Deaths of Homeless People in Northern Ireland - Methodology

Vital Statistics Unit, Northern Ireland Statistics and Research Agency

11 December 2025

Background

Deaths of people experiencing homelessness can be difficult to accurately capture. These new estimates have been derived using a specific process and a definition of homelessness which seeks to identify individuals based on the circumstances at the time of their death rather than whether they were statutorily homeless or not. For the purposes of this publication, the identification of this cohort is achieved by utilising information in death registration data. The deaths identified include people who were using emergency accommodation such as homeless shelters, hostels and temporary homeless accommodation at the time of their death, or those who have died in hospital and have no residential usual address. These statistics do not include those who are sleeping rough, unless this is explicitly stated in the death registration records.

However, sole reliance on information provided to registrars by an informant when a death is registered brings certain known limitations. The informant is usually a family member or a friend, but in some cases they may be a police officer, a social worker, a funeral director or other person. In these cases, the informant may not have all of the information needed to determine whether the person was homeless at the time of their death. Even when the informant is a family member, they may not have been in touch with the deceased for some time and may be unaware of their circumstances, or they may not wish the fact that their relative was homeless to appear on the death certificate.

There are other reasons why counts of people dying while homeless might not be accurate. For example, individuals can present to the Northern Ireland Housing Executive (NIHE) as homeless but not take up an offer of homeless accommodation, or those who are sleeping rough may not officially present themselves as homeless.

In an attempt to address these known issues, a statistical process called ‘capture recapture’ has been employed alongside the death registration data. This process helps estimate the size of a population which is difficult to count directly.

To date, there has been no official publication of data relating to homeless deaths in Northern Ireland. The NIHE have previously published information under the Freedom of Information Act (2000).

Comparability with other UK Regions

Since the aim of this publication is to maximise comparability with other UK regions, this methodology is similar to that of the National Records of Scotland (NRS) and the Office of National Statistics (ONS), with a few differences. For example, NRS carry out two additional searches, searching for cases that have an institution code that specifies ‘hostel’ or ‘shelter’, and searching registrar notes for terms such as ‘no fixed abode’. However, NISRA do not have an institution code that specifies ‘hostel’ or ‘shelter’, and access to registrar notes was not available. While a search of hospital deaths was conducted, this was limited, due to a lack of information on length of stay. Further refinement of searches is planned. The NRS was involved in peer reviewing this work to ensure that this methodology is in line with what is used by NRS.

Rcapture Methodology

The information recorded at death registration does not state or clearly indicate whether an individual was homeless at the time of their death. Therefore, a search of death registration records for any alternative evidence of homelessness at the time of death was needed. The records identified are mainly those people using emergency accommodation such as homeless shelters, hostels, and temporary homeless accommodation at the time of their death. Addresses for such accommodation was provided by the NIHE and identified death registrations were quality assured by NIHE staff supporting the development of these statistics. An additional search was conducted for hospital deaths where the deceased’s usual address was also the hospital, and not a residential address. These addresses were then matched to the database, using the search strategies detailed below.

These searches alone are not guaranteed to capture an accurate estimate of homeless deaths, due to a potential undercount (for example, rough sleepers who have not presented as homeless or those staying with friends or relatives until permanent accommodation is found) and overlap between searches. To address this, the capture-recapture method was used to get an estimation of homeless deaths.

Capture-recapture is a statistical method used to estimate the size of a population when it is impractical or otherwise difficult to count every individual or occurrence directly. The calculations estimate the likely size of an unknown closed population based on multiple captures (searches). The model looks at how many deaths were identified in each of the three searches and the degree of overlap between searches (it is possible for a death to appear in more than one search). It then provides an estimate of the likely number of deaths which were not picked up in these searches to give an estimate of the total number of homeless deaths.

The Chao model was used in the analysis, which is a robust but conservative model, specifically designed for situations with limited data.

First Capture (Search One): A sample of individuals from the population is marked (or identified), and then returned back to the population.

Second and Third Captures (Searches Two and Three): Further samples are captured. Some of these individuals will be marked from the first capture, and some will be unmarked.

Estimation: The proportion of marked individuals in the final sample is used to estimate the total population size.

Search Strategies

Figure 1: Flow Chart of the Search Strategy Procedure

Search One - Place of Residence

The usual address on the death registration record was searched for terms such as “no fixed abode”, “shelter”, “homeless”, “rough sleeper” or “no address known”. Matches were identified against the list of addresses provided by the NIHE.

Search Two - Place of Death

The place of death on the death registration record was searched for terms such as “no fixed abode”, “shelter”, “homeless”, “rough sleeper” or “no address known”. Matches were identified against the list of addresses provided by the NIHE.

Search Three - Hospital Deaths

Deaths where both the place of death and usual address were a hospital, indicating the deceased did not have a residential home at admission (note: there is no facility at present to identify how long the person was in hospital, meaning this search could include long-term patients who are not necessarily homeless).

The list of identified cases from the above search strategies was then sent back to the NIHE for a process of quality assurance, where the NIHE verified whether the address was in use as housing for homeless individuals on the date of death that was identified.


Please note: A lower age limit of 15, and an upper age limit of 74, was applied to the data to keep in line with the methodology of NRS and ONS, who applied these limits to avoid accidental inclusion of elderly people in some institutional settings.

Limitations of these statistics

These statistics are official statistics in development, and represent the best estimate of the number of deaths of people experiencing homelessness, but there are limitations which users should be aware of.

Information provided to registrars by the informant at the time of death registration forms the basis of the process. The informant is usually a family member or a friend, but in some cases the informant may be a police officer, a social worker, a funeral director or other person. In these cases, the informant may not have all of the information needed to determine whether the person was homeless at the time of their death. Even when the informant is a family member, they may not have been in touch with the deceased for some time and may be unaware of their circumstances, or they may not wish the fact that their relative was homeless to appear on the death certificate.

Although multiple searches of available data sources have been carried out to identify homeless people, there are a number of groups that who may have been missed or underestimated in this process.

A small percentage of people present to the NIHE as homeless but never take up an offer of homeless accommodation. They may be able to stay with friends or relatives until permanent accommodation is found for them. They are unlikely to be identified in the searches that have been carried out.

A number of people who are sleeping rough may not present themselves as homeless to the NIHE and are therefore not registered as homeless. If they die while sleeping rough, they may be picked up by these searches but only if the informant provides this information to the registrar when registering the death.

The capture recapture estimation technique used attempts to address this undercount, but the method used is known to produce conservative estimates - the actual number of homeless deaths may be higher.

The method did not allow any estimated deaths to be allocated to local government districts where there were no actual identified deaths of homeless people in the relevant year. This means that a small number of deaths may have occurred in areas that are shown as having no deaths in these figures. Additionally, due to small numbers, local government district figures can fluctuate from year to year and care should be taken in making any inferences from annual changes.

Deaths of homeless people may involve certification by a coroner. Due to the length of time it can take for an inquest to be completed, some of the deaths registered in (for example) 2024 will have occurred in previous years, while some deaths that occurred in 2023 (for example) will not yet be included in the figures.

Improving Methodology

These statistics are experimental, and the methodology is still under development. It is recognised that there may be better methods to estimate homeless deaths or better sources of information to enhance the current methodology. These methods are being published at an early stage to involve users and stakeholders in assessing their suitability and quality. In future, additional data-sharing possibilities between NISRA and the NIHE will be explored. Feedback from users is encouraged on ways in which the methodology or data sources could be developed to improve the quality of these statistics in future years.

Contact Details

- Is this something you find useful?

- Are there alternative or additional analyses you would like to see included?

Any feedback you may have which you can provide through the online survey at the link below:

Vital Statistics User Survey

Please e-mail comments to

NISRA Vital Statistics Unit
Colby House
Stranmillis Court
Belfast
BT9 5RR
Telephone: +44 (0)300 200 7836