There was a modest increase in the number of prison to hospital transfers between 2009 and 2022, though this number was likely constrained by hospital bed availability, and it was low relative to the total prison population. The cohort was predominantly male, with a median age in their early 30 s at the time of first transfer. Over 80% of people transferred from prison to psychiatric hospital units for treatment had a psychotic disorder or bipolar disorder. This suggests people in prison who may warrant intensive psychiatric treatment but are neither psychotic nor acutely behaviourally disturbed are seldom transferred to a psychiatric hospital in New Zealand. People who are unlikely to be transferred from prison to hospital care include those who are suicidal but not psychotic; people at risk of severe self-harm behaviour; and those with an acute crisis related to a personality disorder, adjustment disorder or posttraumatic stress disorder.

The high proportion of Māori (55%) in the cohort reflects their prevalence in the New Zealand prison population, at about 52% [7], compared to 17% in the New Zealand general population. Their true prevalence in our cohort may be even higher than we reported, as Māori are under-counted in health and disability data [21]. The over-representation of Māori in the cohort highlights health and social inequities and marginalization stemming from European colonization in New Zealand over more than two centuries [22]. Colonization has brought about widespread harms for Māori including loss of resources, disconnection from land and social systems, and intergenerational trauma. The 1840 Treaty of Waitangi was intended to provide constitutional protection for Māori, however the continuing inequities show successive New Zealand governments have failed to protect Māori and support their wellbeing and self-determination. Our data highlight a need for strategies to avoid criminalizing Māori with serious mental illness. The policy solutions must address the structural drivers of incarceration for Māori, which span multiple areas of government policy including health, education, justice and economic development. Better resourcing for Māori-focused health services is necessary to improve outcomes for Māori with serious mental illness who experience incarceration [23]. Such services need to be designed and delivered by Māori.

People in New Zealand prisons have access to a range of mental health treatment including medication and psychotherapy. These services are delivered by various providers including the Department of Corrections, non-governmental organisations, private providers, and forensic mental health services. However, the level of access to hospital care for people who are acutely or severely unwell is not equivalent to what is available in the community, with lengthy delays in transfer being reported [24]. Some of these patients are placed in special management units, often with solitary confinement, while awaiting transfer [25]. New Zealand’s regional forensic psychiatric services provide in-reach care to all New Zealand prisons but these services have not expanded in proportion to increasing demand [4]. This increase in demand is driven in part by a steep increase in people on remand while awaiting the outcome of their legal matters. The needs of this group are often higher than they are after sentencing, due to the stress of legal proceedings and dislocation from their usual living situation among other factors. The increasing prevalence of acute methamphetamine-related mental health problems, particularly psychosis, has also driven demand for forensic inpatient care. Over 10% of people in New Zealand prisons have a current (past 12 months) methamphetamine use disorder [12] and the prevalence of methamphetamine use disorder in our sample was 17%.

The challenges for forensic mental health services are not unique to New Zealand, and similar issues have been noted in the provision of prison mental health care in England and Wales [6] and Ireland [5]. While New Zealand has had prison mental health surveys that document the burden of mental illness across the whole prison population [12, 26], there is an urgent need for research into the level of unmet need for treatment, particularly acute mental health care that cannot be delivered in prison. As acutely mentally unwell people in prison are often managed in solitary confinement, research is also needed to estimate the extent to which human rights violations are taking place. Relatedly, there is a need for more transparent and robust processes to decide who is transferred from prison to hospital. Structured decision-making tools to guide admission to forensic units, such as DUNDRUM, can be a useful adjunct to existing decision-making processes [27]. However, these tools have not yet been implemented throughout New Zealand.

Once transferred out of prison, the cohort had a high level of ongoing mental health service utilization and healthcare need. They spent 17% of the follow up period in a psychiatric hospital bed, about 70% of which was in a forensic ward, and half received compulsory mental health treatment in the community. We found an over 40% prevalence of substance use disorder comorbidity in the cohort. However, this is almost certainly an underestimate, as it is far lower than the prevalence in the general prison population [12]. This could be due to diagnostic overshadowing, incomplete recording of substance use disorder diagnoses by clinicians, or incomplete diagnostic data in the PRIMHD dataset – likely it is a combination of these factors. More accurate data on substance use disorders in this population is vital, since substance use is an important modifiable risk factor for adverse outcomes including premature death, as it is for other groups experiencing incarceration [13].

The cohort had a more than fourfold elevation in age and sex-adjusted mortality compared to the general population. While the absolute number of deaths among females in the cohort was small, their risk was elevated ninefold compared to women in the general population. Māori, who made up over half the cohort, also experienced considerably elevated mortality compared to Māori in the general population. The relative increase in the risk of death found for this sample was greater than that found in a recent study of people released from incarceration in New Zealand, which found SMRs of 3.8 for women and 2.7 for men [14]. It was also higher than estimates of increased premature mortality among adults with psychotic disorders using mental health services in New Zealand, who have a standardised mortality rate about three times that of the general population [15].

Analysis of the available cause of death data showed that about 40% of the deaths were from injuries, predominantly overdoses and self-injury including suicide. Cancer and non-communicable cardiorespiratory diseases accounted for most of the non-injury deaths. Many cases were still under investigation by the Coroner, meaning the true ratio of injury to natural deaths is not clear. Recent data from the United States show an immense concentration of suicide deaths among people released from prison [28]. Considering the difficulty in predicting and preventing suicide at an individual level, there is a need for a broad and multisectoral approach that addresses the diverse needs of people released from incarceration [28].

The main strength of this study is that it is a national sample comprising all people transferred under the Mental Health Act from New Zealand prisons to psychiatric hospitals within the study period. The study excludes a small group of people transferred under different kinds of order, for example for Court-ordered inpatient psychiatric reports. The use of linked administrative data is both a limitation and a strength. Available administrative data in New Zealand provide limited detail on why people were transferred or their mental health status at the time. Diagnoses were from clinical coding data and as noted above, this is likely to under-report any disorders that are not accurately recorded in health records. Cause of death information was incomplete for about one third of deaths, chiefly because of delays in Coronial inquiries. Despite this, our use of national data spanning 15 years ensured near-complete ascertainment of mortality in the cohort. Relatively few members of the cohort are likely to have left New Zealand, and therefore to have missing outcome data, but we cannot exclude the possibility a small proportion of the cohort was lost to follow up. To the extent that this occurred, our estimates of mortality in the cohort would be conservative.

In summary, people who are transferred from New Zealand prisons to hospital for mental health treatment are predominantly young men with psychotic disorders or bipolar disorder, with an alarmingly high proportion of Māori, the indigenous people of New Zealand. They are at risk of premature death, which is likely to be driven by both illness factors and a range of social determinants including unstable housing and lack of employment. Such markers of social exclusion appear to be associated with risk of death in a dose–response fashion [29, 30]. The diverse nature of these risk factors implies a need for coordinated, multi-sectoral prevention strategies [13] and is not solely the role of forensic mental health and correctional services [28]. There also needs to be a commitment to providing high quality physical and mental health care for people in contact with the criminal justice system. Policy solutions aimed at this group of people must urgently be developed in partnership with Māori, who currently make up over half this population.