Homelessness statistics user group minutes: May 2026

Attendees and apologies

• Athelstan Croft (Aberdeen City Council)

• Martin Nicoll (Aberdeen City Council)

• Tom Phillips (Aberdeen City Council)

• Alison Logan (Aberdeenshire Council)

• Elaine Lothian (Aberdeenshire Council)

• Paul Murray (Aberdeenshire Council)

• Morven MacIntyre (Argyll & Bute Council)

• Alex Gilbert (Clackmannanshire Council)

• Rhuaraidh Fleming (Crisis)

• James Mullaney (Crisis)

• Susan Heritage  (Dumfries & Galloway Council)

• Alyson Smith (Dumfries & Galloway Council)

• Lynn Ross (East Dunbartonshire Council)

• Susan Morris (East Renfrewshire Council)

• Kevin Hewie (Edinburgh Council)

• Nicholas Laidlaw (Edinburgh Council)

• Leigh Sherwood (Edinburgh Council)

• Helen Dryfe (Falkirk Council)

• Christine Thomson (Falkirk Council)

• Kevin Howe (Glasgow Council)

• Dianne McKendrick (Glasgow HSCP)

• Shona Santiago (Glasgow Council)

• Carrie-Anne Alexander (Highland Council)

• Aileen MacKenzie (Highland Council)

• Donna MacKenzie (Highland Council)

• Jamie Milne (Homeless Network Scotland)

• Alan Baxter (Inverclyde Council)

• Angela Harris (Midlothian Council)

• Matthew McGlone (Midlothian Council)

• Connor McKail (Midlothian Council)

• Martin Gray (Moray Council)

• Apoorva Haston (Moray Council)

• Cara Rollo (Moray Council)

• Beth Watson (National Records of Scotland)

• Liz Brown (North Ayrshire Council)

• Kevin Thomas (North Ayrshire Council)

• Pamela Haughey (North Lanarkshire Council)

• Audrey Hayward (North Lanarkshire Council)

• Hazel McDonagh (Perth & Kinross Council)

• Linda Montgomery (Perth & Kinross Council)

• Steven Robb (Perth & Kinross Council)

• Marlene Paterson (Scottish Borders Council)

• Kenneth Nimmo (Scottish Prison Service)

• John Flannery (Shelter)

• Monique Hunter (Shetland Council)

• Tommy Walterson (Shetland Council)

• Jacqueline Fernie (South Lanarkshire Council)

• Anna McClelland (South Ayrshire Council)

• Lorna Hutton (Stirling Council)Sarah Macaulay (Stirling Council)

• Alison McMaster (Stirling Council)

• John Nish (Stirling Council)

• Josh Tierney (Stirling Council)

• Jennifer MacMahon (West Dunbartonshire Council)

• Kathryn Smith (West Dunbartonshire Council)

• Colin McKillop (West Lothian Council)

• Derek Reid (West Lothian Council)

• Elisabeth Boyling (Scottish Government)

• Ashana Dunstan (Scottish Government)

• Karen Grieve (Scottish Government)

• Matt Howarth (Scottish Government)

• Angela McPhail (Scottish Government)

• Kerry Shaw (Scottish Government)

• Julie Stuart (Scottish Government)

• Ebba Svantesson (Scottish Government)

• Charlotte Walmsley (Scottish Government)

• Dani Barton (organisation unknown)

• Claire Wood (organisation unknown)

• Lisa Patterson (organisation unknown)

• Lesley Lamie (organisation unknown)

Items and actions

Sara welcomed everyone to the meeting. It was explained the delay in holding this meeting was due to uncertainties around work priorities over the past year and it was felt best to have the meeting at a time where meaningful updates were able to be provided.

Agenda item: policy update

Matt and Angela provided an update on homelessness prevention work, and in particular new legislation within the Housing Scotland Act. This places statutory duties on the prevention of homelessness for both local authorities and relevant bodies, with implementation of these to be completed no later than November 2028.

For new duties on relevant bodies, “Ask and Act” pilots are currently underway, with findings from these being used to inform guidance and secondary legislation, to be developed in due course. Policy colleagues will be exploring the best way to work with relevant bodies to ensure they are ready for implementation.

For local authority duties, this extends (current) duties from those who are two months, to those who are six months, away from homelessness. Local authorities must undertake reasonable steps to prevent homelessness from occurring. Work will be undertaken to develop the understanding of reasonable steps, with a view to including them in secondary legislation. 

There is a separate piece of work on-going around Domestic Abuse, with this being implemented earlier than November 2028.  

Karen, Ebba and Charlotte provided an overview from the temporary accommodation policy team. 

Work has started to refresh of the Code of Guidance, with a view to making this more user friendly and fit for purpose, reflecting key legislative updates. It will be co-produced with key stakeholders, such as COSLA and ALACHO, drawing from insights from working groups. The policy team are looking for areas of best practice that can be used to inform this work.

Fourteen local authorities, as well as SG, have declared a Housing Emergency (HE). One of the key focuses of this work is reducing children in temporary accommodation. The action plan published in September 2025 is around supporting local authorities to find a way out of the housing emergency. Funding has been targeted towards local authorities experiencing the most sustained temporary accommodation pressures, enabling them to acquire homes and help reduce the use of unsuitable temporary accommodation. The Housing to 2040 Strategic Board provide oversight of the HE response.

Work is underway related to the temporary accommodation standards framework consultation which closed earlier this year. Analysis of the 41 responses received is in progress with a report of the findings to be published in due course. 

Together with the newly appointed Cabinet Secretary for Social Justice and Housing, consideration is being given to workstream prioritisation following the election. 

Agenda item: publications update

Kelechi provided an overview of the suite of homelessness statistical publications that the team output, with details of future releases. He talked through the new tables and workbooks that had been made available since the last meeting to continue to respond to user needs as best as possible. He also spoke to future developmental work around the use of HL3 as the sole temporary accommodation reporting source and the reduction in Housing Options (PREVENT1) data reporting to support future developments. Further details were provided in the slides – see Annex A.

A request was made for a glossary of methodology/definitions to be made available to data providers so they can understand how to replicate the figures presented e.g. in processing reports. Action: Scottish Government to determine how best to implement.

Clarification was also sought around Housing Options (PREVENT1) and Sara explained that the need for record level data for this specific collection has ceased completely. This is due to this data collection being superseded by new prevention data collection to support the new legislation; this will be brought in under the data review work. An aggregate level collection has been put in place (from 2026-27) to bridge any reporting gap between now and then.

Agenda item: data systems development

Anna provided an overview of the work undertaken over the past year to redevelop current data collection and processing systems to ensure they continue to work within new Scottish Government IT environments, being brought in from January 2027. This will ensure the continued availability of high profile homelessness data.

For the redevelopment of reports, two-thirds of users stated a preference for data back 10 years to be included in the reports with one third stating data back five years. Anna requested that anyone with a need for data further back than this contact her separately to discuss.

Anna highlighted areas of current collections that have been decided not to redevelop as a result of these being replaced as part of data review work, and therefore the resource to redevelop these would outweigh any benefits. These were:

1. The collection of National insurance number (NIN) to be paused until this recommences with the review. This has implications for repeat homelessness calculations, where NIN is used to enhance this.
2. HL2 is being discontinued with HL3 being the sole reporting measure for TA from 1 April 2026 onwards.
3. As noted above, PREVENT1 (Housing Options) will be replaced with a reduced aggregate data collection to bridge the reporting gap re: homelessness prevention until new reporting requirements around the statutory duties are in place.

Further detail is available in the slides presented – see Annex B.

There was discussion around the timing of the policy development work around prevention duties and what this will mean for data collection. It was noted that a data specification for the new collection cannot be finalised until homelessness prevention data items are known. Given the length of time to develop local authority systems, this will mean changes won’t be in place for the new duties commencing in November 2028. It was noted that this risk is recognised within Scottish Government and there will need to be discussion at a relevant point with local authorities around implementation timescales for the new data specification.

A local authority noted the extent of the changes around the operational practice of merging prevention (systems) with homelessness (systems) could lead to complexities.

Scottish Government confirmed that reporting around Ask and Act duties for relevant bodies is outwith the scope of the statistical data collections.

Agenda item: data quality

The homelessness statistics and analysis team presented an overview of enhanced quality assurance activity around previous activity, HL3, “not offered” accommodation and implementation of asylum related categories.  This was detailed in the slides – see Annex C.

It was requested that the HL3 queries (based on data up to Q3) be ran again based on data up to Q4 given this is more up to date and more data cleansing will have taken place for Q4 data. Sara noted this is something that can be done, however, timescales will be dependent on resource, highlighting that the statistics team are reduced in capacity over the summer period, and will also have the annual publication to produce. Action: Scottish Government to determine how and when best to undertake HL3 quality checks for data up to Q4.

Local authorities noted a number of scenarios where there may be genuine cases where the number of children reported on HL3 and HL1 differed. Scottish Government confirmed this was acknowledged and, in these cases, no action was required. The cases provided to local authorities are simply a comprehensive list, where some will require amending, and some will be to confirm only.

Other discussion around data provision confirmed that if local authorities update any data, they can submit this in their returns and Scottish Government will keep the most recent data. It was also confirmed that a new application should be started for a partner where they were previously on another application but had split up from the main applicant.

Agenda item: data review update

Anna provided an update of work to date on the data review. For further details see the content of the slides at Annex D. Anna explained this has been paused for most of the last year due to having to undertake the redevelopment work, but the plan is to recommence this from June. The focus will be on the development of homelessness prevention content, and then to getting all content developed signed off for implementation.

Anna confirmed that trans status will be added in to the collection and provided confirmation that the following protected characteristics information is planned (subject to wider agreement) to be included: 

• Age (from Date of Birth) 
• Gender
• Gender reassignment
• Sexual orientation
• Ethnicity
• Religion or belief
• Pregnancy and maternity
• Disability of a household member

Agenda item: NRS homelessness deaths update

Beth Watson from NRS presented findings on the latest Homelessness Deaths statistics, the methodology behind these and future developments. For further details see slides at Annex E.

One local authority noted the potential use of UPRN to improve quality may be tricky as TA stock changes frequently and so the list would need to be constantly updated as a result.

There was discussion around the reporting being based on place of death, which may not be where the person is from originally. Beth explained there can be ways of looking at an alternative place with a definition of “within 12 months”, ultimately the medical team makes a judgement call. This all highlighted the need for the HL1 data collection to be able to record death.

The wording of the publication was also raised as it was felt this made it sound like people were dying as a result of homelessness, rather than this could be a completely unrelated cause of death and they happen to be homeless. NRS noted they could consider this for future outputs.

Agenda item: AOB

Sara thanked everyone for coming to the meeting and noted the intention to hold another meeting around the same time next year.

Annex A: Publications Update slides

Slide: Publications Overview

• Homelessness in Scotland (HiS)

• Published annually and biannually
• Provides information on:
  • Homelessness applications (HL1)
  • Temporary accommodation (HL2 and HL3)
• Housing Options
• Provides information on homelessness prevention
• Based on the PREVENT1 data collection
• More information: Homelessness statistics - gov.scot (www.gov.scot)

Slide: New Data Tables (HiS)

• Number of applications by eligibility and local authority (annual & biannual)
• Number of closed cases by assessment outcome (annual & biannual)
• Number of closed cases by local authority (annual)
• Number of people, adults and children linked to applications assessed as homeless or threatened with homelessness (biannual)

Slide: Refugee (including UDP) workbook

• Homelessness applications
• Rough sleeping (night before application & three months prior)
• Live homelessness applications
• Applications assessed as homeless or threatened with homelessness
• Number of people, adults and children
• Households in temporary accommodation (including by accommodation type)
• Temporary accommodation not provided
• Breaches of the Unsuitable Accommodation Order
• Closed cases securing settled accommodation

This workbook will be regularly reviewed and on-going need for assessed.

Slide: Future work

• The next releases are scheduled for:
  • Homelessness in Scotland 2025–26 – September 2026
  • Housing Options (PREVENT1) 2025-26 – November 2026
• Developments for 2026-27
  • Replacing ‘headline’ temporary accommodation figures with HL3 data
  • Housing Options (PREVENT1) replaced with reduced data collection
• We welcome feedback and suggestions for further improvements

Annex B: Data Systems Development slides

Slide: Overview

• Scottish Government have made a corporate decision to move to cloud-base servers by the end of December 2026.
• Amazon Warehouse Solutions (AWS) will be the provider.
• Our data collection and processing system will not work on the cloud.
• Scottish Government has also decided to move to open-source software, e.g. R or python.
• Due to these changes we need to re-develop our current systems.
• To ensure the continuation of homelessness statistics work was required re-develop the current data collection into R ahead of the data review collection being implemented, resource from the data review was diverted in June 2025.
• The re-development impacts all homelessness datasets.
  • HL1, HL2, HL3, PREVENT1
• Our aim for the re-development is for this to have little / low impact on local authorities.
• Minimal changes to the collection for HL1, to be discussed later.
• During the re-development extensive testing and refinement of code has been carried out to ensure consistent results to the current system.

Slide: Reports

• Error reports will be combined into one workbook with same information, compared to multiple workbooks currently provided.
• In October 2025 we sent out a questionnaire to inform the content of our quarterly / annual reports. This feedback has helped inform what we include in the reports. The number of tables has been reduced.
• If we have not included information a local authority regularly uses in the reports, please get in touch and we can implement.
• For both reports extensive cross checks have been carried out with the current system.

Slide: Reports time trend data

• We have one further question we would like to ask about the quarterly & annual reports.
• How far back would you like the trend to go?
  1. Back as far as possible
  2. Last 10 years
  3. Last 5 years

Slide: System changes

• During the investigation into the collection and processing system, certain sections were identified as either resource and cost intensive with limited benefit, or likely to be affected by upcoming changes from the data review and new legislation.

• Areas highlighted were:

  • National Insurance Number
  • HL2
  • PREVENT1

Slide: National Insurance Number (NIN)

• Collected for main applicant and partner. Encrypted in VBA code before the data is passed into SAS.
• The VBA code is complex and no expertise in Scottish Government to translate it into another format.
• To convert VBA external contractors would be required.
• There is currently limited use of NIN. The main use is to enhance repeat homelessness by identifying previous application where PREVAPP was not provided.
• Whilst it is important to reduce repeat homelessness, the indicator is not a headline figure.
• The SAS code to enhance PREVAPP and to calculate repeat homelessness is lengthy and complex.
• Plus, definitions will change once the data review collection becomes live.
• Therefore, we have decided to pause the collection of NIN until the new collection under the data review comes into effect.
• Given the extremely limited use of NIN outlined, the resource (time and money) to recreate the encryption for the rebuild of the current system over the interim period would be disproportionate to any benefits realised.
• To note, under the data review collection of NIN will re-commence under a new encryption in ProcXed.

What does this mean for local authorities?

• The HL1 collection still requires the two national insurance number  variables (NINUM1 & NINUM2) for the structure of the data
• For data from 1 April 2026, Local Authorities can either
  1. Leave theses NINUM1 & NINUM2 blank, or
  2. Provide the NIN and Scottish Government will blank before entering the system.

What does this mean for repeat homelessness?

• We are currently assessing the impact of the removal of National Insurance Number on the repeat homelessness calculations.
• Early analysis indicates the impact to be minimal, and if this is the case, repeat homelessness could still be produced by Scottish Government (with a caution when comparing trends).
• Once quarter 4 processing is complete, we will be in a better position to make an assessment and final decision.

Slide: HL2 aggregate temporary accommodation data

• Local Authorities have been asking for HL2 to be discontinued since the start of the collection of HL3
• HL3 is considered better quality and more accurate. Additionally, HL3 has greater analytical power and can provide greater insights.
• Under the data review we have agreed only placement level data will be collected.
• Therefore, we have decided to bring forward the discontinuation of HL2. The 2025/26 collection period is the last collection period for HL2.
• From 1 April 2026, the headline temporary accommodation figures will be based on HL3. Trend data will be backdated to 2017/18 where possible.
• We will issue appropriate communication on the changes.

Annex C: Data Quality slides

Slide: Overview

• We are continuing to carry out targeted data quality work across several key areas of our data collections.
• Our current focus includes:
  • Previous Application (PREVAPP)
  • HL3 (temporary accommodation placements)
  • “Not offered” temporary accommodation cases
  • Asylum-related categories, including property type and reason for homelessness
• The aim is to improve data accuracy, completeness and consistency. We are doing this through a combination of routine validation checks, cross-dataset comparisons, and ongoing engagement with local authorities to resolve identified issues.

Slide: PREVAPP and repeat homelessness

• PREVAPP (previous application number) is used to help calculate repeat homelessness.
• Local Authorities can provide PREVAPP, or
• The processing system looks through previous applications and, where National Insurance Number and household circumstances are the same, populates PREVAPP.
• With the removal of National Insurance Number this won’t be possible.
• This will impact on the quality of the repeat homelessness calculations.
• We have been monitoring the removal of NIN on the impact of finding previous applications whilst re-developing the processing system.
• It shows a large variation between local authorities. Between <0-32%.
• To improve the quality of the future of repeat calculations we are actively encouraging local authorities to complete PREVAPP.
• We will include a PREVAPP completion rate over time in the quarterly annual reports to aid with this.
• Additionally, during the re-development, we have noticed our current system does not always correctly flag the error ‘Previous application not in the system’ (error167).
• For your awareness, error167 will be flagged more in the error reports provided to local authorities. Enhancing the quality of PREVAPP.

Slide: HL3

• Ongoing validation work on HL3 data, which will become the headline measure for temporary accommodation
• Latest findings:
  • Nationally, HL3 is 6% higher (1,066 cases) than HL2
  • Previously, differences ranged from 2–4%
• Aim:
  • Reduce discrepancy as close to 0% as possible
• Routine data quality checks on HL3 will continue

Slide: HL3 additional checks

• In addition, further checks are also being carried out ahead of the move to HL3 as the sole reporting source:
  • All HL3 records must be associated with an HL1 record.
  • Entry date (HL3) vs HL1 closure date
  • Exit date (HL3) vs HL1 closure date
  • Number of children (HL1 vs HL3)

Slide: HL3 records and associated HL1

• All HL3 records must be associated with an HL1 record
  • Checks in place assure all HL3 records contain an HL1 reference.
  • However, not all HL1 references provided can be matched to an existing record.
  • This could indicate an error in the HL1 reference provide and/or an error with a previously provided HL1 application number.

Slide: HL3 entry date vs HL1 close date

• There are cases where the HL3 entry date occurs after the HL1 application closed.
• This should not happen because the closure of HL1 indicates that the homelessness duty has ended. If the household has been placed in accommodation after the HL1 has closed, this is not associated with homelessness duties and should not be included in the HL3 return.
• However, these discrepancies may be due to data entry errors:
  • Dates reported incorrectly
  • Incorrect application references

Slide: HL3 exit date vs HL1 close date

• Some HL3 records associated with a closed HL1 have no exit date or an exit date after the HL1 closure date.
  • We would expect these placements to end when the HL1 case closes.
  • This may indicate that HL3 records have not been updated.

• Note: We recognise that in some these cases, placements may remain open beyond the HL1 closure date for valid reasons, and we are working with local authorities to understand this better

Slide: Children HL1 vs HL3

• HL1 records the number of children associated with homelessness applications.
• HL3 records the number of children in temporary accommodation placements.
• We have undertaken a comparison of these returns and identified cases where the number of children in HL1 and HL3 do not match.
• While we understand there will be genuine differences – which was the purpose of including this data in the HL3 return -  large discrepancies may indicate data reporting issues.

Slide: HL3 QA Action – Local Authorities

• We have identified cases across all quality checks where data   issues are present for each local authority
• A list of affected cases has been shared with each local authority
• Where cases may be genuine, we have:
  • Requested further explanation to improve our understanding

•  Where cases are identified as data errors, we have:
  • Requested corrected data submissions
  • To enable updates to national datasets and publications

Slide: Not offered (temporary accommodation)

• Assess whether reporting practice when temporary accommodation was not able to be offered was consistent.
• Local authorities were asked how they would report against a number of (theoretical) scenarios.
• 24 local authorities responded, with 18 responses being able to be used to draw conclusions.

Evidence suggested reporting practice is largely consistent, and any discrepancy minor.

• 10 LAs reported in line with ‘expected’ responses for all scenarios
• 6 LAs reported in line with ‘expected’ responses for all but one scenario
• 2 LAs reported in line with ‘expected’ response for all but three scenarios
• The scenarios with most discrepancy were those where >1 instance of not being able to offer should be recorded.
• Guidance has been updated to make reporting practice clearer.

• However! Operational practice may vary locally, which could have an (indirect) impact on reporting.
• It is up to individual local authorities to put in place services that meet statutory requirements and best serve the needs of their communities.
• It is not for statistical guidance to inform operational practice.

Slide: Asylum categories

• New categories introduced:
  • Property Type “Asylum accommodation”
  • Reason for Homelessness “Left asylum accommodation”

• Monitoring focus:
  • Completeness of data
  • Ensuring accuracy and consistency of reporting across local authorities i.e., where a household reports “left asylum accommodation” as the reason, then property type must be “asylum accommodation”

Annex D: Data Review slides

Slide: Work to date

• Paused in June 2025 for the re-development work to commence.
• Up to that point, excluding prevention, all content was agreed by working group.
• We were in discussion with working group plus local authority IT suppliers about data layout and format.
• The working group and contractors emphasised that the new specification should incorporate prevention content prior to the commencement of implementation.

Slide: Overview of changes

What have we agreed to collect?

• information on homelessness applications (all stages);
• associated temporary accommodation placements; and
• prevention of homelessness data. This will be aligned with the Housing Act and discussions with policy about what to include have started.

Overarching changes

• Information to be gathered for all household members
• More information on protected characteristics
• Personal information (to enable data linkage)

To help quantify the changes for individual variables, there are:

• 14 with no change
• 20 with change between the current and new specification
• 23 new
• 11 replaced
• 4 removed

For indicators there are:

• 20 Fully comparable
• 2 Partial comparability
• 8 Not comparable

Slide: Example of new variables feasible to collect

• Forename, Surname
• More protected characteristics
• Education Status
• When member joined (household), left and why they left
• Main reason for not taking up the temporary accommodation
• If the accommodation is unsuitable, why is it unsuitable?

Slide: Examples of wholly comparable indicators

• Number of applications and live cases
• Number of people associated with applications (adults and children) assessed as homeless/threatened with homelessness
• Average time (days) through stages of the application and in temporary accommodation
• Households in temporary accommodation and number of children in temporary accommodation
• Number of households entering and exiting temporary accommodation
• Offers of temporary accommodation refused by the applicant
• Instances where households requiring temporary accommodation have not been offered it
• Number of breaches of the Unsuitable Accommodation Order

Slide: Examples of non-comparable indicators

• Sleeping rough
• Main reason for making an application 
• Households re-assessed as homeless (repeat homelessness)
• Property type from which the household became homeless

Slide: Next steps – 2026

• Work on data review is to resume in June 2026.
• Our next steps are focused on development of prevention content.
  • Discussion with policy – June 2026
  • Working group / topic group to discuss and agree content – late August / September 2026
  • External stakeholder engagement – October 2026

• Finalise data specification internally by end of 2026. Working group to review early 2027.

Slide: Next steps – 2027

• 2027 steps are focused on getting sign off on the new data specification and agree an implementation plan.
• This will be a two-step process.
  • Step 1 – all local authorities will be given the opportunity to raise significant concerns over the new data specification and feedback on an implementation plan.
  • Step 2 – all local authorities will be asked to sign off the implementation plan for the new collection.

• By the end of 2027, the aim is to publish the new specification for implementation to begin.

Slide: Implementation

Implementation of the collection

• Feedback from the local authorities suggests that implementation would take a minimum of two years.
• We will revisit this as part of our implementation plans during sign off

During implementation of the collection

• Create new IT and processing systems
• Data governance documents, e.g. data sharing agreements
• Design outputs

Slide: Data review working group

If your local authority is not a part of the data review working group and would like to be, please contact: anna.mackinnon@gov.scot

Annex E: NRS Homelessness deaths

Slide: Deaths of Homeless People in Scotland

[The following main findings were outlined, alongside charts – not included here – that illustrated these.]

• Homeless deaths were at a similar level to the previous year
• Around half of all homeless deaths were people aged under 45 in 2024
• The rate of homeless deaths rate varied across Scotland in 2024
• Around half of all homeless deaths were due to external causes in 2024
• Around a third of homeless deaths were drug misuse deaths in 2024

Slide: Methodology

• Homelessness status not recorded on death certificates
• Multiple searches of death registration records to identify evidence that the deceased was homeless
• Age cut-offs (15 to 74)
• Statistical modelling using capture-recapture technique to estimate a total figure

Slide: Search strategies

1. Place of residence : Terms such as “no fixed abode”, “homeless” or “shelter” or the address of a known homeless hostel or project appearing in the recorded place of residence
2. Place of death : Terms such as “no fixed abode”, “homeless” or “night shelter” or the address of a known homeless hostel or project appearing in the recorded place of death
3. Institution codes : presence of a location code designating a homeless hostel or shelter as the place of death
4. Hospital deaths : where the person died in hospital but no place of residence was recorded
5. Registrar notes : Terms such as “no fixed abode”, “homeless” or “shelter” appearing in the registrar notes on the death registration record

Slide: Local authority involvement

• Temporary accommodation lists
• Quality assurance
• Known deaths data stream

Slide: Limitations

• Best estimate from information available
• Informant may not know that deceased was homeless
• Homeless people staying with friends/relatives unlikely to be picked up
• Rough sleepers more difficult to pick up (although not impossible)
• Estimation model produces conservative estimates – true number may be higher
• Estimate is calculated at Scotland level – the same ratio is applied to sub-national breakdowns

Slide: Next steps

• New or improved data streams
• Improvements to address matching
• Journey to Accredited Official Statistics
• Welcome feedback and suggestions for improving the methodology or new data sources