Research and analysis

MHRA data requirements to support regulatory decision making

Published 2 July 2026

Introduction

The Medicines and Healthcare products Regulatory Agency (MHRA) is the UK regulator for medicines, medical devices and blood components for transfusion. Its primary responsibility is to ensure that these products work as intended and are acceptably safe. The agency’s remit spans the entire product lifecycle - from authorisation and licensing through manufacturing and supply oversight to post-market monitoring of medicines and medical devices.

As part of its statutory duties, the MHRA routinely conducts research on the safety of medical products to inform regulatory actions, such as variations to marketing authorisations or the introduction of additional risk minimisation measures. To do this effectively, the agency requires access to real-world data (RWD) that capture safety outcomes across the UK population. Unlike academic research, which is often curiosity-driven, the MHRA’s research activities are focused on addressing evidence gaps to inform regulatory decision-making. This difference in purpose translates into distinct requirements for data quality. In curiosity-driven contexts, data limitations may introduce acceptable uncertainty, if caveats are clearly acknowledged. In contrast, regulatory research underpins decisions that directly affect patient safety and public health. Consequently, uncertainties in the data translate into risks for patients, meaning that regulatory decisions must be based on high-quality, regulatory grade data.

Despite this statutory remit, the MHRA currently lacks access to sufficient regulatory-grade data to meet all its needs. This limitation was recognised in the Sudlow Review (Sudlow, 2024), which concluded that the UK’s health data systems are not yet fit for purpose and cannot fully support national bodies such as the MHRA. The review recommended the creation of a “a national health data service.” In response, in April 2025 the UK Government announced a £600 million investment to create a UK Health Data Research Service (HDRS). While the details of HDRS are still being developed, its creation offers an opportunity to transform the national health data landscape and strengthen the UK’s position as a leader in life sciences research and regulation. HDRS aims to provide better and faster access to datasets relevant to public health and to catalyse the collection of new, high-value data sources. This would not only benefit the MHRA and other regulators but also enable industry and the wider research community to have better access to data, ultimately leading to better outcomes for patients.

The purpose of this paper is to outline the MHRA’s data needs and use cases, focusing on the limitations of the current UK data landscape and how this a future HDRS could address these gaps.

“Regulatory grade” data

At a high level, regulatory grade data refers to data of sufficient quality and integrity to support regulatory decision-making (Dreyer, 2018; Miksad and Abernethy, 2018). The precise definition is context dependent, but in principle such data must meet the highest possible standards, since regulatory decisions have a direct and immediate impact on patient health and wellbeing. However, there is no single, consensus definition of regulatory grade data in the literature.

One common approach to assessing data quality is to define measurable features that describe the extent to which the data is trustworthy (4,5), i.e. data metrics. Several data quality frameworks have been developed to provide systematic approaches for evaluating data quality and integrity, each with its own distinct set of metrics (7–10). Certain dimensions are consistent across frameworks: for example, Cave et al (6). propose provenance, validity, consistency, and adequacy as key criteria that make data acceptable for decision-making.  Common to these frameworks is the concept that data quality is an intrinsic feature of a dataset, independent of its intended use.

While such metrics help users understand the strengths and weaknesses of a dataset, they are not sufficient on their own to determine whether a dataset is of regulatory grade. There are no universal thresholds that automatically qualify data as regulatory grade. Instead, the metrics must be interpreted in the context of the specific regulatory question – a principle often referred to as fitness for purpose. As Daniel et al.^{[footnote 1]} note:

A data source that is appropriate for one purpose may not be suitable for other evaluations. For example, a large dataset that reveals critical insights about the safety profile of a new psychotropic drug may be inadequate to study potential indication expansions.

This notion of fitness for purpose recognises that even high-quality data may be unsuitable if it is not relevant to the regulatory question at hand, and conversely, that regulatory decisions sometimes mut be made despite imperfect or incomplete data.

The MHRA has published multiple guidance documents on data quality and integrity including its guidance on data integrity^{[footnote 2]} and the use of RWD, which stresses that data “must be demonstrated to be of sufficient quality for the intended use”. Similarly, the European Medicines Agency’s (EMA) Data Quality Framework^{[footnote 3]} provides extensive guidance on intrinsic data metrics but emphasises that quality must always be evaluated in relation to the intended regulatory purpose. The U.S. Food and Drug Administration’s (FDA) guidance^{[footnote 4]} on RWD likewise identifies intrinsic attributes of the data - such as completeness, timeliness, and traceability - as critical to quality, while also stressing the importance of appropriate data-handling and transformation procedures for maintaining integrity.

In a regulatory context, broader strategic considerations also apply. The MHRA’s remit spans the entire UK, encompassing diverse health systems, populations, and care pathways. Medical needs and health outcomes may vary across the four nations, as well as between different socioeconomic and ethnic groups. It is therefore essential that datasets used for regulatory decision-making are representative of the UK population, ensuring that regulatory outcomes are equitable and trustworthy.

In summary, while there is no single, universal definition of regulatory-grade data, a common view is emerging: that data suitability for regulatory use is determined by its intrinsic quality and integrity, its fitness for the intended purpose, and its representativeness of the population served.

MHRA data usage for safety and surveillance

This section outlines the MHRA’s activities that creates its data needs. The focus is on post-market surveillance and the reasons the MHRA requires access to longitudinal, timely, linked, and representative patient-level data that includes information on exposures to medical products, long-term outcomes, and rich covariate detail. The term “medical products” here refers to medicines, vaccines, blood products, and medical devices.

The purpose of MHRA’s post-market surveillance is to detect and assess safety issues associated with medical products that cause patient harm and to measure the effectiveness of regulatory actions (Allen and Donegan, 2017; Guo et al., 2024). Typically, medicines are rigorously tested in clinical trials, and medical devices undergo conformity assessments, prior to authorisation. However, such trials and assessments often involve a limited number of patients over short durations. Most importantly, participants often do not represent the full spectrum of patients who will eventually use, or be treated with, the product. Consequently, uncommon adverse effects, long-latency effects, or effects in subpopulations not represented in pre-authorisation stages often only become apparent during real-world use.

Adverse reactions to medical products cause patient harm and significant costs to the NHS (Osanlou et al., 2022). Safety monitoring and resulting regulatory action can, at least in part, prevent these outcomes and thus improve public health. Beyond direct patient safety benefits, an effective surveillance system enables faster routes to market for novel products, supports an innovation-friendly life sciences ecosystem, and contributes to economic growth. Hence, robust post-market surveillance and good access to data are both public health and economic imperatives.

Data needs arising from signal assessment

In epidemiology, signals are typically statistical associations between an exposure (e.g. a medicine) and an event (e.g. an adverse reaction). Establishing causality may sometimes involve identifying physiological mechanisms, such as biochemical pathways explaining the adverse reaction. Most of the time, however, causality is also understood in a statistical sense – demonstrating that an observed association is unlikely to be due to chance reporting bias, or confounding. Establishing such causality relies on sophisticated epidemiological analyses, either arising from scientific literature or conducted in-house by the MHRA (Donegan et al., 2014, 2013a).

Often, signals are found not to be causal. For example, after HPV vaccination programmes were implemented in the UK, a small number of spontaneous reports to the Yellow Card Scheme described chronic fatigue syndrome (CFS) following vaccination. However, CFS often emerges in adolescence or early adulthood - the same age group receiving the vaccine - confounding the association. The MHRA conducted epidemiological studies using data from the Clinical Practice Research Datalink (CPRD) to compare (i) observed versus expected rates, (ii) incidence rates before and after the start of the vaccination campaign, and (iii) risk in the year post-vaccination compared with other periods. These analyses provided rapid reassurance that there was no association between HPV vaccination and an increased risk of CFS (16). Fast assessment was critical to maintain public confidence in a vaccine that has since almost eliminated cervical cancer among women born since 1995 (Falcaro et al., 2021).

Epidemiological studies addressing regulatory questions typically require patient-level longitudinal data containing exposure information for medical products, health outcomes for exposed and unexposed (control) patients, and covariates such as age, sex, comorbidities, and socioeconomic status. Covariate data are essential both to identify differential impacts across subgroups, and to enable roust epidemiological methods such as propensity scores and double-robust estimators (Funk et al., 2011; Hong et al., 2019).

The most valuable data source currently available to the MHRA is the CPRD. CPRD is based on primary care records linked to secondary care and other health and area-based datasets for patients in England, including Hospital Episode Statistics (HES), Office for National Statistics (ONS) death registration data, and the National Cancer Registration and Analysis Service (NCRAS). It thus provides good coverage of exposures, outcomes, and covariates for medical products prescribed in primary care. However, secondary care prescribing - including anaesthetics, high-cost drugs, chemotherapies, and personalised medicines - is not currently captured. In addition, the electronic Prescribing and Medicines Administration (ePMA) data available through DARS is not patient-level and therefore insufficient for regulatory purposes.  

The future HDRS could address this by providing systematic, real-time linkage between primary and secondary care records, including comprehensive patient-level secondary prescribing data. This capability is essential to understand exposures and outcomes in key areas such as cardiology, oncology, and neurology.

Monitoring impacts of regulation

Beyond signal assessment, the MHRA uses RWD to evaluate the impact of regulatory actions. It is not always clear whether a regulatory action - such as an update to the Summary of Product Characteristics (SmPC) or a variation of indication – achieves its intended outcomes. Monitoring is therefore necessary to confirm effectiveness. For example, following evidence of increased risk of hyperkalaemia, hypotension, and renal impairment when renin-angiotensin system blockers were co-prescribed with one another, regulatory action was taken to advise against combination use. The MHRA monitored the impact of this advice and showed that co-prescribing declined in line with regulatory objectives (Allen and Donegan, 2017).

However, regulatory actions may also have unintended consequences, which also require monitoring. Risk minimisation measures – such as changes to the legal status of a product, introducing a controlled access programme, or issuing a Drug Safety Update (DSU) - may change healthcare or prescribing patterns in unexpected ways. This may potentially lead to undertreatment, use of alternative therapies off-label, increased hospitalisations, treatment delays, missed work or school days, or developmental impacts in children. Effects may also extend to patients not directly receiving the product but affected by alternations to care pathways.

Sodium valproate and pregnancy

Sodium valproate is used to treat epilepsy and bipolar disorder. It is highly effective but known to be teratogenic. The MHRA introduced a series of regulatory actions to restrict its use in women of childbearing potential. This created a need to understand the impact of these regulatory interventions, however, a key limitation was, and remains, the absence of comprehensive and routinely linked records across care settings (primary, secondary and tertiary care), as well as linkage between maternal and child records across datasets.

Despite the MHRA working with NHS England to develop a Medicines in Pregnancy Registry, it remains challenging to evaluate child outcomes that are not well captured in primary or secondary care data (e.g. neurological development, learning difficulties, autism spectrum disorder) in relation to maternal medication use. There are also secondary impacts that require evaluation, such as the outcomes for women who were not prescribed valproate following the restrictions. Addressing such questions requires access to linked datasets which are not currently available in an integrated way. As a result, the MHRA continues to face an incomplete picture when trying to assess whether regulatory actions have achieved their intended impact.

While the MHRA can monitor primary care prescribing through existing datasets, secondary care prescribing and important outcomes remain difficult to capture. The future HDRS could enhance monitoring by enabling richer linkages to socioeconomic and patient-reported data, providing a fuller picture of the real-world impacts of regulation.

Medical device data

Health data on medical devices is often overlooked, despite their central role in healthcare delivery. Common, high-volume, lower-risk devices include syringes, dressings, and in-vitro diagnostics such as glucose and blood tests, used millions of times daily across the NHS. Routine clinical equipment such as ultrasound and X-ray machines is equally integral. The highest-risk devices, such as pacemakers and hip implants, involved over 100,000 procedures annually.

The MHRA ensures that devices comply with the UK Medical Devices Regulations 2002 by reviewing evidence of safety, performance, and quality through monitoring adverse incident reports submitted by manufacturers (via Manufacturer Incident Reports) and from healthcare professionals and the public via the Yellow Card Scheme. However, RWD for devices is often incomplete. Exposure data is inconsistently recorded – for example, the specific device used in a hip replacement may not be traceable to the patient. Consequently, safety monitoring of devices is challenging. Many registries capture outcomes of procedures but not device details; others record device information but lack outcome or covariate data, limiting regulatory value.

A major step forward will be the universal introduction of the Unique Device Identifier, a barcode linking to detailed device information. Including this identifier in patient records in a structured field would enrich electronic health records with device exposure data, which has long been absent. While devices are not the primary focus of the HDRS, their inclusion would, for the first time, enable systematic linkage between exposure and outcome data for medical devices, closing a significant gap in UK health data infrastructure.

Data access

Rapid regulatory action is often necessary to prevent patient harm, which requires access to up-to-date, ideally near real-time, data. However, despite its statutory duty to monitor the safety of medicines and medical devices, the MHRA does not have default access to UK public health datasets, nor a fast-track governance pathway recognising its legal remit and public health mandate. Access procedures are often lengthy, and in some cases, burdensome governance processes have effectively prevented the agency from obtaining the data it needs to meet its statutory obligations.

The value of streamlined access was demonstrated during the COVID-19 pandemic, when emergency procedures enabled the MHRA to establish real-time safety monitoring within weeks (see case study below). Although achieved under exceptional conditions, this experience showed that existing barriers to access are disproportionate to both the risks of data sharing with the regulatory and the harms of delayed surveillance – including preventing patient suffering, lost lives, avoidable NHS costs, and lost economic productivity. Notably, the emergency measures taken during the pandemic did not result in any major data breaches or privacy violations. Lessons from this period should inform how data access can be improved moving forward. The HDRS should therefore provide a single point of access to UK health data with a unified approval process across datasets, and a “trusted user” status for the MHRA as the national regulator.

While a single point of access is essential for timely use, maintaining heterogeneity in the UK data landscape also has value. The collection, management, and presentation of data requires data custodians to make countless decisions, and retaining this data diversity allows for cross-validation of results (Fisher et al., 2023; Gulliford et al., 2020; Jick et al., 2023). Providing access to data through a federated common data model (CDM) that retains the heterogeneity of the original sources achieves an optimal balance between operational efficiency and data richness. Several CDMs exist, including the Sentinel (Brown et al., 2022), TriNetX (Palchuk et al., 2023), and the Observational Medical Outcomes Partnership (OMOP) models. OMOP, widely adopted by the international epidemiological community and supported by the Observational Health Data Sciences and Informatics (OHDSI) network (Booth et al., 2025; Candore et al., 2020), is used by regulators, including the EMA’s DARWIN EU capability. Its adoption by the HDRS would enable joint investigations with peer regulators, ensuring that the UK can contribute to, and benefit from, global regulatory science.

COVID-19 case study

During the height of the COVID-19 pandemic in 2020, many of the usual administrative and hurdles that impede timely data access were temporarily lifted. This allowed the MHRA, in collaboration with NHS England, to establish a near-real time surveillance system in weeks.

In December 2020, the MHRA became the first regulator worldwide to approve a COVID-19 vaccine for use in the general population. COVID-19 vaccines were administered to a large and diverse population at speed and at scale, across a vast range of non-traditional settings. Whilst in normal circumstances, vaccination data are sent back to GPs and entered into the patients’ primary care records, there was a risk that the unprecedented delivery model may have meant this did not occur in a timely or complete manner, especially in the early phases of rollout. Yet the need for rapid, robust safety monitoring was critical to maintain public trust in the vaccination programme.

To address this, the MHRA and NHS England developed a bespoke solution linking a minimum vaccination dataset, adverse event data, and hospitalisation records to CPRD primary care data, creating a near-complete patient-level dataset for monitoring vaccination outcomes. Within 24 hours of a vaccination being administered, all anonymised data were integrated into an analysis-ready dataset accessible to the MHRA for safety signal detection.

The entire end-to-end system – including data sharing agreements and technical infrastructure – was developed and implemented within six weeks, before the start of the national vaccine rollout. Although delivered under emergency conditions, this initiative demonstrated that it is technically and operationally feasible to integrate near-real-time data from multiple sources across the health system for the purpose of safety and surveillance of medicines and medical products.

Finally, access costs are a crucial but over overlooked barrier. Maintaining public health datasets is expensive, and costs are typically passed on to users. However, if access fees are too high, they can effectively block valuable research on which regulatory decisions depend. With the potential to bring widespread public benefits, the HDRS should adopt a funding model reflecting the social value created by its use – prioritising broad access over operation within arbitrary economic constrains.  

Opportunities for new regulatory approaches

In addition to its role as a user of data for safety monitoring, the MHRA has an increasing interest in RWD to support licensing and authorisation decisions, particularly in situations where a randomised clinical trial may be impractical or unethical. In such contexts, the agency does not analyse data itself but has a strong interest in ensuring that other stakeholders, such as academic groups and marketing authorisation applicants, have access to high-quality, well-curated data capable of leading to regulatory-grade evidence.

Clinical trials remain the gold standard for generating reliable evidence on the efficacy of medical products. However, they are expensive, time-consuming, and often poorly representative of real-world patient populations. For example, Tan et al.(Tan et al., 2022) found that multimorbidity was an exclusion criterion in more than 90% of all trials, whereas in real clinical settings, patients commonly present with multiple co-existing conditions.

Real-world evidence derived from routine clinical data can address some of these limitations by capturing broader and more diverse patient experiences. Nevertheless, the quality and completeness of currently available data often remain limiting factors – particularly for studies of effectiveness rather than safety. For example, in evaluating the effectiveness of cancer medicines, all-cause mortality may be measurable through existing datasets, but quality-adjusted life years – a more meaningful measure of benefit – require patient-reported outcomes, which are rarely available in electronic health records. Other important outcomes, such as tumour response rates, imaging data, and other clinical measures, are inconsistently captured unstructured, or unlinked to patient records, making them inaccessible for regulatory-grade analyses.

The future HDRS could enable novel regulatory approaches by making rich, linked data routinely accessible for research. Such data could support earlier and more flexible approvals, reducing development timelines while maintaining rigorous evidence standards.

Selective Serotonin Reuptake Inhibitors (SSRIs) and suicidality

SSRIs are widely prescribed antidepressants used to treat depression and anxiety disorders across all age groups. Concerns about the risk of suicidal thoughts and behaviours, particularly in children, adolescents, and young adults, have been longstanding. Over the past two decades, the MHRA has issued several regulatory communications and label changes based on emerging evidence from clinical trials and real-world data.

However, evaluating the real-world impact of these regulatory actions remains difficult because of the fragmented nature of available data. Primary care records capture most SSRI prescriptions, but relevant outcomes – such as suicide attempts, self-harm incidents, support from Crisis Resolution and Home Treatment Teams, and detention under the Mental Health Act – are typically recorded in specialist mental health service data. These datasets are rarely linked to primary care data at the patient level. Without such linkage, it is impossible to fully determine whether regulatory changes (for example, revised guidance on paediatric prescribing) have led to improvements or unintended consequences.

Furthermore, key contextual factors, such as previous mental health diagnoses, psychosocial stressors, and concurrent medications, are distributed across different data sources and are not routinely linked. Linkage encompassing primary care, mental health trusts, and secondary care would enable more robust longitudinal analyses to assess whether changes in SSRI prescribing patterns correspond to actual reductions in self-harm or suicide risk.

Linked data covering hospital prescribing, primary care records, diagnostic imaging, and physiotherapy data (to capture tendon injuries, for example) would allow the MHRA to track individual patient journeys across the continuum of care. It would also enable analysis of prescribing substitution effects – for example, whether restrictions on fluoroquinolones led to increased use of alternative antibiotics with different risk profiles.

Conclusions

The MHRA has a statutory responsibility to regulate medicines and medical devices, ensuring they are safe and effective. To fulfil this remit, the agency must have access to regulatory-grade data. As recognised in the Sudlow review, the MHRA’s ability to discharge its statutory function is currently constrained by a fragmented health data landscape and complex, often disproportionate, data governance processes.

The creation of the UK HDRS provides a unique opportunity to address these limitations by establishing a single, trusted gateway to the UK’s health datasets. If designed and implemented well, the HRS will become a valuable resource for researchers, clinicians, and regulators alike.

To succeed, it will be essential that the service recognises and accommodates the distinct needs of its diverse user community. Ultimately, a well-designed HDRS is more than a technical infrastructure – it is a foundation for public trust, effective regulation, and a vibrant life sciences ecosystem that delivers better outcomes for patients across the UK.

References

Allen, C., Donegan, K., 2017. The impact of regulatory action on the co‐prescribing of renin–angiotensin system blockers in UK primary care. Pharmacoepidemiol Drug Saf 26, 858–862. https://doi.org/10.1002/pds.4219

Booth, H.P., Connelly, J., Dedman, D., Donegan, K., Cave, A., 2025. A Regulatory Perspective on a UK Federated Data Network for Medicines and Medical Devices: Lessons from a ‘Study-A-Thon.’ Ther Innov Regul Sci. https://doi.org/10.1007/s43441-025-00854-3

Brown, J.S., Mendelsohn, A.B., Nam, Y.H., Maro, J.C., Cocoros, N.M., Rodriguez-Watson, C., Lockhart, C.M., Platt, R., Ball, R., Dal Pan, G.J., Toh, S., 2022. The US Food and Drug Administration Sentinel System: a national resource for a learning health system. Journal of the American Medical Informatics Association 29, 2191–2200. https://doi.org/10.1093/jamia/ocac153

Candore, G., Hedenmalm, K., Slattery, J., Cave, A., Kurz, X., Arlett, P., 2020. Can We Rely on Results From IQVIA Medical Research Data UK Converted to the Observational Medical Outcome Partnership Common Data Model? Clin Pharmacol Ther 107, 915–925. https://doi.org/10.1002/cpt.1785

Donegan, K., Beau-Lejdstrom, R., King, B., Seabroke, S., Thomson, A., Bryan, P., 2013a. Bivalent human papillomavirus vaccine and the risk of fatigue syndromes in girls in the UK. Vaccine 31, 4961–7. https://doi.org/10.1016/j.vaccine.2013.08.024

Donegan, K., Beau-Lejdstrom, R., King, B., Seabroke, S., Thomson, A., Bryan, P., 2013b. Bivalent human papillomavirus vaccine and the risk of fatigue syndromes in girls in the UK. Vaccine 31, 4961–4967. https://doi.org/10.1016/j.vaccine.2013.08.024

Donegan, K., King, B., Bryan, P., 2014. Safety of pertussis vaccination in pregnant women in UK: observational study. BMJ 349, g4219–g4219. https://doi.org/10.1136/bmj.g4219

Dreyer, N.A., 2018. Advancing a Framework for Regulatory Use of Real-World Evidence: When Real Is Reliable. Ther Innov Regul Sci 52, 362–368. https://doi.org/10.1177/2168479018763591

Falcaro, M., Castañon, A., Ndlela, B., Checchi, M., Soldan, K., Lopez-Bernal, J., Elliss-Brookes, L., Sasieni, P., 2021. The effects of the national HPV vaccination programme in England, UK, on cervical cancer and grade 3 cervical intraepithelial neoplasia incidence: a register-based observational study. The Lancet 398, 2084–2092. https://doi.org/10.1016/S0140-6736(21)02178-4

Fisher, L., Hopcroft, L.E., Rodgers, S., Barrett, J., Oliver, K., Avery, A.J., Evans, Dai, Curtis, H., Croker, R., Macdonald, O., Morley, J., Mehrkar, A., Bacon, S., Davy, S., Dillingham, I., Evans, David, Hickman, G., Inglesby, P., Morton, C.E., Smith, B., Ward, T., Hulme, W., Green, A., Massey, J., Walker, A.J., Bates, C., Cockburn, J., Parry, J., Hester, F., Harper, S., O’Hanlon, S., Eavis, A., Jarvis, R., Avramov, D., Griffiths, P., Fowles, A., Parkes, N., Goldacre, B., MacKenna, B., 2023. Changes in medication safety indicators in England throughout the covid-19 pandemic using OpenSAFELY: population based, retrospective cohort study of 57 million patients using federated analytics. BMJ Medicine 2, e000392. https://doi.org/10.1136/bmjmed-2022-000392

Funk, M.J., Westreich, D., Wiesen, C., Stürmer, T., Brookhart, M.A., Davidian, M., 2011. Doubly Robust Estimation of Causal Effects. Am J Epidemiol 173, 761–767. https://doi.org/10.1093/aje/kwq439

Gulliford, M.C., Sun, X., Anjuman, T., Yelland, E., Murray-Thomas, T., 2020. Comparison of antibiotic prescribing records in two UK primary care electronic health record systems: cohort study using CPRD GOLD and CPRD Aurum databases. BMJ Open 10, e038767. https://doi.org/10.1136/bmjopen-2020-038767

Guo, Y., Raventós, B., Català, M., Elhussein, L., López‐Güell, K., Tan, E.H., Prats‐Uribe, A., Dedman, D., Man, W.Y., Omulo, H., Delmestri, A., Lane, J.C.E., Rahman, U., Griffin, X.L., Gao, C., Cole, C., Batty, P., Connelly, J., Booth, H., Cave, A., Donegan, K., Prieto‐Alhambra, D., Burn, E., Jödicke, A.M., 2024. Time Series Methods to Assess the Impact of Regulatory Action: A Study of UK Primary Care and Hospital Data on the Use of Fluoroquinolones. Pharmacoepidemiol Drug Saf 33. https://doi.org/10.1002/pds.70022

Hong, H., Aaby, D.A., Siddique, J., Stuart, E.A., 2019. Propensity Score–Based Estimators With Multiple Error-Prone Covariates. Am J Epidemiol 188, 222–230. https://doi.org/10.1093/aje/kwy210

Jick, S., Vasilakis-Scaramozza, C., Persson, R., Neasham, D., Kafatos, G., Hagberg, K., 2023. Use of the CPRD Aurum Database: Insights Gained from New Data Quality Assessments. Clin Epidemiol Volume 15, 1219–1222. https://doi.org/10.2147/CLEP.S434832

Miksad, R.A., Abernethy, A.P., 2018. Harnessing the Power of Real‐World Evidence (RWE): A Checklist to Ensure Regulatory‐Grade Data Quality. Clin Pharmacol Ther 103, 202–205. https://doi.org/10.1002/cpt.946

Osanlou, R., Walker, L., Hughes, D.A., Burnside, G., Pirmohamed, M., 2022. Adverse drug reactions, multimorbidity and polypharmacy: a prospective analysis of 1 month of medical admissions. BMJ Open 12, e055551. https://doi.org/10.1136/bmjopen-2021-055551

Palchuk, M.B., London, J.W., Perez-Rey, D., Drebert, Z.J., Winer-Jones, J.P., Thompson, C.N., Esposito, J., Claerhout, B., 2023. A global federated real-world data and analytics platform for research. JAMIA Open 6. https://doi.org/10.1093/jamiaopen/ooad035

Sudlow, C., 2024. Uniting the UK’s Health Data: A Huge Opportunity for Society. https://doi.org/10.5281/zenodo.13353746

Tan, Y.Y., Papez, V., Chang, W.H., Mueller, S.H., Denaxas, S., Lai, A.G., 2022. Comparing clinical trial population representativeness to real-world populations: an external validity analysis encompassing 43 895 trials and 5 685 738 individuals across 989 unique drugs and 286 conditions in England. Lancet Healthy Longev 3, e674–e689. https://doi.org/10.1016/S2666-7568(22)00186-6

Wong, J., Donegan, K., Harrison, K., Jan, T., Cave, A., Tregunno, P., 2025. Implementation and Results of Active Vaccine Safety Monitoring During the COVID-19 Pandemic in the UK: A Regulatory Perspective. Drug Saf.

1. Daniel, G, C Silcox, J Bryan, M McClellan, M Romine, and K Frank, 2018, Characterizing RWD Quality and Relevancy for Regulatory Purposes, Duke Margolis Center for Health Policy, accessed 23/08/2025, https://healthpolicy.duke.edu/sites/default/files/2020-03/characterizing_rwd.pdf. ↩

2. https://www.gov.uk/government/publications/guidance-on-gxp-data-integrity ↩

3. https://www.ema.europa.eu/en/documents/regulatory-procedural-guideline/data-quality-framework-eu-medicines-regulation_en.pdf ↩

4. https://www.fda.gov/regulatory-information/search-fda-guidance-documents/real-world-data-assessing-electronic-health-records-and-medical-claims-data-support-regulatory ↩