Overview

This study examines deaths among 12,766 people aged 18–59 who were released from prison in Northern Ireland (NI) between April 2011 and March 2020. It includes deaths occurring up to March 2021 that were registered by March 2022.

The research examines the scale, timing and causes of death following the most recent release, with a particular focus on the first year. It also compares patterns of mortality among people released from prison with those in the Northern Ireland population.

Key findings

Substantially higher risk of death after release from prison

People released from prison in Northern Ireland faced a markedly elevated risk of death.

  • Among the 5,520 people with a single committal, 79 people died. This was approximately 1.4% of this group.

  • The total number of deaths among former prisoners with a single committal observed within one year of release was more than 11 times the number expected based on mortality rates in the Northern Ireland population, taking account of differences in age.

  • Of the 12,766 people in the study, 384 died within one year of their latest release. This was approximately 3.0% of the people included.

  • The total number of deaths among former prisoners observed within one year of release was more than 24 times the number expected based on mortality rates in the Northern Ireland population, taking account of differences in age This estimate is based on each person’s latest release and may overestimate average mortality across all releases.

Non-natural causes drive post-release deaths

Most of the 384 deaths within one year of release were from non-natural causes:

  • 54.4% were drug-related.
  • 27.3% were due to suicide.

Share of Northern Ireland drug-related deaths

People in the study accounted for more than one quarter of all drug-related deaths in Northern Ireland during the study period:

  • 14.6% of all drug-related deaths in Northern Ireland involved people from the study within one year of their latest release.
  • 27.5% involved people at any time after release during the study period.

High-risk period immediately after release

Deaths were particularly concentrated in the period immediately following release:

  • There were 44 deaths within 14 days of release.
  • The average number of deaths per day during the first two weeks was about three times the average for the first year.
  • Of these 44 deaths, 70.5% were drug-related.

Strong links with deprivation and disadvantage

Deaths were heavily concentrated among people associated with areas of high deprivation.

  • Of all deaths within one year of latest release, 70.3% involved people who had at least one recorded pre-release address in the 10% most deprived areas of Northern Ireland.

  • Compared with those who survived for at least one year, people who died were more likely to have:

    • multiple previous committals
    • longer periods on remand
    • property-related offences (property-related offences comprise burglary, fraud, criminal damage and theft)

Substance use a major driver

Drug-related deaths accounted for an increasing proportion of deaths over the study period:

  • The percentage of deaths within one year of latest release that were drug-related rose from 26.7% in 2011/12 to 72.0% in 2019/20.

In the analysis of specific drugs, opioids and benzodiazepines were often mentioned on death registrations. Among the deaths examined:

  • opioids were mentioned in 37.7%, compared with 4.8% of Northern Ireland deaths;
  • benzodiazepines were mentioned in 33.2%, compared with 3.5%; and
  • both were mentioned in 28.4%, compared with 3.0%.

Drug-related deaths are concentrated in the most deprived areas:

  • 76.6% of drug-related deaths involved people with a recorded pre-release address in the 10% most deprived areas.

Drug-related deaths were particularly prominent among younger people:

  • Over 70% of the 96 deaths among people released aged 18–24 were drug-related.

  • Of all drug-related deaths among young people in Northern Ireland, 28.7% involved people within one year of release from prison.

Clear opportunity for early intervention

  • The concentration of deaths shortly after release suggests that the period leading up to release and the transition from custody to the community represent a critical window for action.

  • Reducing preventable deaths will require coordinated responses across government, with a focus on continuity of care and support during transition from custody to the community.

Background

International research consistently shows that people recently released from custody face a substantially higher risk of death compared with the general population. The risk is particularly elevated during the immediate post-release period, with non-natural causes of death, such as drug-related deaths and suicides, being especially prominent. This is often associated with changes in drug-tolerance and a concentration of pre-existing vulnerabilities among people who have experienced imprisonment (see references 1, 2, 3).

This study examines mortality among people who have been in prison in Northern Ireland focusing on scale, timing, causes of death, and the role of drugs. The research addresses a key evidence gap for this hard-to-reach population and is one of the first studies of its kind in Northern Ireland.

A further aim is to demonstrate the value of securely linking administrative data from multiple organisations and making them available for accredited researchers. Such data linkage enables more robust analyses of outcomes among people who have experienced imprisonment. The findings show important links, but they do not show that imprisonment alone causes these outcomes. Instead, they help improve understanding and support more effective policy and better joined-up services.

What we did

This research project linked prisoner records from the Northern Ireland Prison Service with death registrations from the Northern Ireland General Register Office. The linkage was carried out under the data sharing provisions of the Digital Economy Act 2017 which enables data sharing for research in the public interest.

The prison dataset included each individual’s most recent release, with information such as age at release, committal date, time spent in custody and principal offence category. Summary information on previous incarcerations, such as total number of committals, was also provided. For people linked to a death registration, additional data included time from release to death, underlying cause of death, and mentions of specific substances on the death certificate.

To comply with the Digital Economy Act 2017 and ensure confidentiality, some variables were restricted. For example, sex could not be included because Northern Ireland has only one female prison, which would risk disclosure of the establishment.

The anonymised dataset contained 12,766 people aged 18–59 years who had been released from prison between April 2011 and March 2020. The analysis was restricted to this age range to ensure sufficient sample size and robustness of estimates for comparisons across groups. Deaths occurring up to March 2021 and registered by March 2022 were included, ensuring at least one year of follow-up after each person’s latest release. People who had been extradited or had no recorded home address were excluded because linkage to death registrations could not be reliably completed.

This proof-of-concept study was designed to demonstrate to policymakers and data owners that secure data transfer, linkage, and analysis can be carried out effectively. Although based on a restricted dataset (see Limitations section), the findings provide important insights and demonstrate the value of further analytical work.

What we found

Scale and causes of mortality

Between April 2011 and March 2020, 12,766 people aged 18–59 were released from prison in Northern Ireland. Of these, 384, or 3.0%, died within one year of their latest release. By March 2021, a total of 1,022 had died following their latest release.

Mortality in the year following release was substantially higher than expected compared with the general population (standardised mortality ratio (SMR) = 24.3, 95% CI 21.9–26.8), indicating that deaths occurred at more than 24 times the expected rate.


Post-prison mortality in context

To understand mortality after release from prison, researchers focused on the first year following each person’s latest release, providing the same length of follow-up for everyone in this analysis.

Post-prison deaths are compared with those among people of the same age in the Northern Ireland population using a standardised mortality ratio (SMR). The SMR compares the number of deaths observed (in the post-prison population) with the number expected if the group had the same age-specific death rates as the Northern Ireland population. This provides a measure of how much higher or lower the risk of death is relative to the general population. A 95% confidence interval (CI) is also presented; this represents the range within which the true value is likely to lie. Further information on the calculation of SMRs is provided in the Technical section.

This part of the analysis focuses on individuals’ latest release from prison; earlier periods of custody and prior releases are not included. As a result, the all-cause SMR reflects mortality following latest release and may overestimate average mortality across all releases. Cause-specific SMRs have not been produced for the full study population due to this limitation, which is discussed in detail in the Protecting identities section of the research limitations.

Mortality in the first year following latest release from prison was approximately 24.3 times higher than expected, based on age-specific mortality rates in the Northern Ireland population, indicating an extremely elevated risk during this period.

Note: This analysis is based on latest releases only and may therefore overestimate overall mortality risk across all release episodes.

Direct comparison with studies in other countries is difficult because the study populations, follow-up periods and methods differ. For context, the estimate in this study is higher than the SMR of 3.3 reported for Scotland (see reference 4) and higher than many estimates reported internationally. International estimates range from 1.0 to 9.4 for men and from 2.6 to 41.3 for women (see reference 3). These comparisons should be interpreted cautiously, particularly as sex was not available in the Northern Ireland data. No comparable research has been identified for the Republic of Ireland. No directly comparable all-cause SMR had been identified for England and Wales, where recent studies report cause-specific mortality rates only (see reference 5).


The study allowed detailed exploration of the underlying cause and circumstances of deaths within one year of release, with comparisons to all deaths in Northern Ireland. All coding is described in Annex 1.

In the first year following latest release, there were 384 deaths, of which:

  • 54.4% were drug-related
  • 27.3% were due to suicide
  • 10.9% were drug-related suicides (included in both categories above)
  • 6.8% were alcohol-specific
  • 2.6% were due to assault

Figure 1 shows that causes of death differ markedly between people who had been in prison and the Northern Ireland population of the same age.


Figure 1. Underlying cause of death within one year of latest release from prison (April 2011 – March 2021)

Disproportionate burden of non-natural deaths among persons after release from prison compared to Northern Ireland. Source: Research data and Vital Statistics Unit, NISRA.
Chart

Stacked column chart showing the distribution of the cause of death of of post-prison community (who were released from prison between April 2011 and March 2020) within one year of latest release. The detail is discussed in the commentary above.

Data
Cause of death Deaths of persons within one
year of release from prison
(n=384) 		All deaths
Northern Ireland
(n=19,000)
Drug-related 54.4% 7.5%
Suicide 27.3% 10.5%
Alcohol-specific 6.8% 9.8%
Assault 2.6% 0.6%
Other 19.8% 73.8%



Annex 2 highlights the most prevalent underlying causes of death by specific ICD-10 codes (International Statistical Classification of Diseases and Related Health Problems 10th Revision).

Contribution to deaths in Northern Ireland

After release from prison, people in the study who died within one year accounted for 1.9% of all deaths in Northern Ireland between April 2011 and March 2021; those who had died following their latest release during the study period accounted for 2.7% of all NI deaths. (See Figure 2.)

However, they made up:

  • 14.6% of all drug-related deaths within one year
  • 27.5% of all drug-related deaths when considering deaths following latest release across the full study period


Figure 2. Cause of post-release deaths as a proportion of Northern Ireland deaths (April 2011 – March 2021)

Significant contribution of post-release deaths to overall mortality in Northern Ireland, by cause. Source: Research data and Vital Statistics Unit, NISRA.
Chart

The proportion of deaths by cause to the post-prison community as a proportion of all NI deaths. The detail is discussed in the commentary above.

Data
Cause of death Post-prison deaths within one year of release
% (n) 		Deaths in Northern Ireland
18 - 60 years (n) 		Post-prison deaths during full study period
% (n) 		Deaths in Northern Ireland
18 - 69 years (n)
Alcohol-specific 1.3% (26) 1,941 5.0% (125) 2,500
Drug-related 14.6% (209) 1,435 27.5% (409) 1,488
Suicide 5.2% (105) 2,022 9.2% (202) 2,198
Assault 8.5% (10) 118 17.4% (23) 132
Other 0.5% (76) 15,300 1.0% (336) 32,541
All causes 1.9% (384) 20,381 2.7% (1022) 38,399



Note: Age groups for the post-prison population are based on age at release, whereas Northern Ireland age groups are based on age at death. To allow for people getting older between release and death, the upper age limit of each Northern Ireland age group was increased by one year when examining deaths within one year of release, and by 10 years when examining deaths across the full study period.

Over a quarter (28.7%) of drug-related deaths among people aged 18–25 in Northern Ireland between April 2011 and March 2021 were among people in the study who had left prison aged 18–24 and died within one year of their most recent release. The Northern Ireland comparison includes those aged 18–25 because people released aged 18–24 could be up to one year older when they died. Figure 3 illustrates the contribution of the post-prison population to suicide and drug-related deaths in Northern Ireland, broken down by age group.


Figure 3. Deaths within one year of release as a proportion of Northern Ireland deaths, by age at release and cause of death (April 2011 – March 2021)

Significant contribution of post-release deaths to suicide and drug-related mortality in Northern Ireland, by age group within one year of latest release. Source: Research data and Vital Statistics Unit, NISRA.
Chart

The proportion of drug and suicide deaths by age group to the post-prison community as a proportion of all NI deaths. The detail is discussed in the commentary above.

Data
Cause of death: Age group Deaths within one year of release
% (n) 		All Northern Ireland deaths
(n)
Drug-related deaths: Released aged 18–24; NI deaths aged 18–25 28.7% (68) 237
Drug-related deaths: Released aged 25–29; NI deaths aged 25–30 16.8% (46) 274
Drug-related deaths: Released aged 30–34; NI deaths aged 30–35 14.9% (43) 289
Drug-related deaths: Released aged 35–39; NI deaths aged 35–40 10.8% (27) 249
Drug-related deaths: Released aged 40–59; NI deaths aged 40–60 4.6% (25) 542
Suicide: Released aged 18–24; NI deaths aged 18–25 7.0% (25) 356
Suicide: Released aged 25–29; NI deaths aged 25–30 8.2% (26) 319
Suicide: Released aged 30–34; NI deaths aged 30–35 7.6% (23) 302
Suicide: Released aged 35–39; NI deaths aged 35–40 4.2% (13) 312
Suicide: Released aged 40–59; NI deaths aged 40–60 2.0% (18) 918
All: Released aged 18–24; NI deaths aged 18–25 9.4% (96) 1,024
All: Released aged 25–29; NI deaths aged 25–30 7.2% (72) 1,000
All: Released aged 30–34; NI deaths aged 30–35 5.8% (72) 1,237
All: Released aged 35–39; NI deaths aged 35–40 3.1% (49) 1,572
All: Released aged 40–59; NI deaths aged 40–60 0.6% (95) 16,421



Note: Age groups for the post-prison population are based on age at release, whereas Northern Ireland age groups are based on age at death. To allow for people getting older between release and death, the upper age limit of each Northern Ireland age group extends one year beyond the corresponding release-age group.


Timing of death after release

Deaths were particularly concentrated in the period immediately following release.

  • 44 deaths occurred within the first 14 days.
  • The average number of deaths per day during the first two weeks was about three times the average for the first year (see Figure 4).

This pattern (elevated risk in the weeks after release) is consistent with international evidence (see reference 6) and aligns with findings reported by the Prisoner Ombudsman for Northern Ireland.


Figure 4. Deaths by time from latest release (April 2011 – March 2021)

Mortality is concentrated in the period immediately following release from prison. Source: Research data
Chart

Data
Time after latest release Count of deaths Average daily deaths
Days 0 - 14 44 3.1
Days 15 - 30 34 2.1
Days 31 - 60 44 1.5
Days 61 - 90 37 1.2
Days 91 - 120 35 1.2
Days 121 - 150 23 0.8
Days 151 - 180 25 0.8
Days 181 - 365 142 0.8



Among people aged 18–24 years at release, 31.8% of deaths in the first year occurred in the first 14 days.

Drug-related causes dominate early deaths:

  • 31 of the 44 deaths (70.5%) in the first 14 days were drug-related.

Demographic and social patterns

In the first year after release, there were 96 deaths among people aged 18–24 years old at the time of release, 70.8% of which were drug‑related. In comparison, among the 859 deaths in the general Northern Ireland population aged 18–24 years, 22.5% were drug‑related. Similar patterns are evident across other age groups.

There was a higher proportion of deaths due to suicide for those leaving prison than in the general population, except for the age group 18–24 where the proportion of deaths due to suicide in the general population was higher (34.7%) than those following prison release (26.0%). Among those aged 40–59, there were 95 post‑release deaths, with 18.9% due to suicide. This is higher than the 6.0% of the 15,040 deaths in the general Northern Ireland population aged 40–59 that were due to suicide. Further detail is presented in Figure 5.


Figure 5. Age profile of post-release deaths due to drugs and suicide (April 2011 – March 2021)

Post-release deaths due to drugs and suicide by age group compared to Northern Ireland. Source: Research data and Vital Statistics Unit, NISRA.
Chart

This chart shows deaths by drugs and suicide broken down by age group. The detail is discussed in the commentary above.

Data
Deaths of persons after latest release from prison
All deaths Northern Ireland
Cause of death 18-24 25-29 30-34 35-39 40-59 18-24 25-29 30-34 35-39 40-59
Drug-related 70.8% 63.9% 59.7% 55.1% 26.3% 22.5% 28.0% 24.8% 17.3% 3.5%
Suicide 26.0% 36.1% 31.9% 26.5% 18.9% 34.7% 33.1% 25.4% 20.3% 6.0%
All causes
(n) 		96 72 72 49 95 859 813 1,012 1,276 15,040


Over two-thirds (70.3%) of deaths within one year of latest release involved people who had at least one recorded pre-release address in the 10% most deprived areas of Northern Ireland. The proportions were 76.6% for drug-related deaths and 63.8% for deaths due to suicide. See Figure 6.

Note: A person’s deprivation group was based on the most deprived area associated with any address held for them in the prison data before release. This was not necessarily their address immediately before custody or the address to which they returned on release. No address information after release was available. Deprivation deciles were assigned using the Northern Ireland Multiple Deprivation Measure 2017.

Additional tables are provided in Annex 3 showing the profile of post-release deaths by socio-economic classification and the proportion involving people with a recorded pre-release address in the 10% most deprived areas over time.


Figure 6. Pre-release area deprivation profile of drug-related deaths and deaths due to suicide (April 2011 – March 2021)

Post-release deaths due to drugs and suicide by deprivation decile of recorded pre-release address. Source: Research data.
Chart

Chart showing the deaths due to drugs and suicide by deprivation decile. The detail is discussed in the commentary above.

Data
Area deprivation decile All post-release deaths
(n=384) 		Drug-related deaths
(n=209) 		Suicides
(n=105)
1 - Most deprived decile 70.3% 76.6% 63.8%
2 - Second most deprived decile 12.5% 10.0% 14.3%
3-10 - Least deprived deciles 17.2% 13.4% 21.9%



The vast majority (84.4%) of those who died within one year of their latest release from prison had a marital status of single, compared with 43.0% of those who died in the same period in the Northern Ireland population. Figure 7 provides more details.


Figure 7. Marital status profile of post-release deaths (April 2011 – March 2021)

Marital Status of those who died within a year of latest release compared to all deaths in Northern Ireland. Source: Research data and Vital Statistics Unit, NISRA.
Chart

Deaths by marital status. Key differences are discussed in commentary above.

Data
Marital Status Deaths of persons within one
year of release from prison
(n=384) 		All deaths
Northern Ireland
(n=19,000)
Single 84.4% 43.0%
Married 8.9% 41.1%
Divorced/Widowed 6.8% 15.9%


Circumstances of death

Among deaths occurring within one year of latest release, 16.4% occurred in hospital, compared with 48.0% of deaths in the relevant Northern Ireland comparison population. See Figure 11.


Figure 11. Profile of place of death (April 2011 – March 2021)

Deaths less likely to occur in institutional settings compared with Northern Ireland population. Source: Research data and Vital Statistics Unit, NISRA.
Chart

Deaths less likely to occur in institutional settings compared with Northern Ireland population. The detail is discussed in the commentary above.

Data
Place of death Deaths of persons within one
year of release from prison
(n=384) 		All deaths
Northern Ireland
(n=19,000)
Hospital – including psychiatric, NHS and other 16.4% 48.0%
Other 83.6% 52.0%


Characteristics of those who died and survived

Comparing the characteristics of those who died within one year of release with those who remained alive beyond one year can help identify groups that appear to be at increased risk following release. In this section, “survivors” refers to people with no linked death record within one year of their latest release.

Figure 12 outlines the characteristics of those who died in the first year after latest release compared to the characteristics of survivors. Groups over-represented among those who died include:

  • Repeat contact with the justice system
    • People with five or more total committals (45.8% deaths compared with 16.8% survivors).
  • Area deprivation
    • People with a recorded pre-release address in the 10% most deprived areas (70.3% deaths compared with 51.4% survivors).
  • Offence profile
    • People with property-related principal offences (37.5% deaths compared with 24.4% survivors). Property-related offences comprise burglary, fraud, criminal damage and theft.
  • Custodial experience
    • People with longer cumulative time in custody (for example those with a total number of days on remand of over 30 (62.5% deaths compared with 37.2% survivors)).
  • Probation Board supervision
    • People who were not supervised by the Probation Board for Northern Ireland (92.4% deaths compared with 80.2% survivors).

Additional analyses were undertaken for people discharged from prison with their license requiring them to be under the supervision of the Probation Board for Northern Ireland. These findings are presented in Annex 7 and include cause of death, timing of death following release, and characteristics associated with mortality.


Figure 12. Characteristics of post-release deaths compared with survivors (April 2011 – March 2021)


Deaths of people with a single committal

Given the limitations of analysing mortality solely on the basis of the latest release from prison, this section focuses on people who had a single committal during the study period. For this group, the dataset provides a more complete measure of time at risk following release because no further committals were recorded during the study period. This allows mortality to be estimated more reliably than in the analysis based on latest release alone. Analyses based on latest release only do not take account of earlier periods following release during which people were living in the community.

Between April 2011 and March 2020, 5,520 people aged 18–59 were released from prison following a single committal in Northern Ireland. Of these, 79, or 1.4%, died within one year of their latest release. By March 2021, a total of 262 had died following release.

Mortality in the first year after release was significantly higher than expected compared with the Northern Ireland population (SMR = 11.7, 95% CI 9.1–14.3), indicating a notably elevated period of risk for people released after a single committal.


Post-prison mortality in context

Mortality in the first year following release from a single committal was approximately 11.7 times higher than expected based on age-specific mortality rates in the Northern Ireland population, indicating a high-risk period for this group.


Cause of death

Within one year of release from prison, looking at the 79 deaths that were identified:

  • 36.7% were drug-related
  • 35.4% were due to suicide
  • 13.9% were alcohol-specific

Figure 13 profiles the underlying causes of death for people within one year of release from prison compared with all deaths in Northern Ireland for 18–59 year olds between April 2011 and March 2021.

Note: Due to rules around statistical disclosure control, deaths due to assault have been included in the ‘other’ category.


Figure 13. Cause of death within one year of release from prison – single committal (April 2011 – March 2021)

Disproportionate burden of non-natural deaths among persons after release from prison compared to Northern Ireland. Source: Research data and Vital Statistics Unit, NISRA.
Chart

Disproportionate burden of non-natural deaths among persons after release from prison compared to Northern Ireland. The detail is discussed in the commentary above.

Data
Cause of death Deaths of persons within one
year of release from prison
(n=79) 		All deaths
Northern Ireland
(n=19,000)
Drug-related 36.7% 7.5%
Suicide 35.4% 10.5%
Alcohol-specific 13.9% 9.8%
Other 21.5% 74.5%



Cause-specific mortality varied substantially. Drug-related mortality was 35.7 times higher than expected (SMR 35.7, 95% CI 22.7–48.7). Mortality due to suicide was 26.1 times higher than expected (SMR 26.1, 95% CI 16.4–35.8), while alcohol-specific mortality was 18.6 times higher than expected (SMR 18.6, 95% CI 7.6–29.5). Mortality from all other causes was also elevated, at 3.8 times the number expected (SMR 3.8, 95% CI 2.0–5.6).


Post-prison mortality in context

This part of the analysis focuses on people with a single committal, providing a more complete measure of time at risk. All-cause and cause-specific SMRs are presented but should be interpreted with caution due to the small numbers involved.

  • Drug-related deaths: Mortality in the first year following release was 35.7 times higher than expected, based on age-specific mortality rates in the Northern Ireland population, representing an extremely elevated risk.

  • Suicide: Mortality in the first year following release was 26.1 times higher than expected, also representing an extremely elevated risk.

  • Alcohol-specific deaths: Mortality in the first year following release was 18.6 times higher than expected, indicating a very high elevation in risk.

  • All other causes: Mortality in the first year following release was 3.8 times higher than expected, indicating a substantially elevated risk.

Note: These cause-specific SMRs are based on small numbers and should therefore be interpreted cautiously.

The SMRs for drug-related deaths and deaths due to suicide are higher than similar figures relating to all releases reported internationally (drug-related SMR 15.58; see reference 7 and suicide SMR range 2.6 to 13.5; see reference 3). No comparable research has been identified for people following a single committal or for alcohol-specific deaths. Differences in cohort definitions, follow-up periods, and methodological approaches mean that these comparisons should be interpreted with caution.


Among people released after a single committal, mortality was higher than expected across all cause groups, with particularly high SMRs for drug-related deaths and deaths due to suicide (see Figure 14).


Figure 14. Standardised mortality ratios (SMRs) by underlying cause of death for those with a single committal (April 2011 – March 2021)

Significantly higher risk of death across all causes. Source: Research data and Vital Statistics Unit, NISRA.
Chart

Significantly higher risk of death across all causes. Key differences are discussed in commentary above.

Data
Cause of death Number of deaths Standardised Mortality Ratio
(95% Confidence Interval)
All causes 79 11.7
( 9.1 - 14.3)
Drug-related deaths 29 35.7
(22.7 - 48.7)
Suicides 28 26.1
(16.4 - 35.8)
Alcohol-specific 11 18.6
( 7.6 - 29.5)
All other causes 17 3.8
( 2.0 - 5.6)


Timing of death

Within the first year after latest release, 79 deaths were recorded. In the first 30 days, there were 12 deaths, nearly twice the average monthly figure of 6.6.

This sharp early peak, followed by a gradual reduction over time is shown in Figure 15. This pattern also indicates that the time immediately following release is particularly important for support for people released after a single committal to prison.


Figure 15. Deaths by time from release after a single committal (April 2011 – March 2021)

Mortality is concentrated in the period immediately following release from prison. Source: Research data.
Chart

Mortality is concentrated in the period immediately following release from prison. Key findings are discussed in commentary above.

Data
Time after latest release Count of deaths Average daily deaths
Days 0 - 30 12 0.4
Days 31 - 90 17 0.3
Days 91 -180 18 0.2
Days 180 - 365 32 0.2


Circumstances of death

Among deaths occurring within one year of release after a single committal, 15.2% occurred in hospital, compared with 48.0% in the relevant Northern Ireland comparison population. See Figure 16.


Figure 16. Profile of place of death for those with a single committal (April 2011 – March 2021)

Deaths less likely to occur in institutional settings compared with Northern Ireland population. Source: Research data and Vital Statistics Unit, NISRA.
Chart

Deaths less likely to occur in institutional settings compared with Northern Ireland population. The detail is discussed in the commentary above.

Data
Place of death Deaths of persons within one
year of release from prison
(n=79) 		All deaths
Northern Ireland
(n=19,000)
Hospital – including psychiatric, NHS and other 15.2% 48.0%
Other 84.8% 52.0%


Characteristics of those who died and survived: single-committal group

Comparing the characteristics of those who died within one year of release with those who remained alive beyond one year can help identify groups that appear to be at increased risk following release. In this section, “survivors” refers to people with no linked death record within one year of release.

For those with a single committal, Figure 17 compares the characteristics of those who died within one year of release with survivors. Groups over-represented among those who died include:

  • People who had a disposal of remand (i.e. held in custody pending trial or sentencing) (58.2% deaths compared with 40.4% survivors).
  • People with a recorded pre-release address in the 10% most deprived areas (49.4% deaths compared with 38.3% survivors).
  • People with property-related offences (31.6% deaths compared with 21.2% survivors). Property-related offences comprise burglary, fraud, criminal damage and theft.
  • People with a total number of days remanded of 8–29 days (24.1% deaths compared with 13.9% survivors).


Figure 17. Characteristics of single committal post-release deaths compared to survivors (April 2011 – March 2021)


Why it matters

This report fulfils two complementary aims.

First, it presents new evidence on mortality among people who have left prison in Northern Ireland, providing a clearer picture of the scale of death, when deaths occur and what the underlying causes are.

Second, it demonstrates the practical and strategic value of linking administrative data across organisational boundaries. This research shows that sensitive information can be securely transferred, processed and analysed in line with legal and ethical requirements, and that doing so generates insights that would not otherwise be visible.

Key findings and their policy relevance

The findings reinforce a consistent international evidence base showing that people leaving prison face an elevated risk of death compared with the general population.

Crucially, this risk is not evenly distributed over time. A disproportionate number of deaths occur shortly after release, highlighting the transition period as a particularly vulnerable stage. This has direct relevance for public agencies, including the Prisoner Ombudsman for Northern Ireland, who may investigate deaths occurring within 14 days of release where concerns are raised about care received in custody.

Within the first year following release, non‑natural causes account for the vast majority of deaths, with drug‑related fatalities most prominent and suicide the second leading cause. Existing literature indicates these are likely long-standing issues with a substantial proportion of people entering prison with pre‑existing substance use issues as well as a significant number using crisis and mental health services whilst in prison (see reference 8). The contribution of specific drugs, particularly opioids and benzodiazepines, underscores the need for coordinated public health and criminal justice responses to substance dependence both before and after release.

The analysis also shows that higher mortality risk is not confined to people with repeated or long‑term engagement with the justice system. People with a single committal still experience elevated risk, suggesting that broader health, social and economic disadvantages - rather than custodial criminal history alone - shape post‑release outcomes. This pattern mirrors wider research pointing to the influence of factors such as employment, social connection and stable housing (see reference 5). While some of these factors are visible in the comparison between those who survived and those who died, further analytical work is needed to better understand causal pathways and opportunities for intervention.

Implications for policy and practice

The findings, particularly the contribution of post-release deaths to the overall death figures in Northern Ireland, underscore that the criminal justice system - courts, prisons and probation services - has a critical opportunity to identify risk and intervene early, but that responsibility for improving outcomes extends beyond justice alone. The circumstances faced by people leaving custody cut across health, housing, employment, education, finance and community support.

While this study does not model future outcomes, the findings show that people released during the study period faced a substantially elevated risk of death, particularly in the first weeks after release. If similar patterns continue, the immediate post-release period is likely to remain an important point for preventive support. Many of these deaths are likely to arise from preventable causes, most notably drug-related deaths and suicide. The concentration of risk in the immediate post-release period highlights a predictable and time-critical window where intervention could have the greatest impact.

This research highlights the importance of:

  • identifying the characteristics and circumstances associated with greatest risk, supported by more comprehensive, up-to-date linked data, to inform targeted prevention;
  • continuity of care in the period leading up to release and throughout the transition from custody to the community, particularly for people with substance use or mental health needs;
  • supporting cross-government action, recognising that post‑release mortality is shaped by structural factors not confined to the justice domain.

Conclusion

By evidencing the scale and nature of post-prison mortality in Northern Ireland, this research provides important new evidence to support policy development and operational decision-making. It demonstrates that data linkage can play an important role in improving understanding of risk, informing the development of interventions and supporting a more coordinated, whole-system approach.

Ultimately, these insights can support efforts to reduce preventable deaths, enable safer transitions from custody to the community, and contribute to improved health, wellbeing, and social outcomes for people with experience of the criminal justice system.

What next?

The research team is engaging with key stakeholders, including the Department of Justice, Northern Ireland Prison Service, the Department of Health, the Department for Communities and the Housing Executive Northern Ireland to maximise the policy and operational relevance of this work.

The success of this proof-of-concept project demonstrates the potential of safe and secure data sharing to generate insights of real policy value. Administrative Data Research Northern Ireland (ADR NI) hopes this will encourage greater collaboration with data owners and policy leads within the justice system on future research.

Adherence to the legislative framework that enables such work, most notably the Digital Economy Act 2017, remains critical. At the same time, there are several opportunities to strengthen future research datasets while ensuring that individual identities remain fully protected through statistical disclosure control measures:

  • Expanding data sources. Linking to a population spine or other administrative datasets could improve verification of life status and enhance data quality. Greater use of health, education, employment and benefits data would provide a richer understanding of the circumstances and factors associated with pathways into, and outcomes following, involvement with the justice system.
  • Capturing all releases. Providing data across multiple release episodes (not just latest release) would enable more robust analysis of time at risk, while disclosure risks could be managed through careful post-linkage engineering (e.g. aggregation, reduced granularity).
  • Early researcher involvement. Engaging researchers throughout the dataset design process would ensure that data are structured in a way that supports answering key research questions, with sufficient metadata to assess utility.
  • Improved geography. Providing more consistent and timely address information, subject to disclosure controls, would support deeper analysis of place-based patterns and inform more targeted policy responses.

While the findings presented here are necessarily high-level due to dataset limitations, implementing these recommendations would substantially enhance the utility of future linked data. This would allow policymakers and practitioners to gain richer insights into the characteristics and circumstances of those most at risk of death following release from custody.

In parallel, ADR NI colleagues at Queen’s University Belfast are undertaking research on pathways from prison to community living, examining the intersection between health and criminal justice.

Together, these initiatives will provide a stronger evidence base to support interventions that reduce harm and improve outcomes for people leaving prison.

Research limitations

This research provides new, high-level insights in Northern Ireland on overall heightened risk of death for people who have left prison, especially in the period immediately following release and the prominence of non-natural causes of death.

As with any study based on administrative data linkage, a number of limitations should be acknowledged when interpreting the findings. These relate to the completeness and accuracy of the source data, the requirements of protecting individual confidentiality, and the constraints of using a “latest release only” dataset.

Data sources

  • Evidence of life. The absence of a linked death record following release does not necessarily indicate that an individual is alive and residing in Northern Ireland. This limitation is common to similar studies internationally.

  • Evidence of death. Not all deaths may have been successfully linked. Differences in identifying information (e.g. name, date of birth, address) between datasets can prevent linkage, and delays in death registration, particularly for non-natural causes, may mean some relevant deaths were not captured. The study is restricted to people released from prison between April 2011 and March 2020, linked to deaths occurring up to March 2021 that were registered by March 2022. While figures for overall deaths are presented, the analysis focuses on mortality within one year of release. Deaths registered after March 2022 that occurred during the study period were not included; these are listed in Annex 4 but cannot be attributed to the people in this study with certainty.

  • Timeliness of the data. The findings describe mortality among people released from prison between April 2011 and March 2020 and may not fully reflect subsequent changes in the prison population, services, release arrangements or patterns of substance use. More recent linked data would be needed to assess the current scale and nature of post-release mortality.

  • Evidence of cause of death. The Review of Suicide Statistics in Northern Ireland (published May 2022) resulted in methodological changes, most notably the reclassification of many drug-related deaths from “undetermined” (and thus within the suicide definition) to “accidental” (outside the suicide definition). Researchers considered the implications and concluded that this would not materially alter the overall findings. Additional information is provided in Annex 5.

Protecting identities

Given the relatively small scale of prison committals and deaths in Northern Ireland, stringent disclosure control measures were required to protect the identity of individuals and institutions, in line with the Digital Economy Act 2017. These measures included restricting the dataset to latest releases only, aggregating geographical detail, and suppressing certain variables (e.g. sex). While necessary, these safeguards reduce the scope for disaggregated analysis.

  • Latest release only. The dataset is structured around individuals’ most recent release, with only summary information on prior committals. This prevents robust estimation of the complete “time at risk” for people who have left prison and limits exploration of mortality risk in sub-groups. Analyses based on latest release only may overstate overall mortality risk by not accounting for repeated imprisonment and time spent in the community. Researchers considered reconstructing a release-by-release dataset using available summary information but concluded that the assumptions required would bias denominators and risk misinforming policy interventions. More details are provided in Annex 6.

  • Geography. Geographical information was available only at a broad level, covering Local Government District and deprivation decile. A person’s deprivation group was based on the most deprived area associated with any address held for them in the prison data before release. This was not necessarily their address immediately before custody or at release, and no address information after release was available. The measure may therefore not reflect where a person lived or their circumstances on returning to the community.

  • Sex. International evidence consistently highlights differences in post-release mortality by sex (see reference 6). However, because Northern Ireland has only one female prison and the Digital Economy Act prohibits identifying establishments, sex could not be included as a variable. Consequently, analysis of sex-specific mortality was not possible.

Taken together, these limitations meant that researchers could identify important high-level patterns but could not model the factors independently associated with mortality, determine precisely which characteristics were independently linked with greatest risk, or assess whether the scale and nature of risk have changed since the study period.

Where to go for help

If you are struggling to cope, please call one of the organisations below. There is help available around the clock, every single day of the year, providing a safe place for anyone struggling to cope, whoever they are, however they feel.

Minding Your Head - Find out more about mental health and the issues that can affect it; early warning signs that a mental health issue may be developing; tips on how to maintain good mental health.

Lifeline - A free 24 hour crisis response helpline for people who are experiencing distress or despair, where trained counsellors will listen and help immediately on the phone and follow up with other support if necessary. Phone 0808 808 8000

Samaritans – A registered charity aimed at providing emotional support to anyone in emotional distress, struggling to cope, or at risk of suicide throughout UK and Ireland, often through their telephone helpline or online chat. Freephone 116 123

Information for the media There is strong evidence that sensationalist media reports about suicide and the nature of suicide deaths can lead to subsequent additional suicidal behaviours (suicides and suicide attempts) or indeed increase the likelihood of copycat deaths. Media professionals should exercise caution in reporting on suicide, balancing the public’s “right to know” against the risk of causing harm. It is therefore important that those reporting on suicide statistics adhere to the guidelines of safe reporting from the World Health Organisation (WHO), the International Association for Suicide Prevention (IASP) and the Samaritans.

Technical notes

Standardised mortality ratios (SMRs)

Standardised Mortality Ratios (SMRs) were calculated as the ratio of observed deaths to expected deaths.

Expected deaths represent the number of deaths that would be expected if people released from prison experienced the same age-specific mortality rates as the Northern Ireland population.

Observed deaths were those occurring within one year of an individual’s latest release from prison.

Age-specific mortality rates for the Northern Ireland population were calculated using all deaths among people aged 18-59 years occurring between April 2011 and March 2021, divided by the cumulative age-specific mid-year population estimates for 2011 to 2021.

Expected deaths in the post-prison population were calculated by applying these Northern Ireland age-specific mortality rates to the corresponding person-years at risk accumulated by those released from prison. Expected deaths were then summed across all age groups.

The SMR was calculated as:

\[ SMR = \frac{\text{Observed deaths}}{\text{Expected deaths}} \]

An SMR greater than 1 indicates that mortality among people released from prison was higher than expected based on mortality rates in the Northern Ireland population, while an SMR less than 1 indicates lower than expected mortality.

The following age groups were used:

  • 18 - 24
  • 25 - 29
  • 30 - 34
  • 35 - 39
  • 40 - 44
  • 45 - 49
  • 50 - 54
  • 55 - 59

Age was defined as age at latest release from prison. As information on attained age during follow-up was unavailable, individuals remained in the age group assigned at release throughout the one-year follow-up period.

Small counts and grouping data

Counts below 10 were suppressed or grouped to meet disclosure control standards.

Acknowledgements

Administrative Data Research - Northern Ireland (ADR-NI) takes privacy protection very seriously. All information that directly identifies individuals/organisations will be removed from the datasets by trusted third parties before researchers access them. All researchers using ADR-NI are trained and accredited to use sensitive data safely and ethically, they will only access the data via a secure environment and all their findings will be vetted to ensure they comply with strict confidentiality requirements. The help provided by the staff of the Administrative Data Research Centre - Northern Ireland (ADRC-NI) and the ADR-NI support officers within NISRA Research Support Unit (RSU) is acknowledged. ADR-NI is funded by the Economic and Social Research Council (ESRC). The authors alone are responsible for the interpretation of the data and any views or opinions presented are solely those of the author and do not necessarily represent those of ADR-NI. Data from the Department of Justice and the General Register Office (GRO) has been supplied for the sole purpose of this project.

About ADR UK

ADR UK (Administrative Data Research UK) is a partnership transforming the way researchers access the UK’s wealth of public sector data, to enable better informed policy decisions that improve people’s lives. By linking together data held by different parts of government and facilitating safe and secure access for accredited researchers to these newly joined-up and de-identified data sets, ADR UK is creating a sustainable body of knowledge about how our society and economy function – tailored to give decision makers the answers they need to solve important policy questions. ADR UK is funded by the Economic and Social Research Council (ESRC), part of UK Research and Innovation.

Contact Details

Published by: Administrative Data Research Unit, Northern Ireland Statistics & Research Agency

Lead Statistician: Carmel Colohan

Telephone: 028 90 388465

Email: carmel.colohan@nisra.gov.uk

Accessibility contact

Please contact Dissemination Branch for assistance with accessibility requirements or alternative formats. Contact details are:

Email:

Telephone: +44 (0)300 200 7836

Dissemination Branch
NISRA
Colby House
Stranmillis Court
BELFAST
BT9 5RR

Annexes

Annex 1. Coding and definitions

Table 1. Underlying cause of death recode – deaths after all latest releases from prison

Charactersitic Underlying.cause.of.death ICD10.Codes
Cause of death Alcohol-specific E244, F10, G312, G621, G721, I426, K292, K70, K852, K860, Q860, R780, X45, X65, E24, G31, G32, G62, G72, I42, K29, K85, K86, R78, X45, Y15, K73, K74
Cause of death Drug-related F11-F16, F18-F19, X40-X44, X60-X64, X85, Y10-Y14
Cause of death Drug-related suicide X60-X64, Y10-Y14
Cause of death Assault X85-X99, Y00-Y09
Cause of death Suicide X60-X84, Y87.0, Y10-Y34, Y87.2
Cause of death Other All other ICD10 codes


Table 2. Principal offence recoding

Principal Offence Category Principal Offence Group Principal Offences
Other Drugs Drugs
Fraud & Miscellaneous Miscellaneous
Motoring Motoring
Public Order & Possession of Weapons Possession of Weapons
Public Order & Possession of Weapons Public Order
Person Burglary & Robbery Robbery
Violence Against the Person & Sexual (including Misc. Sexual) Other Miscellaneous Sexual
Violence Against the Person & Sexual (including Misc. Sexual) Sexual
Violence Against the Person & Sexual (including Misc. Sexual) Violence Against the Person
Property Burglary & Robbery Burglary
Criminal Damage Criminal Damage
Fraud & Miscellaneous Fraud
Theft Theft


Principal Offence: If an individual has only one offence, this will be their principal offence. Where a person is in custody for two or more offences, the principal offence is the offence with the statutory highest maximum sentence. If an individual is both sentenced and remanded for different offences, then the principal offence will always be a sentenced offence.


Table 3. Disposal groups recoding

Disposal Groups Disposal
Fine defaulter Adult Fine Defaulter
Juvenile Fine Defaulter
Young Offender Fine Defaulter
Remand Adult Awaiting Trial
Adult Immigration Detainee
Adult Remand
Juvenile Awaiting Trial
Juvenile Remand
Young Offender Awaiting Trial
Young Offender Immig Detainee
Young Offender Remand
Sentenced Adult Appellant
Adult Determinate Cust Sent
Adult Extended Cust Sent
Adult Indeterminate Cust Sent
Adult Lifer
Adult S.O.S.P.
Adult Sentenced
Juvenile Determinate Cust Sent
Juvenile Sentenced
Young Off Determinate Cust Sen
Young Off Extended Cust Sent
Young Offender Appellant
Young Offender S.O.S.P.
Young Offender Sentenced


Disposal groups: Disposal is the custodial category on the day of release


Table 4. Final release status recoding

Final release status group Final release status
Bail D.A.C. - Bail To Appeal
DAC - On Bail
Released - On Bail
Released-On Bail(E.Monitoring)
Left Northern Ireland (excluded from research data) Deported
Released - By Immigration
Released - Extradited
Released - Foreign National ER
Released - Hm Immigration Bail
Released - On GB Parole
Released - On Licence (GB)
Released - Repatriated
Released - To Imm Hold Centre
Released - To Uk Border Agency
Transfer To UK Prison
Time-served and release on licence D.A.C. - Time Served
DAC - Time Served On Remand
Release On CER Licence
Released - By Hospital Order
Released - Cjo Licence
Released - Discharge Warrant
Released - Life licence
Released - On Licence
Released - Order Of Court
Released - Secretary of state
Released - Time Served
Transfer To Juv Just Centre
Charges withdrawn and Non-custodial outcome D.A.C. - Withdrawn
DAC - Non Custodial
DAC - Not Guilty
DAC - On Acquittal
Released - Charges Withdrawn
Released - Committed In Error
Released - Fine Paid
Released - Non Custodial

Probation Board for Northern Ireland

Individuals who were discharged from prison with their license requiring them to be under the supervision of the Probation Board for Northern Ireland were identified as those with the following disposals and/or final release statuses:

  • Disposal:
    • Adult Determinate Cust Sent
    • Adult Extended Cust Sent
    • Adult Indeterminate Cust Sent
    • Adult Lifer
    • Young Off Determinate Cust Sen
    • Young Off Extended Cust Sent
  • Final release status:
    • Released - Life licence
    • Released - On Licence

Annex 2 - Deaths by specific ICD 10 codes

Table 5. Profile of underlying causes of death with a year of all latest releases from prison by individual ICD 10 codes with counts over 10. Source: Research data
ICD10 code and description Count of deaths % of all deaths Research coding
X44: Accidental poisoning by and exposure to other and unspecified drugs, medicaments, and biological substances (unintentional) 69 18.0% Drug-related
X42: Accidental poisoning by and exposure to narcotics and psychodysleptics (hallucinogens) (unintentional) 63 16.4% Drug-related
X70: Intentional self‑harm by hanging, strangulation, and suffocation 54 14.1% Suicide
X41: Accidental poisoning by and exposure to antiepileptic, sedative‑hypnotic, antiparkinsonism and psychotropic drugs (unintentional) 20 5.2% Drug-related
I25: Chronic ischemic heart disease (non-billable category); specifically, atherosclerotic heart disease of native coronary artery without angina (I25.10) 17 4.4% Other
Y12: Poisoning by and exposure to narcotics and psychodysleptics [hallucinogens], undetermined intent 15 3.9% Drug-related, Suicide, Drug-related suicide
K70: Alcoholic liver disease (non-billable category, includes alcoholic fatty liver, hepatitis, fibrosis, cirrhosis, hepatic failure) 14 3.6% Alcohol-specific
F19: Other psychoactive substance–related disorders, covering misuse of psychoactive drugs not classified elsewhere (e.g., polysubstance use) 12 3.1% Drug-related

Annex 3 - Additional analysis

Table 6. Socio-economic position of those who died within a year of latest release compared to all deaths in Northern Ireland. Source: Research data and Vital Statistics Unit

Socio-economic classification Deaths of persons within one
year of release from prison
(n=384) 		All deaths
Northern Ireland
(n=19,000)
Managerial and Intermediate (NS-SEC: 1-3) 4.2% 21.1%
Small employers and supervisory (NS-SEC: 4-5) 7.6% 9.7%
Routine and semi-routine (NS-SEC: 6-7) 24.5% 26.5%
Never worked or long-term unemployed (NS-SEC: 8) 12.2% 4.5%
Not classified (NS-SEC: 9) 39.8% 24.3%
Not applicable (NS-SEC: NA) 11.7% 14.0%


Deaths within a year of latest release from prison are markedly less concentrated in higher and intermediate occupation groups compared to deaths in the Northern Ireland population (4.2% vs 21.1%). Table 6 also shows that while the distribution across routine and semi-routine occupations is broadly similar, persons after latest release from prison are over-represented in the not-classified category (39.8% vs Northern Ireland 24.3%) and among those who have never worked or were long-term unemployed (12.2% vs Northern Ireland 4.5%).

Table 7. Deaths for those with a recorded pre-release address in the 10% most deprived areas. Source: Research data

Year of release Number of deaths to people in most deprived areas Total number of deaths within one year of latest release Percentage of deaths in most deprived areas
2011/2012 22 45 48.9%
2012/2013 23 39 59.0%
2013/2014 19 33 57.6%
2014/2015 33 47 70.2%
2015/2016 30 39 76.9%
2016/2017 32 41 78.0%
2017/2018 37 48 77.1%
2018/2019 31 42 73.8%
2019/2020 43 50 86.0%

Annex 4 - Deaths not included in research data linkage

Table 8. Number of deaths in Northern Ireland by cause, April 2011 – March 2021, registered after March 2022 so not included in the research data. Source: Vital Statistics Unit.

Deaths not included in linkage between people who had left prison between April 2011 and March 2020 to death registrations in Northern Ireland between April 2011 and March 2021 so not included in research data. Source: Vital Statistics Unit, NISRA.
Cause of death Number of deaths
Alcohol 5
Drug-Related 29
Assault 28
Suicide 32
Other 93


Annex 5 - Review of Suicide Statistics Northern Ireland

Deaths registered between 2015-2020 were reviewed with subsequent registrations seeing the review recommendations applied. The deaths data used in this research was shared on 24 June 2022 and contains the updated cause of death, as a result of the review, for the majority of deaths registered in 2015 – 17 and for all deaths from 2018 onwards. However, a subsequent release in November 2022 (specifically 75 unreviewed registrations in 2015–2017) was not available when the linkage took place so any changes to cause of death as a result of this have not been included in this research. We are unable to determine whether any of these 75 registrations relate to people released from prison. It is known that just over half were moved from intentional to accidental categorisations but all remained within the non-natural causes of death.

Annex 6 - Latest release only

In order to protect the privacy of people released from prison and abide by the legislative requirements for sharing de-identified data under the Digital Economy Act 2017, any data shared and made available to researchers must be managed carefully. In this bespoke data build primary information is available that relates to a person’s latest release, with only summary information available on any previous committals. This means that researchers could not identify all periods following earlier releases or determine when a person had returned to prison. As a result, time spent in the community after all releases could not be measured accurately. Using only the most recent release may underestimate the total time at risk across all releases and therefore overestimate mortality. Focusing on shorter follow-up periods reduces the chance of there being a previous incarceration in the relevant period, although it should be noted that it does not eliminate this risk. This needs to be balanced against the statistical disclosure policy and the challenge of providing useful characteristic breakdowns with limited numbers available. Researchers have focused on a one-year period following each person’s latest release.

Researchers have explored the feasibility of utilising summary information on any previous releases to synthesise data for all releases in an attempt to better quantify mortality risk. Unfortunately, in the absence of the detailed characteristics related to any previous committals, resulting assumptions could lead to large potential errors which may misrepresent mortality risk and result in policy interventions being incorrectly targeted. For example: in the absence of age information for all releases, assuming everyone remains in the same age category for all releases fails to take into account age progression. This could overestimate the number of releases assigned to older age groups, leading to an underestimate of mortality risk at older ages and an overestimate at younger ages.

Annex 7 - Probation Board Analysis

Between April 2011 and March 2020, 2,420 individuals were discharged from prison, with their license requiring them to be under the supervision of the Probation Board for Northern Ireland at their most recent release. By March 2021, 150 deaths were linked to this cohort, with 67 occurring within one year of latest release.

Table 9. Underlying cause of death for persons supervised by Probation Board within one year of latest release from prison (April 2011 – March 2021). Source: Research data and Vital Statistics Unit

Cause of death Deaths of persons supervised by Probation Board within one
year of release from prison
(n=67) 		Deaths of persons within one
year of release from prison
(n=384) 		All deaths
Northern Ireland
(n=19,000)
Drug-related 65.7% 54.4% 7.5%
Suicide 22.4% 27.3% 10.5%
Other 22.4% 29.2% 84.3%


Table 9 shows that in the first year following latest release, there were 67 deaths among people, with their license requiring them to be under supervision of Probation Board, of which:

  • 65.7% were drug-related

  • 22.4% were due to suicide


Table 10. Deaths of persons supervised by Probation Board by time from latest release (April 2011 – March 2021). Source: Research data

Time after latest release Count of deaths Average daily deaths
Days 0 - 30 15 0.5
Days 31 - 60 10 0.3
Days 61 - 180 20 0.2
Days 181 - 365 22 0.1


Table 10 shows the period immediately following release is a time of particularly elevated risk for people discharged from prison with their license requiring them to be supervised by Probation Board:

  • 15 deaths occurred within the first 30 days


Table 11. Characteristics of post-release deaths compared to survivors for persons supervised by Probation Board (April 2011 – March 2021). Source: Research data

Characteristic Deaths
(n = 67) 		Survivors one year of
latest release from prison
(n= 2,353)
Age group
18-24 19.4% 18.3%
25-29 20.9% 21.6%
30-34 28.4% 18.3%
35-59 31.3% 41.7%
Total committals
1 to 2 19.4% 60.6%
3 to 4 22.4% 19.8%
5 to 7 28.4% 11.9%
8 + 29.9% 7.6%
Area deprivation
1 - Most deprived decile 77.6% 56.0%
2 to 10 22.4% 44.0%
Principal offence category
Other 50.7% 43.2%
Person 29.9% 41.9%
Property 19.4% 14.8%
Principal offence group
Burglary & Robbery 26.9% 13.8%
Public Order & Possession of Weapons 29.9% 15.6%
Violence Against the Person & Sexual (including Misc. Sexual) 14.9% 32.9%
All other offences 28.4% 37.6%
Total days remanded
1 - 179 days 25.4% 63.4%
180 days + 74.6% 36.6%
Total days other
0 days 38.8% 67.1%
1 - 7 days 26.9% 16.4%
8 days + 34.3% 16.5%
Sentence length
1 - 179 days 28.4% 16.9%
180 days + 71.6% 83.1%
Year of latest release
2011/12 to 2013/14 20.9% 26.6%
2014/15 22.4% 14.7%
2015/16 14.9% 13.0%
2016/17 - 2017/18 22.4% 22.2%
2018/19 - 2019/20 19.4% 23.5%
Local Government District
Rest of NI 40.3% 55.6%
Belfast 59.7% 44.4%


Table 11 presents the characteristics of individuals whose licence required supervision by the Probation Board, comparing those who died within one year of their most recent release from prison with those who remained alive one year after release. Groups over-represented among those who died include:

  • Persons aged 30–34
    • People aged 30–34 years old at latest release [28.4% deaths compared to 18.3% survivors]
  • Repeat contact with the justice system
    • People with 5–7 total committals (28.4% deaths compared to 11.9% survivors)
    • People with eight or more total committals (29.9% deaths compared to 7.6% survivors)
  • Area deprivation
    • People with a recorded pre-release address in the 10% most deprived areas (77.6% deaths compared to 56.0% survivors)
  • Offence profile
    • People with a principal offence of burglary or robbery (26.9% deaths compared to 13.8% survivors)
    • People with a principal offence relating to public order or possession of weapons (29.9% deaths compared to 15.6% survivors).
  • Custodial experience
    • People with longer time spent on remand (over 180 days) (74.6% deaths compared to 36.6% survivors)


References

  1. Binswanger, I. A., Stern, M. F., Deyo, R. A., et al. (2007). Release from prison — A high risk of death for former inmates. New England Journal of Medicine, 356(2), 157–165. https://doi.org/10.1056/NEJMsa064115

  2. Merrall, E. L. C., Kariminia, A., Binswanger, I. A., et al. (2010). Meta-analysis of drug-related deaths soon after release from prison. Addiction, 105(9), 1545–1554. https://doi.org/10.1111/j.1360-0443.2010.02990.x

  3. Zlodre, J., & Fazel, S. (2012). All-cause and external mortality in released prisoners: systematic review and meta-analysis. American Journal of Public Health, 102(12), e67–e75. https://doi.org/10.2105/AJPH.2012.300764

  4. Graham, L., Fischbacher, C.M., Stockton, D., Fraser, A., Fleming, M. and Greig, K. (2015) ‘Understanding extreme mortality among prisoners: a national cohort study in Scotland using data linkage’, European Journal of Public Health, 25(5), pp. 879–885. https://doi.org/10.1093/eurpub/cku252

  5. Slade, K., Justice, L., Baguley, T., Bowen, E., Shorter, G.W., Adamson, L., Beck, A. and Borschmann, R. (2025) ‘Mortality after prison release in England and Wales, 2019–2021: a comparative analysis of cause-specific death rates and risk profiles’, Social Science & Medicine, 369, 117821. https://doi.org/10.1016/j.socscimed.2025.117821

  6. Borschmann, R., Kinner, S. and MARIC team (2024) “Rates and causes of death after release from incarceration: an individual participant data meta-analysis from 1,471,526 people in eight high- and middle-income countries”, International Journal of Population Data Science, 9(5). https://doi.org/10.23889/ijpds.v9i5.2489

  1. Cooper, JA, Onyeka, I, Cardwell, C, Paterson, E, Kirk, R, O’Reilly, D & Donnelly, M 2023, ‘Record linkage studies of drug-related deaths among adults who were released from prison to the community: a scoping review’, BMC Public Health, vol. 23, 826 . https://doi.org/10.1186/s12889-023-15673-0

  2. Cooper, J., Ross, E., Maguire, A., O’Reilly, D., Kirk, R., Kelly, T., McGarrigle, S., Gray, R., Cardwell, C., Murphy, S. and Donnelly, M. (2025) “Mental health, substance addiction and mortality risk among people in prison”, International Journal of Population Data Science, 10(4). https://doi.org/10.23889/ijpds.v10i4.3227