An exploratory analysis into the mortality of offenders

Ong Su-Wuen
Principal Strategic Analyst, Department of Corrections

Ella Lynch
Research Adviser, Department of Corrections

Author biographies:
Su-Wuen joined the Department of Corrections Research and Analysis Team in 2012. He worked as a hydrologist in the first half of his career. Since 2001 he has worked for the Land Transport Safety Authority, the Ministry of Research, Science and Technology and the Ministry of Justice.

Ella joined the Department of Corrections Research and Analysis Team in 2015. Ella graduated from Victoria University in 2016 with an honours degree in Criminology. Prior to joining the Department, she worked in a research role at Wellington Rape Crisis.


Introduction

The Department of Corrections is increasingly seeking new knowledge and insights into the offenders it manages through research using Statistics New Zealand’s Integrated Data Infrastructure (IDI). This study seeks to shed light on the association between criminal careers and mortality rates of offenders.

Research on the effect of imprisonment and/or offending on mortality and life span relative to the general population is limited (Pridemore, 2014; Rosen, Schoenbach & Wohl, 2008). Most existing research has focused on the issue of prisoner (and ex-prisoner) suicide, as suicide is consistently found to be a leading cause of death for prisoners in several countries (Sattar & Killias, 2005), although Pratt and colleagues (2006) suggest studies do not often consider the issue of suicide in the post-release period. Evidence from several international studies indicates that prisoners and ex-prisoners (male and female) have higher mortality rates than the general population (Kariminia, Jones & Law., 2012; Pratt, Piper, Appleby, Webb & Shaw., 2006; Pridemore, 2014; Pritchard, Cox & Dawson, 1997; Rosen et al., 2006; Van Dooren, Kinner & Forsyth, 2013).

Kariminia and colleagues’ (2012) exploration of the increased mortality of indigenous people during and after their release from prison in New South Wales (1998-2002) highlighted this difference, finding the mortality rate to be 4.8 times higher for Aboriginal men and 12.6 times higher for Aboriginal women than that of New South Wales residents of the same age and sex, also highlighting a gender difference in the rate of death. Offenders have also been found to have markedly lower life spans than the general public, with the median age of death often being in the early to mid-thirties* (Kariminia et al., 2007; Kariminia et al., 2012; Sattar & Killias, 2005). In Sattar and Killias’ (2005) study of the death of offenders in Switzerland, the mean age of death was found to be 33.5 years with 45.5% of total deaths occurring in the 25-34 year old age band. This low life span is explained as a result of higher proportions of unnatural deaths occurring in younger age bands. Pridemore (2014:215) concluded, from interpretation of data from the Russian Family Study (male only) that “incarceration has durable effects on illness, [and] that its consequences extend to a greater risk of early death”.

Several reasons have been put forward to explain the differential in mortality. Rosen and colleagues (2008:278) consider that mortality rates could reflect the impoverished communities which many prisoners come from, as well as prisoners’ participation in risky behaviours, activities and lifestyles (e.g. substance abuse, violence and greater exposure to situations involving risk of assault or homicide) which are “illegal and harmful to health”. It is also likely that the greater number of deaths from natural causes may be a result of limited access/engagement with healthcare providers in the community and issues surrounding greater alcohol consumption and smoking, as well as dietary factors (Kariminia et al., 2012). Pridemore (2014) also considers that offenders’ negative health outcomes and early death compared with the general population could be exacerbated or caused by exposure to infectious diseases from the prison environment, stress from imprisonment and reintegration, and broken or damaged relationships with families and partners.

Ex-prisoners also appear to be at heightened risk of death soon after their release (Farrell & Marsden, 2007; Pridemore, 2014; Rosen et al., 2006). This finding was reinforced in Kariminia and colleagues’ (2007) Australian research which found that the first year of release coincided with the highest excess mortality for both male and female ex-prisoners. In particular, suicide was three times more likely to occur within the first year of follow up than after three years (Kariminia et al., 2007). Similarly, an analysis into suicide rates of recently released prisoners in England and Wales (Pratt et al., 2006) found that of the suicides that occurred within a year of release, 21% of ex-prisoner suicides occurred within the first 28 days, with suicide rates for all age bands of recently released prisoners being higher than the general populations’. Strikingly, Farrell and Marsden’s (2007) investigation into drug-related deaths of recently released prisoners in England and Wales, found that male and female prisoners were more likely to die in the week after their release (29 times and 69 times, respectively) from prison compared to the death rate of the general population during this time.

The literature is consistent in the chief causes of prisoner and ex-prisoner death. These include: substances (namely drug overdoses/accidental poisoning), suicide, homicide and accidents/injury (Van Dooren et al., 2013; Farrell & Marsden, 2007; Kariminia et al., 2007**; Kariminia et al., 2012; Pratt et al., 2006; Pridemore, 2014; Rosen et al 2008; Sattar & Killias, 2005). Unnatural deaths, particularly from drug overdose, suicide and homicide, were found by Kariminia and colleagues (2012) to be more frequent in those younger than 45 years. Sattar and Killias’ (2005:317) analysis into the death of offenders in Switzerland confirmed that death as a result of unnatural causes is “rather common” for offenders, with 65.4% of deaths being classified as unnatural. A high rate of drug related death was found in Farrell and Marsden’s (2007) study of newly released prisoners in England and Wales. From a sample of nearly 49,000 male and female prisoners released between 1998 and 2000, whose mortality was tracked over three years, 59% of deaths were recorded as drug related. Drug overdose also played a significant role in deaths of offenders, with drug overdose overall being accountable for 29% of excess deaths from Aboriginal men and 39% for Aboriginal women (Kariminia et al., 2012).

Regarding death from natural causes, ex-prisoners appeared to have excess mortality and more commonly died from cardiovascular, respiratory, digestive and liver related diseases as well as infectious diseases (Farrell & Marsden, 2007; Kariminia, Jones & Law, 2012; Pridemore, 2014; Rosen et al., 2008). In an Australian study (Kariminia et al., 2012) which explored the mortality of Aboriginal offenders in New South Wales, the “leading” cause of death for men was cardiovascular disease (23%). However, the central cause of death for women was recorded as being from mental or behavioural disorders (23%) which were all deemed to result from drug addiction. Sattar and Killas’ (2005) Swiss study found that adult convicted prisoners who died between 1984 and 2000 most commonly died of natural causes (i.e. illness and disease in 34.6%). However, this was closely followed by unnatural deaths from drug overdose (28.6%) and suicide*** (28.2%). High excess mortality from chronic conditions, particularly cardiovascular and respiratory conditions were considered to be a possible outcome of higher rates of smoking, alcohol use and dietary issues among this population (Kariminia et al., 2012). As expected, natural deaths were more strongly associated with an older age group (Kariminia et al., 2012; Sattar & Killias, 2005).

On the basis of such findings, research has advocated for more drug treatment programmes and mental health support, and for this support to extend beyond imprisonment and into the community (Kariminia et al., 2007; Rosen et al., 2008). In addition, the role of prison in addressing offender health needs is noted, with Kariminia and colleagues (2012:278) arguing that prison “provides an important (but underutilised) public health opportunity to screen for chronic diseases and assess treatment needs of offenders who are likely to have limited interaction with the health system when in the community”. Rosen and colleagues (2008:2278) also found that the excess of ex-prisoner deaths from “injuries and medical conditions common to prison populations highlight ex-prisoners’ medical vulnerability and the need to improve correctional and community preventive health services”.

Offender mortality study

Research into offender mortality in New Zealand has (until recently) been limited which meant it could not be conclusively demonstrated that lower life expectancy was an issue amongst the offender population. This is in part due to limitations in data collection. For instance, the Department is unable to record reliable data on offender deaths unless the event occurs in prison; death may be recorded when it occurs before the end of a community sentence, but this is not always done.

With the advent of the Integrated Data Infrastructure (IDI), many new kinds of statistical analysis are now possible within any given sub-population, including offenders. The IDI facility allows data from various government agencies to be brought together and linked according to individual identity. A major advantage of using data in the IDI is that it is not susceptible to systemic bias in recording of cases. Whilst data may not be 100% complete in either the Department of Internal Affairs’ (DIA) or Ministry of Health’s (MoH) mortality data, it does not bias for or against any one class of individual (in this case, offenders). As a result, it offers the potential to generate many new insights into sub-populations of interest.

Methodology

Data from Births, Deaths and Marriages (DIA) and the MoH’s mortality data have been matched within the IDI against offender identities held by Corrections. The current analyses sought to identify the following:

  1. the proportion of deaths within an age-range of offenders
  2. the rate of deaths within the age-range
  3. the life expectancy of offenders
  4. the causes of deaths;

and to compare rates with those recorded across the New Zealand population as a whole.

The first three analyses used data derived from the DIA data tables. The DIA table also contains the causes of death. However these are in free text form and are, therefore, quite difficult to analyse efficiently. To remedy this, MoH mortality data was used instead for cause of death. This data covers people who died either in a hospital, or had the cause of death recorded by a hospital. It presumably excludes the deaths of the elderly in their own home where there are no suspicious circumstances. For the offender population, nearly 80% of all deaths had a MoH cause of death recorded. The proportion for the general public was only slightly less at 76%.

Results of the study

Corrections holds records of around 365,000 persons. Since 2005, matched data indicates that around 16,200 offenders have died. In the same time period, the total number of deaths in New Zealand was about 346,000. Figure 1 shows the distribution of mortality by age group, for both offenders and non-offenders.

The difference in the pattern of mortality is striking: peak mortality for “all persons” is in the above-75 group, while for offenders it is in the 46-55 and 56-65 year age-groups. A logical reason accounts for the differential in distributions: most offenders are relatively young at time of sentencing, and Corrections’ data is patchy and incomplete for those dealt with prior to the mid-1970s. Together, these reasons mean that the pool of known offenders who now are (or would have been) elderly is actually quite small.

Figure 1:
Distribution of mortality by age-group

This is graphically shown in Figure 2. Given the disparity of the over-75 group, it is unsurprising that this is reflected in the mortality percentages of offenders.

A better way of investigating the mortality of offenders is via rates of mortality within distinct age groups. Given the differences between the offender and non-offender population in terms of gender and ethnicity distribution, this was done for four categories:

  1. male and Mäori/Pacific ethnicity
  2. female and Mäori/Pacific ethnicity
  3. male and “other” (including NZ European) ethnicity
  4. female and “other” (including NZ European) ethnicity.

From these figures, ratios for offender vs non-offender rates can be calculated; these “odds ratios” are shown in Figure 3 and can be thought of as the multiplier effect of being an offender on mortality.

What these show is quite remarkable: the highest ratios are non-Mäori/Pacific males and females in the youngest age-bands (i.e. from 17 to 45 age groups). Offenders in these age bands are six to seven times more likely to die within that age band than are non-offenders of the same age, sex and ethnicity.

Figure 2:
Distribution of groups still alive

Figure 3:
Ratio of mortality rates for offenders, relative to non-offenders.

Cause of death

MoH morbidity tables list the causes of deaths in New Zealand. The distribution and ranking of cause of all deaths for all New Zealanders, and for offenders, are shown in Tables 1 and 2. Also shown are the mean age for the cause of mortality and relevant percentages.

The most common causes of death are heart disease and cancer, which applies equally to both offenders and the wider public.

However, the third most common cause of death for offenders was found to be intentional self-harm (suicide). Suicides thus account for 8% of all offender deaths, compared to just 1.6% across the wider public. In relation to suicide, the mean age of mortality is 36.1 years. Other causes that feature more prominently amongst offenders are accidents (transport, non-transport) and assaults (i.e., homicides).

Table 1:
Distribution of causes of mortality: All New Zealand

Category (ICD-10)

Mean age

Percent

Heart and circulatory system diseases

79.0

40.9

Neoplasms (malignant, in situ, benign)

71.9

26.1

Respiratory system diseases

79.7

7.6

Nervous system diseases

74.9

3.7

Endocrine, nutritional and metabolic diseases

72.5

3.4

Mental and behavioural disorders

86.7

2.9

Digestive system diseases

78.1

2.6

Other external causes of accidental injury

65.4

2.4

Intentional self-harm

40.9

1.6

Transport accidents

40.9

1.5

Conditions originating in the perinatal period

0.1

1.4

Genital, urinary system diseases

82.9

1.4

Skin diseases

80.7

1.0

Musculoskeletal system diseases

77.1

0.9

Congenital malformations, deformations
and chromosomal abnormalities

15.6

0.9

Infectious and parasitic diseases

71.0

0.7

Assault and other

46.7

0.4

Not elsewhere classified

55.1

0.3

Blood and blood-forming organ diseases

74.4

0.2

Pregnancy, childbirth

34.5

0.0

Unknown

80.4

0.0

Table 2:
Distribution of causes of mortality: Offenders

Category (ICD-10)

Mean age

Percent

Heart and circulatory system diseases

57.8

35.7

Neoplasms (malignant, in situ, benign)

58.5

22.6

Intentional self-harm

36.1

7.9

Respiratory system diseases

62.0

6.4

Transport accidents

38.5

5.9

Other external causes of accidental injury

43.6

5.2

Endocrine, nutritional and metabolic diseases

55.7

4.8

Digestive system diseases

57.2

3.0

Assault and other

39.5

2.4

Nervous system diseases

51.9

2.0

Certain infectious and parasitic diseases

52.2

1.1

Skin diseases

65.4

0.7

Mental and behavioural disorders

60.5

0.7

Genital, urinary system diseases

59.1

0.6

Musculoskeletal system diseases

53.8

0.3

Congenital malformations, deformations and chromosomal abnormalities

43.3

0.2

Not elsewhere classified

46.6

0.2

Discussion

This analysis provides statistical evidence for what until this point could only be hypothesised; that having a criminal history is associated with a shortened life span. The data matching exercise using Corrections, MoH, and Department of Internal Affairs data shows, on average, a person with a criminal history in New Zealand has a life expectancy of 64 years, which is between 10 and 15 years less than that enjoyed by the average New Zealand citizen. This could have important implications for future decision making and practice surrounding offenders’ health and wellbeing.

Cause of death data largely confirm that lifestyle, risk-taking and psychological issues are key to understanding this disparity. People with criminal histories have a higher chance of dying in car accidents, being killed by another person, and are more likely to take their own lives.

While consistent with research on offender mortality internationally, these are an uncomfortable set of findings. Further research is required to unpick the specific reasons for the different rates of death between groups of offenders and between offenders and non-offenders, but understanding these differences is important in informing where healthcare services and resources are directed in order to increase the lifespan of offenders, to target preventable illnesses and to provide sufficient mental health care, particularly post-release support.

Mental health services may be particularly important given this study found suicide to be a leading cause of offender mortality, a marked contrast to the general population.

This finding aligns with the Department’s piloting of increased availability of mental health services to both prisoners and offenders in the community. Corrections already works to support offenders to address their health and mental health needs, especially as unmet needs can impact on rehabilitation. For example, in 2012, the Department introduced a mental health screening tool to identify prisoners’ mental health needs and improve their care. Prisoners undergo a number of other checks and assessments for their mental health needs during their time in prison. These include drug and alcohol screening, psychological evaluation, and assessments to check if they are at risk of self-harm or suicide. In addition, all prisoners are seen by a registered nurse following their arrival to prison. Healthcare staff engage prisoners in over 100,000 consultations a year.

Following a comprehensive survey of prisoner mental health needs (Bowman, 2015) Corrections is investing further in mental health services, including packages of support, counselling, post-release family services and supported accommodation. Corrections is also working on additional alcohol and other drug (AOD) aftercare services to be delivered over the next two years, including AOD aftercare workers to support graduates of our Drug Treatment Units and Intensive AOD Treatment Programmes.

While the implications of the research findings presented here will require further analysis and consideration, one implication is to reinforce the value of offender rehabilitation. Arguably, effective rehabilitation not only benefits society by reducing crime, but may directly benefit offenders through improved health outcomes and increased lifespan.

As such, the findings of this analysis may serve as useful motivational information for engaging offenders in the process of rehabilitation. The findings also have relevance to current work relating to managing risk of self-harm amongst those on community sentences and orders.

This analysis opens the door for future valuable research and analysis to show the extent and direction of correlation between mortality and other variables such as, type of offences, length of criminal career, and time spent in prison.

Conclusion

This analysis demonstrates that offenders have different mortality patterns to the general public. For some age and ethnic groups, the likelihood of offenders dying is much higher than non-offenders of the same age and ethnic cohort. In addition, self-harm is a leading cause of mortality for offenders.

Understanding these different patterns will enable relevant agencies to develop policy responses to improve the health and well-being of persons with a criminal history.


* It is possible that this is also a result of, or determined by, socio-economic factors.
** Disease related causes of mortality were not considered in this study
***Sattar and Killias (2005) noted that their finding that suicide was the third cause of death was unusual and lower compared to other studies of imprisonment and death which have found suicide to be the top cause of prisoner death. They explained that their finding of a comparatively lower rate of suicide was likely due to the fact that their sample did not include pre-trial detainees


References

Bowman, J. (2015) Co-morbidity research – Part one. Practice, The New Zealand Corrections Journal, Vol 3, Issue 2, 33 – 34. Department of Corrections

Farrell, M. & Marsden, J. (2007). Acute risk of drug-related death among newly released prisoners in England and Wales. Addiction, 103, pp.251-255.

Kariminia, A., Law, M. G., Butler, T. G., Corben, S. P., Levy, M. H., Kaldor, J. M. & Grant, L. (2007). Factors associated with mortality in a cohort of Australian prisoners. European Journal of Epidemiology, 22, pp.417-428.

Kariminia, A., Jones, J. & Law, M. (2012). Increased mortality among Indigenous persons during and after release from prison in New South Wales. Australian and New Zealand Journal of Public Health. 36(3), pp.274-80.

Pratt, D., Piper, M., Appleby, L., Webb, R. & Shaw, J. (2006). Suicide in recently released prisoners: a population-based cohort study. Lancet, 368, pp.119-23.

Pridemore, W. A. (2014). The Mortality Penalty: Evidence from a Population-based Case-control Study of Working-age Males. Journal of Health and Social Behaviour, 55(2), pp.215-233.

Pritchard, C., Cox, M. & Dawson, A. (1997). Suicide and ‘violent’ death in a six-year cohort of male probationers compared with pattern of mortality in the general population. Journal of the Royal Society of Health, 117(3), pp.180-5.

Rosen, D. L., Schoenbach, V. J. & Wohl, D. A. (2008). All-Cause and Cause-Specific Mortality Among Men Released From State Prison, 1980-2005. American Journal of Public Health, 98(12), pp.2278-2284.

Sattar, G. & Killias, M. (2005). The Death of Offenders in Switzerland. European Journal of Criminology, 2(3), pp.317-340.

Van Dooren, K., Kinner, S. A. & Forsyth, S. (2013). Risk of death for young ex-prisoners in the year following release from adult prison. Australian and New Zealand Journal of Public Health, 37(4), pp.377-382.