Abstract
Aim:
Data suggest mortality rates of those under community justice services such as probation or parole have been increasing year on year. Little is known about why and how these individuals are dying. This scoping review explores the causes and contributing factors of mortality in those under community justice supervision.
Methods:
Studies published between 2011 and 2021 were identified across CINAHL, Embase, Global Health, Ovid Medline and PsycINFO. Articles were included if they presented original data on either mortality rates among those under community justice supervision or risk factors associated with the mortality of those under community justice supervision.
Results:
Searches identified 101 unique articles of which 13 were included in the review. Articles were representative of five countries. All articles were either retrospective reviews or retrospective cohort studies. The studies fell into the categories of all-cause mortality, self-inflicted deaths or drug-related deaths.
Conclusion:
Mortality rates of those under community justice supervision were found to be consistently higher than mortality rates for the general population regardless of cause of death. Factors identified as affecting mortality included history of drug use, history of self-harm and previous imprisonment including length of time in custody and experience of hospitalisation or solitary confinement while in custody.
Introduction
In many countries, deaths occurring while in state custody are closely scrutinised and rates monitored, as per the requirement of international humanitarian and human rights law to ensure all undergo formal independent investigation.1,2 Such investigation has enabled global learning around the mortality of those in state custody, influencing factors such as mental health, the care of those of advanced age in prison, risk factors for self-inflicted deaths and homicides.3–7 Deaths occurring while under supervision of community justice services (such as under parole or probation for the purposes of serving community orders/sentences or under justice monitoring following release from prison) do not necessarily fulfil the same legal criteria for an equivalent level of independent scrutiny, and therefore, the factors affecting mortality of this group are much less understood.
Mortality rates can indicate the health of a population. 8 Although few countries publish mortality rates of those under community justice services, the limited data available demonstrate a clear upwards trend. Figures from the USA suggest exits from probation services due to death represent 12% of all known probation exits. 9 In 2020, deaths under probation in the USA saw an overall increase of 6%, with actual numbers increasing in 28 out of 38 reporting states. 9 In the UK, the number dying while under community justice supervision in England and Wales has risen steadily over the past decade from 557 recorded deaths in 2014/2015 to 1343 recorded deaths in 2020/2021. 10 These trends have continued during the COVID-19 pandemic with the Ministry of Justice (MOJ) 10 for England and Wales reporting an increase of 34% in deaths of offenders under community supervision from the previous year.
The high number of people dying while under community justice supervision appears to have gone largely unnoticed by health or public health professions despite repeated calls for greater scrutiny from a criminal justice and human rights perspective.11–13 Furthermore, from a public health perspective, it is important to understand the mortality risks of this cohort in order to estimate their health needs, inform preventive strategies, help monitor the effectiveness of interventions, and help advocate for resources or action. This is particularly important if the use of community justice measures is to increase, as is recommended in order to better support rehabilitation and reintegration as well as alleviating issues associated with prison overcrowding. 14
The aim of this study was to undertake a scoping review of the international literature to explore mortality rates and identify the main causes and contributing factors affecting mortality in those under community justice supervision.
Method
The Population, Intervention, Comparison, Outcome (PICO) framework was used to formulate a literature search strategy and guide search terms. The population considered were those under community supervision of justice services, the intervention was probation or community supervision and the outcome was death (all causes). Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines 15 were used to develop the scoping review protocol. Literature searches were performed in June 2021 across five electronic databases (CINAHL, Embase, Global Health, Ovid Medline, PsycINFO) for all fields of articles using the following search terms: offender$ OR criminal$ OR ex-prison* OR ex-detainee* OR perpetrator$ OR probation* OR incarcerate* AND probation OR community ADJ1 OR supervision OR parole AND death OR mortality OR suicide. Articles were limited to those published in English within the last 10 years (2011–2021). Articles exploring children or youth were excluded. The lead author conducted the initial identification and deduplication. Eligibility screening by title, abstract and full article was conducted by all authors.
Summary of Search Results
After removing duplicates, the initial search of the literature identified 101 articles (Figure 1). Following a review of titles, abstracts and full articles, the final number of articles included in the review was 13 (Table 1). Articles were representative of five countries: Australia (n = 1), Denmark (n = 1), Sweden (n = 1), UK (n = 1) and USA (n = 9). All articles were either retrospective reviews or retrospective cohort studies. Findings were considered according to mortality rates and factors associated with mortality. The articles identified fell clearly into three categories: all-cause mortality, self-inflicted deaths and drug-related deaths.
PRISMA diagram of study selection
Overview of included studies.
HR: hazard ratio; CI: confidence interval; IRR: incidence rate ratio; OR: odds ratio; OAT: opioid agonist treatment; CDC: Centers for Disease Control and Prevention.
Findings
Mortality rates of those under community justice supervision
The mortality rates of those under community justice supervision were found to be consistently higher than mortality rates of the general population, regardless of cause of death. In several instances, the mortality of those under community justice supervision was demonstrated as higher than the mortality rates of those in custodial settings. A comparison of age-standardised mortality ratios of people on probation in the USA against the general US population found that those on probation died at a rate 3.42 times higher than people in jail and 2.10 times higher than the general US population. 28 These findings were similar to those of a review of the community death index alongside community supervision and incarceration histories that found the adjusted risk for all-cause death was 3.61 (95% confidence intervals (CI) 3.48–3.73) times higher among former prisoners within the community than that among those who had never been imprisoned. 18 This latter review found unintentional injury (driven by overdose) to be the leading cause of death among offenders within the community, followed second by cardiovascular disease, and third, cancer.
Mortality rates were also documented for self-inflicted deaths where a review of suicides among community offenders in the UK found the suicide rate to be higher than that in the general population by a factor of 8.67, with an even higher risk for female community offenders, and higher than the prison population by a factor of 1.42. 24 Drug-related deaths for community offenders who had previously been incarcerated were found to be 5.29 (95% CI 4.93–5.65) times higher than those in the general population, 25 with an increased risk of mortality in the weeks following release. 17
Factors affecting mortality of those under community justice supervision
Factors identified as affecting the mortality of those under community justice supervision were limited to previous imprisonment (including total length of time spent within custody), history of hospitalisation while in custody and experience of solitary confinement. We were unable to find any articles that focused on community justice cohorts with no history of imprisonment even though a significant proportion of those serving community sentences will fall within this category. Other factors identified from the scoping review included history of drug use and previous suicide attempts. Reference was made by several articles to race, gender and age as factors.
Factors associated with previous imprisonment
Two studies considered the impact of previous imprisonment on the mortality of community offenders with conflicting results. Patterson 23 conducted a retrospective analysis of those who had been in prison in New York State from 1989 to 1993 and released on parole, following outcomes until 2003. The analysis suggested that the length of incarceration was associated with an increased risk of death, with each month served in prison increasing the odds of dying upon release by 1.7%. This was countered by the odds of dying decreasing by 2% for each month survived while on parole. Analysis used administrative data and could not account for behaviour factors which might influence mortality. The data used were old in terms of wider factors affecting mortality likely to have changed since the late 1980s and early 1990s. Binswanger et al. 17 conducted a similar retrospective cohort review of inmates released from 1999 to 2003, which found increasing length of incarceration was associated with a decreased risk of death for all-cause (hazard ratios (HR) 0.95, 95% CI 0.91–0.99) and overdose mortality (HR 0.80, 95% CI 0.68–0.95). The authors suggest that those in prison for extended periods may access medical treatment and health support while in prison which may reduce the risk of death but acknowledge that further investigation is needed.
One study indicated the postrelease mortality of community offenders was found to be higher among those who had been hospitalised while in prison than among those not hospitalised (adjusted hazard ratios (AHR) 2.44, 95% CI 2.01–2.96). 26 For those who had been hospitalised, deaths were mostly resulting from chronic conditions. Overall hospitalisation rates were believed to be one-sixth to one-eighth the rate of the community population, suggesting that hospitalisation for those in custody only took place for those most acutely unwell. These patients may be more likely to have ongoing health problems increasing their risk of mortality upon release. The retrospective review did not include details on the reasons for hospitalisation, so links between cause of hospitalisation and cause of death could not be made.
Solitary confinement while in prison, something consistently associated with poor health outcomes,29,30 was found to be associated with elevated risk of mortality within the five years after release from prison (AHR 2.34, 95% CI 1.53–3.59), particularly those who were younger and serving longer sentences. 27 The retrospective Danish study acknowledged the causal effect of solitary confinement on mortality could not be estimated due to the many variables and confounders that may also influence mortality. Nevertheless, results suggest additional support on release may be valuable for those who have experienced solitary confinement during their prison stay.
Influence of substance misuse history
Those misusing substances are at increased risk of prison or community supervision due to the illicit nature of the substances being used. It is also well known that the weeks following release from prison pose an increased risk of overdose for those who used substances before incarceration but are released with reduced opiate tolerance. Pizzicato et al. 25 found the adjusted standardised mortality rate (SMR) for overdose deaths in the two weeks following release from prison was significantly higher than that of community residents (SMR 36.91, 95% CI 29.92–43.90), whereas non-overdose deaths were not statistically different from those in a matched community population (SMR 0.96, 95% CI 0.67–1.26).
Literature from the scoping review specifically highlighted the probation and parole caseload as a high-risk population for substance-related deaths. A review of 180 opiate-related deaths in Sweden found 28% had been under the supervision of the prison and probation service in the year prior to death. 16 Of these, 8% had served a prison sentence in the year before death. Similarly, a retrospective cohort study from the USA used the National Death Index to explore overdose mortality for those sentenced to probation from 2003 to 2006. 19 Nearly a third of deaths (28.8%) occurred among those on probation, and 12.8% among those on parole. The first month after probation ended was associated with a higher overdose mortality rate than other periods on probation (incidence rate ratio (IRR) 1.96, 95% CI 1.06–3.34), signifying the vulnerability of transitional periods after any engagement with justice services, not just upon release from prison.
Influence of previous suicide attempts
The rate of self-inflicted deaths while on probation was found to be significantly higher (118 per 100,000, 95% CI 99–137) than rates among those in prison (83 per 100,000, 95% CI 66–100) or rates of the general community (13.6 per 100,000). 24 Only one article was found to look specifically at self-inflicted deaths among those under community justice, focusing on the relationship between prior suicidal behaviour and mortality. 20 The US-based study used data on deaths of those under a specific task force of community corrections 2002–2007. History of a suicide attempt was independently associated with mortality (odds ratio (OR) 2.21, 95% CI 1.25–3.88) and had the second greatest effect after gender (male) (OR 2.48, 95% CI 1.36–4.55).
Association between demographics and mortality while under community justice supervision
The number of men on community supervision typically exceeds that of women. Results from the scoping review largely indicated that women are at greater risk of death, particularly for overdose18,25 but also for suicide. 24 Rosen et al. 26 suggested that women in prison are more likely to be hospitalised than men (AHR 0.74, 95% CI 0.67–0.81) which may influence their risk of death on release. Not all authors agree; a retrospective review of risk factors for early deaths following early release from prison found women had the same risk of death as men. 17
The median age of death for those dying after release from prison was 45 years according to Binswanger et al. 18 While most studies considered age of death within the wider cohort of deaths by substance misuse or suicide, we could not find results specific to those under community supervision. The median age of opioid fatalities in Sweden was 34.5 years; however, age breakdown was not available for the proportion of these deaths that occurred while under probation supervision. 16
Ethnicity as a factor influencing mortality was considered by several studies, again for the wider cohort of deaths examined but not always reported by those under community justice supervision. Krawczyk et al. 22 found black adults engaged with community justice services were less likely to die of overdose than white adults (adjusted odds ratio (AOR) 0.45, 95% CI 0.22–0.89). While this was the only study reporting ethnicity directly linked to those under community supervision, it is consistent with the findings of other studies of those recently released from prison where African Americans (AHR 0.82, 95% CI 0.74–0.91), Hispanic (AHR 0.65, 95% CI 0.56–0.76) and Asians (AHR 0.61, 95% CI 0.42–0.88) were found to have a statistically significant reduced risk of all-cause, overdose and opioid-related deaths after release from prison compared to white individuals. 18 Drug-related mortality on release from prison was found to be significantly lower for indigenous ex-prisoners than for non-indigenous (HR 0.34, 95% CI 0.21–0.53) but greater by alcohol-related causes (HR 1.9, 95% CI 1.1–3.1). 21 Binswanger et al. 18 highlighted important points about local reporting which could risk under-reporting non-white deaths; therefore, local reporting issues need to be considered for each study. An earlier study by Binswanger et al. 17 found Latinos were at increased risk of early death after release from prison compared with whites. Where mortality was related to time served in prison, Hispanics were found to have higher odds of death (OR 1.69, 95% CI 1.45–1.97) than non-Hispanic whites. 23
Gaps in Literature and Next Steps
To our knowledge, this is the first scoping review of the literature for factors affecting mortality for those under community justice supervision. Within the UK and other nations, mortality rates for this population have been increasing, yet only a relatively small number of articles were found on this subject. No articles were identified on mortality rates for those with no previous experience of imprisonment. Due to the nature of the eligibility criteria, we acknowledge that some evidence in languages other than English was excluded; however, this was outside the remit of this scoping review.
Most of the evidence focused on drug-related or self-inflicted deaths. Robust evidence of deaths by other causes, including those that may be termed as a natural cause, for this population could not be found. The use of community justice measures has increased in light of the COVID-19 pandemic and following the United Nations Office on Drugs and Crime (UNODC) recommendations; therefore, understanding the factors affecting mortality of those under probation, to inform strategies to protect them from premature mortality, should be a public health priority.14,31
Community justice populations are demographically diverse and vary in the type of supervision they may be subject to (parole, probation, community licence and so on). Robust studies exploring mortality rates according to the type of justice supervision could not be found; therefore, we are unable to draw conclusions about mortality rates between types of supervision. Of the 13 articles included in the review, there was some basic, but limited, exploration around gender, age and ethnicity. The complete absence of other protected characteristics from the literature such as disability, religion and sexual orientation was stark. The small sample sizes and local concerns about accurate reporting of characteristics such as ethnicity make it difficult to draw conclusions between these and mortality. Nevertheless, indications of disparities between some characteristics and mortality from the small numbers of studies identified demonstrate the need for standardised local data reporting that includes basic demographics and, going forward, protected characteristics.
The impact of factors on mortality such as previous experience of hospitalisation 24 or segregation 23 was only considered by single studies, indicating their impact on mortality in this cohort remains conceptual. Substance use and history of suicide attempts perhaps unsurprisingly appear to increase the risk of mortality.25–28 Evidence around previous experience of imprisonment as a risk factor17,23 was mixed with suggestions that access to healthcare through previous imprisonment might, in some cases, provide a protective factor against early mortality on release. 17 Understanding this relationship in more detail will help inform how prison health services can help prevent premature mortality upon release and indeed how community justice services can support the continuation of health interventions initiated by prison health services during the transition from prison back to community.
All the studies identified were retrospective reviews or cohort studies that acknowledged many issues with data quality, both in the data that were available and issues matching data between health and justice datasets. Future understanding would be enhanced by robust prospective mortality surveillance designed to capture understanding of mortality risk, characteristics and context of deaths among different subsets of community justice populations. We acknowledge this may be methodologically challenging and will likely require triangulation of health and justice datasets. Death audits could also be used, and in countries where clinical records are not accessible, methods such as verbal and social autopsies could be considered. Real-time mortality surveillance using agreed national measures would help improve data quality, enabling trends to be monitored and the impact of future interventions to be measured. Tomczak and Mulgrew 32 propose tenets to guide prisoner death data collection to ensure it is comprehensive, disaggregated and contextualised. We advise their suggested data variables are considered more broadly for anyone who dies under justice supervision, including those under community justice supervision who may never have been held in state custody.
The high mortality rates among this population suggest community justice services should have a role not just in mortality surveillance but also in supporting those under supervision to engage with healthcare services. Community justice services may also be appropriate settings to deliver health interventions to prevent premature mortality. Such interventions could improve equity in health provision for this population, benefitting from a wider community gain. For example, the concept of specialist nursing services for ex-offenders has been previously suggested, 33 but to the authors’ knowledge, it has never been implemented. Public health services working with imprisoned populations should think more broadly by including community justice populations, focusing efforts towards improving health and reducing premature mortality for anyone in touch with justice services. Health policy for those imprisoned should also be widened to include community justice populations.
Conclusion
This scoping review explored mortality rates and factors affecting the mortality of community justice populations. It outlined an increasing number of deaths globally, yet a paucity of evidence was available to really understand the factors affecting mortality in this cohort. Mortality surveillance for those under community justice services needs to be established in order to understand risks, estimate needs, inform the development of interventions and to monitor the effectiveness of responses. Such surveillance should include demographics and protected characteristics but should also enable understanding of context around the death. Given the rising mortality rates, there is clearly a need for public health organisations to consider this population with a view to implementing interventions and steering policy to reduce premature mortality. Community justice services should equally establish partnerships with public health organisations to initiate actions to reduce mortality and improve health. Where public health teams are already working with imprisoned populations, this should be broadened to include those under community justice. This is even more imperative if the use of community justice measures continues to increase.
