Official Statistics

Annex 2 - Community Accommodation Service Tier 3 - guidance

Published 31 July 2025

Applies to England and Wales

1. Background: Community Accommodation Service Tier 3 (CAS3)

On 19 July 2021, HM Prison and Probation Service (HMPPS) launched a new transitional accommodation service, Community Accommodation Service Tier 3 (CAS3), providing up to 12 weeks’ basic accommodation for those leaving prison at risk of homelessness.

The service was initially launched in five probation regions (Yorkshire and the Humber; North West; Greater Manchester; Kent, Surrey and Sussex; and East of England). Delivery of CAS3 in Wales, through grant funding arrangements with Welsh Local Authorities, commenced in June 2022, and mobilisation of the service in the remaining six probation regions in England commenced in May 2023, with full national coverage from December 2023. Therefore, 2024/25 is the first year with full national coverage.

2. Data quality

2.1 Change to CAS3 Digital Service

2021/22 is reported as a partial performance year from July 2021 onwards, as the CAS3 service did not exist prior to this point.

From December 2023 all Probation Regions used a new Digital Service to make referrals and manage CAS3 bookings.  Prior to this, all CAS3 referrals and bookings were managed through manual processes.

All referrals and bookings in England are now managed through the Digital Service. Due to operational differences, Wales continues to use the manual system for referrals, but subsequent bookings are managed by HPTs on the Digital Service.  

The new Digital Service now enables Probation Practitioners and Regional Probation Homelessness Prevention Teams (HPTs) to access an online platform through which they make their CAS3 referrals and manage bookings. The platform pulls through data directly from the HMPPS Case Management System (NDelius) therefore improving the efficiency of the service and data quality. Referrals are completed by Probation Practitioners via the Digital Service and submitted to Regional HPTs who assess the referrals, make bookings, and manage the bookings in the Digital Service.

The CAS3 data from the regional Management Information systems is merged, updated and quality assured monthly.

Because these cases are extracted from a live digital service, in a minority of cases retrospective changes to referral and placement dates may be made. For this reason, the total volume of placements for each performance year may vary from previously published figures.

3. CAS3 definitions and processes

CAS3 placement volumes provided count the number of eligible referrals to CAS3 that have a recorded start date, meaning that the person on probation has begun their CAS3 placement, and has physically moved into a CAS3 property.

Where an individual has been through the CAS3 service multiple times they will be counted once for each placement.

People on probation are eligible for CAS3 if they are recorded as being at risk of homelessness upon release from either prison, Community Accommodation Service Tier 1 (CAS1), or Community Accommodation Service Tier 2 (CAS2). The relevant performance year for each case is based on the recorded start date of the CAS3 placement.

Figures provided come from merging historical CAS3 data and current Wales referrals data, collected through manual process, with referral and bookings data collected through the Digital Service. For a given performance year, the number of placements and the number of outcomes will differ, as performance year for placements is reported based on start date, while outcomes are reported based on end date. Individual’s may therefore have a placement recorded in one performance year, and an outcome recorded in the subsequent performance year.

4.  Gender recording

Gender breakdowns for CAS3 residents are calculated based on the recorded gender of residents in the CAS3 case management system that their stay was recorded in.

Prior to the introduction of the CAS3 Digital Service, a CAS3 resident’s gender identity was manually inputted into a CAS3 referral form by a Probation Practitioner. This data would either be sourced from the person on probation’s NDelius record - or if the Probation Practitioner believed this data was not up to date, they would ask the person on probation to confirm their gender identity and submit this on the CAS3 referral.

Since the introduction of the CAS3 Digital Service, all gender identity data is automatically populated onto the CAS3 referral information directly from NDelius. Where gender identity data is missing for a referral, sex data has been used as a proxy.

Across both manual and digital systems, all cases recorded as either Male or Transgender Male are aggregated as ‘Male’, and all cases recorded as Female or Transgender Female are aggregated as ‘Female’.

Due to small case numbers, cases recorded as non-binary, transgender with no specific gender identity recorded and blank cases are aggregated as ‘Other’.

Recording of gender within the CAS3 service may differ from recording within other Official Statistics, due to differences in categorisation and recording point. The recording and aggregation method used for gender recording in CAS3 will be kept under review and may be subject to changes in future publications.

5. Accommodation outcomes recording

Figures count accommodation move-on outcomes at the end of a CAS3 stay. CAS3 move-on outcomes data is added to the CAS3 resident’s nDelius record at the end of their stay. Outcome options include:

Settled accommodation
(i) Any accommodation that provides a permanent, independent housing solution, including:
- as a householder or house owner
- rental accommodation (tenant) – private or social rental
- living with friends and family (settled) – the individual can reside in that home and is able to return to that home
- living with a friend with a bedroom available for the individual’s use and access to domestic facilities
- a caravan or boat that is viewed by the individual as their permanent home (ii) Supported housing - regardless of the duration or residency (iii) Long-term residential healthcare.

Transient accommodation
(i) Accommodation that is not intended to be used for more than 13 consecutive weeks. This can be any type of accommodation but is unlikely to provide a long-term solution to a housing need. For example, this may include, but is not limited to:
- tenancies available for less than 13 weeks
- bed and breakfast accommodation - hostels
- crisis houses - staying with friends/ family on a temporary basis - the individual has moved onto CAS1 or CAS2 accommodation.

Homeless
The individual does not have any accommodation that they are entitled to occupy, or they have accommodation they are entitled to occupy but they cannot be reasonably expected to occupy it for any reason.
Either:
- the individual occupies a property with no legal authority or permission to do so, such as squatting
- the individual may reside in night shelters, emergency hostels or campsites

Rough sleeping
(i) The individual does not have any accommodation they are entitled to occupy, or they have accommodation they are entitled to occupy but they cannot be reasonably expected to occupy it for any reason. -(ii) The individual sleeps in open air (such as on the streets, in tents, doorways, parks, bus shelters or encampments) or in other places not designed for occupation (such as stairwells, barns, sheds, car parks, cars, derelict boats or stations)

Recalled
(i) The individual has been recalled to prison for breaching licence conditions. (i) The individual is unlawfully at large.

Prison
(i) The individual is subject to a further custodial event, separate to recall. (ii) The individual is being held at an immigration detention centre

Pending
The individual has completed their CAS3 stay but the outcome status has not yet been updated.

Unknown / blank The accommodation status is unknown, either because:
- the individual is awaiting assessment
- no status is recorded
- there are errors in the record.

Some outcomes have been grouped into overarching categories. These include:

• ‘Homeless’ and ‘Rough sleeping’ - combined into single category termed ‘Homeless / Rough sleeping’
• ‘Recalled’ and ‘Prison’ - combined into a single category termed ‘Recalled / Further custody’
• ‘Unknown’ and missing data grouped as ‘Unknown / blank’.

Figures for those whose outcome is ‘deceased’ have not been included in the overall figures as this is not classed as an accommodation outcome.