing condition also means that outcomes inevitably take time to manifest and are also likely to be highly dependent on the relationships that service users have with support or treatment staff. While the expectation is that programmes such as ADDER/Accelerators may in time show benefits in terms of reducing drug use and levels of drug-related crime, these are unlikely to be observable within a 2-year or even 3-year timescale (see Annex N).

5.5 ASB and re-offending

There is not enough evidence to determine whether Project ADDER contributed towards reductions in anti-social behaviour and re-offending.

5.5.1 Hypothesis and fidelity

The programme hypothesised that communications campaigns funded by Project ADDER would lead to reductions in ASB and re-offending. Local projects were permitted to use funding for various communications campaigns targeted to schools, dealers, or users, depending on the needs or context of the local area, some of which sought to tackle ASB and some of which sought to tackle re-offending.

Not all local areas set out to conduct communications work, but among those that did (4 ADDER areas and 3 Accelerators areas), the campaigns were largely implemented as intended. Some local projects did experience delays related to issues with the recruitment and retention of communications officers, and difficulties aligning with the plans of external partners, such as schools or Crimewatch.

It was hypothesised that where these activities took place, and depending on the specific aims of the communication campaigns, the area would go on to see a reduction in ASB (such as street dealing and use of drugs in open public spaces) or a reduction in drug-motivated offending. This theory was considered plausible and was supported by programme- and local-level project stakeholders at the outset of the evaluation.

To investigate and test this hypothesis, the following evidence was used:

• qualitative insights from year 1 and year 2 stakeholder and contribution analysis interviews

Information related to ASB or re-offending was not collected via MI data and as such was not available for descriptive analysis.

QED analysis was also not conducted on PRC data as ASB records were not considered a reliable measure of crimes, due to high levels of underreporting. Similarly, there is variation in how police forces record re-arrests and there is some time lag in recording re-offending (such as through MoJ data on known re-offending).

5.5.2 Evidence of outcomes

Local project stakeholders reported that ASB, including street dealing and public use of drugs, may have decreased as a result of Project ADDER, but were often unsure as to what extent this had occurred. This uncertainty was due to a range of factors, including the large range of offences that constitute ASB, the effect of the COVID-19 pandemic, the effect of the cost-of-living crisis, and the seasonal nature of ASB offending.

Of the stakeholders that did report a possible reduction in ASB, most did not attribute this directly to communications campaigns. Instead, stakeholders attributed this to greater police presence, enforcement action, or diversion into treatment. Treatment stakeholders also reported a reduction in ASB and re-offending amongst those receiving support, and this was supported by some service users, who reported that their own offending behaviour patterns had changed following treatment.

Case study: ASB and re-offending in Newcastle

In Newcastle, a dedicated criminal justice team was resourced to work with Northumbria Police officers, probation, courts, and local prisons. This also had a dedicated women’s worker to link with the Women’s Hubs and a dedicated Family Intervention Worker to support the family members affected. Arrest referral workers were employed to support the custody-based assessments and improve processes for offenders as early in the pathway as possible, supporting the re-introduction of DToA.

Strategic leads reported that reduced caseloads within the team have allowed the delivery of focused and quality interventions focused on key outcomes (including reducing re-offending, improving quality of life, and a range of health outcomes), and for more flexible approaches to be carried out (including assessments within probation offices, in-reach to prisons, and closer working with police teams). They felt that closer alignment of this Criminal Justice Intervention Team (CJIT) focus (which includes a refreshed focus on Integrated Offender Management locally) had brought a stronger pathway for offenders and their families, where resources can meaningfully target areas of need both for individuals/offenders, while also aligning resources to any operational focus within policing.

In an effort to focus on the student community, enforcement activity has been combined with an education and communications approach (with the 2 education awareness officers who work in the Violence Reduction Unit) by establishing pathways and in-reach from the drug and alcohol provider. The focus of this work has been to prevent and disrupt drug use and drug-related offending among the university student population. This work is also based on the understanding that the student community may be particularly vulnerable to being targeted by OCGs and drug suppliers, given previous drug-related deaths among the student community. The packages include information on the risks of popular drugs, and how a criminal conviction would impact their ability to continue their degree course, travel and find employment. During university freshers’ week, the areas which have a high density of the student population were targeted by the police, to communicate these messages to students.

“We’re making it not punitive, giving them increased awareness to be more informed. It’s not just about pursue and enforcement – [it] is about protect, prepare and prevention.”

Strategic, enforcement

Case study: ASB and re-offending in Bristol

Several interventions in Bristol aimed to reduce ASB and re-offending. For example, Bristol Accelerators funded drug education within schools, carried out by a dedicated drugs worker. This drug education aimed to give young people the skills to reduce the risks of problematic drug use, and also involved upskilling teachers and professionals in how best to communicate with young people.

In addition, Crimestoppers carried out a communications campaign in particular areas of the city. This included messaging on posters and buses to encourage the local community to report concerns about drug use or drug-related crime to the police. The campaign highlighted the harm reduction and safeguarding focus of the project.

“In south Bristol we’re putting out posters that communicate a public health approach, to show that the police aren’t just here to arrest drug users, they are more interested in organised crime and want to support vulnerable people and get them into treatment.”

Operational, Enforcement

Stakeholders felt activities carried out under Accelerators, including the communications work, had been effective in encouraging reporting to the police. Stakeholders stated that community reporting to the police had increased during the life of the project, providing the police with increased intelligence. This, in turn, supports targeted enforcement activity to close down properties that are sites of drug dealing and cuckooing activity.

“In Southmead we have noticed that the public are a lot more willing to talk to the police, who have closed down 9 sites of dealing in the area in the last year. People are noticing this … People will call the police and say I can hear shouting or banging. Unfortunately, that’s not going to get a police response on a Friday night at 2 am, even though it’s making their life hell over many months, so being able to come in and kick everyone out, changes their life for the better.”

Managerial, Enforcement

Stakeholders identified numerous other Accelerator activities that may have contributed to a reduction in ASB and re-offending. This included partnership working between the police and the local authority’s ASB team and work carried out by treatment services.

“I’ve been clean for over a year now. And I’ve not been in trouble with the law either. The help from [the drug treatment service] has helped me do that and stay out of trouble.”

Service user

5.5.3 Conclusion and future reflections

Overall, there is not enough suitable evidence to determine whether or not Project ADDER interventions had an effect on ASB or re-offending.

Although not all local areas set out to conduct communications work, typically the areas that did were able to conduct them as planned, though sometimes implementation was delayed.

Most stakeholders were unable to comment on whether or not ASB and re-offending had reduced in their area. Stakeholders did not feel it was possible to directly attribute a reduction in ASB and re-offending to communication-related interventions funded by Project ADDER. They also did not feel it was possible to attribute an impact to Project ADDER activities more generally, due to the number of alternative factors known to influence ASB. This evaluation did not assess whether these interventions were in line with current evidence and guidance, such as that from the Advisory Council on the Misuse of Drugs.

A few stakeholders mentioned that increased police presence and more intensive treatment could have decreased ASB and re-offending, and also that ASB and re-offending may have reduced among those receiving Project ADDER treatment or support services. However, this was not conclusive.

Beyond stakeholder evidence, there was no suitable quantitative data available on ASB and re-offending for quasi-experimental impact analysis or descriptive trend analysis. Whilst MoJ data on re-offending is made publicly available,^{[footnote 29]} the data period (up to December 2021) did not align with the Project ADDER delivery period, so as to compare data before and after the programme. Future evaluations could consider using future reports of this data to analyse trends in re-offending.

Going forwards, it may be useful for the Home Office to make funding available for local projects that are conducting communications campaigns to commission specialist communication evaluations. This would allow for additional evidence to be collected to test the ToC for ASB and re-offending. Furthermore, evidence for additional causal mechanisms in the ToC could be considered, namely the effect of support services on ASB and re-offending among service users.

6. Diversion

6.1 Referral into treatment from the CJS

Project ADDER activities related to DToA, OOCDs, and improving integrated care pathways are likely to have contributed towards an increase in the number of drug users referred into support services from the CJS.

6.1.1 Hypothesis and fidelity

Project ADDER was expected to lead to an increase in the number of people using drugs referred into treatment or support services from the CJS, as a result of activities that aim to divert individuals into treatment, specifically DToA, expansion of OOCDs, and integrated and improved pathways into and through treatment.

Overall, areas largely delivered DToA and OOCD interventions as planned. Some areas experienced delays in mobilising integrated pathways into treatment, largely due to recruitment issues, delays in setting up data sharing agreements, and the amount of time required to align enforcement, treatment cultures and working practices. Therefore, according to the underlying programme theory, it is reasonable to expect that Project ADDER has contributed towards an increase in the number of people using drugs diverted into treatment and support services from the CJS.

To investigate and test this hypothesis, the following evidence was used:

• descriptive analysis of MI data on DToA and OOCD provision (the main OOCDs used being community resolutions or conditional cautions)
• QED analysis of NDTMS and WNDSM on referrals into treatment from the CJS
• QED analysis of PRC data for the numbers of drug possession offences that resulted in a community resolution^{[footnote 30]}
• qualitative insights from year 1 and year 2 stakeholder and contribution analysis interviews

6.1.2 Evidence of outcomes

Stakeholders reported that DToA, treatment outreach in custody suites and prisons, and continuity of care outside of prison had facilitated more referrals into support and treatment services, and that the use of multi-disciplinary teams and partnership working had made the referral process smoother and faster for service users. Anecdotally, stakeholders mentioned that more individuals subjected to cuckooing were also being referred into support services.

However, stakeholders also reported that they did not see as many referrals from the CJS as expected. Some stakeholders noted that this may have been due to delays in implementation, as mentioned above, and also due to some areas directing judiciaries to use more DRRs. Some stakeholders also commented that DToAs were only testing for cocaine and opiates, which could have limited referrals.

MI data reveals that there was an average of 306 arrestees tested for drugs per quarter in ADDER areas, and an average of 1,425 per quarter in Accelerators areas. A majority of tests were positive in ADDER and Accelerators areas, with 65%^{[footnote 31]} and 51%^{[footnote 32]} positive tests in ADDER and Accelerators areas, respectively. Overall rates of DToA remained stable through implementation, although there was some area-level variation: for example, there was an increase in arrestees tested for drugs in Norwich and Newcastle, and a decrease in Hackney and Tower Hamlets.

The use of OOCDs in ADDER areas increased over the course of the programme, with a 65% increase in individuals being offered OOCD/Diversion schemes since April to June 2021. In Accelerators areas, the number of OOCDs largely remained stable over the delivery period, with an average of 1,990 OOCDs offered per quarter.^{[footnote 33]}

Table 11: Average use of DToA and OOCDs per quarter in MI data

Average use of DToA and OOCDs per quarter		ADDER areas	Accelerators areas
DToA	Number per quarter	306	1,425
	Percentage change	+11%	-15%
OOCDs	Number per quarter	507	1,990
	Percentage change	+65%	+9%

In line with evidence from stakeholder interviews and MI data, analysis of PRC data indicates that there was a significant increase in the proportion of drug possession offences resulting in community resolutions (a type of OOCD) as a result of the programme for Accelerators areas in year 2 (Annex J, Table 1.3). This is a significantly greater increase than the counterfactual, which also increased over the programme period, which may be indicative of national-level changes in the approach to drug offences and community resolutions.^{[footnote 34]} However, the results for ADDER areas demonstrate a different picture, with evidence of a lower proportion of community resolutions in year 1 than the counterfactual, and no evidence of a difference from the counterfactual in year 2.

Figure 6: Proportion of drug possession offences resulting in a community resolution

Notes:

1. PRC and outcomes data published quarterly by the Home Office, taken from the period April to June 2016 to the period October to December 2022.

Analysis of NDTMS and WNDSM data also supports the assertion that Project ADDER has increased referrals into treatment from the CJS. As illustrated in Figure 7, there is evidence of an increase in referrals into treatment in both ADDER areas and Accelerators areas in year 2, though the evidence is comparatively weaker in Accelerators areas (Annex I, Table 1.5). The increase in ADDER areas was largely driven by an increase in Middlesbrough in year 2. The later realisation of impacts in year 2 could have been due to delays in mobilising OCCDs, as reported by stakeholders in some areas.

Figure 7: Number of patients referred into treatment from the CJS per 100,000 people

Notes:

1. NDTMS and WNDSM (Wales) data published from the period January to March 2015 to the period January to March 2023 shared by OHID and NHS Wales.

However, some stakeholders did report that referrals in from the CJS were not as high as they were initially expecting, for a variety of reasons, including the following: OOCDs not always being suitable for heroin or crack users; entry into treatment, or engagement with support, not being mandated; and delays or challenges in establishing referral pathways.

Case study: referring people who use drugs into treatment or support services from the CJS in Norwich

Norwich used ADDER funding to increase the use of DToA and offer improved care pathways by placing treatment workers in custodial settings. Stakeholders reported that they believed DToA had increased diversion from the CJS into treatment because it identified people who would not otherwise have sought treatment, or even have considered themselves to have a drug problem. Workers were also able to engage and divert arrestees in this key moment by offering assessments and booking appointments. MI data shows a clear upward trend in both the number of arrestees being drug tested and people who use drugs being offered OOCDs. Norwich had the highest number of DToAs compared to other ADDER areas, drug testing 1,106 arrestees for the presence of cocaine and Class A opiates over the ADDER evaluation period.

Stakeholders reported that referrals from CJS pathways had increased, and this was supported by impact data which showed a significant increase in criminal justice referrals into drug treatment in year 1.

Norwich also improved pathways between the CJS and treatment by running SMART Recovery groups, psychosocial sessions, and one-to-one appointments in prison. Crucially, partnership working meant that prisons began giving detailed information to treatment services about people prior to release (including date of release and prescription dosages) and, as a result, treatment services were better able to engage prison leavers and get them on the right prescription straight away.

“[Prison leavers] are attending here on release and engaging with us … they are more trusting in the process and that there is going to be support when they’re released.”

Managerial, Treatment

6.1.3 Conclusions and future reflections

Overall, Project ADDER activities to increase DToA and OOCDs and improve integrated care pathways are likely to have contributed towards an increase in the number of people using drugs and vulnerable individuals referred into support services, in at least some areas.

Stakeholders reported that activities conducted under Project ADDER had led to an increase in referrals of individuals from the CJS into treatment or support. The MI data on DToA reveals that these largely remained stable in ADDER and Accelerators areas over the course of the programme. In Accelerators areas, OCCDs also remained stable. However, ADDER areas saw a large increase in OCCDs (65%) throughout the implementation period.

Stakeholder views are supported by significantly greater referrals into treatment compared to the counterfactual in NDTMS/WNDSM data across Project ADDER. There was also a significantly greater proportion of drug possession offences resulting in community resolutions in Accelerators areas in year 2, though notably there was no significant difference in ADDER areas from the counterfactual in the same period. Some stakeholders also reported that referrals in from the CJS were not as high as they were initially expecting.

Broader literature supports the conclusion that integrated care pathways are likely to have contributed towards an increase in the number of people using drugs and vulnerable individuals referred into support services, particularly DToA. However, evidence also suggests that DToA is least likely to engage entrenched opiate users, who comply because they ‘have to’.

Broader literature further states that the implication of the high levels of attrition in getting criminally involved problem users into treatment is that expectations about timescales for effects to emerge must be realistic. Only a minority will take up their first referral to assessment, and then accept treatment, and then stay in treatment, but over time the numbers actually receiving treatment will increase.

7. Treatment

7.1 Treatment summary

• Project ADDER activities related to DToA, OOCDs, and improving care pathways may have contributed towards an increase in the number of people using drugs referred into support services from the CJS
• Project ADDER activities may be contributing towards an increase in the number of people who use non-opiate drugs referred into treatment and wider support services
• there was no observable positive effect of Project ADDER on increasing the number of people who use opiate drugs referred onto treatment and wider support services, and some evidence of a negative impact for this group
• Project ADDER may be contributing towards more service users being supported to sustain a life that is no longer dependent on drugs, but further evidence would be needed to determine whether this is the case
• Project ADDER may be contributing towards a reduction in the prevalence of drug use, but further evidence would be needed to determine whether this is the case
• there was no observable effect of Project ADDER on drug-related deaths

7.2 Referral into treatment and support services

Project ADDER activities may be contributing towards an increase in the number of non-opiate users referred into treatment and wider support services

There was no observable effect of Project ADDER on the number of opiate users referred into treatment and wider support services.

7.2.1 Hypothesis and fidelity

It was expected that Project ADDER activities would lead to an increase in the number of people using drugs referred into treatment and wider services, including housing and employment support. Relevant Project ADDER activities included increasing the provision of pharmacological, psychological and harm reduction programmes, as well as providing greater treatment capacity (for example, reducing casework workloads) and offering integrated and improved care pathways.

Overall, areas were able to implement these activities as intended. However, in many areas, the mobilisation of integrated and improved care pathways was slower than expected due to a number of challenges experienced during implementation, including issues with recruiting and onboarding new staff (especially clinical psychologists and non-medical prescribers/nurses), developing multi-agency partnerships, and engaging individuals who have been traditionally disconnected from services.

To investigate and test the hypothesis, the following evidence was used:

• QED analysis of NDTMS and WNDSM on numbers of new opiate patients entering treatment and new non-opiate patients entering treatment
• qualitative insights from year 1 and year 2 stakeholder and contribution analysis interviews
• descriptive analysis of MI data on individuals reported as having meaningful contact with a support or treatment service

7.2.2 Evidence of outcomes

Stakeholders commonly reported that referrals into treatment had increased, both from CJS pathways and as a result of targeted outreach activities for specific groups: for example, sex workers, individuals experiencing homelessness, and those experiencing co-use or co-morbidities. Stakeholders shared that the referrals were often facilitated by more integrated care pathways, information sharing between partners, the use of multi-disciplinary teams, and co-location of support services. The presence of new, evidence-led referral pathways, and reductions in treatment worker caseloads, was thought to have allowed for more individuals to be referred into more forms of treatment.

Service users of Project ADDER corroborated this and reported that they had noticed greater and faster access routes to a range of support services that were appropriate for their needs.

The QED analysis of NDTMS/WNDSM data did not show any evidence of an increase in the number of new non-opiate patients entering treatment in ADDER or Accelerators areas. However, the trend in ADDER areas was consistently higher than the counterfactual from midway through year 1, and showed promising signs of an overall average increase.

Figure 8: Number of new non-opiate patients entering treatment per 100,000 people (chart on left shows results for ADDER sites; chart on right shows results for Accelerator sites)

Notes:

1. NDTMS/WNDSM data (from January to March 2015 to January to March 2023).

Conversely, NDTMS data indicates that in ADDER England areas,^{[footnote 35]} there is evidence of a reduction in the number of opiate patients entering treatment, equivalent to 247 fewer patients (or 48 patients per 10,000 people) in year 2 – significantly lower than the counterfactual (Annex I, Table 1.4). These estimates are driven by decreases in Blackpool and Norwich. There was no significant difference between Accelerators areas and the counterfactual, although there is evidence of a reduction in Tower Hamlets.^{[footnote 36]}

Figure 9: Number of new opiate patients entering treatment per 100,000 people (chart on left shows results for ADDER sites; chart on right shows results for Accelerator sites)

Notes:

1. NDTMS/WNDSM data (January to March 2015 to January to March 2023)

A number of explanations were given by stakeholders as to why referrals into treatment among opiate users might have been lower than expected. Firstly, they reported that the focus of Project ADDER treatment activities was on the quality of referrals rather than quantity, seeking to target people using drugs who had historically been the most challenging to access. Given that opiate users are more likely than other groups to have entrenched drug use patterns, this approach may have increased referrals, but at a relatively low rate compared to outreach conducted for other groups. Stakeholders also reported that opiate users are particularly likely to require significant preparatory work or support before they can formally enter treatment. This form of support would not be recorded in NDTMS data: for example, unstructured treatment services, mentoring, and tier 2 support.

Beyond these explanations, analysis of MI data gives reason to be optimistic about the impact of Project ADDER on referrals going forward. In Accelerators areas, there was a 35% increase in instances of ‘meaningful contact’ with various types of outreach activity since the start of the programme, with increases recorded across 5 of the 6 areas.^{[footnote 37]} The same data for ADDER areas reveals a negative trend, although 3 of the 5 areas reported increases in meaningful contact over the implementation period.

Table 12: MI data on instances of meaningful contact

Instances of ‘meaningful contact’ recorded in MI data	ADDER areas	Accelerators areas
Average instances per quarter	2,014	3,479
Change over project (%)	-20%	+35%

Case study: referral into treatment and support services in Swansea

Stakeholders in Swansea aimed to increase referrals into treatment via 3 new support packages.

Firstly, the Rapid Access Prescribing Programme (RAPP) provided pharmacological support to sex workers and individuals at risk of overdose. RAPP increased referrals into treatment because it did not require people who use drugs to be part of the CJS and it used assertive and targeted outreach to engage people who use drugs with significant complexity and vulnerabilities (as opposed to putting the onus on them to actively seek support). Through word of mouth, more individuals became aware of the service and started to self-refer, with some even making false disclosures of sex work or accidental overdose in an attempt to access it. To ensure the service reached the people it intended to, workers took measures to verify the claims, such as checking with SWAN, a partner agency dedicated to supporting sex workers in the area, and checking GP records.

“Having the capacity to do outreach is definitely a complement to our services, because you’re not expecting people to come in. These are individuals that 11 o’clock on a Tuesday means nothing to them. They don’t have schedules; they don’t have structure. With experience as an outreach worker, you know where to locate people … once you’ve got that ground knowledge it pays dividends for tracking people down.”

Managerial, Treatment

Additionally, 2 treatment workers were based in police custody suites to provide early intervention to people who use lower-level drugs (not crack cocaine or heroin) identified via DToA as needing support. Stakeholders reported that this contributed to better referrals as these individuals would not have been eligible for specialised treatment services prior to ADDER. The fact that they were offered an assessment and appointment in custody, rather than simply signposting them to treatment, was also thought to have increased engagement and therefore referrals.

“It’s [early intervention] probably the most sought-after service at the moment. If you think of the 85 individuals who come through to early intervention … they previously wouldn’t have been eligible for Dyfodol interventions because at the time Dyfodol was just looking at crack cocaine and opiates.”

Managerial, Treatment

Finally, Mental Health Treatment Requirements gave court-ordered mental health support to offenders with substance misuse issues. Stakeholders believed that this supported referrals into treatment by reaching people who would previously not have been eligible for any mental health support due to not meeting the ‘crisis threshold’ (that is, level of distress). Stakeholders reported receiving feedback from service users that this mental health support is something they have been looking for but have previously been unable to access.

7.2.3 Conclusion and future reflections

Project ADDER may be contributing towards an increase in the number of people using non-opiates referred into treatment and wider support services. There is no observable positive effect on referrals of opiate users.

Stakeholders spoke positively about the effect of Project ADDER on supporting people using drugs into treatment and support services. However, stakeholders did report that delays in implementation led to fewer referrals than originally expected.

Whilst the QED estimates showed no evidence of an increase in the number of non-opiate patients entering treatment, the positive trend compared with the counterfactual in ADDER areas looks promising. There was not a significant difference between Accelerators areas and the counterfactual for referrals of new non-opiate patients; however, the qualitative evidence shows signs of improvement in the number of new patients, and the MI data shows an increase in meaningful contact in Accelerators areas over the course of implementation. Surprisingly, in ADDER areas, there was a reduction in new opiate users referred into treatment compared to the counterfactual. There was no significant difference between the counterfactual and Accelerators areas.

Stakeholders pointed out that opiate users tend to be more challenging to refer into treatment as they are more likely to have entrenched patterns of drug use over a long period of time. This group are also more likely to require referrals into unstructured or new forms of referral that are not captured in NDTMS data. Whilst these comments may explain why there was not an increase in referrals, they do not explain a negative trend for opiate users. Further monitoring of this data over time will be beneficial, as well as potentially further primary research at local level to understand why there are positive trends for non-opiate users compared with opiate users.^{[footnote 38]}

7.3 Being supported to sustain a life no longer dependent on drugs

Indicatively, Project ADDER may be contributing towards more service users being supported to sustain a life that is no longer dependent on drugs, but further evidence would be needed to determine whether this is the case

7.3.1 Hypothesis and fidelity

It was expected that Project ADDER activities would lead to more individuals being supported to sustain a life that is no longer dependent on drugs, given that there would be an increase in referrals into treatment, and that more and better-quality support services would be funded.

Project ADDER may be contributing towards an increase in non-opiate entry into treatment and is also likely to have increased treatment entry from the CJS but there was a decrease in treatment entry for opiate patients, compared with the counterfactual, and therefore no observable increase as hypothesised in the ToC. Therefore, it is reasonable to expect to see some evidence of more service users, particularly those with a history of using non-opiates, being supported to sustain a life that is no longer dependent on drugs. This is providing that there is evidence of an increase in the quality of support and the number of service users deriving benefit from treatment as a result of Project ADDER.

To investigate and test this hypothesis, the following evidence was used:

• QED analysis of NDTMS data on failure to complete treatment, successful completion of treatment, deaths in treatment, and self-reported measures of being in work, being in education, and housing need
• qualitative insights from year 1 and year 2 stakeholder and contribution analysis interviews
• qualitative insights from service user interviews

7.3.2 Evidence of outcomes

Stakeholders reported that service users were deriving significant benefits from support and treatment services provided under Project ADDER, as support services were now more data-driven, holistic, intensive and tailored. This included support for service users to break out of the ‘revolving door’– a colloquial term used to refer to the pattern whereby individuals regularly drop out of treatment and support, only to re-present at a later date.

Service users also provided evidence that they were being supported to reduce drug dependency, reporting that their treatment or support service under ADDER has:

• improved their engagement with support and support communities
• given them greater stability as they have access to essential documents, property or equipment required for education, employment and housing
• improved mental and physical health
• improved personal relationships, including re-connecting with their children and families
• improved chances of finding employment, training or volunteering
• coming out of a cycle of exploitation or criminality.

Stakeholders perceived that the benefits of treatment for service users had been enabled or augmented by the additional funding under Project ADDER. Stakeholders reported that the dedicated resources for the implementation of Project ADDER had resulted in improved service user experiences, including the following: reducing the number of individuals a service user needs to interact with; providing specialist new services; increasing the intensity of support provided; and speeding up referral or diversion to appropriate support services.

This was corroborated by service users, who also reported that they had recently seen noticeably greater and faster access to a range of support services. It is therefore reasonable to conclude that the benefits of treatment experienced by service users was facilitated by the improved access to and intensity of supported provided under Project ADDER. However, it is not definitive what impact this had on their drug use.

On balance, NDTMS and Treatment Outcomes Profile (TOP) data does not support the view that more service users are being supported to reduce dependence on drug use (Annex K, Table 1.1). In ADDER areas, the DiD analysis of rates of ‘failure to complete treatment due to drop-out or death’ in NDTMS data demonstrates weak evidence that patients are less likely to drop out of treatment compared to the counterfactual.

However, in Accelerators areas, there is some evidence that patients in treatment are more likely to drop out of treatment and are also less likely to complete treatment, compared to the counterfactual. There is no difference with regards to rates of deaths in treatment, self-reported education and employment outcomes, or self-reported housing needs across Project ADDER, compared with the counterfactual.^{[footnote 39]}

Figure 10: Percentage point difference in the likelihood of patients in Project ADDER areas achieving treatment outcomes, compared with the counterfactual (chart on left shows results for ADDER sites; chart on right shows results for Accelerator sites)

Notes:

1. NDTMS journey level and TOP data. Failure to complete treatment: if the patient dropped out (died or had an unplanned discharge).

Figure 11: Percentage point difference in the likelihood of patients in Project ADDER areas achieving self-reported outcomes, compared with the counterfactual (chart on left shows results for ADDER sites; chart on right shows results for Accelerator sites)

Notes:

1. NDTMS journey level and TOP data. Employment: if the user recorded at least one day in paid or unpaid work in the past 4 weeks; education: if the user recorded at least one day at college or school in the past 4 weeks; homelessness: if the user recorded ‘acute housing problem’ in the past 4 weeks.

Stakeholders reported that that were some external factors that proved a challenge to programme implementation that may have affected the extent to which people using drugs were supported to live a life that is no longer dependent on drugs. These included a lack of control over wider systematic factors, such as housing supply, and prevalent stigma against people using drugs from employers causing barriers to accessing employment. Stakeholders also reported that treatment or support may need to continue for extended periods of time to establish engagement routines and for benefits to be realised. Some stakeholders also reported that engagement with OOCDs had not been as high as hoped, possibly because attendance is not mandated.

Case study: being supported to sustain a life that is no longer dependent on drugs in Blackpool

Stakeholders reported that Project ADDER allowed for the provision of more holistic support to meet the core and interrelated needs of people who use drugs, via collaboration across financial, housing and mental health services. This, combined with a reduction in casework, meant caseworkers could provide more time and attention to individual service users. This reportedly provided them with a better experience of services and facilitated longer-term engagement. The lived experience team was particularly highlighted as playing an important role in being a tangible source of inspiration and increasing aspiration for the service users.

The attentiveness of the caseworkers in Blackpool was consistently noted across the service user interviews. Individuals emphasised the importance of the holistic nature of the support provided within the programme, and they referenced the benefits of the provision of financial, sexual health, and housing advice. There was also mention of the benefits from the support networks that were created among service users within the programme. Here, activities such as boxing, going to the cinema, and poetry were highlighted as positive engagements for the service users.

“They’ve never let me down – I feel prioritised. They drive me all over or offer to get me to places in a taxi. I don’t miss appointments anymore.”

Service user

Case study: being supported to sustain a life that is no longer dependent on drugs in Hackney and Tower Hamlets

In Hackney, Project ADDER has allowed for a greater focus on enabling and encouraging treatment services to actively divert and support individuals into education, training and employment (ETE) activities. Before the programme, recovery workers were mainly focused on medical treatment related to drug usage, with relatively little emphasis placed on how ETE activities and other opportunities that are available could also potentially support the recovery process. Through ADDER, a more collaborative and ‘whole systems’ working practice was adopted, which allowed for treatment services to work directly with ETE providers to divert individuals quickly and easily. Consequently, the quality and impact of discussions with service users was reportedly much greater than before. Stakeholders reported that this also supported more engagement in meaningful ETE activities.

“We had some individuals referred from treatment pathways into learning to play African drum as part of structured learning sessions, which has then led to these individuals becoming more job-ready. This wasn’t happening previously, and we need to do more of this in future.”

Managerial, Treatment

In Tower Hamlets, ETE and IPS projects have reportedly allowed for service users experiencing disruption to establish a routine and structure in their day-to-day life.

“With ETE projects, users tend to have very chaotic lifestyles, but [increased levels of engagement have] led to improved job sustainability for individuals, who are now holding on to a job for more than 6 months or a year. It has really helped their recovery process that they got support, not just in pre-employment stage, but also continued to receive this when they are also in job.”

Managerial, Treatment and diversion

“I felt quite supported because … when [a support worker] asked me about what I want to do for work, it wasn’t just like pushing me onto any given job that was out there, like other providers had done before.”

Service user

7.3.3 Conclusions and future reflections

Project ADDER may be contributing towards individuals being supported to sustain a life that is no longer dependent on drugs, but further evidence would be needed to determine whether this is the case.

Areas experienced some delays to the mobilisation of some critical interventions, but, despite these, activities were largely conducted as intended by year 2 of the programme.

Stakeholder interviews indicated that Project ADDER interventions and activities had improved the intensity and variety of support for service users, which they perceived had improved engagement and led to a greater likelihood of service users deriving benefits from treatment. Service users corroborated this view, mentioning that they had seen a variety of benefits from the support provided, in terms of their physical or mental health; stability; employment, training or volunteering; engagement with support communities; and access to critical documents, property or equipment. Stakeholders and service users considered that Project ADDER had facilitated these treatment benefits through greater access to, and intensity of, support.

The QED analysis did not present any evidence for an impact of Project ADDER on successfully completing treatment, failure to complete treatment due to drop-out or death, being in paid/unpaid work (in the past 28 days), being in education (in the past 28 days), being homeless (in the past 28 days), or death during treatment. Stakeholders reported that treatment or support may need to continue for extended periods of time to establish engagement routines, and for benefits to be realised.

Whilst there was no quantitative evidence of an impact at the programme level, the qualitative evidence does show evidence of improvements in quality of life at the individual level. It is reasonable to assume that these impacts could lead to average improvements across the programme over time. Stakeholders generally felt that living a life that is no longer dependent on drugs is a long-term outcome which requires addressing multiple and complex barriers to realise. Therefore, to measure the impact of Project ADDER, it is recommended that these outcomes are measured again in the longer term, several years after treatment, to observe any long-term effects, and it is recommended that the ToC is updated to show that this outcome is a longer-term impact. It is also recommended that the ToC is amended to show a clearer distinction between increased entries into treatment, and improvements in quality of treatment, which are linked to the level of benefits experienced by service users.

7.4 Prevalence of drug use

There was no observable effect of Project ADDER on drug-related death in year 1. There is not enough evidence to determine whether Project ADDER contributed towards a reduction in drug-related death in year 2.

7.4.1 Hypothesis and fidelity

Project ADDER activities were expected to increase the number of individuals being supported to live a life that is no longer dependent on drugs, and to reduce drug supply to and within local areas, and subsequently to lead to a reduction in the prevalence of drug use in Project ADDER areas.

The evidence suggests that Project ADDER may be supporting more individuals to live a life that is no longer dependent on drugs, and may be reducing drug supply to and within local areas, but further evidence would be needed to determine if this is the case. Therefore, it may not be expected that the programme would have an observable effect on the prevalence of drug use.

To investigate and test the hypothesis, the following data was used to measure the prevalence of drug use in Project ADDER areas:

• QED analysis of NHS data on hospital admissions due to drug poisoning and hospital admissions for drug-related mental and behavioural disorders
• QED analysis of self-reported drug use in NDTMS TOP data

The following data was also used to measure the prevalence of drug use among Project ADDER service users specifically:

• qualitative insights from year 1 and year 2 stakeholder and contribution analysis interviews
• qualitative insights from service user interviews

Since the focus of this outcome was on the prevalence of drug use in the general population in Project ADDER areas, the qualitative evidence on service users’ drug use was considered as contextual evidence for Project ADDER’s impact on drug use. The quantitative area-level evidence was therefore weighted more strongly when drawing conclusions.

7.4.2 Evidence of outcomes

Stakeholders highlighted that early interventions were critical for addressing drug harms, and that many of the Project ADDER interventions are attempting to undo many years of entrenched drug use. Therefore, it may be unrealistic to expect all service users to be completely abstaining from drugs from the outset of treatment, or to anticipate very low rates of relapse. This is supported by analysis of NDTMS TOP data, which does not suggest a significant effect of Project ADDER on self-reported drug use in Project ADDER areas compared to the counterfactual.

Figure 12: Percentage point difference in the likelihood of patients in Project ADDER areas self-reporting drug use, compared with the counterfactual

Notes:

1. NDTMS journey level and TOP data. Drug use: if the user recorded any substance use in the past 4 weeks (excluding alcohol).^{[footnote 40]}

Despite this, there were some anecdotal reports that service users had reduced or eliminated drug use as a result of the support service they had received, with depot buprenorphine-related^{[footnote 41]} interventions particularly highlighted as being effective. Consistent with recent research that confirms the clinical superiority of buprenorphine over methadone, this was corroborated by a small number of service users, who reported that they had either reduced drug use or were no longer using drugs at all. Wider evidence also supports the effectiveness of depot buprenorphine.^{[footnote 42]}

Furthermore, QED analysis indicates that in ADDER areas the programme significantly reduced the number of hospital admissions for drug poisonings as a proxy for drug use in the general population in year 2 (Annex I, Table 1.2).^{[footnote 43]} As illustrated in Figure 14, there is evidence of a reduction compared to the counterfactual in these areas equivalent to 331 fewer admissions (or 37 per 100,000 people) in year 2. The decrease was evidenced in Hastings, Neath Port Talbot and, to an extent, Swansea. There was no significant difference between Accelerators areas and the counterfactual, suggesting that the programme in those areas did not have a significant impact on hospital admissions for drug poisoning. One possible explanation for this could be the relatively short implementation period for Accelerators areas compared to their ADDER counterparts.

Figure 13: Number of hospital admissions for drug poisonings per 100,000 people (chart on left shows results for ADDER sites; chart on right shows results for Accelerator sites)

Notes:

1. Quarterly data from the period January to March 2015 to the period October to December 2022 shared by OHID and NHS Wales from the HES (England) and the PEDW.

However, there was also no significant difference between Project ADDER areas and the counterfactual on rates of hospital admissions for drug-related mental and behavioural disorders in ADDER or Accelerators areas, as illustrated in Figure 14.^{[footnote 44]}

Figure 14: Number of hospital admissions for drug-related mental and behavioural disorders per 100,000 people (chart on left shows results for ADDER sites; chart on right shows results for Accelerator sites)

Notes:

1. Quarterly data from the period January to March 2015 to the period October to December 2022 shared by OHID and NHS Wales from the HES (England) and the PEDW.

Stakeholders noted that levels of supply and demand in local areas are significant drivers in the prevalence of drug use, and also that reducing or eliminating drug use is not always a realistic or useful goal for every service user, particularly those needing ongoing support over several years.

Case study: prevalence of drug use in Swansea

Swansea used ADDER funding to create the Rapid Access Prescribing Service (RAPS) to support more individuals to sustain a life that is no longer dependent on drugs via the offer of depot buprenorphine. This was highlighted by both stakeholders and service users as key to reducing heroin use.

Both stakeholders and service users felt RAPS led to a reduction in heroin use because it eliminated the need for daily pick-ups, enabling service users to make changes to the structure of their lives (such as taking courses or taking up employment), while also helping them avoid other people who use drugs or people looking to exploit them. The prescriptions also stop cravings and painful withdrawal symptoms, and render heroin ineffective if used together.

“The [depot buprenorphine] injection is the only thing really that has made a huge difference and a really positive impact in giving people the space to make changes to the structure of their days.”

Managerial, Treatment

“I didn’t crave heroin and I didn’t look back. It’s the best invention ever. [Methadone and Subutex] are [rubbish] because you can use on top.”

Service user

Furthermore, RAPS took a client-centred approach whereby they would rearrange missed appointments (even if they were missed multiple times) or go out to meet service users in community locations or their home. This helped service users to stay on their prescriptions or get back on them quickly if they had fallen off, therefore reducing illicit drug use.

Swansea also offered a long-term prescribing programme called Recovery+, through which people on statutory orders were given access to an additional 12 months of prescriptions. Stakeholders explained that this intervention reduced the time pressure for service users, allowing more of them to taper off their medication and live a drug-free life.

“We’re having quite a few self-directed detoxes – more than we initially thought, which is good news.”

Managerial, Treatment

Case study: prevalence of drug use in Wakefield

In Wakefield, Accelerator funding was used to expand opportunities for residential rehabilitation, counselling, group support, specialist mental health support, particularly for those with experience of the CJS. Wakefield also prescribed depot buprenorphine to criminal justice clients, co-located IPS employment officers with treatment partners and offered bespoke targeted support to female clients through its partnership with Wakefield’s Well Women Centre.

Though evidence is limited on the impact of treatment on the prevalence of drug use in Wakefield, the area has seen a slight upward trend in opiate users entering treatment.

Accelerator stakeholders felt the depot buprenorphine pilot was a ‘game-changer’ in terms of helping users in their long-term recovery. Stakeholders felt the change in approach (with monthly injections as opposed to more frequent methadone treatments) meant service users had more control over their daily lives and felt less dependent on the treatment method and the need to meet frequent appointments. Stakeholders reported that this had allowed some to spend more time focusing on obtaining and sustaining employment, or rebuilding family relationships.

“[Depot buprenorphine has] provided stability for a huge group of clients in terms of engaging them, keeping them in treatment, and allowing them to start making the changes.”

Managerial, Treatment

The inclusion of specialist support was seen as crucial to sustaining long-term recovery. This focused on wider needs beyond direct drug treatment, including the Well Women Centre (with an emphasis on the needs of women who may have experienced domestic violence), and specialist employment support from IPS workers.

7.4.3 Conclusion and future reflections

There was no observable effect of Project ADDER on reduction in the prevalence of drug use in Project ADDER areas.

Stakeholders anecdotally reported that service users of Project ADDER interventions had reduced or eliminated their drug use, and this was directly corroborated by some service users. Depot buprenorphine interventions in particular were reported as being influential in reducing or eliminating use.

There was a significant reduction in hospital admissions for poisoning due to drug misuse in ADDER areas in year 2, compared to the counterfactual. There is no evidence of an impact in Accelerators areas for hospital admissions for poisoning due to drug misuse. Conversely, there is no evidence across Project ADDER for an effect of the programme on drug-related mental and behavioural disorders. It should be noted that whilst these measures are related to drug use, they are indirect proxies and not an exact measure of the true levels of drug use in the general population. These proxy measures could change irrespective of any change in the prevalence of drug use, because they are a product of the number of individuals who use drugs, as well as the degree of risk associated with their drug use.

Self-reported drug use in NDTMS data is a more direct measure of the prevalence of drug use, noting the limitations posed by social desirability bias. However, there was no evidence of an impact on self-reported drug use.

Stakeholders did report that reducing or eliminating drug use was not always a realistic or useful goal for every service user and that in some cases individuals may need ongoing support over several years for this to begin to overcome the harms caused by entrenched use. Levels of demand and supply for various forms of drugs in local projects are also likely to have influenced results and it was concluded that more evidence would be required to ascertain the impact of Project ADDER on drug supply.

Much research suggests that similar programmes found that self-reported crime fell steeply once people using drugs were admitted to treatment and engagement was sustained. In particular, programmes addressing use of heroin, methadone and benzodiazepines have elicited results after 5 years of follow-up, although relatively less evidence exists for an affect after this timeframe for crack cocaine use. However, there are some more recent studies, with more rigorous methodologies, that have reached more pessimistic conclusions.

7.5 Drug-related deaths

7.5.1 Hypothesis and fidelity

Project ADDER activities were expected to lead to a reduction in the prevalence of drug use, and subsequently to result in a reduction in drug-related deaths.

There was no observable effect of Project ADDER on the prevalence of drug use. Therefore, it is unlikely that the programme would be seen to have the intended effect on drug-related deaths.

To investigate and test the hypothesis, the following evidence was used:

• QED analysis of ONS data on drug-related deaths, where the underlying cause of death is listed as drug abuse, drug dependence and/or drug poisoning by a coroner, for year 1 of Project ADDER; this includes any substances under the Misuse of Drugs Act 1971
• qualitative insights from year 1 and year 2 stakeholder and contribution analysis interviews
• qualitative insights from service user interviews

7.5.2 Evidence of outcomes

Stakeholders were unable to comment on the extent to which Project ADDER had reduced rates of drug-related deaths in the area. Some stakeholders noted that it can take many years of drug use to lead to a drug-related death. Even where the programme is targeting entrenched use, such as 5 years or more, the preventive outcome of avoiding drug-related deaths might not be visible for another 15 years, when it could be ‘expected’ to occur.

This view is supported by QED analysis of ONS data, which revealed no significant difference in the number of drug-related deaths across Project ADDER areas compared to the counterfactual (Annex J, Table 1.4).^{[footnote 45]} It is important to note that most drug-related deaths have to be certified by a coroner. The time taken to hold an inquest causes a significant delay between the date of death and the date of registration (when the data is available for ONS). In 2022 the median registration delay for drug poisoning deaths was around 7 months.^{[footnote 46]} These outcomes therefore include data points up to around one year of delivery in ADDER areas and up to 6 months of delivery in Accelerators areas.

Figure 15: Number of drug-related deaths in Project ADDER areas compared to the counterfactual

Notes:

1. Data provided by ONS from coroners’ reports. Due to small counts and volatility of trends, data was aggregated by semester, from April to September 2016 to October to March 2022.

However, some stakeholders were able to give anecdotal evidence of a small number of cases where Project ADDER may have avoided an instance of drug-related death, largely via the provision or expansion of naloxone. For example, a number of police forces were provided with naloxone and naloxone training, and a few instances were reported of naloxone being used to revive people using drugs from a near-fatal overdose. MI data corroborates stakeholders’ views on the increased use of naloxone, with a 200% increase in the number of naloxone kits distributed in ADDER areas (3,710 kits distributed up to June 2023), and a 42% increase in Accelerators areas over the course of the programme (9,740 kits distributed up to June 2023). One service user also reported how their involvement in Project ADDER support services had directly prevented life-ending circumstances for them.

Case study – drug-related deaths in Middlesbrough

In Middlesbrough, ADDER funding was used to provide police and outreach workers with naloxone, both to carry with them and to distribute in communities. Stakeholders reported that naloxone distribution to the police had been a major success in the area, attributing this to strong partnerships between enforcement and treatment partners. Police carry naloxone kits during their patrols, especially in areas that are hotspots for drug use, and have received training on their use.

For service users who are not ready to access formal treatment services or initiate their recovery journeys, the police also provide needle exchanges. This not only enables access to clean needles but also helps build trust between service users and the police, which ultimately results in easy referrals to relevant treatment services.

“Drug users are seeing that police are carrying naloxone because their lives matter, whereas previously they would have felt that police don’t care because they are drug users. It’s been a real game-changer.”

Managerial, Enforcement

In addition to this, police officers received training about approaching service users in a manner that is more grounded in victim support, rather than punishment. Outreach workers, non-fatal overdose workers, and harm minimisation complemented police initiatives by providing users with efficient access to A&E in cases of overdose. These activities reportedly saved a number of lives, and built rapport with potential service users.

“We have great relationships with the police, who have become our key partner. Absolutely amazing relationship with the chief inspector who has massively supported us on rolling out the naloxone.”

Managerial, Treatment

However, stakeholders noted that opioid users had continued to be hard to engage, as a result of entrenched use.

Case study: drug-related death in Merseyside

A treatment service in Merseyside has used ADDER funding to engage in several different areas of activity to reduce drug use among the treatment population, many of which involved reinstating practice and activities that had been halted in recent years due to lack of funding. Stakeholders believe that these activities have led to a reduction in drug-related deaths among the treatment cohort.

One example is the increased intensity and tailoring of support for individuals who use benzodiazepines. An influx of supply of this particular type of drug had led to a spike in drug-related deaths, with 4 occurring in the area related to benzodiazepine use over a period of just 10 days in 2022. As a result, the treatment service provided more regular face-to-face appointments with individuals who were known to be using this drug, gave them information about the risks, and took measures to try to reduce the risk they were exposed to, such as changing the day they picked up their substitution prescription.

Treatment outreach to hostels and the homeless population successfully engaged more individuals in treatment who were previously resistant to the offer and helped them gain longer-term stability in their lives. This in turn helped them cease, reduce or stabilise their drug use. Treatment services also provided needle exchanges, with one in a hostel dealing with over 2,000 pieces of equipment per month.

Treatment services in Merseyside also increased naloxone distribution and naloxone training to partner organisations. In one particular hostel in Wirral, the rate of fatal overdoses was perceived to have fallen as a result of these harm prevention measures, such as naloxone distribution and training.

“The overdose rate in hostels has dropped dramatically to what it was in the previous year. Naloxone is a life-saving piece of equipment and we have 9 or 10 hostels with kits now.”

Operational, Treatment

Stakeholders reported that the new funding has also provided the capacity for regular close attention, system scrutiny and analysis to be given to near-fatal overdoses, which has provided valuable learning for the team.

The treatment service has also supported individuals to engage with depot buprenorphine as a treatment. Stakeholders and service users reported that depot buprenorphine had enabled individuals to maintain abstinence and stability in their lives.

“Everyone on [depot buprenorphine] has stayed engaged, they haven’t dropped out of treatment … we had one service user who works full time, so he was having to take time off work to collect prescriptions, but with [depot buprenorphine], which is a monthly injection, it reduces the need to take time off work. It’s also less stigmatising as they don’t have to go to the chemist each week.”

Managerial, Treatment

7.5.3 Conclusion and future reflections

There was no observable impact of Project ADDER on drug-related deaths over the course of the evaluation. The impact analysis showed no evidence of an impact on ONS-recorded drug-related deaths in year 1, and, due to a data lag on recorded drug-related deaths, it was not possible to come to a conclusion regarding the impact on drug-related deaths in year 2 of the programme. Stakeholders were unable to comment on the extent to which Project ADDER may have prevented deaths due to drug misuse, particularly at any scale.

However, there are some indications that Project ADDER could go on to have an impact on drug-related deaths in the long term, via a reduction in drug-related poisonings. For example, stakeholders anecdotally reported instances where drug-related deaths were likely to have been averted among individuals receiving treatment or support under the project, most often as a result of the expansion of naloxone programmes.

The literature review highlights that overdose-related deaths are rare events within small areas, and admissions for overdose that have a risk of death will be hard to detect. However, the increased use of naloxone is likely result in better outcomes for such cases.

8. Reflections and recommendations

The final section of this report aims to bring together all the available evidence to provide an overall reflection on the impact of Project ADDER. Whilst it is inevitably challenging to generalise about a programme that was designed to be tailored and flexible in order to meet local area needs, a programme-level perspective is important to contribute to evidence gaps, and to identify suitable recommendations for future policymaking and evaluation at a national level.

8.1 Policy recommendations

Overall, the evaluation found that local projects were able to successfully implement the enforcement, diversion, and treatment and recovery activities over the duration of the initial funding period. A variety of factors, from the outbreak of COVID-19 and pandemic prevention measures to skills shortages in relevant occupations, and local factors such as a history of partnership working, drug demand and service user demographics, combined to cause some challenges and delays. However, by the close of the evaluation, these had largely been overcome.

This report does not go into implementation in detail (that is covered in a separate Project ADDER Evaluation: Practitioner Report); however, it is worth noting that recruitment difficulties were the key challenge for the programme and led to most of the delays in implementation. Delays are likely to have affected the extent to which outcomes were achieved over the course of the evaluation period (by shortening the window for results to materialise). The importance of rebuilding the professional workforce is acknowledged within the government’s 10-year Drugs Strategy and, as such, there is learning from this evaluation that should be addressed by the Home Office and OHID. This includes anticipating that some specialist posts may be hard to fill, tailoring job descriptions to adapt to the existing skills base, anticipating additional delays for accreditation, and/or ring-fencing resources for technical training to bring candidates up to required levels of expertise.

The findings from this evaluation support the mainstreaming of integrated enforcement, treatment and recovery efforts: the WSA. Where we see evidence that outcomes occurred and that Project ADDER made a contribution to their achievement, this tends to be those shorter-term outcomes that are directly related to the activities and interventions delivered. For example, increased arrests and seizures were seen to result from additional intelligence gathering resources (for example, drug testing equipment or digital forensic tools) and better information sharing, and therefore enforcement action that was more frequent, proactive, targeted and efficient than previously.

Similarly, in some areas referrals into treatment from the CJS were enhanced though greater use of DToAs, OOCDs and client-centred practice, resulting in more rapid diversion of service users towards treatment services. This suggests that the integrated approach is working, and that multi-disciplinary teams and partnerships should become business as usual. The Home Office should share learning from the implementation strand of this evaluation to support other areas to embed the WSA. Good practice includes setting up strong partnerships and multi-disciplinary teams; designating funding for ADDER-specific personnel, services and equipment; and establishing a trauma-informed and client-centred approach. The Home Office could also consider extending funding to support local areas in embedding whole systems working for the longer-term.

There is evidence that flexibility in central funding is required to meet local needs and to avoid artificially inhibiting the potential impact of a programme. Local areas welcomed the change to the policy design to allow funding to be spent on preventative interventions: for example, activities aimed at young people. Further programmes are likely to benefit from funding that allows for a greater focus on prevention (drawing on established National Institute for Health and Care Excellence (NICE) and World Health Organization (WHO) guidance on early years-, peer- and family-focused approaches), as well as including interventions relating to alcohol or non-Class A drugs that are often co-used with Class A drugs.

We found less evidence that Project ADDER has achieved its intended results in regard to the other outcomes that lie further along the casual pathway towards impact. It may be that these outcomes need longer to materialise, and broader literature suggests that effects on the prevalence of drug use and drug-related deaths are likely to take an extended period of time to be realised. However, wider evidence also suggests that to achieve these outcomes there is likely to be more work to be done to overcome attrition along the treatment journey. Whilst Project ADDER may have contributed towards an increase in the number of people who use non-opiate drugs referred into treatment and wider support services, entries to treatment had not increased. In addition, there was no positive impact on referrals or entries for opiate patients, which is significant, given the focus of the programme.

Stakeholders reported that the focus of Project ADDER interventions was on improving the quality of treatment, rather than increasing the number of people in treatment. However, to see a change in the measures of interest for the Home Office (that is, an overall reduction in drug use, and in crimes committed by people using drugs) both will need to happen. The evaluation found that the WSA had led to improved treatment capacity and quality (as described in the accompanying Project ADDER Evaluation: Practitioner Report) and, to at least some extent, more referrals being made. Now Home Office and partner departments should focus policy efforts on facilitating those next steps in the causal chain to increase entry and engagement in treatment by those referred from the CJS, particularly entrenched opiate users.

It was suggested that we might see some unintended consequences from Project ADDER, such as displacement of crime (including dealing of drugs) to other areas, and the creation of a ‘postcode lottery’ of treatment, but we observed little evidence of this.

8.2 Evaluation recommendations

Given that Project ADDER has been extended until March 2025, there is value in continuing to evaluate programme impact. In particular to:

1. Allow further time for monitoring impacts on key outcomes, such as drug use and drug deaths. Qualitative evidence from this evaluation, plus a growing wider evidence base, highlights the potential for naloxone and depot buprenorphine to make a significant impact in these areas; therefore, these outcomes may benefit from more evidence being collected over a longer period of time.
2. Explore the trend of divergence in outcomes between opiate use and non-opiate use service users, in line with broader evidence that opiate users typically have more entrenched use and experience relatively greater disruption to their lives, which affects the extent to which they realise benefits from support or treatment. It would be useful to monitor this dynamic going forwards, to understand negative findings on treatment entries, to ascertain the extent or patterns of divergence, and to provide evidence regarding which methods are most effective for achieving beneficial outcomes for this group.
3. Identify alternative/additional metrics for key outcomes: this includes re-offending data (that is, MoJ proven re-offending statistics) and investigating scope for accessing relevant indicators of drug supply (such as analysis of wastewater, or price metrics).^{[footnote 47]}
4. Map the various programmes that are active, or have recently been active, in this policy space, both at a national and local level. This would allow for further interrogation of alternative explanations for any impacts or outcomes observed.
5. Further test and revise the ToC. The outcome relating to ‘people being supported to live a life that is no longer dependent on drugs’ should be reframed to be more clearly related to the benefits associated with improved treatment quality, as distinct from additional treatment entries/completions (covered elsewhere). It was not possible to evidence the reach and impact of communications activities on drug use and drug-related crime, so that element of the Project ADDER ToC remains largely untested. There was also an additional pathway that was identified in qualitative evidence between the experience of support or treatment services and reductions in re-offending/ASB. Such pathways should be tested first through a targeted literature review and then, where relevant, followed up with additional commissioned research.

There is also wider learning for future evaluations of drug (and other) interventions:

1. Involving specialist evaluators as early as possible is essential for robustly evaluating the impact of a policy or programme; early involvement of evaluators may allow for the collection of bespoke baseline data that can go on to demonstrate the effects of a programme against key metrics.
2. Early evaluator involvement is also useful for interrogating the intended theory behind the programme, with local and national stakeholders; this should be visualised in an agreed ToC and causal pathways and mechanisms should be described in detail, to form the basis for research materials and analysis; ToCs should be reviewed regularly to account for developments and iterative changes made to local delivery plans or designs.
3. Create an early dialogue between evaluators, the Home Office and local stakeholders, to co-create monitoring data requirements and the approach to reporting; this would facilitate consistency across and within local areas, ensure it is feasible for practitioners to gather the data requested, and reduce instances of error and reporting burden; this should include the timetable, guidance, training and support, and could even include dedicated funding/posts for data collection and reporting, to ensure its prioritisation by local partners.
4. Integrating QED and rubrics within the design of theory-based analysis in alignment with qualitative work can be useful for conducting rigorous evaluation in highly complex or dynamic policy environments; the approach to contribution analysis applied in this study has involved tailoring the weight we give to these different sources of evidence (that is, QED analysis, qualitative data, MI) to the outcome of focus, rather than applying a blanket approach to say, for example, that QED evidence is always more reliable or robust than other sources; in some cases, the QED analysis draws on secondary data that is directly relevant to the outcome of interest (for example, drug-related death) but in others it may be more of a proxy, there may be data quality issues, and/or the qualitative data or MI might be more directly relevant; for more details on the iterative mixed-method design for this evaluation, please refer to the methodology and Annexes C to E.
5. Additional data gathering approaches might also be useful: for example, social media monitoring; the 2020 Black Review Evidence Pack cites increasing anecdotal evidence that young people are sourcing recreational drugs via social media (predominantly cannabis but also cocaine, ecstasy and Xanax).
6. In seeking to conduct ethical research with groups, such as service users, who may experience significant disruption or vulnerability, a number of steps can be taken to ensure that they are not overburdened by the research process, whilst also facilitating as much participation and insight as is possible. For more details on the specific steps taken in this evaluation please refer to Annex E.

Annex A: Programme-level ToCs

See the programme-level Theory of Change for enforcement and treatment and diversion here: Annex A

Annex B: Local project activities

Blackpool – enforcement

• creation of ADDER taskforce
• analytical work identifying the repeat locations of drugs overdoses/near-fatal overdoses
• analytical work focusing on the vulnerabilities of those involved with county lines in Blackpool
• analytical work focusing on the hierarchy and structure of drug supply within Blackpool
• researching police and partner agencies systems to identify intelligence opportunities
• researching near-fatal overdoses and dissemination to Public Health England
• enhanced activity on financial investigations and money flows, including a focus on seizing proceeds of crimes
• high-visibility policing
• police engagement and support with the young ADDER cohort
• acting on police intelligence to develop the knowledge around the county lines groups operating in Blackpool
• increased police presence via ADDER taskforce
• communications
• utilisation of Drug Dealing Telecommunication Restriction Orders to close county lines and communicate public safety messages to service users
• communicating to Public Health England when high-risk drugs are seized, to facilitate public safety messages

Blackpool – treatment and iversion

• ADDER taskforce carry naloxone
• arm reduction and buprenorphine injections
• physical health management and preventative care
• identification of crack and heroin users who are not in treatment and with a history of criminal activity, through cells, probation, drug-related death and non-fatal overdose panel
• pathways between services strengthened and improved
• outreach/key workers
• assertive outreach and lived experience team engaging with service users throughout treatment journey
• expanding OOCDs

Hastings – enforcement

• creation of ADDER taskforce
• enhanced capacity to analyse seized mobile phones
• enhanced drug analysis capacity and fingerprint DNA analysis
• enhanced crime enforcement analysis targeting criminal drug supply associations, upstream suppliers and organised crime group (OCG) networks implemented
• intelligence-led policing targeting line holders and street-level dealing
• high-visibility policing – directed uniformed patrol activity targeting visible signs of drug harm, including anti-social behaviour
• execution of drug warrants
• cuckooing checks
• social media campaign; out-of-home marketing; design of physical and digital assets

Hastings – treatment and diversion

• creation of a delivery model which encompasses a WSA, providing a holistic package of care
• specialist outreach ADDER team (Seaview) created and pathways to treatment established
• specialist outreach team identify at-risk substance mis-users and engage with them to support into treatment
• care navigation workers establish referral pathways to key services which address the root causes of substance misuse
• Criminal Justice Intervention Team post to work alongside the aged 18 and under treatment provider
• substance heroin and crack keyworkers providing specialist support (assessment, referrals, case management, key working) to substance mis-users
• SWIFT specialist nurse (part of outreach) delivers a consultation and screening clinic to early help staff and service users in local children’s centres on a weekly basis
• partnering with Change, Grow, Live to ensure naloxone is readily available through community partners and Hastings custody
• increased residential rehab places and in-patient detox places
• clinical psychology treatment delivered to patients in rehab
• OOCDs for adults and young people, including DToA expanded to 7 days a week

Middlesbrough – enforcement

• effective investigation strategies, especially for drug arrests and county line; including the use of software like Chorus that enables police officers to conduct basic county line investigation
• coordination between police units and ADDER staff in enforcement; organised crime team had been supporting the arrest of high-harm individuals; the new violence reduction unit also works with ADDER staff to reduce drug-related violence
• the police conducted Operation Arrow – uncovered child exploitation, as well as disrupted operations of a major OCG; the police deployed cover activities like renting different cars
• the intelligence unit also works with prisoners to inform intelligence about county lines, including juveniles, as well as having a dedicated intelligence analyst since year 1 on Project ADDER
• 250 to 300 police officers have been trained on carrying naloxone kits to reduce the risk of overdoses and on distributing clean needles to reduce the risk of blood-borne viruses; increased distribution of naloxone kits and clean needles on streets and in custody
• activities in tandem with the treatment services; DToA; the Introduction of many diversionary schemes, supported by police and senior management (to reduce re-offending)
• Automatic Number Plate Recognition equipment was bought in year 1 and has been used for gathering intelligence
• police also conduct outreach in schools to raise awareness about drugs and drug-related harm through these collaborations
• police patrols for surveillance (operation stay safe), along with neighbourhood policing working with the community
• increased engagement with the community by the neighbourhood policing team
• police identifying people in custody and those committing low-level crimes, to refer them to diversionary services; partnerships have also facilitated the reporting of issues in a timely manner, and sharing intelligence; more intelligence on drug supply, exploitation, and criminal networks
• identifying premises involved in the supply of controlled drugs or associated with ASB and liaising with local authority to close these
• Crimestoppers (independent agency) campaigns related to county line awareness, and encourages people (vulnerable and exploited) to report criminal activity to the police
• get connected scheme: this scheme has been implemented in the hotspot area, whereby agencies conduct various workshops on partnership working and WSA

Middlesbrough – treatment and diversion

• creating a website that provides information to a wider audience, such as open-source information on drug harm, exploitation, early signs of drug misuse and accessing early intervention workers for advice; telephone support team who triages the caller to the right service, such as the housing team, substance use
• information and brief advice training is a part preventative work stream where the trainer recruited provides training to various organisations, such as schools, professionals, voluntary and community sector organisations, family and peer support groups on drug harm, early signs of use, and the remit of treatment services and referral services that can be accessed; trainings are tailored according to the organisation
• assertive outreach workers and collaboration with Barnardo’s workers to identify young people in custody or prisons
• the outreach team distribute cards across localities, pharmacies, GP surgeries, police stations, and primary care to get more referrals from organisations and reach people using drugs through various areas and services; they then refer the users to respective services based on need, like housing, mental health, employment, and recovery rehabs; alongside this they also deliver referrals pathways and support to the families of people using drugs
• individuals who have suffered loss or removal of children are supported through rapid access and are provided with enhanced harm minimisation work to reduce the risk of drug-related deaths
• a lived experience worker conducts fieldwork in custody, prisons and other hotspots in the community to encourage users to get involved in treatment
• a lung health clinic has also been established: it provides primary care to those who smoke drugs, along with health check-ups, lung health advice and chronic obstructive pulmonary disease tests; it also refers people with more severe health issues to smoking cessation pathways; however, top-level services with the NHS specialist have not been implemented; still engaging in dialogue with the trust to revisit
• harm minimisation team has also created a mobile needle exchange programme (additional element), which involves distributing clean needles through a mobile vehicle and a dedicated phone line for people to send in requests for equipment
• the 2 harm minimisation nurses have also been onboarded as a part of outreach to provide primary care to service users in the community and refer them into treatment when necessary; however, they do not have the ability to prescribe medicines and refer people to the NHS foundation instead
• the harm minimisation team has also recruited near/non-fatal overdose workers, who have been working with primary healthcare to identify service users who have had near-miss overdoses, to provide recommendations for treatment as well as to refer them to other relevant services; they are also in the process of developing more pathways for people who visit A&E due to overdose
• the council has also onboarded a charity called Seen Heard Believed, which is delivering more trauma-intensive practices to service users (youth and adults), which include basic counselling and further referrals
• the transformation worker has been delivering services to service users through caseloads and consistent support, along with liaising with mental health services, GP surgeries, hospitals, domestic abuse services, rehabs; the worker handles complex cases such as sex workers, individuals with high levels of drug use, the homeless population, and youth with safeguarding and social care needs
• a nutrition health clinic (for youth and elderly in care homes) was established in partnership with Recovery Connections; they have trained/ upskilled their staff, who are taking forward the initiative
• young people outreach worker goes into custody to get referrals and to provide advice for young people

Norwich – enforcement

• using data and technology to target police activity on the identification of and enforcement against line supply lines into the county
• utilising counter reconnaissance behavioural detection tactics
• deploying police officers to transportation hubs, county lines recruitment areas, drug misuse and drug-related hotspots

Norwich – treatment and diversion

• targeted communications to users
• enhanced outreach using engagement workers from education and training providers and local employers
• voluntary community interventions for young people not already in statutory services or early help
• specialist support for vulnerable women, including assessment, referral and support
• assertive and targeted outreach team engaging with service users throughout treatment journey
• recovery support for people aged 18 to 25
• OOCDs for minor and first offences
• drug testing on arrest or in custody, with referral to Change Grow Live in-reach workers
• engaging children aged 10 to 17 who are arrested and taken to Police Investigation Centre
• court disposals via Community Sentence Treatment Requirements (CSTRs)
• working with other organisations to engage individuals who are not in treatment
• enhanced recovery support
• providing naloxone in police and community hotspots and opiate substitute therapy to those that use illicit heroin
• improving pathways between the criminal justice system and drug treatment

Swansea – enforcement

• intelligence-led targeted activity/visible presence within known hotspot areas
• disrupting organised crime/gang-related criminality by targeting county lines, street gangs and middle market dealers
• proactively pursuing any escalation in threats/violence amongst the wider drug markets
• working closely with other forces, agencies and proactive teams to disrupt criminal activity both locally and nationally
• working with British Transport Police and National Express (or similar) to target the train and road networks
• effective and ethical use of stop and search
• proactively patrolling areas that are most affected by drug criminality
• conducting regular unannounced checks on those subject to cuckooing
• supporting the use of ancillary orders
• purchasing new ICT equipment
• using media to promote the success of the operational teams whilst providing consistent messages to reassure the public
• hiring an analyst and 2 case file builders (admin roles)

Swansea – treatment and diversion

• contributing to community safety through sharing intelligence with partners in order to focus on vulnerability and provide support to build resilience with those at risk
• RAPS
• Post-Criminal Justice Prescribing Maintenance Programme (now called Recovery +)
• Community Mental Health Treatment Requirement (MHTR)
• Service User Involvement Coordinator position (developed peer-to-peer naloxone programme, needle and syringe exchange and dried blood spot testing for blood-borne viruses)
• naloxone carried by police and peer volunteers
• additional Dyfodol worker (x 2) in police custody suites and pre-custody to provide early intervention to those identified (via DToA) as needing support.

Bristol – enforcement

• targeted local enforcement and an increase in targeted drug warranty activity, including interception of illicit drug supply to HMP Bristol
• intelligence gathering, sharing and analysis to inform targeted enforcement activity
• safeguarding and reassurance policing
• DToA (started October to December year 1)
• enhanced partnership and whole systems working
• targeted communications and awareness raising approach to schools, dealers, users and local community
• ASB ADDER workers support property closures and ASB orders where diversions to treatment fail
• ADDER detainee investigators specialise in drug investigations and boost charge/remand rates

Bristol – treatment and diversion

• employment and training support for people using drugs not in structured treatment
• enhanced benefits and employment support for people in recovery
• enhanced outreach and engagement
• enhanced/piloted nasal naloxone provision
• piloted buprenorphine provision
• enhanced psychological therapeutic support
• intensive family support
• respiratory and renal service risk assessment for opiate dependents
• prison in-reach and community working
• drug testing for people on DRRs
• criminal justice substance misuse workers
• joint investigations of drug-related deaths and non-fatal overdoses
• housing support for people using drugs who are not in structured treatment
• trauma-informed workforce training
• naloxone roll-out to police (300 in Bristol)
• deferred prosecution scheme (Call-In) for those involved in drug supply
• working with schools to ensure safeguarding leads have a good understanding of drug misuse and how to refer
• enhanced mentoring offering for young people (and those on remand in prison)

Hackney – enforcement

• supporting drug-related deaths panels and non-fatal overdose partnership investigations
• enhanced partnership working and whole systems working, including through monthly multi-agency meetings
• improved systems to support information sharing, with dedicated staff working alongside local authority and other partners and develop pathways into treatment and out-of-court initiatives
• greater intelligence gathering, sharing and analysis to inform targeted enforcement activity
• targeting the market drivers of drug activity through the use of electric pedal cycles and vehicle car hire, reflecting local organised drug profile
• funded overtime for police investigators and officers to support drug trafficking investigations and reduce drug-related offending through the recovery of greater quantities of drugs and assets
• training for and acquisition of Cell Site Analysis Suite bulk data analytical software to upgrade police ability to analyse organised criminal networks through interrogation of bulk telephony data

Hackney – treatment and diversion

• new forensic/clinical psychologist to help improve and shape health and criminal justice pathways
• enhanced recovery support: provision of legal support services, peer mentoring, and in-reach initiatives
• delivery of enhanced prison in-reach – is progressing but behind on delivery objectives
• psychosocial interventions and key working delivered by new criminal justice recovery workers
• enhanced harm reduction provision, targeting groups that are at risk

Merseyside – enforcement

• increase in targeted drugs warranty activity
• enhanced activity on financial investigations and money flows; drug market profiling, dark web analysis
• increase in use of ANPR, drug wipes and testing kits to disrupt trafficking
• workforce training to increase understanding of services offer and referral pathways
• increase in reassurance policing and safeguarding
• whole systems coordination and collaboration between treatment, diversion and enforcement
• targeted communications approach to schools, dealers and users
• working with Crimestoppers to increase intelligence on drug-related activity
• increased DToA and OOCDs, diversionary support
• procurement of drug testing equipment

Merseyside – treatment and diversion

• enhanced outreach and engagement
• expanding naloxone provision, needle exchange programmes (Knowlsey and Liverpool)
• increased residential rehab (Liverpool) and housing support (Wirral)
• enhanced prison release pathway support
• piloted/enhanced buprenorphine provision
• enhanced diversion from criminal justice pathways
• enhanced offer for dual diagnosis
• improved continuity of care from non-criminal justice settings (for example, hospitals)
• enhanced role of community pharmacists to support health interventions and harm reduction (Wirral)
• physical health clinics (Wirral)
• development of/strengthened recovery community and peer support network (Knowlsey and Liverpool)
• enhanced support for families of people using drugs (Liverpool and Wirral)
• whole systems coordination and collaboration

Newcastle – enforcement

• working with a range of key partners to increase actionable intelligence linked to the harms associated with drugs
• increase in targeted drugs warrants based upon intelligence, data and evidence-led approach
• increased police visibility in high-harm areas linked to harms associated with drugs
• increased proactive use of ANPR
• development of covert proactive tactics to target serious and organised crime
• creation of a detailed partnership drugs market profile
• re-introduction of DToA within custody
• improve education awareness within schools around the harms associated with drugs, alcohol and improved understanding of vulnerability
• enhanced activity linked to financial intelligence and maximising investigative opportunities to identify and seize criminal property
• targeted communications to drug dealers and users
• provision of nasal naloxone to officers

Newcastle – treatment and diversion

• family intervention workers supporting the family and carers of people using drugs – dedicated posts and pathways
• Breaking Free online platform access for people using drugs
• additional focus and capacity for outreach and engagement
• increased buprenorphine provision
• established post to improve pathways into residential rehab
• dedicated CJIT team which included carer/family support, prison interface, probation, courts worker and women’s worker
• creche budget for parents in recovery groups
• psychosocial role – counselling support and other clinical capacity
• OOCD pathway and assessments of DToA cases
• payment to support and expand recovery community, including recovery coaching
• short-term accommodation clients staffing support to encourage engagement in recovery services
• enhanced naloxone provision
• complex needs team for those with co-occurring mental and physical health needs
• a focus on training and workforce development
• physical healthcare team – focusing on outreach, low-threshold interventions, increasing naloxone, BBV testing, physical healthcare checks, smoking cessation
• research report on changes in drug supply and criminal exploitation and serious violence

Tower Hamlets – enforcement

• supporting drug-related deaths panels and non-fatal overdose partnership investigations
• enhanced partnership working and whole systems working, including through monthly multi-agency meetings
• improved systems to support information sharing, with dedicated staff working alongside local authority and other partners and develop pathways into treatment and out-of-court initiatives
• greater intelligence gathering, sharing and analysis to inform targeted enforcement activity
• targeting the market drivers of drug activity and through the use of electric pedal cycles and vehicle car hire, reflecting local organised drug profile
• funded overtime for police investigators and officers to support drug trafficking investigations and reduce drug-related offending through the recovery of greater quantities of drugs and assets
• training for and acquisition of CSAS bulk data analytical software to upgrade police ability to analyse organised criminal networks through interrogation of bulk telephony data

Tower Hamlets – treatment and diversion

• improved key working and case management interventions led by the Community Safety Response Team (CSRT) to provide intensive care to the aged 18 to 24 cohort
• enhanced Recovery Support: New Youth Coach, Case Manager and Sessional meeting/group accommodation to help implement Education, training and employment support and specialist training for people aged 18 to 24 within the treatment system
• delivery of enhanced prison in-reach to support transition from HMP Thameside back into Tower Hamlets, and using Through the Gate Workers to provide support to residents due for release into community services
• improved continuity of care from non-criminal justice settings, ensuring treatment continued between community drug treatment services and other settings (for example, hospitals), led by women’s criminal justice pathway coordinator

Wakefield – enforcement

• procurement of drug testing equipment (DToA)
• creation of Neighbourhood Impact Team
• overt operation focused on disruption and deterrence
• covert intelligence operation targeting wholesale market and drug dealing
• increased use of ANPR for disruption
• digital investigative tool for recovering and analysing iPhone data; detainee visits; analyst for analysing information from police and partners in support of operations

Wakefield – treatment and diversion

• homelessness and accommodation outreach (including training in use of naloxone)
• pilot use of buprenorphine for criminal justice clients
• residential rehabilitation
• specialist mental health support for people using drugs in the criminal justice system
• increased clinical capacity, prescribing, and treatment resources focusing on prison leavers and interventions for criminal justice system clients
• provision of group work interventions to criminal justice cohort to deliver across various settings
• court liaison to assess and encourage appropriate use of Drug Rehabilitation Requirement, including training and education of court staff
• prison in-reach to support release into the community
• increased use of OOCDs

Annex C: Overview of evaluation activities

In the first year of the evaluation, the timeline for Accelerators and ADDER areas differed because programme implementation started later in Accelerators areas compared to ADDER areas. At the outset of the second year of the evaluation, the timelines for evaluating the programme in both ADDER and Accelerators areas were aligned.

Table 13: Activities in year 1 of the evaluation

	1. Project inception	2. Area-level immersion	3. Process evaluation	4. Impact evaluation
ADDER area dates	March 2021 to August 2021	June 2021 to July 2021	August 2021 to January 2022	January 2022 to June 2022
Accelerators area dates	September 2021 to November 2021	November 2021 to December 2021	January 2022 to March 2022	April to June 2022
Evaluation activities	Inception meetings and project inception document agreed with the Home Office. Document and bid review. Stakeholder mapping workshop. Development of programme-level ToC and workshops. Initial development of the evaluation framework.	In-depth interviews with strategic leads. Development of area-level ToCs and workshops. Final evaluation framework delivered.	In-depth interviews with strategic leads and area-level stakeholders. Interim descriptive analysis of MI data.	Interim quasi-experimental analysis of outcomes. Interim contribution analysis and workshops with stakeholders. Year 1 evaluation presentation and report.

Table 14: Activities in year 2 of the evaluation

Stage of the evaluation	5. Reflection	6. Process evaluation	7. Impact evaluation
Dates	July 2022	August 2022 to April 2023	May 2023 to September 2023
Evaluation activities	Review of ToCs and evaluation framework. Stakeholder mapping.	In-depth interviews with area-level stakeholders. In-depth interviews with service users. Analysis of final MI data.	Quasi-experimental analysis of outcomes. Contribution analysis and follow-up in-depth interviews with strategic leads. Year 2 ‘impact story’ workshops with Home Office stakeholders.

Scoping phase and process evaluation methodology

Annex D: Year 1 Scoping and process evaluation

Project scoping and area-level immersion

The process evaluation began with a systematic review of project documentation and delivery plans. Following on from this review, 2 programme-level ToCs – one for ADDER areas and another for Accelerators areas – were developed following a co-creation workshop with Home Office stakeholders. The final ToCs are available in Annex A.

Verian then made initial contact with strategic leads for each area and conducted area-level stakeholder mapping. Typically, each area had 2 strategic leads: one was responsible for treatment and diversion interventions and was likely to be an employee of the relevant council; the other strategic lead was responsible for enforcement interventions and was typically a high-ranking individual within the police. Strategic leads were asked to provide a list of relevant individuals responsible for the implementation of various interventions or activities in the area. Verian reviewed these lists to map out appropriate respondents for the evaluation at an area level.

In-depth interviews were then held with strategic leads in each area. The purpose of these interviews was for the evaluation team to begin developing a relationship with the strategic leads, gain a deeper understanding of the context of each area, the interventions they were running, emerging facilitators and barriers, and the current status of delivery.

Insights from bid documentation and interviews with strategic leads were then used in combination with the programme-level ToC to create draft area-level ToC models. These draft ToCs were then shared with strategic leads in a series of workshops to co-produce final versions.

The full list of local-level activities is available in Annex B.

Year 1 stakeholder interviews

The subsequent case study research included further in-depth interviews with a range of stakeholders identified through the stakeholder mapping process, including senior local authority leads, health and social care representatives, housing and employment services, police and court personnel, and voluntary sector representatives. These stakeholders were a mixture of operational and managerial staff, and efforts were made to ensure that the final stakeholder profile allowed for all activities being run in the local area to be discussed. Strategic leads were also re-interviewed to provide more detailed information on implementation progress, to update any insights from immersion interviews, and to triangulate with information provided by other stakeholders.

Table 15: Number of ADDER areas stakeholder interviews in year 1 of the evaluation

	Blackpool	Hastings	Middlesbrough	Norwich	Swansea
Number of stakeholder and strategic lead interviews	17	19	11	20	12

Table 16: Number of Accelerators areas stakeholder interviews in year 1 of the evaluation

	Bristol	Hackney	Tower Hamlets	Merseyside	Newcastle	Wakefield
Number of stakeholder and strategic lead interviews	18	18	18	33	17	12

These interviews focused mainly on the implementation journey, including how activities were implemented, challenges and facilitators experienced, and what could have been improved. We also used these interactions to probe stakeholders for any anecdotal evidence about emerging outcomes and to regularly review the assumptions about change that the local-level ToC represented.

Due to delays in implementation, year 1 stakeholder interviews took place at an earlier stage in implementation journeys than was originally intended. All interviews were completed by March 2023 and therefore stakeholder experiences that occurred after this time fall outside the scope of the evaluation findings.

Standardisation and analysis

All interviews and workshops across all audiences were conducted based on programme-level discussion guides to ensure consistency across areas. However, each local area lead in the evaluation team was given the discretion to add probes to their discussion guides to address area-specific questions according to the local context. All discussion guides were designed to be used in a flexible manner to follow the cognitive flow of the participant.

All interviews and workshops were transcribed and analysed using an analysis framework to identify key themes and insights.

Annex E: Year 2 Process evaluation

Year 2 stakeholder interviews

Ahead of the year 2 stakeholder interviews, stakeholder mapping conducted in year 1 was reviewed and updated in collaboration with strategic leads in each area. As in year 1, year 2 interviews consisted of in-depth interviews with a range of stakeholders, an ongoing review of project-related MI data submitted by each area, and regular attendance at project board meetings.

These interviews mainly focused on emerging impacts and outcomes of Project ADDER. However, they also covered updates or changes to the implementation experience, and functioned as a review of the local-level ToC. Additionally, these interviews were used to obtain contact details for relevant support workers employed within various interventions, in preparation for service user interviews.

Table 17: Number of ADDER areas stakeholder interviews in year 2 of the evaluation

	Blackpool	Hastings	Middlesbrough	Norwich	Swansea
Number of stakeholder and strategic lead interviews	9	9	12	11	10

Table 18: Number of Accelerators areas stakeholder interviews in year 2 of the evaluation

	Bristol	Hackney	Tower Hamlets	Merseyside	Newcastle	Wakefield
Number of stakeholder and strategic lead interviews	13	10	10	17	12	15

All area-level stakeholders were recommended to Verian by the strategic lead in each area. This increased the risk that stakeholders involved in abandoned or less successful activities were less likely to be referred in to contribute to the evaluation by strategic leads or were less likely to engage with it if they were referred. There were some individuals Verian was unable to get in contact with due to capacity, turnover of staff, or other reasons specific to the individual.

Year 2 service user interviews

In addition to stakeholder interviews, in-depth interviews were also conducted with service users of the interventions. The aim of these interviews was to build a comprehensive picture of project implementation and outcomes by understanding the interventions from a recipient perspective. Service users were asked about how they were referred into support, the nature of the support they were receiving, and how this might be improved in relation to need, access, quality, or other factors. They were also asked to comment on the extent to which the support they were receiving had an impact on their life or circumstances.

Service users were recruited via the support workers for the service they were receiving. Support workers were provided with information sheets and recruitment forms to distribute, and they were asked to collect completed forms and return these to Verian. Service users of only enforcement-related activities, such as DToA, drug warranty, or enhanced enforcement activity, were not approached to contribute to the evaluation. This was due to additional logistical and ethical considerations related to interviewing and incentivising individuals in police custody or prison. However, service users who had been diverted from the CJS and were currently receiving a support service under the treatment strand of the evaluation were eligible for interview.

Verian aimed to achieve between 5 and 12 interviews per area; however, in some areas fewer or no interviews were achieved due to a lack of engagement from service users, or low capacity from support service workers.

Table 19: Number of service user interviews conducted in ADDER areas

	Blackpool	Hastings	Middlesbrough	Norwich	Swansea
Number of service user interviews	9	12	6	4	7

Table 20: Number of Accelerators areas stakeholder interviews in year 1 of the evaluation

	Bristol	Hackney	Tower Hamlets	Merseyside	Newcastle	Wakefield
Number of service user interviews	9	0	2	9	12	0

A number of provisions were put in place for these interviews to ensure ethical and robust conduct throughout, given the sensitive nature of the subject. These included:

• service users were able to select how they wanted to be contacted (by email, by phone, or by text) and interviews were offered in-person, online and via telephone
• all materials were reviewed by Verian’s expert partner, Revolving Doors
• researchers were briefed on using appropriate language and flexible use of research tools in interviews to ensure that service users did not have to speak about things they did not want to speak about
• information about the service user relevant to researcher or service user safety was shared with Verian in advance by support workers
• all service users were offered a break and water in each interview – other refreshments were permitted in the interview if the participant provided them
• researchers were briefed on signs of significant distress, drug or alcohol use, and language that indicated imminent harm to themselves or others; code words were agreed ahead of each interview so that researchers could communicate with each other to flag concern for the safety of a service user
• consent was gathered from service users for support staff to be made aware of any safety or wellbeing concerns, or adverse events
• an established process was set up for researchers to follow if they became concerned for the participants’ safety or wellbeing, or if an adverse event was mentioned, including informing the service users’ support worker and, if necessary, contacting the appropriate emergency services
• support workers were in close proximity, but not present, in the interview room. Service users were able to bring in another support person or animal on request
• a separate process was enacted for interviewing service users aged 17 and under, whereby parental consent was gathered beforehand, and various other measures were instituted in line with the Market Research Society (MRS) children and young people guidelines; more information on this process can be made available on request

For service users, the risk of sample selection bias was present, with the possibility of bias towards service users:

• with positive experiences of support
• experiencing less disruption to their personal lives
• with greater access to technology and communication tools who were using support services
• engaging with staff who were more able to engage with the evaluation

In all areas, engaging service users to participate in the evaluation was challenging. Service users may experience a high complexity of needs and more disruption to their lives – such as homelessness, addiction, or exploitation – as compared to the general public, which may affect their propensity to engage with research. Service users with positive experiences of support could feasibly experience relatively less disruption and complexity in their lives, and therefore be more able to participate. Drop-outs among service users who had originally expressed interest in participating, and non-attendance at arranged interviews, were common. A further bias towards service users with positive experiences of support may have occurred at the referral stage, given that support staff may have chosen not to refer an individual experiencing significant disruption in their lives, or who they thought might give critical feedback on the programme.

Service users also have varying levels of access to technology and communications tools. Verian made pre-emptive efforts to collect service users’ communication preferences, offering telephone calls, texts or emails to arrange interviews. However, not all service users had access to credit for telephone calls and texts, or devices capable of receiving emails. Some service users are also known to share communication devices with others, meaning that Verian researchers had to regularly validate the identity of the service user in order to avoid accidentally revealing personally identifiable information to someone other than the potential participant. This added extra points of friction throughout the recruitment process, increasing the likelihood of drop-out or non-attendance. There may therefore have been a degree of bias towards service users with greater access to technology and communication tools.

Furthermore, the process of recruitment for service users was burdensome for support service workers. This burden was necessary to ensure that interviews were recruited and conducted ethically, and that all data was stored in line with relevant legislations. Significant efforts were made by Verian to limit this burden as much as possible: for example, using one recruitment form per person, providing dedicated assistance to support workers to securely transfer personally identifiable information, arranging for in-person interviews to be conducted on the same day at the same venue where possible, and taking over the recruitment process once a support worker had returned completed forms. However, in some cases the burden for support service workers was such that they were unable to refer in as many service users as was originally hoped, alongside conducting their usual day-to-day work. This could have led to a secondary sample bias for service users, whereby service users being supported by highly burdened staff may have been less likely to be referred into the evaluation compared to others.

Overall, there is a risk that qualitative insights are biased towards the most successful activities or most positive experiences. In order to pre-emptively address this, requested both strategic leads and support workers to refer in as many relevant individuals as possible, including those who may have had less success or less positive experiences, in order to get a holistic view of the programme. However, it is not possible to eliminate these risks entirely, and this should be taken into account when interpreting the evaluation results.

MI data

Monthly MI data submissions were shared by strategic leads in each area with the Home Office, and this data was subsequently passed on to Verian. For the process evaluation, this data was used in the evaluation to inform area-specific probes and questions.

Standardisation and analysis

All interviews and workshops across all audiences were conducted based on programme-level discussion guides to ensure consistency across areas. However, each local area lead in the evaluation team was given the discretion to add probes to their discussion guides to address area-specific questions according to the local context. All discussion guides were designed to be used in a flexible manner to follow the cognitive flow of the participant. Programme-level discussion guides can be provided on request.

All interviews and workshops were transcribed and analysed using an analysis framework to identify key themes and insights.

Impact evaluation methodology

Annex F: Impact evaluation overview

Analytical methods

There were broadly 2 types of model that were used to estimate the impacts of Project ADDER: GSC and DiD. The choice of model depended on the type of data available. This Annex describes the approaches taken for each outcome. Further information about the data sources used, and details of the impact estimates, can be found in Annexes G to M.

For each outcome, the impact is estimated by comparing the outcome observed in Project ADDER areas against an estimate of what that outcome would have been in the absence of Project ADDER (the counterfactual). It is important to recognise that there were other initiatives and activities taking place over the same timeframe which also aimed to influence the same outcomes. To the extent that these activities affected outcomes in different areas equally, this does not pose a problem for the impact analysis. However, it is impossible to rule out the possibility that any apparent difference in outcomes is partly or wholly due to other factors, rather than the effects of Project ADDER.

Each model was fit separately for ADDER areas and Accelerators areas, recognising the fact that the start of Project ADDER was different for the 2 groups. The set of comparison areas was made up of areas which were not part of Project ADDER (neither ADDER areas nor Accelerators areas).

For the purpose of these models, the start of Project ADDER was taken to be October 2020 in ADDER areas and April 2021 in Accelerators areas. Where possible, impacts were estimated across 2 time periods: the first 12 months after the start of Project ADDER, and any subsequent time periods after these 12 months. For example, the models in ADDER areas using PRC data estimated impacts separately for: (i) October 2020 to September 2021, and (ii) October 2021 to December 2022. This is because it was expected that impacts might be smaller in the first 12 months, while activities were being developed and becoming established.

GSC

For many of the outcomes in this report, impacts were estimated using GSC models. This was the case for outcomes related to:

• the number of crimes recorded by police (with the exception of homicides, see below)
• the number of drug trafficking offences recorded by police which resulted in a charge or summons
• the total number of patients entering drug treatment
• the number of patients entering drug treatment through criminal justice referrals
• the number of drug-related hospital admissions

These models compare the trend for a given outcome in Project ADDER areas against an estimate of what would have been observed without Project ADDER. The method works by estimating the counterfactual for all ADDER and Accelerators areas – what would have happened in those areas without the programme – as a weighted average of the outcomes in the areas where Project ADDER was not delivered. The weights were derived such that the weighted trend for an outcome in comparison areas prior to the start of Project ADDER is as close as possible to the trend observed in Project ADDER areas. The intuition is that if the trends are closely matched before the programme started, it is reasonable to attribute any difference after the start of Project ADDER to the effects of the programme. In practice, areas which are not useful for predicting the historic trend prior to the start of the programme are given little weight. In this way, the models do not require the a priori selection of a subset of areas which are considered a suitable match for the treatment area.

As an illustration, Table 21 shows the weights assigned to each of the first 20 comparison areas (alphabetically) in the GSC model to measure the impact of Project ADDER on acquisitive crime in ADDER areas. The model applies these weights to the counts of offences in each area in the post-intervention period (from October 2020 for ADDER areas) and aggregates them to create a counterfactual count of offences that captures what would have happened in the absence of the intervention.

Table 21: Estimated weights of each of the 20 first control group units (alphabetically) for each of the treatment group units (acquisitive crime GSC model)

Comparison area	Blackpool	Hastings	Middlesbrough	Neath Port Talbot	Norwich	Swansea
Adur	-0.57	0.70	-0.08	-0.40	-0.30	0.81
Allerdale	0.14	-0.36	-0.30	0.51	0.16	-0.42
Amber Valley	0.12	-0.20	-0.21	0.36	-0.14	-0.10
Arun	-0.22	0.55	0.08	-0.33	-0.42	0.76
Ashfield	0.01	-0.17	0.08	-0.08	0.03	-0.18
Ashford	-0.16	0.27	0.23	-0.44	-0.32	0.44
Barking and Dagenham	-0.35	0.42	0.21	-0.58	-0.26	0.54
Barnet	-0.52	0.95	0.46	-1.08	-1.01	1.48
Barnsley	0.16	-0.44	-0.04	0.24	0.81	-0.91
Barrow-in-Furness	0.07	-0.12	-0.33	0.47	-0.05	-0.08
Basildon	0.41	-0.45	-0.28	0.73	0.58	-0.76
Bassetlaw	-0.29	0.51	0.18	-0.48	0.08	0.41
Bath and North East Somerset	0.04	0.02	-0.23	0.32	0.06	-0.01
Bedford	0.06	-0.57	0.04	-0.01	-0.90	0.04
Bexley	-0.52	0.77	0.47	-1.09	-1.09	1.37
Birmingham	-0.22	-0.41	0.17	-0.32	0.62	-0.77
Blaby	0.32	-0.49	-0.12	0.46	0.77	-0.92
Blackburn with Darwen	0.26	-0.45	-0.04	0.27	-0.09	-0.35
Blaenau Gwent	-0.08	0.41	-0.02	-0.06	0.03	0.35
Bolsover	-0.05	-0.03	-0.10	0.07	-0.13	0.05

Sensitivity analysis

The GSC model described above is an extension of the classical synthetic control method (SCM). GSC is more efficient when the treated unit consists of several geographic areas (in this case, ADDER and Accelerators areas) because it is able to estimate the treatment effect of multiple areas simultaneously. GSC is also more flexible: SCM restricts the possible weights so that (i) they are all positive, and (ii) the weights add up to one. By relaxing these restrictions, GSC can more closely fit the historic trend in Project ADDER areas, although this can lead to a type of overfitting: models that fit the historic data closely but do not capture the underlying dynamics of this trend well, which can lead to relatively poorer predictions of the counterfactual outcomes.

Therefore, as a sensitivity check, impacts were also estimated using SCM for each outcome for each area. In the vast majority of cases, both estimates (SCM and GSC) were consistent in the sense that the SCM estimate was within the 95% confidence interval for the GSC estimate. However, there were a few instances where the SCM estimate was outside the GSC confidence interval. In these cases, both estimates are reported (the SCM estimate is given in parenthesis) and the results should be interpreted with greater caution.

DiD

For other outcomes, impacts were estimated using DiD models. These compare the change in a given outcome observed in Project ADDER areas against the change observed in areas which were not part of Project ADDER. The change in these comparison areas represents the counterfactual. In other words, the change in these comparison areas is assumed to be the same as the change that would have happened in Project ADDER areas in the absence of Project ADDER.

DiD was used for relatively rare outcomes, where GSC (as described above) would not be reliable, specifically:

• the number of homicides recorded by police
• the number of deaths due to drug misuse
• the number of hospital admissions for assault with a sharp object

In each of these cases, the model can be written as:

    𝑌_𝑎𝑡 = β_0𝑎 + β_1t +β₂𝑃𝑒𝑟𝑖𝑜𝑑_𝑡 * 𝑃𝑟𝑜𝑔𝑟𝑎𝑚𝑚𝑒_a + ε_𝑎𝑡

where 𝑌 is the outcome variable for area a at time point 𝑡, 𝑃𝑒𝑟𝑖𝑜𝑑 indicates whether or not time point 𝑡 was after the start of Project ADDER, and 𝑃𝑟𝑜𝑔𝑟𝑎𝑚𝑚𝑒 indicates whether or not area a is an area in which Project ADDER was implemented. The impact of Project ADDER is estimated as β₂, which represents the extent to which the change observed in ADDER areas (or Accelerators areas) since the start of Project ADDER was different to the change in other areas.

The terms β_0𝑎 and β_1𝑡 represent differences between individual areas and individual time points, respectively. These terms allow the model to account for (i) underlying differences between areas (to the extent that these are consistent over time), and (ii) external factors which affected all areas at specific times. For example, the effects of the COVID-19 pandemic on outcomes are accounted for in the models to the extent to which these effects were consistent across areas. However, one limitation of the models is that they cannot account for factors which have different effects on different areas at different times.

Sensitivity analysis

For each of the 3 outcomes described above, impacts were also estimated using a variation of the model described above: a quasi-Poisson DiD model. This is a very similar model, but it is specifically intended for count data. This was done to address the risk that the models would perform poorly with rare outcomes where for any given time point there may be very few cases recorded in some areas. As the results were very similar, this report presents the results for the simpler linear model described above.

The key assumption of the DiD model is that the change observed in comparison areas after the start of the programme is the change that would have been observed in treatment areas in the absence of Project ADDER. This is often referred to as the ‘parallel trends’ assumption. It is impossible to test this assumption directly. A common test is to compare the trends before treatment starts: if the treatment and comparison groups moved in parallel before the start of the programme, that gives greater confidence that they would have continued to move in parallel afterwards if Project ADDER had not taken place. However, examining the trends prior to the start of Project ADDER, it is clear that the treatment and comparison groups often did not move in parallel in the past. One possibility for this is that some of the comparison areas may not be good points of comparison for ADDER or Accelerators areas. To check this, each of the models was refit numerous times with different subsets of the comparison areas:

• first, taking a single comparison area for each ADDER (or Accelerator) area; these were the comparison areas where the pre-ADDER trend most closely matched that of the treatment areas
• second, taking the 2 most similar comparison areas for each treatment area
• this process was repeated until all possible comparison areas were included in the model

The results were generally similar across these different models, indicating that restricting the analysis to a smaller set of comparison areas would not change the findings. In other words, although there is some risk that the parallel trends assumption does not hold, the results were not sensitive to the choice of which areas should be included in the comparison group.

Quasi-Poisson DiD

An extension of the DiD model described above was used for one outcome: the proportion of drug possession offences resulting in a community resolution. This is because this outcome was a different type of data to the outcomes above: a proportion rather than a count of incidents. Instead, outcomes were estimated using a quasi-Poisson DiD. The model can be written as:

    𝑙𝑜𝑔( μ_𝑎𝑡 ) = 𝑙𝑜𝑔( τ_𝑎𝑡 )+β_0𝑎 + β_1𝑡 + β₂ 𝑃𝑒𝑟𝑖𝑜𝑑_𝑡 * 𝑃𝑟𝑜𝑔𝑟𝑎𝑚𝑚𝑒_𝑎

This is similar to the simpler DiD described above, with 2 main changes:

𝑙𝑜𝑔(μ_𝑎𝑡 ) represents the log of the number of drug possession offences resulting in a community resolution recorded in area a at time point 𝑡; 𝑙𝑜𝑔(τ_𝑎𝑡 ) represents the log of the total number of drug possession offences recorded in area a at time point 𝑡 (regardless of the outcome recorded for that offence). In technical terms, this is known as an ‘offset’.

In practice, the model is estimating the impact of Project ADDER on the rate of drug possession offences which resulted in a community resolution. The total number of offences recorded affects the number of community resolutions that should be expected. This specification accounts for the ways in which the total number of drug possession offences recorded varies over time and between areas.

The impact of Project ADDER is again estimated as β₂, which here represents the extent to which the change observed in ADDER areas (or Accelerators areas) since the start of Project ADDER was different to the change in other areas. The assumptions behind this model are very similar to the assumptions behind the simpler DiD. In particular, the core assumption is that the proportional change in the outcome in comparison areas is the same as what the proportional change would have been in ADDER or Accelerators areas in the absence of the programmes.

Sensitivity analysis

The trends in crime outcomes are closely related to policies and practices within police force areas. On examination of the pre-ADDER data, it was clear that there were large differences between local authorities in different police force areas. For example, the number of drug possession offences resulting in a community resolution was very low in all Merseyside areas before 2018, then gradually increased after this point. This is a different pattern to that observed in other areas.

The original analysis plan was to conduct this analysis using all areas in England and Wales, as was done for the DiD described previously. However, these large differences between police force areas undermine the core assumption that the proportional change observed in comparison areas is the same as the proportional change that would have been observed in treatment areas in the absence of Project ADDER. Therefore, the results presented in this report are based on models restricted to the local authorities within the same police force areas as the treatment areas. Comparison areas in other police force areas are not included.

Entropy balancing with repeated cross-section DiD

The data from OHID (for England) and NHS Wales (for Wales) provides information about the progress and outcomes of individuals’ drug treatment journeys. The outcomes analysed were:

• the likelihood of dropping out of formal drug treatment
• the likelihood of dying during treatment
• the likelihood of being in education in the past 28 days
• the likelihood of being in paid or unpaid work in the past 28 days
• the likelihood of having an acute housing need in the past 28 days
• the likelihood of reporting having used any substance in the past 28 days (excluding alcohol)

One key difference for these outcomes as compared with other outcomes in the evaluation is that data was available for individuals and not just for areas. This allowed a more detailed approach to estimating impacts: specifically, extending a DiD design by using a combination of a matching approach called entropy balancing. In this approach, the entropy balancing method creates a set of weights that balance (matches) the characteristics of individuals in Project ADDER areas and other areas. This is to account for observed differences between the 2 groups. After accounting for these differences, it is more reasonable to think any remaining difference in outcomes is due to the impacts of Project ADDER, rather than being due to differences in the types of individuals in drug treatment in different areas.

A key threat to the validity of designs using repeated cross-sectional data concerns changes in the composition of samples over time. If the profile of people in treatment changes, this could lead to a change in outcomes which is not due to Project ADDER. Therefore, as well as matching between the treatment and comparison groups at each stage, entropy balancing weights were generated to match between time periods. In practice, this means there were 3 stages of matching:

• first, matching journeys in comparison areas before the start of Project ADDER to journeys in Project ADDER areas at the same time
• second, matching journeys in Project ADDER areas after the start of Project ADDER to journeys in the same areas before the start of the programme
• third, matching journeys in comparison areas after the start of Project ADDER to journeys in Project ADDER at the same time

This process is shown below in Figure 16.

Figure 16: Visualisation of the process of matching used for the evaluation of Project ADDER

The entropy balancing approach estimates a set of weights that minimise the difference in individuals’ characteristics between journeys in the 2 groups being matched (see Hainmueller, 2012). Compared to traditional matching methods, such as nearest neighbour propensity score matching and other similar methods, entropy balancing is developed specifically so as to not drop observations. Creating weights generally circumvents a reduction in the sample size, therefore generally increasing the precision of estimates. Furthermore, the approach balances directly on respondents’ characteristics, rather than indirectly via a propensity score (or other pseudo-statistical measures). This means it often provides more convincingly balanced groups (for example, see Black et al., 2022; Matschinger et al., 2020; Hainmueller, 2012).

The following characteristics were used in the matching:

• age (whether aged 30 and above at the start of treatment, or not)
• country (whether living in England or Wales)
• sex (whether male or female)
• ethnicity (whether White or any other ethnicity)
• housing problem (whether any housing problem was recorded at the point of entering treatment)
• employment status (whether employed or unemployed at the point of entering treatment)
• referral source (whether referred by family or friends, or self-referred)
• opiate use (whether the patient reported using any opiates in the 28 days prior to their entry)
• injecting status (whether recorded as having previously injected at the point of entering treatment)

The weights were derived separately for each individual outcome, accounting for different types of missing data in the sample. Table 22 gives an example of the effects of the balancing weights for one outcome. While there are some differences between the group in ADDER areas and the unweighted group in comparison areas, the 2 groups are very similar after the balancing weights are applied.

Table 22: Example of differences in characteristics between ADDER group and comparison group before and after matching (weighting) for sample for outcome related to drop-outs

Variable	ADDER group	Comparison group (unweighted)	Comparison group (weighted)
Age (over 30 years)	0.82	0.82	0.82
Country (England)	0.94	0.86	0.94
Sex (male)	0.64	0.58	0.64
Ethnicity (White)	0.97	0.96	0.97
No housing problem	0.83	0.81	0.83
Employment status	0.53	0.54	0.53
Referral	0.73	0.62	0.73
Opiate patient	0.31	0.38	0.31
Not previously injected	0.76	0.77	0.76

Following estimating the entropy balancing weights, the impacts were estimated with weighted repeated cross-section DiD. For the outcomes related to (i) dropping out of treatment, and (ii) death in treatment, the model can be written as:

    𝑌_𝑖 ω_𝑖 = β₀ + β₁𝑃𝑟𝑜𝑔𝑟𝑎𝑚𝑚𝑒_𝑎 * 𝑃𝑒𝑟𝑖𝑜𝑑_𝑖 + β₂𝑋_𝑖 + μ_𝑎 + ε_𝑖

Where 𝑌_𝑖𝑡 is the outcome 𝑌 for journey 𝑖, and ω_𝑖 represent the entropy balancing weight for journey 𝑖. The impacts of Project ADDER are estimated by β₁, which represents the difference in the change between the pre- and post-ADDER periods between Project ADDER areas and comparison areas. The term μ_𝑎 accounts for differences between areas.

For other treatment outcomes – education, employment, housing, and self-reported substance use – outcomes were recorded at multiple time points: an initial assessment shortly after referral (the entry TOP), with additional assessments every 3 to 6 months, and a final assessment at the point of leaving treatment (exit TOP). The models for these outcomes focused on the change between the entry TOP and the latest recorded outcome for a given treatment journey. The DiD was therefore extended to account for the fact that there were 2 outcome measurements for each journey:

    𝑌_𝑖𝑡 ω_𝑖 = β₀ + β₁𝑃𝑟𝑜𝑔𝑟𝑎𝑚𝑚𝑒_𝑎 * 𝑃𝑒𝑟𝑖𝑜𝑑_𝑖 * 𝑆𝑡𝑎𝑔𝑒_𝑖𝑡+β₂𝑋_𝑖𝑡 + μ_𝑖+ε_𝑖𝑡

Where 𝑆𝑡𝑎𝑔𝑒_𝑖𝑡 is a binary variable identifying whether an outcome was recorded in the latest TOP or the entry TOP. In these models, the term μ_𝑖 accounts for the differences between individual treatment journeys. The impact of Project ADDER was again estimated by β₁.

From the data received from OHID and NHS England it was necessary to restrict the sample of journeys for this analysis. First, in the Welsh data, journeys where treatment outcomes are referred to as ‘referred’ or ‘moved on’ are not included in reported completion or drop-out rates; to ensure consistency between the Welsh and English data, these cases were omitted from the analysis in both countries.

Second, due to inconsistencies in the recording of data, journeys were also excluded where: (i) the date of the assessment of entry outcome measure precedes the start date of a journey; (ii) an individual’s journey does not include an outcome assessments on entry to their drug treatment journey plus at least one other follow-up assessment; (iii) journeys started and ended on the same day (as it was unclear whether for some journeys there are errors in these entries); (iv) a journey is an exact duplicate of another record; (v) there is more than one entry or exit outcome assessment. Additionally, journeys were excluded where data was not available for any one of the characteristics used for matching in Table 22.

MI Data

Monthly MI data submissions were shared by strategic leads in each area with the Home Office, and this data was subsequently passed on to Verian. For the impact evaluation, descriptive-only analysis of this data was used to provide additional context to quasi-experimental data.

MI data metrics utilised within this report include:

• arrests
• cash and assets seizures
• drug seizures
• total number of OOCD disruptions
• arrestees tested for drugs
• total number of individuals offered an OOCD or diversion scheme
• instances of meaningful contact with an outreach team

Key limitations to the use of MI data include the following:

• no baseline data was established, meaning that none of the metrics are able to demonstrate whether the current rates are different to what was occurring before the implementation of Project ADDER
• no adjustment has been made to account for differences in area size or context
• there was inconsistency between and within areas with regards to how different metrics were filled out and defined
• in some cases data for specific areas was missing or not submitted
• some metrics were difficult to record accurately: for example, asset seizures; drug seizures were measured in instances of seizure, rather than weight of drugs seized; for OCG disruptions, there may have been some variability in how each area defined different degrees of disruption, or the length of time disruption persisted^{[footnote 48]}

Contribution Analysis

In year 1 of the evaluation, key causal chains were workshopped with the Home Office and local project strategic and managerial stakeholders. The purpose of these workshops was to test the ToCs and to identify the underlying assumptions and which data or evidence sources were best placed to evidence each assumption. The ToCs were then revised ahead of year 2 of the evaluation.

After the data collection period closed for year 2 of the evaluation, the full ToCs were mapped to identify the key hypotheses for testing. The key hypotheses defined were as follows:

Enforcement:

• targeted communications to dealers, users and young people under Project ADDER would disrupt and prevent drug use, leading to a reduction in ASB (such as street dealing and public use of drugs) and prolific offenders motivated by drugs desisting
• Project ADDER activities related to drug testing capabilities and enforcement activity would result in an increase in arrests of high-harm individuals involved in drug supply, and seizures of drugs, assets and cash
• where Project ADDER activities had resulted in an increase in arrests and seizures of drugs, assets and cash, a reduction in the drug supply would follow
• where Project ADDER resulted in a reduction in drug supply, increased arrests and a reduction in ASB/re-offending, a reduction in drug-related offending would follow

Treatment and diversion:

• using Project ADDER funding to increase DToA, expand the use of OOCDs, and enhance care pathways would increase the number of individuals referred into treatment or support services from criminal justice pathways
• using funding for Project ADDER to increase the provision of pharmacological, psychological and harm reduction programmes, expand treatment capacity, and integrate or improved care pathways would result in an increase in the number of individuals referred into treatment and support
• increasing the number of individuals referred into support services, combined with more and greater quality support interventions and services, would increase the number of people using drugs engaging with support and deriving benefit from doing so; this in turn would mean that more individuals would be supported in sustaining a life that is no longer dependent on drugs
• where Project ADDER increased the number of individuals being supported in sustaining a life that is no longer dependent on drugs, a reduction in the prevalence of drug use would follow
• where Project ADDER reduced the prevalence of drug use, a reduction in drug-related deaths would follow

Verian then worked closely with the Home Office to re-visit the sources of evidence that should be used to test each hypothesis. Thereafter, Verian collated all the evidence from the evaluation to produce a local-level contribution analysis story. Final interviews were then conducted with the strategic leads for each area, during which the local contribution analysis story was shared, and strategic leads were invited to elaborate on alternative explanations or data limitations. Following on from these interviews, local-level contribution analysis stories were revised, and conclusions were drawn according to rubrics that mapped the key contribution analysis questions (see ‘Impact evaluation approach and contribution analysis’) against the 5 possible conclusions.

Within the rubrics, each conclusion for each hypothesis at a local level was given a score of zero to 3, as outlined in Table 23.

Table 23: Rubric scoring by conclusion type

Score	Conclusion
0	Not enough evidence to determine whether Project ADDER contributed towards the outcome
0	There has been no observable effect of Project ADDER on the outcome
1	Project ADDER may be contributing towards the outcome, but further evidence would be needed to determine if this is the case
2	Project ADDER is likely to have contributed towards the outcome, but there are some results or explanations that weaken the strength of the conclusion
3	Project ADDER has contributed towards the achievement of the outcome

Local-level contribution analysis scores for each hypothesis were then averaged to produce an overall programme-level score and conclusion. The programme-level contribution analysis story was presented to key stakeholders at the Home Office and OHID to identify further alternative explanations and to gain additional feedback on the conclusions at a programmatic level.

Annex G: Summary of quasi-experimental data result

Table 24: Summary of quasi-experimental data estimates

Outcome	ADDER areas	Accelerators areas
Enforcement-related outcomes
Drug trafficking offences	Evidence of an increase in drug trafficking offences.	No strong evidence of an impact.
Drug possession offences	No strong evidence of an impact.	No strong evidence of an impact.
Drug trafficking charges	No strong evidence of an impact.	Higher than the counterfactual in year 1 and year 2.
Serious violent crime offences	No strong evidence of an impact.	No strong evidence of impact.
Homicides	Weak evidence of a reduction in homicides.	No strong evidence of an impact.
Acquisitive crime offences	No strong evidence of an impact.	No strong evidence of impact.
Neighbourhood crime offences	No strong evidence of an impact.	Evidence of an increase in recorded neighbourhood crime offences in Accelerators areas in year 1 of implementation.
Hospital admissions for assault with sharp object	No strong evidence of an impact.	Evidence of a reduction in hospital admissions for assault with a sharp object in Accelerators areas in year 2.
Diversion-related outcomes
Community resolutions for drug offences	Weak evidence of a reduction in the proportion of drug possession offences that result in a community resolution.	Evidence of an increase in the proportion of drug possession offences that result in community resolutions.
Criminal justice referrals into drug treatment	Evidence of an increase in CJS referrals into treatment – ADDER England only.	Evidence of an increase in CJS referrals into treatment.
Treatment and health-related outcomes
New opiates patients	Evidence of a reduction in the number of opiate patients entering treatment – ADDER England only.	No strong evidence of a positive impact. Evidence of a negative impact.
New non-opiates patients	No strong evidence of a positive impact.	No strong evidence of a positive impact.
Admissions for poisoning due to drug misuse	Evidence of a reduction in the number of poisoning hospital admissions in year 2.	No strong evidence of an impact.
Admissions for drug-related mental and behavioural disorders	No strong evidence of an impact.	No strong evidence of an impact.
Deaths related to drug misuse	No strong evidence of an impact.	No strong evidence of impact.

Annex H: Quasi-experimental outcomes and data sources

Police recorded crime outcomes

Source

Data is recorded by police forces and then provided to the Home Office monthly. Verian used publicly available data released in May 2023.

A separate dataset was supplied for the ADDER Norwich area as this does not map on exactly to local authority boundaries.

Outcomes

For each lower-tier local authority:

a). The number of drug trafficking offences recorded per 10,000 people. For the purposes of this analysis, drug possession offences include the following offence subgroup: trafficking in controlled drugs (crime code 92A).

b). The number of serious violence offences recorded. For the purposes of this analysis, this includes the following offence subgroups: attempted murder (crime code 2); assault with intent to cause serious harm (crime code 5D).

c). The number of homicides recorded per 10,000 people, consisting of: murder (crime code 1); manslaughter (crime code 4/1); corporate manslaughter (crime code 4/10); infanticide (crime code 4/2).

d). The number of drug possession offences recorded per 10,000 people. For the purposes of this analysis, drug possession offences include the following offence subgroups: possession of controlled drugs (excluding cannabis) (crime code 92D); possession of controlled drugs (cannabis) (crime code 92E); other drug offences (crime code 92C).

e). The number of acquisitive crime offences recorded per 10,000 people. For the purposes of this analysis, acquisitive crime includes the following offence subgroups:

• robbery: robbery of business property (crime code 34A); robbery of personal property (crime code 34B)

• burglary: burglary – residential (crime code 28E, since 2017/18); burglary – business and community (crime code 30C, since 2017/18); domestic burglary (prior to 2017/18); on-domestic burglary (prior to 2017/18)

• theft: theft from the person (crime code 39); bicycle theft (crime code 44); shoplifting (crime code 46); other theft (crime code 49)

• vehicle offences: aggravated vehicle taking (crime code 37/2); theft from a motor vehicle (crime code 45); theft or unauthorised taking of a motor vehicle (crime code 48); interfering with a motor vehicle (crime code 126)

f). The number of neighbourhood crime offences recorded per 10,000 people. For the purposes of this analysis, neighbourhood crime includes the following offence subgroups:

• robbery of personal property (crime code 34B)

• burglary – residential (crime code 28E, since 2017/18); domestic burglary (prior to 2017/18)

• theft from the person (crime code 39)

• vehicle offences: aggravated vehicle taking (crime code 37/2); theft from a motor vehicle (crime code 45); theft or unauthorised taking of a motor vehicle (crime code 48); interfering with a motor vehicle (crime code 126)

Timeframe

The data covers the period from April to June 2015 to October to December 2022. The dataset takes the form of quarterly counts of each crime type recorded within each local authority.

Exclusions

• some records did not have a local authority recorded and these are excluded from the analysis (recorded area ‘unassigned’ or ‘not known’)

• records from Devon and Cornwall Police Force were removed because the police force was not able to provide data for July to September 2022/23

• City of London and City of Westminster were excluded due to the high number of offences recorded and the small population size

• the areas within Broadland and South Norfolk not covered by the ADDER Norwich area were excluded from the analysis

Limitations

• the dataset covers the number of offences recorded by police; this is not the same as the number of offences occurring; an offence might not be recorded if it is not detected by or reported to police; a change in the number of offences recorded could therefore be due to a change in the number of offences occurring and/or a change in the proportion of offences recorded by police

• while the Home Office Counting Rules provide national standards for the recording of offences, the practices of different police forces still vary to some extent; differences between areas in the number of offences recorded could be in part due to the practices of different police forces

• additionally, since 2014, a number of improvements have been made to the reliability and consistency of the recording of offences; further information can be found in the ONS User Guide to Crime Statistics for England and Wales; these changes will have varied between police forces and so a difference in the trends between areas could be in part due to these changes in recording practices

• a small number of quarterly crime counts were negative; according to the Home Office, this may be due to: a police force transferring or cancelling several offences from one period to the next; the transition from police forces recording fraud offences to Action Fraud; or the creation of a back series of data on ‘making off without payment’ offences; these negative values were included in the analysis since they would have been dissipated when aggregating crime subgroups

Patients entering drug treatment and criminal justice referrals into treatment

Source

The datasets for this analysis were supplied by OHID (for England) and NHS Wales (for Wales). For England, the NDTMS is managed by OHID. The NDTMS is a collection of administrative datasets holding information about individuals receiving drug and alcohol treatment in England. For Wales, similar data is collected through the WNDSM.

Outcomes

For each lower-tier local authority:

• the number of new opiates patients (per 100,000 people)

• the number of new non-opiates drug treatment patients (per 100,000 people)

• criminal justice referrals into opiate and non-opiate treatment (per 100,000 people)

Following the usual reporting practice for the NDTMS, patients are counted as an opiate user if they cite use of opiates at any point during treatment, regardless of any other substances they may have used. Non-opiates patients are those who are only recorded as having used non-opiates.

Timeframe

• the dataset for England covers the period January 2015 to March 2023 (monthly)

• the dataset for Wales cover from April to June 2014 to September to December 2022 (quarterly)

Exclusions

• due to incomplete records for 2022 in Wales, the analysis for local authorities in England did not include data for Welsh local authorities; the analysis for local authorities in Wales included both Welsh and English local authorities, although was restricted to a shorter time period, for which more complete records were available

• City of London and Rutland were removed due to missing data in several quarters

Limitations

• data for Norwich corresponds to the Norwich local authority (local authority code E07000148), which is a more limited area than that covered by the Norwich ADDER programme

• in Wales some agencies (including Neath Port Talbot and Swansea) moved to a new reporting system in April 2022, and have not submitted complete data from April 2022 onwards

Hospital admissions

Source

The datasets for this analysis were supplied by OHID (for England) and NHS Wales (for Wales).

• NHS England and NHS Wales record information about hospital episodes. In England, the HES is a database covering all admissions, A&E attendances and outpatient appointments at NHS hospitals in England

• in Wales, the PEDW records information about all in-patient, day case and maternity care cases in NHS trusts in Wales, as well as for Welsh residents treated in NHS trusts in England

• for both England and Wales, data is collected from the hospitals and other medical sites where care is provided; for each case, information about diagnoses is recorded according to the International Statistical Classification of Diseases and Related Health Problems (ICD10)

Outcomes

For each lower-tier local authority:

• the number of hospital admissions where drug-related mental and behavioural disorders were a factor (per 100,000 people)

• the number of hospital admissions for poisoning by drug misuse primary (per 100,000 people)

Timeframe

• the dataset for England covers the period from January to March 2015 to October to December 2022 (quarterly)

• the dataset for Wales covers the period from April to June 2014 to January to March 2023 (quarterly)

Exclusions

• some records did not have a local authority recorded and these were excluded from the analysis (recorded area ‘unknown’, ‘outside England’, ‘outside Wales’, ‘no fixed abode’)

• Isles of Scilly and City of London were missing a number of time periods of data and thus were excluded from the analysis

Limitations

• there are some differences in what information is recorded and how it is recorded between the HES (in England) and PEDW (in Wales); it is not expected that these differences will have a substantial effect on the results of this analysis

• there were some limitations in the format of the data available for England; specifically, all values in England were rounded to the nearest 5, and all values lower than 5 (and greater than zero) were suppressed

Drug-related deaths

Source

• the dataset for this analysis was supplied by ONS. It contains provisional number of deaths related to (i) drug poisoning and (ii) drug misuse by local authority and month of occurrence

• ONS publishes deaths related to drug poisoning annually in August for deaths registered in the previous calendar year; the publication is based on all drug poisoning deaths, and separate figures are provided for drug misuse deaths

• drug poisoning is a broad definition, including accidents, suicides and assaults from drug abuse and drug dependence; the deaths involve a wide spectrum of substances, including legal and illegal drugs, prescription drugs, and over the counter medications; some deaths may also be the result of complications arising from drug abuse (such as deep vein thrombosis or septicaemia resulting from intravenous drug use, or heart disease due to chronic cocaine use), rather than an acute drug overdose

• drug misuse is smaller subset of deaths caused by drug poisoning; for these deaths the underlying cause was drug abuse or drug dependence, or the underlying cause was drug poisoning and the substance is controlled under the Misuse of Drugs Act 1971; drug misuse accounts for around two-thirds of all drug poisoning deaths

• Swansea Bay and Greater Norwich areas were customised to match ADDER geographical areas

Outcomes

The number of deaths related to drug misuse for each local authority.

Timeframe

• the dataset covers the period from January 2016 to December 2022

Exclusions

• due to lags in reporting, provisional counts from April 2022 to December 2022 were substantially lower than for earlier time periods; this will be primarily because deaths over this timeframe have not yet been recorded by coroners; these time periods were therefore excluded from the analysis

Limitations

• it is important to note that most drug-related deaths have to be certified by a coroner; the time taken to hold an inquest causes a significant delay between the date of death and the date of registration (when the data is available for ONS); in 2022 the median registration delay for drug poisoning deaths was around 7 months;^{[footnote 49]} these outcomes therefore include data points up to around one year of delivery in ADDER areas and up to 6 months of delivery in Accelerators areas

• drug misuse figures require additional processing to identify relevant substances recorded on death certificates; presently, ONS only completes this processing once per year for the annual death registration statistical release in August; this means that some deaths due to drug misuse will not have been recorded in the available data

Police recorded crime outcomes

Source

• a dataset was supplied by the Home Office with the number of offences recorded by the police and the outcome of the crime, broken down by Community Safety Partnership and by month

• a separate dataset was supplied for the ADDER Norwich area as this does not map on exactly to local authority boundaries

Outcomes

• the number of drug possession offences resulting in a community resolution (per 100,000 people)

• the number of drug trafficking offences resulting in a charged/summonsed outcome (per 100,000 people)

Timeframe

The data covers the period from April to June 2016 to October to December 2022 (quarterly).

Exclusions

The data received excludes records from Staffordshire, West Midlands, Humberside and British Transport Police. The outcomes for offences recorded in 2022 in these police forces are not up to date due to technical issues.

Limitations

Data at the Community Safety Partnership level has not been quality-assured with police forces.

Hospital admissions for assault with sharp object

Source

The datasets for this analysis were supplied by OHID (for England) and NHS Wales (for Wales). For more detail about this data see ‘Hospital admissions’ section above.

Outcomes

For each local authority (England) and Local Health Board (Wales): the number of hospital admissions for sharp object assault (per 100,000 people).

Timeframe

The data covers the period 2017 to 2022 (calendar year counts).

Exclusions

• all areas with at least one data point supressed (n=211) were excluded from the analysis; the comparison pool for the analysis consisted of 92 areas without suppressed data points

• ADDER and Accelerators areas did not have suppressed data points, except for Hastings; the 2 years where the counts were supressed in Hastings (2018 and 2022) were replaced with 5

Limitations

• in order to protect patient confidentiality, counts of less than 7 are suppressed and replaced with an asterisk in HES data; in the England data, all sub-national counts are rounded to the nearest 5

• hospital admissions for assault with sharp object are rare, hence data was aggregated by OHID and NHS Wales to calendar years to minimise the level of data suppression

• England data from April 2022 onwards are provisional and may be incomplete or contain errors for which no adjustments have yet been made; counts produced from provisional data are likely to be lower than those generated for the same period in the final dataset; it is also probable that clinical data is not complete; there may also be errors due to coding inconsistencies that have not yet been investigated and corrected

• there are some differences in what information is recorded and how it is recorded between the HES (in England) and PEDW (in Wales); it is not expected that these differences will have a substantial effect on the results of this analysis

Drug treatment outcomes

Source

• the datasets for this analysis were supplied by OHID (for England) and NHS Wales (for Wales); 2 datasets were supplied per country: drug treatment journey data; and TOP data

• these datasets were extracted from the NDTMS and the WNDSM

• both datasets of each country were merged using a patient and a journey identifier; each journey had at least one associated TOP; a patient should have a TOP within 14 to 20 days of the start of the journey, a review TOP every 3 to 6 months, and an exit TOP within 14 days of discharge (although this was not consistent in the dataset)

Outcomes

For patients in ADDER and Accelerators areas:

• likelihood of dropping out of treatment (died or unplanned discharge); source: Journey data

• likelihood of successfully completing treatment; source: Journey data

• likelihood of dying in treatment; source: Journey data

• likelihood of being in education in the past 28 days (from the date of the TOP); source: TOP data

• likelihood of being in paid or unpaid work in the past 28 days (from the date of the TOP); source: TOP data

• likelihood of having an acute housing need in the past 28 days (from the date of the TOP); source: TOP data

• likelihood of having used any substance excluding alcohol in the past 28 days (from the date of the TOP); source: TOP data

Timeframe

• English data covers the period January 2016 to March 2023

• Welsh data covers the period April 2018 to March 2023

Exclusions

• journeys where treatment outcomes are referred to as ‘referred’ or ‘moved on’ are excluded

• the analysis was limited to journeys with at least 2 TOP assessments

• a journey where the date of the entry TOP assessment precedes the start date of a journey is excluded

• journeys that started and ended on the same day were omitted (as it was unclear whether for some journeys there are errors in these entries)

• exact duplicates of journeys recorded twice or more in the data were omitted

• journeys where there is more than one entry or exit TOP assessment were omitted

• journeys where characteristic or journey information is not available for matching were omitted

Limitations

• the journeys did not have an ADDER or Accelerators flag; the ‘treatment variable’ is whether the journey/TOP corresponds to a patient in an ADDER/Accelerator area and the journey started after the introduction of ADDER/Accelerators (that is, after October 2020 or April 2021)

• education, employment, housing need and drug use are self-reported and therefore might be subject to social desirability bias and recall bias

• ‘completed treatment’ is not necessarily a positive outcome if the patient needs more treatment; it also leaves out people that are still in treatment (which is not a negative outcome); this is why the likelihood of dropping out of treatment is used as the main measure of success in the impact evaluation

• data inconsistencies exist in some of the variables across English and Welsh datasets

Annex I: Average impact estimates all ADDER and Accelerators areas – GSCG

Annex I tables can be found here: Annex I

Annex J: Average impact estimates for all ADDER and Accelerators areas (DiD models)

Annex J tables can be found here: Annex J

Annex K: Average treatment effect of the treated (ATT) for all ADDER and Accelerators areas (ordinary least squares (OLS) fixed-effects models)

Annex K tables can be found here: Annex K

Annex L: ADDER Area-level post-treatment impact estimates – GSC

Annex L tables can be found here: Annex L

Annex M: Accelerators area-level post-treatment impact estimates – GSC

Annex M tables can be found here: Annex M

Annex N: Drug treatment and criminal justice: selective literature review

Authors: Mike Hough and Bina Bhardwha; ICPR at Birkbeck, University of London

1. Introduction

This Annex has been prepared as a source document for Verian’s final impact report on the evaluation of the ADDER/Accelerator programme. It starts by setting out the policy context and background in England and Wales, including Dame Carol Black’s 2-part Review of Drugs. It then summarises the body of research evidence that has accumulated since the early 1990s on the effectiveness of interventions to tackle the problems associated with heavy illicit drug use. It ends with a section extrapolating from this body of work to assess the potential outcomes for ADDER/Accelerators.

2. The Policy Context

Drug-related offending and drug-driven crime started to emerge as policy issues in the early 1990s, at the time that the first heroin epidemic became well established. Researchers and drug workers were reporting extensive levels of crime committed by individuals who were dependent on heroin, crack cocaine or both. Crimes included shoplifting, burglary, robbery, offences associated with sex work and the offences of drug possession and supply. There were initiatives to increase the level of provision of drug treatment for dependency, as well as ones to increase routes of referral to treatment from CJS agencies. Drawing on UK and international research (Russell, 1994; Hough, 1996), in the 1990s the Home Office set up a Central Drugs Prevention Unit, with a remit to develop drug prevention initiatives. In 1997 this was transferred to the Cabinet Office and renamed the Central Drugs Coordination Unit, which was placed under the leadership of ‘drugs tzar’ Keith Hellawell. The unit published a 10-year drug strategy in 1998.

This unit was effective in promoting drug treatment, and in particular in advocating referral mechanisms within the CJS to treatment agencies. It promoted inter-agency joint working: for example, through the establishment of Drug Action Teams. It also built a network of researchers and practitioners who developed proposals for drug treatment and testing orders (DTTOs), enacted in the Crime and Disorder Act 1998. The institutional structures for tackling illicit drug use were further strengthened with the establishment of the National Treatment Agency for Substance Misuse (NTA) within the NHS in 2001. The programme shifted along the continuum from coerced to compulsory treatment with provisions in the 2003 Drugs and Crime Act.

As a Cabinet Office priority, significant funding was provided for the Drug Prevention Initiative (DPI), reshaped in 2003 into the Drug Intervention Programme (DIP). Police, prison, health services and local authorities were encouraged to set up multi-agency CJITs. The Home Office argued at the time that drug prevention strategies accounted for much of the observed fall in recorded crime.

The encouraging crime trends and the pressures for austerity flowing from the global financial crisis in 2007/08 combined to reduce government commitment to fund the DIP. The NTA was abolished in 2010. The Home Office suspended DIP funding in 2013. Since then, whether as a result of this disinvestment or as a consequence of new patterns of drug distribution, such as aggressive county lines marketing, there has been clear evidence of rising levels of illicit drug use, drug deaths and rising drug-related crime, including systemic violence associated with drug markets.

Dame Carol Black’s 2-phase review, completed in 2021, marked an important return to the drug strategy of the noughties. Her 2 reports chart trends in drug use, drug-related crime and drug markets, and investment in drug services. Overall, the review paints a picture of reversals in a number of positive trends seen in the first decade of the century:

• the use of opiates and crack is increasing significantly

• the use of other drugs by adults and children is also increasing significantly
• drug misuse deaths are reaching a record high
• fewer people that need treatment are receiving it and there are poorer treatment outcomes for those that do receive it
• there is a clear divide between the north and south of the country in levels of problematic use and harm

The review offers a range of explanations for these trends, including more aggressive marketing of drugs, the availability of new drugs, disinvestment in drug treatment services and the erosion of robust arrangements for local inter-agency partnership.

The review’s recommendations, very largely accepted by the government, include rebuilding the drug service workforce, rebuilding treatment services, rebuilding recovery support, and diverting more offenders into treatment services. The review also recommended a WSA to combatting drugs and the setting up of a cross-government coordinating body, which now exists as the Joint Combatting Drugs Unit, located in the Home Office.

3. The research evidence on treatment effectiveness

Since the mid-1990s a significant international evidence base has built up on the outcomes of drug treatment. Here we summarise findings for the UK under the following headings:

• treatment provided by drug agencies and health services
• using the CJS as a referral mechanism into treatment
• court-mandated treatment

3.1 Treatment provided by drug agencies and health services

Overall, the research evidence over 30 years suggests that ‘good-quality’ drug treatment can have a positive impact on reducing drug use, risks of overdose, spread of blood-borne viruses, and can, for some, result in long-term abstinence (for summaries, see, for example, NTA (2006: 7); McSweeney et al. (2008); Department of Health (2017)).

There have been 2 major national studies of treatment effectiveness and cost effectiveness in England and Wales: the National Treatment Outcome Research Study (NTORS) carried out in the late 1990s, and the Drug Treatment Outcome Research Study (DTORS) covering 2007 to 2009. NTORS found that rates of abstinence from illicit drugs increased both in the short term and over a 5 year follow-up. Reductions in drug use were found for heroin, methadone and benzodiazepines, and injecting and sharing of injecting equipment was also reduced. Treatment for users of crack cocaine – who formed a small minority of people using drugs at the time – was less effective. Other benefits included reductions in criminal convictions.

DTORS updated the findings of NTORS, reflecting changes in approaches to treatment, particularly in relation to crack cocaine. Again, the findings were largely positive. Those who were retained in treatment showed reductions in drug use and improvements in social function. Treatment was found to be highly cost effective: for every £1 spent on treatment, there were savings of £2.50, and treatment was cost effective in 80% of cases. A similar follow-up study in Scotland, DORIS, reached broadly similar conclusions.

NTORS, DTORS and DORIS all found that self-reported crime fell steeply once people using drugs were admitted to treatment. Unsurprisingly, retention in treatment was important in securing these benefits, and sustaining the motivation of those in treatment was a critical factor. The ability of drug services to achieve this emerged as highly variable.

Whilst NTORS, DTORS and DORIS are strongly indicative of treatment effectiveness, they did not have experimental designs with experimental and control groups, and thus cannot rule out benefits associated with spontaneous remission. This problem is a feature of much prospective research in this field.

The current treatment picture draws on NDTMS data published by OHID (2023). This tracks individual treatment outcomes from their point of entry to their 6 month review: for example, it demonstrates a reduction in the average number of days clients used opiates. The most recent data shows that at the start of treatment, users reported using, on average, for 23 days in the previous 28 days; this fell to 8.6 days at their 6-month review.

3.2 Using the CJS as a referral mechanism into treatment

From the mid-1990s, with the encouragement of the Central Drugs Coordination Unit, police forces began to experiment with arrest referral schemes as part of the DPI. Evaluation of early pilot schemes found that, following arrest, people using drugs could be cajoled or persuaded into taking up treatment opportunities. Typically, arrest referral workers in police stations screened arrestees, and those that they assessed as in need of treatment were encouraged to attend appointments with appropriate services.

Early evaluations found – albeit with methodological limitations – that those who took up treatment opportunities showed reductions in drug use and in offending to support drug habits (cf. Harvey et al., 2007).

However, more recent research has painted a more mixed picture. For example, Hayhurst et al. (2017) mounted an international systematic review and meta-analysis of community-based diversion programmes for people who use Class A drugs. This concluded that diversion programmes reduced drug use, but that there was little evidence of reductions in associated acquisitive crime. The review noted that studies included in the review were predominantly from North America, where – unlike the UK – most diversion programmes were court-mandated.

A UK study by Pierce et al. (2018) examined a large sample of heroin-dependent offenders, all of whom had been assessed as suitable for treatment. Those who initiated treatment were compared with those who did not, and no evidence was found that those who initiated treatment went on to have lower offending rates than those who did not take up treatment.

However, more positive results have been reported for a Turning Point randomised controlled trial of diversion in the West Midlands, in which those in the treatment group showed reduced offending, relative to the control group (Slothower et al., 2017).

A review by Stevens and colleagues provides a useful conceptual framework for making sense of the complexity of findings in this field (Stevens et al., 2022). In making sense of the variable research on the impact of diversion to treatment, it is important to recognise the considerable attrition that occurs at the point of referral, at the start of treatment and in the course of treatment. Connor et al. (2005) found that only a small percentage of those who tested positive for a Class A drug on arrest subsequently engaged with treatment, but the numbers effectively engaging did increase year on year. The study highlights that DToA can positively impact the numbers being identified, assessed and quickly referred to treatment, including the identification of opiate and crack users who may not have been identified through any other means. However, the study showed that DToA was least likely to engage entrenched opiate users who complied because they ‘had to’.

The implication of the high levels of attrition in getting people who have used drugs and have experience of the CJS into treatment is that expectations about timescales for effects to emerge must be realistic. Only a minority will take up their first referral to assessment, and then accept treatment, and then stay in treatment. But, over time, the numbers actually receiving treatment will increase.

It must also be remembered that the Project ADDER areas are – to varying degrees – rebuilding a treatment workforce. As discussed below, if the CJIT evaluation is a guide, the skills, competences and working practices of teams will initially be variable, and it will take time for best practice to emerge.

The DPI was reshaped into the well-resourced DIP in 2003, with CJITs tasked to provide inter-agency case management approach to dealing with people using drugs from point of entry into the CJS through to ‘aftercare’ and ‘re-settlement’. The Home Office funded a large-scale evaluation of the programme, whose findings are relevant to ADDER/Accelerator.

The main findings of the evaluation were as follows:

• offenders recruited into the study showed significant reductions in both drug use and offending, compared to their patterns of behaviour prior to being taken on as part of the CJIT caseload
• the cost-reducing effect of CJITs on the cost of crimes appeared to be broadly offset by the cost of providing the service
• in the initial stages of delivery of case management by CJITs, there was considerable variation in practices between sites but over the course of the evaluation a level of effective practice was achieved
• in common with preceding schemes (such as arrest referral), CJITs experienced considerable attrition between initial contacts with drug-misusing offenders and successful completion of an appropriate assessment
• in contrast to preceding schemes, CJITs successfully encouraged a very high proportion of those taken on to the CJIT caseload into treatment
• for reasons beyond the control of the evaluators, the sample sizes of offenders in the outcome evaluation were smaller than planned, and as a result the findings were less definitive than expected

The evaluation concluded that further progress in ensuring better coordination between agencies would continue to pay dividends, and that more integrated schemes would be more effective. In particular, the work highlighted that in many areas there was a need for better communication and information sharing between prison drug services (CARATs) and community-based support for drug-misusing offenders.

A significant Home Office analysis (Morgan, 2014) concluded that falls in numbers of people using drugs might account for between a quarter and a third of the fall in acquisitive crime. The report argued that the number of people using drugs fell because the epidemic cohort aged, received treatment, quit drug use, or died.

3.3 Court-mandated treatment

DTTOs were piloted in 1998, rolled out nationally in 2000 and were replaced (in England and Wales, but not in Scotland) by Community Orders with DRRs in 2005. Regular drug testing was/is central to both DTTOs and DRRs. In the English 2003 evaluation, those who completed their order successfully were in a minority. There were high rates of revocation and high re-conviction rates. The evaluation’s researchers suggested that these disappointing results may have reflected the fact that the orders targeted high-risk offenders and that revocation criteria in some sites were inflexible and unforgiving. A parallel Scottish evaluation found lower revocation rates and lower re-conviction rates. DRRs have not yet been evaluated.

More recent research has used MoJ records of offenders on probation sentenced to DRRs between August 2018 and March 2022, linking this to NDTMS data. A recent report (MoJ and OHID, 2023) found around a third engaged with treatment services on the date they were sentenced or after being sentenced. This figure fell to 20% 3 weeks after the sentence date.

Court-mandated treatment illustrates in extreme form the dilemmas involved in routing people using drugs to treatment via the CJS. Early arrest referral systems relied largely on persuasion, albeit within the coercive framework of the criminal process. DTTOs and DRRs impose legally enforceable requirements on people using drugs, who almost by definition suffer from a relapsing condition. It is very difficult to strike the right balance between toleration of relapses and the punishing of non-compliance: trust between drug worker and offender is usually a prerequisite for effective outcomes, and this is rapidly lost when offenders are taken back to court for breach of conditions.

4. Potential Outcomes for ADDER/Accelerators

As described above, in the UK and elsewhere across the globe, the CJS has decades of experience in diverting offenders into treatment services. For reasons discussed above, following the global financial crisis the UK government disinvested in an extensive programme of work with criminally involved people using drugs. Following Dame Carol Black’s recommendations, set out in her (2021) Review of Drugs, the government has returned to this approach. What sorts of benefits can be expected, and within what timescale?

4.1 Reductions in drug-related crime

The past history of UK policy on drug-related crime gives grounds for cautious optimism. The reasons for optimism are that:

• previous investment in similar policies appears to have reduced numbers of criminally involved people using drugs, and - though the evidence is less clear – the volume of drug-related crime
• there is plenty of headroom at present in drawing more people using drugs into treatment as there has been a sharp fall in the number of heroin and crack users entering drug treatment through criminal justice pathways – noticeably, the number of crack users (Black, 2021; PHE and Home Office, 2019)

The reasons for caution are that:

• rebuilding a workforce with the necessary skills will take time, and different areas will progress at different speeds
• in particular, treatment is a human process, with outcomes dependent on the quality of relationships between drug workers and their clientele; maintaining these relationships within the coercive framework of criminal justice creates real dilemmas
• the nature of drug dependence as a relapsing condition means that progress inevitably takes time (“two steps forward and one step back”)
• treatment interventions are just part of a much wider range of social and environmental factors that can influence outcomes (cf. Gossop, 2005)
• a consistent feature of many of the evaluations of diversion schemes has been the high level of attrition at each stage of the process
• the emergence of new drugs, and new forms of drug marketing, could derail progress; the US experience with fentanyl, especially when marketed with cocaine, is a salutary warning
• the links between social deprivation and problem drug use (Black 2020; 2021). This is most recently evidenced by the deepening cost-of-living crisis, which is having an impact on drug use; according to a YouGov poll commissioned by the Forward Trust,^{[footnote 50]} almost a third (32%) of adults reported relapse into addiction or knew someone close to them who had relapsed; anxiety, stress and trauma caused by the cost-of-living crisis was stated as a contributing factor for almost two-thirds (61%) of respondents
• harm reduction interventions (for example, needle exchanges, naloxone distribution) are aimed at injecting opiate users, with few interventions aimed at addressing the needs of crack users (Reuter and Pollack, 2006)

The wider literature, including research from North America, suggests that we are unlikely to observe reductions in the number of people using drugs and levels of drug-related crime over a 2 to 3 year programme. Evidence from similar diversionary programmes has shown mixed success in terms of crime reduction. However, the evidence for the impact of treatment programmes on reducing crime is somewhat more promising. Future crime reduction might therefore depend on the extent to which impact is demonstrated on the treatment side.

Moreover, it is perhaps more useful to move away from a linear notion of cause-and-effect and to consider, for example, the role of social capital, which can aid ‘natural recovery’ without treatment (Granfield and Cloud, 2001), or to deploy a more expansive definition of ‘recovery’, moving away from metrics that are synonymous with a reduction in drug use and offending behaviour (Neale et al., 2014).

4.2 Greater use of community resolutions, less use of charging with possession offences

It should be straightforward to track trends in the use of community resolutions, and one might expect positive effects to emerge in ADDER/Accelerators areas. However, a DiD analysis may not show clear effects, because it would be hard to prevent ‘spillage’ in police support for use of community resolutions into comparison sites.

There is a paucity of research examining the impact of community resolutions. What we do know is that community resolutions can be effective in reducing re-offending^{[footnote 51]} and some descriptive data suggests that community resolutions can offer a positive alternative to criminalisation for young people (Marshall, 2023). However, there is very little published research that has looked specifically at the efficacy of community resolutions for drug possession offences.

One such early evaluation of the use of community resolutions as part of a Drugs Diversion Pilot run by Thames Valley Police demonstrated some positive outcomes (Spyt et al. 2019). Of 55 cases diverted (from December 2018 to March 2019), almost half (25; 45%) attended an initial assessment with a treatment service following referral by the police and, bar 2 (23; 92%), all who attended the initial assessment went on to complete the programme (3 sessions). Comparative data demonstrated that in the absence of the diversion scheme, there was limited or no opportunities to address drug use for those caught in possession of drugs and one in 3 (39%) in the control group went on to reoffend, although it was perhaps too early in the pilot for any firm conclusions to be drawn on re-offending rates.

To account for the absence in the literature of research examining the impact of community resolutions on drug possession offences, it is perhaps more useful to draw on the wider, qualitative literature on police-led diversion schemes and restorative approaches, to provide a better indicator of how effective such measures are. For example, Barberi and Taxman (2019) found that whilst people who used drugs saw diversion schemes positively, they experienced negative attitudes and stigma from police and the CJS more broadly. Police culture was cited as a barrier to deploying alternatives to arrest and diversionary activity. Enthusiasm from police leads did not always filter down to officers, who were often ‘pro-arrest’, which they viewed instrumentally as being beneficial for their careers, and practically as being less time-consuming than diverting people using drugs into treatment.

4.3 Supporting individuals following release from prison

Those newly released from prison often lack aftercare or professional support in the community and are likely to be amongst those people using drugs who suffer from overdose and drug-related deaths (Farrell and Marsden, 2008; Lloyd et al., 2017).

An evaluation of the impact of drug recovery wings in prisons which followed 109 people using drugs in the 6-month period post-release found that many were released without any aftercare provision in place. However, although acquisitive crime was still common, there was evidence of reductions in drug and alcohol use and self-reported offending (Lloyd et al., 2017). The evaluation highlighted the need to rebuild ‘recovery capital’ for people using drugs prior to release from prison, which was seen as pivotal to sustaining reductions in drug use. This is also emphasised by the ACMD’s recommendations regarding ensuring better custody-to-community transitions.^{[footnote 52]}

4.4 Reductions in hospital admissions

Speculatively, we will not see a dip in hospital admissions until enough time has elapsed for real reductions in the number of people using drugs.

4.5 Reductions in overdose-related deaths

As overdose-related deaths are pretty rare events within small areas, admissions for overdoses that have a risk of death will be hard to detect – but the increased use of naloxone will result in better outcomes for such cases. We can confidently say that if naloxone is deployed more widely – by medical practitioners and others – its value in saving lives will be self-evident.

5. Morbidity and mortality-related metrics to gauge Project ADDER impacts

Shrinking the population of people using drugs, as measured by a reduction in hospital admissions for new patients (opiate and non-opiate) and admissions for poisoning due to drug misuse and drug-related mental and behavioural disorders, is likely to be a slow process and contingent on several factors:

• the number of people using drugs accessing and engaging with structured treatment for opiate and crack use
• the availability of harm reduction services which lead to improved health outcomes, including lessening the transmission of disease and related mortality (Strang et al., 2012)
• the availability of aftercare and ‘recovery’ support
• the distribution of naloxone and training for bystander responses
• better integrated care for dual diagnosis
• an ageing cohort of heroin users with multiple and complex physical and mental health needs^{[footnote 53]}
• the supply of potentially more harmful drugs in circulation, such as fentanyl

5.1 Hospital admissions and reducing drug-related deaths

According to Kalk et al. (2018:188), the “Prediction of drug-related death is scientifically very difficult. Such deaths are rare events and multiple factors are at play in a heterogenous group exposed to many risks.” Those who are most likely to experience an overdose are those with an opioid dependency and opioid injectors.^{[footnote 54]} As witnessed in the US and Canada, the risk of overdose has increased in recent years, due to the circulation of drugs laced with high-potency synthetic opioids, such as fentanyl.^{[footnote 55]}

In policy terms, according to the European Monitoring Centre for Drugs and Drug Addiction,^{[footnote 56]} the key approaches for reducing opioid-related deaths include:

• reducing vulnerability: for example, a public health approach; empowering people using drugs to protect themselves; and enabling environments which remove the barriers to service provision and outreach and low-threshold/accessible services
• reducing risk of overdose: for example, overdose awareness; overdose risk assessment in treatment facilities and prisons; and retention in opioid substitution treatment
• reducing fatal outcome of overdose: for example, take-home naloxone programmes for an improved bystander response, and supervised drug consumption

This has been similarly echoed in the UK by the ACMD in their recommendations^{[footnote 57]} to the Home Secretary in 2016, including:

• continued provision and investment in opiate substitution treatment (OST) at optimal dosage and duration
• HAT (drug consumption rooms) in areas with high levels of injecting drug use
• wider provision for naloxone

5.2 OST

OST – in the form of methadone or other opiate substitutes – is “the most effective way of reducing drug-related deaths”, although in the UK there has been a fall in the number of people entering treatment for opiates over the last 10 years.^{[footnote 58]} The potential benefits of OST can, however, be undermined by poor quality of provision (Strang et al., 2012).

Further, Pierce et al. (2016) found that opioid users who received only psychological support were at a heightened risk of overdose in comparison to those who received opioid-agonist pharmacotherapy.

Critically, it has been argued that an over-reliance on experimental evidence with randomised controlled trials, championed as the ‘gold standard’ for evaluating the effectiveness of interventions, could lead to the dismissal of valuable observational, experiential and qualitative evidence (Hough, 2011; Smith and Pell, 2003). For example, despite observational evidence highlighting the potential strengths of introducing overdose prevention centres (OPCs) to reduce drug-related deaths, and no evidence to suggest OPCs increase crime or drug use, their introduction is opposed by the UK government in the absence of experimental evidence.^{[footnote 59]}

5.3 Naloxone

The effectiveness of naloxone (including take-home naloxone) to prevent opioid-related overdose deaths is widely documented in the research literature (Kerensky et al., 2017; Chimber et al., 2018; Nielsen and Van Hout, 2016). There is a growing evidence base that demonstrates the benefits of providing training to opioid users on overdose management and the use of take-home naloxone to prevent overdose events (Bennett and Holloway, 2012; Gaston et al., 2009). A realist review of the available evidence has drawn attention to the need to understand the contextual factors and mechanisms that lead to positive naloxone-based interventions (Miller et al., 2022). Miller et al. (2022) found that training delivered in non-judgemental harm reduction settings where there is in-group acceptance of people who use drugs can promote willingness to engage in a bystander response. The study also demonstrated the effectiveness of peer-to-peer models of naloxone training.

6. Literature review references

Barberi, D., & Taxman, F. S. (2019). ‘Diversion and Alternatives to Arrest: A Qualitative Understanding of Police and Substance Users’ Perspective’, Journal of Drug Issues, 49(4), 703 to 717.

Bennett, T. & Holloway, K. (2012). ‘The impact of take-home naloxone distribution and training on opiate overdose knowledge and response: An evaluation of the THN Project in Wales’, Drugs: Education, Prevention and Policy, 19(4), 320 to 328.

Black, C. (2020). ‘Review of drugs executive summary’. London. Available: Microsoft Word - SummaryPhaseOne+ foreword 200219 (publishing.service.gov.uk).

Black, C. (2021). ‘Review of drugs part two: Prevention, treatment, and recovery’. London. Available at: Review of drugs part two: prevention, treatment, and recovery - GOV.UK (www.gov.uk).

Chimbar, L., Moleta, Y. (2018) ‘Naloxone Effectiveness: A Systematic Review’, Journal of Addictions Nursing, 29(3), 167 to 171, 7/9.

Clinical Guidelines on Drug Misuse and Dependence Update 2017 Independent Expert Working Group (2017) ‘Drug misuse and dependence: UK guidelines on clinical management’. London: Department of Health.

Connor, M., Green, G., Thomas, N., Sondhi, A., & Pevalin, D. (2020) ‘Drug testing on arrest-who benefits?’ Health & Justice, 8(1), 3.

Farrell, M. and Marsden, J. (2008). ‘Acute risk of drug-related death among newly released prisoners in England and Wales’. Addiction, 103 (2), 251 to 5.

Gaston, R. L., Best, D., Manning, V., & Day, E. (2009). ‘Can we prevent drug related deaths by training opioid users to recognise and manage overdoses?’, Harm Reduction Journal, 6, 26.

Granfield, R. and Cloud, W. (2001) ‘Social context and ‘‘natural recovery’’: The role of social capital in the resolution of drug-associated problems’, Substance Use & Misuse, 36(11), 1543 to 1570.

Harvey, E., Shakeshaft, A., Hetherington, K., Sannibale, C., & Mattick, R. (2007). ‘The efficacy of diversion and aftercare strategies for adult drug-involved offenders: A summary and methodological review of the outcome literature’, Drug and Alcohol Review, 26, 379 to 387.

Hayhurst , K . P., Leitner, M., Davies, L., Millar, T., Jones, A., Flentje, R., Hickman, M., Seena, F., Mayet, S., King, C., Senior, J., Lennox, C., Gold, R., Buck, D., & Shaw, J. (2019). ‘The effectiveness of diversion programmes for offenders using Class A drugs: a systematic review and meta-analysis’. Drugs: Education, Prevention and Policy, 26(2), 113 to 124. https://doi.org/10.1080/09687637.2017.1398715 .

Hough, M. (2011) ‘Criminology and the role of experimental research’, in C. Hoyle and M. Bosworth (eds.) What is Criminology? Oxford: Oxford University Press.

Kalk, N.J., Robertson, J.R., Kidd, B. et al. 2017. ‘Treatment and Intervention for Opiate Dependence in the United Kingdom: Lessons from Triumph and Failure’, Eur J Crim Policy Res, 24, 183 to 200.

Kerensky, T., Walley, A.Y. (2017). ‘Opioid overdose prevention and naloxone rescue kits: what we know and what we don’t know’, Addict Sci Clin Pract 12, 4.

Lloyd, C., Page, G. W., McKeganey, N., Russell, C., & Liebling, A. (2017). ‘The evaluation of the drug recovery wing pilots: Final report’. Policy Research Programme, Department of Health.

Marshall et al. (2023) ‘The implementation and delivery of community resolutions: the role of youth offending services’ (justiceinspectorates.gov.uk).

McSweeney, T., and Turnbull, P. J. & Hough, M. (2008) ‘The treatment and supervision of drug-dependent offenders: a review of the literature prepared for the UK Drug Policy Commission. Project Report’. London, UK: Institute for Criminal Policy Research.

Miller, N. M., Waterhouse-Bradley, B., Campbell, C., & Shorter, G. W. (2022). ‘How do naloxone-based interventions work to reduce overdose deaths: a realist review’, Harm Reduction Journal, 19(1), 18.

MoJ and OHID (2023) ‘Pathways between probation and addiction treatment in England: Report’. London: Ministry of Justice and Office for Health Improvement & Disparities. Pathways between probation and addiction treatment in England: report - GOV.UK (www.gov.uk).

Morgan, N. (2014) ‘The heroin epidemic of the 1980s and 1990s and its effect on crime trends - then and now: Technical report’. London: Home Office.

Neale, J., Finch, E., Marsden, J., Mitcheson, L., Rose, D., Strang, J., Tompkins, C., Wheeler, C., & Wykes, T. (2014). ‘How should we measure addiction recovery? Analysis of service provider perspectives using online Delphi groups’ Drugs: Education, Prevention and Policy, 21(4), 310 to 323.

Nielsen, S. and Van Hout, M. (2016) ‘What is known about community pharmacy supply of naloxone? A scoping review’, International Journal of Drug Policy, 32, 24 to 33.

NTA (2006) ‘Models of care for treatment of adult drug misusers: Update 2006’. London: National Treatment Agency for Substance Misuse.

OHID (2023) ‘Adult substance misuse treatment statistics 2021 to 2022: report’. London: Office for Health Improvement & Disparities. Adult substance misuse treatment statistics 2021 to 2022: report - GOV.UK (www.gov.uk).

Pierce, M., Bird, S.M., Hickman, M., Marsden, J., Dunn, G., Seddon, T., Millar, T.(2018) ‘Effect of initiating drug treatment on the risk of drug-related poisoning death and acquisitive crime among offending heroin users’, International Journal of Drug Policy 51, 42–51. doi: 10.1016/j.drugpo.2017.09.017.

PHE and Home Office (2019) ‘Research and analysis: Increase in crack cocaine use inquiry: summary of findings’..

Slothower, M., Neyroud, P., Hobday, J., Sherman, L., Ariel, B., Neyroud, E. & Barnes, G. (2017) ‘Presentation. The Turning Point Project: Implementation and 2-year Outcomes’. West Midlands Police and University of Cambridge.

Smith, C. S. & Pell, J. P. (2003) ‘Parachute use to prevent death and major trauma related to gravitational challenge: Systematic review of randomised controlled trials’, British Medical Journal, 327, 1459 to 1461.

Spyt, W., Barnham, L., & Kew, J. (2019). ‘Diversion: Going soft on drugs?’ Thames Valley Police Journal, 4, 44 to 58.

Strang, J., Babor, T., Caulkins, J., Fischer, B., Foxcroft, D. & Humphreys, K., (2012) ‘Drug policy and the public good: evidence for effective interventions’, The Lancet, 379(9810), 71 to 83.

Stevens, A., Hughes, C. E., Hulme, S., & Cassidy, R. (2022). ‘Depenalization, diversion and decriminalization: A realist review and programme theory of alternatives to criminalization for simple drug possession’, European Journal of Criminology, 19(1), 29 to 54. https://doi.org/10.1177/1477370819887514.

Reuter, P. and Pollack, H. (2006) ‘How much can treatment reduce national drug problems?’ Addiction, 101, 341 to 347.

1. For more information about cuckooing, see: https://paceuk.info/criminal-exploitation/county-lines-slang/. ↩

2. For more information about ‘trigger offences’, see: https://www.gov.uk/government/news/drug-testing-on-arrest-expanded-to-help-cut-crime. ↩

3. For more information about IPS, see: https://www.centreformentalhealth.org.uk/what-ips. ↩

4. For more information about middle market targets, see: http://eprints.lse.ac.uk/13878/1/Middle_market_drug_distribution.pdf. ↩

5. For more information about organised crime groups, see: Serious Crime Act 2015. ↩

6. Whilst definitions of ‘high harm’ varied by area, this term was usually used to refer to individuals involved in the higher ranks of drug trafficking organisations, or those who commit extreme violence. ↩

7. The full Dame Carol Black Review: https://www.gov.uk/government/collections/independent-review-of-drugs-by-professor-dame-carol-black. ↩

8. Government response to the independent review of drugs by Dame Carol Black - GOV.UK. ↩

9. For more information on Dame Carol Black’s recommendations in the report ‘From harm to hope: A 10-year drugs plan to cut crime and save lives’ see: https://www.gov.uk/government/publications/from-harm-to-hope-a-10-year-drugs-plan-to-cut-crime-and-save-lives . ↩

10. Formerly Public Health England. ↩

11. https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/1079147/From_harm_to_hope_PDF.pdf. ↩

12. For the full Magenta Book, see: https://www.gov.uk/government/publications/the-magenta-book and for the full Green Book, see: https://www.gov.uk/government/publications/the-green-book-appraisal-and-evaluation-in-central-governent. ↩

13. https://www.3ieimpact.org/sites/default/files/2017-11/Working_Paper_3.pdf. ↩

14. In 2009, Howard White set out 6 principles for successful application of theory-based approaches. The full working paper is available online: https://www.3ieimpact.org/sites/default/files/2017-11/Working_Paper_3.pdf. ↩

15. The Global Evaluation Initiative has set out 6 steps for producing a credible contribution analysis story. The full report is available online: https://www.betterevaluation.org/methods-approaches/approaches/contribution-analysis. ↩

16. In the initial analysis phase, further conclusions were considered to examine the potential for the programme to cause negative effects. These conclusions were discounted in later phases as the available evidence did not support them. ↩

17. For further details please refer to the separate Pactitioner report. ↩

18. In order to interpret this data it is important to note that the data does not distinguish between ‘high-harm’ and ‘low-level’ individuals arrested. Furthermore, one quarter of data was not available for Hackney and Tower Hamlets. ↩

19. In order to interpret this data it is important to note that seizures of assets and cash were recorded in ‘instances’ of seizure rather than by weight or value. ↩

20. Police officer uplift, England and Wales, quarterly update to 31 March 2023 - GOV.UK (www.gov.uk). ↩

21. Police officer uplift, England and Wales, quarterly update to 31 March 2023 - GOV.UK (www.gov.uk). ↩

22. Operation Orochi and Project Medus are part of the Metropolitan Police Taskforce’s County Lines Programme. More information at: County Lines Programme overview - GOV.UK. ↩

23. A disruption occurs when the Police impacts an OCG or an individual involved in serious organised crime, resulting in them being unable to operate at their usual level of criminality. ↩

24. OCG disruptions are law enforcement activities carried out against OCGs, here limited to drugs OCGs. They are categorised as either major, moderate or minor, depending on their impact (significance and longevity). A disruption can include events such as a cash seizure or the conviction of an individual. The classification of each disruption is assessed after the fact and is subject to a periodic moderation process. ↩

25. Defined as having a minimal and/or short-term disruptive impact on the capability of an OCG, individual or vulnerability. ↩

26. Serious violent crime is a combination of 2 crime groups: attempted murder and assault with intent to cause serious harm. ↩

27. Wider evidence indicates that there is a link between drugs and acquisitive crime (Understanding organised crime: estimating the scale and the social and economic costs (publishing.service.gov.uk) Table A3.2), serious violence (Home Office – Serious Violence Strategy, April 2018 (publishing.service.gov.uk)), and homicides (Homicide in England and Wales - Office for National Statistics). ↩

28. Neighbourhood crime is a specific subset of acquisitive crime that includes the following sub-groups: robbery of personal property, domestic and residential burglary, theft from the person, and vehicle offences. ↩

29. Proven reoffending statistics: October to December 2021 - GOV.UK (www.gov.uk). ↩

30. It is not possible to disaggregate conditional cautions from simple cautions in PRC data, so community resolutions were the sole PRC diversion metric focused on. ↩

31. Data aggregated from Hastings and Norwich only. No data available for other ADDER areas. ↩

32. Data aggregated from Bristol, Met Police, Knowsley and Liverpool. No data available for other Accelerators areas. ↩

33. In order to interpret this data it is important to note that this data does not provide details on the outcomes of DToAs or OOCDs. ↩

34. Since 2022, police forces have been moving to a new framework for delivering OOCDs. The new approach is intended to create more meaningful and proportionate consequences: in essence, moving away from warnings and simple cautions. In relation to drug offences, this involves greater use of conditional cautions and community resolutions, while phasing out cannabis warnings and simple cautions. See [Reforms to the adult out of court disposals framework in the Police, Crime, Sentencing and Courts Bill: Equalities Impact Assessment - GOV.UK. ↩

35. Data for the full time period was not available in Swansea due to errors in reporting, and these results have a high degree of uncertainty, which means strong conclusions cannot be drawn for this area. ↩

36. It was not possible to develop impact estimates for Swansea due to incomplete records in Welsh data for 2022. ↩

37. In order to interpret this data it is important to note that no specific definition of ‘meaningful contact’ was given. It would be possible for an individual receiving more than one support service to be double-counted, as ‘contacts’ were recorded separately for each outreach activity. ↩

38. The data was not available in time to allow this analysis. ↩

39. This data is self-reported by service users on sensitive subjects, and is likely to be subject to significant response and social desirability bias. ↩

40. Substances in the questionnaire: opiates, crack, cocaine, amphetamines, cannabis. ↩

41. Depot buprenorphine products (for example, Buvidal and Sublocade) are injected weekly or monthly for the treatment of opioid dependence. ↩

42. See, for example, Marsden, J., Kelleher, K., Gilvarry, E., Mitcheson, L., Bisla, J., Cape, A. et al. (2023) Superiority and cost-effectiveness of monthly extended-release buprenorphine versus daily standard of care medication: a pragmatic, parallel-group, open-label, multicentre, randomised, controlled, phase 3 trial, eClinicalMedicine, Volume 66, 102311. ↩

43. There are some limitations in using NHS data to measure the impact of Project ADDER. For example, there are differences in how hospital admissions are recorded in England in Wales, and also some differences in how data is rounded in these datasets. ↩

44. There are some limitations in using NHS data to measure the impact of Project ADDER. For example, there are differences in how hospital admissions are recorded in England in Wales, and also some differences in how data is rounded in these datasets. ↩

45. Due to small counts and volatility of trends, data was aggregated by semester, from April to September 2016 to October to March 2022. Data from April to December 2022 was removed due to low counts (incomplete records). ↩

46. See Deaths related to drug poisoning in England and Wales - Office for National Statistics (ons.gov.uk). ↩

47. The other outcome where quantitative data was lacking was ASB, but this is unlikely to be useful as a source of impact data given that ASB is not necessarily related to drugs, and much ASB is not recorded. ↩

48. The definitions provided by the Home Office as guidance for data returns included the following definitions: major: a significant and/or long-term disruptive impact on the capability of the OCG, individual or vulnerability; moderate: a noticeable and/or medium-term disruptive impact on the capability of the OCG, individual or vulnerability; minor: a minimal and/or short-term disruptive impact on the capability of the OCG, individual or vulnerability. ↩

49. Deaths related to drug poisoning in England and Wales - Office for National Statistics (ons.gov.uk). ↩

50. Taking Action On Addiction - A campaign website to bring more understanding and help reduce the stigma surrounding addiction. ↩

51. Does Community Resolution reduce reoffending? (birmingham.ac.uk) [Accessed: 24/07/2023]. ↩

52. https://www.gov.uk/government/publications/acmd-report-custody-community-transitions [Accessed 24 July 2023]. ↩

53. Ageing_cohort_of_drug_users.pdf (publishing.service.gov.uk) [Accessed 24 July 2023]. ↩

54. Prevention of drug-related deaths – topic overview www.emcdda.europa.eu [Accessed: 24 July 2023]. ↩

55. Prevention of drug-related deaths – topic overview www.emcdda.europa.eu [Accessed: 24 July 2023]. ↩

56. Ibid. ↩

57. ACMD-Drug-Related-Deaths-Report-161212.pdf (publishing.service.gov.uk) [Accessed: 24 July 2023]. ↩

58. Another New High for Drug Deaths in England & Wales – Here’s What Needs To Change (talkingdrugs.org) [Accessed: 24 July 2023]. ↩

59. Overdose prevention centres in the UK (thelancet.com) [Accessed: 24 January 2023]. ↩