Incidence of physical non-communicable disease in people who have experienced imprisonment compared with the general population in high-income countries: a systematic review and meta-analysis Authors: Beth Dorrans, Martha Tindall, Rachel M Thomson, Shona Mackinnon, Philip Broadbent, Valerie Wells, Andrea Elizabeth Williamson, Srinivasa Vittal Katikireddi Published: 2026 ABSTRACT AbstractBackgroundNon-communicable diseases (NCDs) represent a significant cause of morbidity and mortality globally. Despite a strong evidence base outlining increased risk of communicable diseases and poor mental health in people who have experienced imprisonment, the risk of NCDs has not been quantified in previous reviews. This study aimed to describe the incidence of morbidity and mortality of physical NCDs in people who have experienced imprisonment compared with the general population in high-income countries.MethodsWe searched MEDLINE, Embase, CINAHL and SocINDEX (PROSPERO CRD42024540173) for comparative studies of morbidity or mortality from NCDs. Articles were assessed against predefined criteria to select studies that included people with experience of imprisonment; measured NCD incidence; compared incidence to a general population; and were conducted in a high-income country. Two reviewers conducted independent risk of bias assessment using an adapted version of the Newcastle-Ottawa Scale. Data were synthesised using random-effects meta-analyses and vote counting based on effect direction. Meta-regression and subgroup analysis investigated heterogeneity.ResultsOf 3085 articles screened, 32 were included providing 341 datapoints. Meta-analysis showed a higher risk of mortality from several diseases, including respiratory disease (risk ratio (RR) 2.38 (95% CI 1.18 to 4.80), I2=97%), cardiovascular disease (RR 1.80 (95% CI 1.32 to 2.46), I2=99%), liver disease and cirrhosis (RR 2.50 (95% CI 1.08 to 5.77), I2=99%), digestive disease (RR 2.92 (95% CI 1.09 to 7.78), I2=99%), neurological disease (RR 1.94 (95% CI 1.09 to 3.44), I2=92%), head and neck cancer (RR 21.31 (95% CI 4.32 to 105.14), I2=97%), liver cancer (RR 4.07 (95% CI 2.34 to 7.08), I2=94%), cervical cancer (RR 3.95 (95% CI 3.11 to 5.01), I2=0%) and lung cancer (RR 1.95 (95% CI 1.34 to 2.85), I2=0%). Synthesis without meta-analysis also indicated imprisonment was associated with increased morbidity from multiple NCDs.ConclusionsSeveral NCDs, including multiple cancers, cardiovascular disease, respiratory disease and liver disease, are substantially more common in people who have experienced imprisonment compared with the general population, highlighting a need for primary prevention and NCD management.Keywords: EPIDEMIOLOGY, HEALTHCARE DISPARITIES, Health inequalities, META-ANALYSIS, PREVENTIVE MEDICINE WHAT IS ALREADY KNOWN ON THIS TOPIC Experience of imprisonment, along with other forms of social exclusion, is associated with high all-cause mortality. Several systematic reviews have highlighted a high burden of infectious disease and mental health morbidity in people who have experienced imprisonment, but evidence on the incidence of non-communicable diseases (NCDs) is scarce. Experience of imprisonment, along with other forms of social exclusion, is associated with high all-cause mortality. Several systematic reviews have highlighted a high burden of infectious disease and mental health morbidity in people who have experienced imprisonment, but evidence on the incidence of non-communicable diseases (NCDs) is scarce. WHAT THIS STUDY ADDS This systematic review for the first time provides strong evidence of increased risk of mortality from respiratory disease, cardiovascular disease, liver disease and cirrhosis, digestive disease, neurological disease and several cancers among people who have experienced imprisonment, adding to the existing evidence base which focuses on infectious diseases and mental health. This systematic review for the first time provides strong evidence of increased risk of mortality from respiratory disease, cardiovascular disease, liver disease and cirrhosis, digestive disease, neurological disease and several cancers among people who have experienced imprisonment, adding to the existing evidence base which focuses on infectious diseases and mental health. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY Policy and practice responses are required to address NCDs as well as infectious diseases and poor mental health. Interventions to reduce exposure to causes of NCDs should be considered, such as smoke-free prisons, alongside inclusive primary care provision and measures to address the underlying social conditions that predispose to disease. Policy and practice responses are required to address NCDs as well as infectious diseases and poor mental health. Interventions to reduce exposure to causes of NCDs should be considered, such as smoke-free prisons, alongside inclusive primary care provision and measures to address the underlying social conditions that predispose to disease. Introduction Imprisonment leads to physical exclusion of prisoners from society.1 People who experience imprisonment may also experience social exclusion throughout their lives alongside significant health inequalities with markedly increased mortality.2 A large body of evidence reports a high burden of depression, post-traumatic stress disorder and substance use, with suicide the leading cause of death within prisons globally.3 4 Morbidity and mortality from infectious disease in people with current or past experience of imprisonment is also common, including hepatitis C virus, hepatitis B virus, HIV and tuberculosis.3 However, the burden of non-communicable diseases (NCDs) is less well understood.3 5 6 NCDs are responsible for approximately three-quarters of deaths worldwide, the majority due to cardiovascular disease, cancer, chronic respiratory disease and diabetes.7 NCDs are closely linked to socioeconomic deprivation, which in turn is also associated with imprisonment and other forms of social exclusion.7 8 Previous review articles outline possible increased incidence and prevalence of cardiovascular disease, cancer, diabetes and asthma in this population. However, this evidence is limited by restriction of the population of interest to older prisoners, or to specific countries. No systematic reviews or meta-analyses have synthesised data on the risk of NCDs in people who have experienced imprisonment in high-income countries, and therefore the burden of NCDs in this population is relatively unknown.9,12 This may limit the recognition of health needs, planning of services and development of policies to improve the health of this population. Descriptive epidemiology aiming to quantify the burden of disease, without necessarily establishing causation, can be used to understand increased risks in specific populations and allow planning of interventions to improve their health. Therefore, the aim of this systematic review was to describe the incidence of morbidity and mortality from physical NCDs in people who have experienced imprisonment compared with the general population in high-income countries. Methods A review protocol was developed (PROSPERO CRD42024540173).13 This review is reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses and Synthesis Without Meta-analysis (SWiM) reporting guidance.14 15 Search Strategy Searches were conducted using MEDLINE, Embase, CINAHL and SocINDEX on 29 April 2024. The search strategy, developed with an information scientist, combined synonyms related to “prisoners”, “non-communicable diseases” and “incidence” to identify relevant results with no date or language restrictions (online supplemental tables 1-4). Backwards citation searching was carried out. Eligibility criteria The population of interest was people with current or previous experience of imprisonment of any gender (online supplemental table 5). We excluded studies with participants who were prisoners of war, detained in immigration centres, psychiatric services and police custody as these populations may experience different circumstance-specific health outcomes that are not generalisable to the wider population experiencing imprisonment. We excluded studies where the majority of participants were under 18 years old at the time of recruitment. The outcomes of interest were any physical NCD as defined by the Global Burden of Disease Study (online supplemental table 6).16 We excluded mental and substance use disorders, communicable diseases, injuries, congenital abnormalities, NCD risk factors and precancerous conditions. Included studies had to measure and compare incidence of morbidity (diagnosis) or mortality from NCDs to a comparison group from the general population in a high-income country (defined by the World Bank, online supplemental table 7).17 Restricting studies to high-income countries was undertaken given the varying sociopolitical approaches to imprisonment worldwide, aiming to provide less heterogenous and more interpretable results. Studies were also restricted to those published in English due to availability of translation services. Study selection All titles, abstracts and full-text papers were independently screened by BD and a second reviewer (MT, RMT or SM) with conflicts resolved following discussion with a third reviewer (SVK). Data extraction and risk of bias assessment Data were extracted using an Excel template by BD (online supplemental table 8).18 All data were checked independently by a second reviewer (MT or PB). Effect estimates extracted were measures of morbidity or mortality comparing people who have experienced imprisonment to the general population. If a study provided multiple models of adjustment for an effect estimate, the simplest model including age and sex was extracted to ensure descriptive results could be interpreted appropriately.19 This was justified as this study aimed to describe the burden of NCDs in this population and did not aim to investigate causality. Therefore, in descriptive epidemiology (unlike causal epidemiology), analytical adjustment is not always essential.19 In fact, further analysis to include possible confounders, for example psychiatric morbidity or substance use, was not considered appropriate as they may mask the inequality between people who have experienced imprisonment and the general population, limiting the results of this descriptive review. Study authors were contacted to provide further data if required. If overlapping data were presented in multiple studies, data were extracted from the study in which the data were most complete for meta-analysis. An adapted version of the Newcastle-Ottawa Scale was used to assess risk of bias (RoB) (supplemental material p17-19).20 Modifications included adaptation to the review population, removal of questions that implied causative relationships between exposure and outcome, and identifying domains of particular importance to the review question. Reviewers assessed the representativeness of cohorts, the methods of exposure measurement, the comparability of cohorts, the methods of outcome assessment and adequacy of follow-up. RoB assessment was undertaken independently by BD and a second reviewer (MT or PB). Conflicts were resolved following discussion with a third reviewer (MT, PB or SVK). Data synthesis Meta-analysis Random-effects meta-analysis was conducted given the anticipated heterogeneity across studies in terms of population characteristics. Between-study variance was estimated using restricted maximum likelihood estimator and the DerSimonian and Laird method was used to calculate CIs for the random effects estimate. Results were displayed using forest plots. Standardised mortality ratios (SMRs), standardised incidence ratios (SIRs) and risk ratios (RRs) were synthesised. Synthesis of ORs, HRs and incidence rate ratios with SMRs, SIRs and RRs was not conducted, and these were included in SWiM. The I2 value was used to report heterogeneity between studies. Prediction intervals were included in forest plots to illustrate heterogeneity further, indicating the spread of effect estimates in included studies.21 These were used when there were more than 10 effect estimates included in the meta-analysis and no clear funnel plot asymmetry.21 Heterogeneity was investigated using meta-regression and subgroup analysis which was planned a priori. Where there were at least 10 effect estimates available, meta-regression investigated differences by RoB category, participant sex, location of study and temporality of imprisonment (ie, whether study participants were imprisoned at the time of the research being conducted or had been imprisoned in the past). These variables were believed to be relevant to potential heterogeneity between studies. Where meta-regression showed evidence of a difference between subgroups, subgroup meta-analyses were carried out. Non-reporting bias was investigated qualitatively through funnel plots and quantitatively using the Egger test.22 Investigation of non-reporting bias was only carried out for meta-analyses of outcomes with at least 10 effect estimates.21 Synthesis without meta-analysis Vote counting based on direction of effect was used to synthesise effect estimates that could not be included in meta-analysis due to incomplete reporting or incompatibility. As an acceptable method of data synthesis outlined by the Cochrane Handbook for Systematic Reviews, this method did not take account of statistical significance reported in studies.21 Effect estimates were simply categorised as ‘beneficial’ or ‘harmful’, disregarding their effect size and statistical significance. The number of effect estimates that were beneficial was then compared with the number that were harmful, giving a standardised metric to understand the results of studies included in SWiM. Following this method, as per the Cochrane Handbook for Systematic Reviews, a sign test was used to assess the probability of observing the results presented, with a lower p-value indicating a lower probability of observing this result if imprisonment truly had no effect on NCD risk.21 SWiM results were displayed in a harvest plot. Heterogeneity was not explored due to the small number of available effect estimates. All statistical analyses were carried out using R V.4.1.1, meta package.23 Results Following deduplication, 3081 studies were screened, with 179 full texts assessed for eligibility (figure 1). We excluded 147 studies (online supplement p20-27) leaving 32 included studies.24,55 This provided 341 datapoints; effect estimates of the risk of NCD morbidity or mortality in people who have experienced imprisonment compared with the general population. PRISMA flow diagram of study selection. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses. Characteristics and results of the included studies are reported in online supplemental tables 9 and 10. Included studies were published between 1987 and 2024. Of the included studies, 31 were cohort studies, and one was a case–control study. The most commonly included NCDs were cancer, cardiovascular disease, digestive diseases and respiratory disease. 19 studies investigated mortality, 10 studies investigated morbidity: three investigated both. Morbidity measures were from recorded hospitalisations or cancer registry reporting. Most studies were carried out in the USA (n=17); other locations included Australia (n=4), the UK (n=3), Canada (n=2), France (n=2), the Netherlands (n=1), New Zealand (n=1), Norway (n=1) and Sweden (n=1). The time period of imprisonment varied, with 11 studies focusing on those who had been imprisoned in the past, 10 focusing on those who were currently imprisoned and 11 focusing on those who had experience of past or current imprisonment. 21 studies investigated both men and women. Six studies investigated men, and one study focused on women. One study did not report the sex or gender of participants. A range of ages was included in the studies but was not reported consistently. Nine studies were rated low RoB, 19 moderate and seven high (online supplemental tables 11 and 12). Common methodological concerns included selection bias in the exposed cohort and concerns regarding comparability of the exposed and non-exposed cohorts with adjustments made for covariates other than age and sex. Although this may have been appropriate within original studies, this may make synthesis less interpretable when describing the relative disease burden. Meta-analyses Effect estimates from 19 studies were included across 25 meta-analyses Cancer morbidity and mortality The RR for diagnosis of all types of cancer was 1.06 ((95% CI 0.73 to 1.54), I2=97%, online supplemental figure 1). There was an increased risk of diagnosis of liver cancer (RR 5.58 (95% CI 2.02 to 15.47), I2=78%, online supplemental figure 2), lung cancer (RR 1.99 (95% CI 1.53 to 2.59), I2=83%, online supplemental figure 3) and head and neck cancer (RR 1.86 (95% CI 1.53 to 2.27), I2=0%, online supplemental figure 4) in those who had experienced imprisonment. The risk of colorectal cancer diagnosis was reduced in people who have experienced imprisonment (RR 0.83 (95% CI 0.71 to 0.96), I2=0%, online supplemental figure 5). The RR for mortality from any form of cancer was 1.18 (95% CI 0.90 to 1.53, I2=98%, figure 2). Forest plot of cancer mortality among people who experienced imprisonment compared with the general population. Meta-regression showed effect estimates for mortality from any type of cancer compared with the general population were higher in studies in the USA compared with England and Wales; however, only one study was conducted in England and Wales (online supplemental table 13, online supplemental figures 23 and 24). There were also higher effect estimates for mortality from any form of cancer in studies that included people with a history of past imprisonment compared with studies only including those currently experiencing imprisonment (online supplemental table 13). There was insufficient evidence from subgroup analysis that risk of cancer mortality in studies with participants with past experience of imprisonment was increased (RR 1.58 (95% CI 0.93 to 2.68), I2=99%, online supplemental figure 25). Risk in studies with participants with either past or current experience was increased (RR 1.29 (95% CI 1.14 to 1.45), I2=80%, online supplemental figure 26). Risk in participants who were currently imprisoned was unclear (RR 0.59 (95% CI 0.27, 1.29), I2=98%, online supplemental figure 27). Stratifying by cancer type, there was an association between imprisonment and mortality from head and neck cancer (RR 21.31 (95% CI 4.32 to 105.14), I2=97%, online supplemental figure 6), liver cancer (RR 4.07 (95% CI 2.34 to 7.08), I2=94%, online supplemental figure 7), lung cancer (RR 1.95 (95% CI 1.34 to 2.85), I2=98%, online supplemental figure 8) and cervical cancer (RR 3.95 (95% CI 3.11 to 5.01), I2=0%, online supplemental figure 9). There was no conclusive evidence of an increased risk of mortality from colorectal cancer (RR 2.27, (95% CI 0.54 to 9.61), I2=97%, online supplemental figure 10) or prostate cancer (RR 3.70 (95% CI 0.15 to 89.43), I2=96%, online supplemental figure 11). Cardiovascular disease morbidity and mortality Experiencing imprisonment was associated with an increased risk of cardiovascular disease (RR 1.80, 95% CI 1.32 to 2.46, I2=99%, figure 3). Forest plot of cardiovascular disease mortality among people who experienced imprisonment compared with the general population. Meta-regression (online supplemental table 13) and subgroup analyses provided different effect estimates for mortality from cardiovascular disease in men (RR 1.20 (95% CI 0.85 to 1.69), I2=99%, online supplemental figure 28) and women (RR 3.21 (95% CI 2.32 to 4.45), I2=77%, online supplemental figure 29), compared with the general population. There was no conclusive evidence of an increase in risk of mortality from ischaemic heart disease (RR 2.26 (95% CI 0.94 to 5.45), I2=100%, online supplemental figure 12), cerebrovascular disease (RR 1.50 (95% CI 0.55 to 4.10), I2=98%, online supplemental figure 13) or hypertension (RR 10.20 (95% CI 0.85 to 122.78), I2=94%, online supplemental figure 14), or morbidity from hypertension (RR 1.31 (95% CI 0.93 to 1.86), I2=33%, online supplemental figure 15) for people who have experienced imprisonment compared with the general population. Respiratory disease morbidity and mortality There was a higher risk of mortality from respiratory disease in people who have experienced imprisonment (RR 2.38 (95% CI 1.18 to 4.80), I2=97%, figure 4) compared with the general population. Forest plot of respiratory disease mortality among people who experienced imprisonment compared with the general population. Meta-regression suggested the risk of mortality from respiratory disease in people imprisoned in the past may be higher compared with those who were currently imprisoned at the time of the study being conducted (online supplemental table 13). Subgroup analysis showed that the magnitude of association was reduced in people currently imprisoned (RR 1.28 (95% CI 0.53 to 3.10), I2=97%, online supplemental figure 30), and higher in the single study that investigated respiratory disease mortality in those who were previously imprisoned (RR 41.49 (95% CI 25.15 to 68.46), online supplemental figure 31). Morbidity and Mortality from Other NCDs There was an increased risk of mortality from liver disease (RR 2.50 (95% CI 1.08 to 5.77), I2=99%, online supplemental figure 16), digestive system disease (RR 2.92 (95% CI 1.09 to 7.78), I2=99%, online supplemental figure 17) and neurological disease (RR 1.94 (95% CI 1.09 to 3.44), I2=92%, online supplemental figure 18). Evidence was insufficient to conclude an increased risk of mortality from diabetes (RR 1.46 (95% CI 0.41 to 5.12), I2=98%, online supplemental figure 19), genitourinary disease (RR 1.48 (95% CI 0.97 to 2.22), I2=0%, online supplemental figure 20) and endocrine, nutritional, metabolic and immune diseases (RR 1.42, 95% CI 0.69 to 2.92, I2=92%, online supplemental figure 21). One study with multiple datapoints identified an association with morbidity from dementia (HR 2.30 (95% CI 2.18 to 2.43), I2=0%, online supplemental figure 22). Meta-regression did not provide evidence of difference in results depending on RoB categories (online supplemental table 13). No evidence of non-reporting bias was identified (online supplemental figures 32-35). Evidence gaps Evidence gaps identified included data on several variables, for example, temporality of imprisonment, age, ethnicity and time since release from prison. Several NCDs were underrepresented within the data, including cervical cancer, prostate cancer, diabetes and neurological disorders. The majority of included studies were conducted in the USA with male participants, and therefore there is less evidence outlining NCD incidence in other locations and genders. Synthesis without meta-analysis 117 datapoints covering 16 outcomes were included in SWiM (figure 5, online supplemental table 14). Experience of imprisonment was associated with increased risk of morbidity from cardiovascular disease (p=0.02), neurological disease (p=0.008), endocrine and nutritional disease (p=0.008), and skin disease (p=0.008). There was not enough evidence to conclude increased morbidity from any cancer (p=1), lung cancer (p=0.38), respiratory disease (p=0.11), asthma (p=0.125), digestive disease (p=0.73), gallbladder disease (p=0.5), blood disease (p=1), musculoskeletal disease (p=0.73), and dental caries (p=0.5) or mortality from cardiovascular disease (p=1), respiratory disease (p=1). These results should be interpreted with caution given the small number of datapoints included. Harvest plot of effect direction for health risks among people who have experienced imprisonment compared with the general population. Discussion Our comprehensive systematic review provides evidence of considerably higher physical NCD morbidity and mortality in people who have experienced imprisonment compared with the general population in high income countries. A meta-analysis found increased risk of morbidity and mortality from lung cancer, liver cancer and head and neck cancer. There was also an increased risk of mortality from cervical cancer, cardiovascular disease, respiratory disease, liver disease and cirrhosis, digestive system disease and neurological disease. On subgroup analysis, the excess risk of cardiovascular disease compared with the general population was higher in women experiencing imprisonment than men. SWiM provided further evidence of increased risk of morbidity from cardiovascular, neurological, skin, endocrine and nutritional diseases. Previous reviews have indicated an increased prevalence of cardiovascular disease, cancer, diabetes, asthma and dental caries in prison populations in cross-sectional studies.9 10 12 The findings of our review report consistent evidence of increased risk of cardiovascular disease mortality and multiple cancers. Despite extracted effect estimates suggesting increased risk of diabetes, asthma and dental caries, the evidence was not substantial enough to conclude this. Previous studies have investigated outcomes for specific ages. However, due to a lack of standardised reporting of results by age, it was not possible to explore differential outcomes across the life course. In the wider context of inclusion health, two reviews have investigated morbidity and mortality in people who experience imprisonment among other causes of social exclusion.2 56 Most studies originated in the USA, and studies reporting NCDs in these populations were scarce. Our finding of increased risk of mortality from cardiovascular disease in women who have experienced imprisonment compared with men echoes the findings of a previous systematic review investigating morbidity and mortality in socially excluded populations, which reported higher summary SMRs in females experiencing social exclusion compared with males.2 Women were the minority in many of the studies included in our review, but the relative risks of women who experience imprisonment were high. As outlined by Aldridge et al, this may reflect women’s needs within prisons being relatively neglected (given they comprise a relatively small proportion of imprisoned people), and the lower incidence rates experienced by women overall for many physical NCD outcomes meaning higher relative risks are more readily observed.57 It is beyond the scope of this review and meta-analysis to determine the causal effect of imprisonment itself on NCDs. It is likely the increased risk of several NCDs is due to a combination of many socio-political, environmental and behavioural factors. Social support and socioeconomic status are considered fundamental causes of disease, influencing the resources that people can access to improve their health.58 This can be seen through the disproportionate burden of NCDs in areas of socioeconomic deprivation and may shed light on why imprisonment is associated with higher NCD risk, through social exclusion and socioeconomic disadvantage.59 Regardless of the presence or absence of a causal relationship between imprisonment and NCDs, descriptive epidemiology can be used to identify health inequalities in order to address them.19 Furthermore, decisions about whether to incarcerate are unlikely to be primarily motivated by the health outcomes of those who are imprisoned, but policy and practice responses to improve health should still be informed by an understanding of available data. This systematic review has multiple strengths, being the first to investigate the comparative morbidity or mortality from NCDs in people who have experienced imprisonment. Investigation of both morbidity and mortality as separate measures of incidence provides important insight into the burden of NCDs in this population. NCDs may be undiagnosed until fatality from the condition occurs, contributing to incidence. The drivers of this, including reduced healthcare access, both in prison and on release, are complex. Best practice guidance for conducting and reporting systematic reviews was followed throughout.14 15 21 Inclusion of both morbidity and mortality was used as measures of risk as high heterogeneity between studies was anticipated. Subgroup analysis was planned a priori to further investigate this heterogeneity focusing on RoB category, participant sex, location of study and temporality of imprisonment. Alongside the limitation of high heterogeneity, other limitations include a small evidence base to fully examine several of these variables and others, including temporality of imprisonment, age, ethnicity and time since release from prison. Several NCDs were underrepresented within the data and therefore evidence was limited to fully understand their incidence, including cervical cancer; prostate cancer; diabetes; and neurological disorders, for example dementia. Other limitations included high RoB identified in seven studies. The review protocol excluded studies not published in English; however, none were identified at screening, and so no studies were excluded for this reason. This review only included studies from high-income countries. The majority of the studies were conducted in the USA, and the majority of participants were male and therefore the generalisability of these results to other contexts may not be appropriate. Prior to publication, a further brief search of the most recent literature was conducted to consider any more recent published papers that may have been concluded. Two papers of interest were identified.60 61 Dhimal et al compared age-adjusted cancer-related mortality rates in incarcerated individuals in Louisiana to the general population of Louisiana and the USA. However, wide CIs for the estimates limit the ability to draw strong conclusions based on this work alone.61 Lüchtenborg et al conducted a population-based cohort study to compare cancer incidence, treatment and survival between patients diagnosed in prison and the general population. This provided evidence of higher cancer incidence in people in prisons, with lower survival when compared with the general population. These studies may reflect a growing focus on cancer outcomes in people who have experienced imprisonment, providing further evidence of health inequalities affecting this population. Our results have implications for future research. Further primary research with consistent and standardised reporting of participant demographics including ethnicity is required to allow more complete data synthesis. There should also be increased focus on specific NCDs under-represented in this work, including certain cancers, diabetes and neurological disorders. Given that the sociopolitical context of a country will impact both imprisonment and health, further research in a wider range of countries could strengthen the evidence base, as would research investigating differences in risk over time. Recognising the association between imprisonment and other sources of social exclusion, further research could consider the effect of multiple disadvantages and multimorbidity on NCD incidence. This review has provided evidence that people who experience imprisonment are at higher risk of morbidity and mortality from multiple NCDs. Cancer, cardiovascular disease, respiratory disease and diabetes are responsible for 74% of all deaths globally. Given this high baseline incidence, a small increase in relative terms may represent a large increase in the rate of NCDs.7 Our review suggests that the focus of prison healthcare should expand beyond mental health, substance dependency and communicable diseases to include physical NCDs as a priority. In particular, healthcare providers working with people who have experienced imprisonment should focus efforts on prevention, diagnosis and early treatment of respiratory disease, cardiovascular disease, liver disease and several cancers (specifically lung, liver, head and neck and cervical cancer). The evidence presented in this review strengthens the case for measures to reduce exposure to NCD risk factors for people who have experienced imprisonment, recognising their higher risk of morbidity and mortality from multiple conditions. This could include, but is not limited to, ‘smoke-free’ policies in prisons and continued smoking cessation support on release, recognising the continued high risk of respiratory disease mortality in later years.62 Given the high risk of liver cancer found in this review, continued efforts to reduce alcohol consumption in those with harmful alcohol intake and hepatitis C infection in high-risk groups are important. The high risk of cardiovascular disease, particularly in women, provides a rationale for more targeted policies that address risk factors of cardiovascular disease, including smoking, diet and hypertension. Individual level interventions serve a purpose; however, they do not account for the wider experience of social exclusion that many people who experience imprisonment face, limiting their ability to take part in these interventions both in prison and following release. Therefore, policy makers must continue to address the fundamental causes of disease that are strongly linked to imprisonment, ‘lifestyle’ risk factors and incidence of NCDs.58 This review addresses a key research gap and highlights the burden of NCDs in people who have experienced imprisonment, adding to the previous evidence emphasising the burden of communicable diseases and mental illness. This increased risk, along with the well-documented challenges of accessing healthcare both in prison and following release, should be recognised for people who have experienced imprisonment.63 64 Previous reviews have called for a more horizontal approach to inclusion health, one that does not focus on single risk factors or types of disease but takes a holistic view of the health of people who experience social exclusion.2 In practice, this may mean continued work to provide a strong investment in primary care interventions to reduce and manage NCDs in prisons and following release. Supplementary material Acknowledgements The authors also gratefully acknowledge Dr Erik Igelström for providing guidance on statistical synthesis and the use of the ‘meta’ package in R. Footnotes Funding: BD, RMT, MT, PB, VW and SVK acknowledge funding from the Medical Research Council (MC_UU_00022/2) and the Scottish Government Chief Scientist Office (SPHSU17). PB acknowledges Wellcome Trust PhD funding for research on multimorbidity. Provenance and peer review: Not commissioned; externally peer reviewed. Patient consent for publication: Not applicable. Ethics approval: Formal ethics approval was not required for this research. However, several ethical issues were considered, including possible stigma associated with the terms ‘social exclusion’ and ‘imprisonment’. The use of the term ‘social exclusion’ in this context reflects the impact of imprisonment on social inclusion due to many factors outside an individual’s control that are often unavoidable and fundamentally unjust. Social exclusion in this context is not related to the personal ability of an individual to take part in social activities and form social networks following or during a period of imprisonment. Person-centred language was used aiming to reduce further marginalisation of the population of interest. Presented at: This research will be presented as a poster at the 18th European Public Health Conference in November 2025, and the abstract will be published in the conference Abstract Supplement. Correction notice: This article has been updated since it first published. The article is now open access. Data availability statement No data are available. References Associated Data Supplementary Materials Data Availability Statement No data are available.