Pregnancy, prison and perinatal outcomes in New South Wales, Australia: a retrospective cohort study using linked health data

New South Wales (NSW) has the largest overall population of the Australian states and territories. The state accounts for almost one-third of Australia’s births annually and 30% of the Australian female prisoner population [1]. In NSW and Australia as a whole, the majority of imprisoned women are of childbearing age. Women in NSW prisons typically experience a series of short sentences and/or periods of remand, which constitutes ‘a form of serial institutionalisation’ [2]. The average daily number of women in full-time custody in NSW in 2012 was 661 [3]. Most are held in three female-only institutions, with smaller numbers in female-only units, work camps and transitional centres. In the state of NSW, the standard term of reference for Indigenous Australians is Aboriginal peoples, ‘in recognition that Aboriginal people are the original inhabitants of NSW’ [4]. Aboriginal women are severely over-represented in the NSW prisoner population, comprising 30% of female prisoners in NSW [1] but only 2% of the general population. Their rate of imprisonment in NSW rose by 12.5% between 2006 and 2009 alone [5].

Women in prison have a higher number of pregnancies and a disproportionate number have given birth before the age of twenty, when compared with the general birthing population of NSW. The NSW Inmate Health Survey is the only source of data on the birthing histories of female prisoners in the State. Eighty-two percent of women who participated in the 2009 NSW Inmate Health Survey had been pregnant at least once, including 27% who reported having had five or more pregnancies [6]. In contrast, between 2002 and 2006, 1.6% of women in the general NSW population who gave birth had five or more previous pregnancies of at least 20 weeks duration [7]. Whilst the figures are not directly comparable, these published data also show that more than half (51%) of women in the NSW Inmate Health Survey with a history of childbirth gave birth before twenty years of age, whilst the proportion of pregnancies of at least 20 weeks duration in the NSW population which were to women aged less than twenty years was only 4% between 2002–2006 [6, 7].

Studies in the United States (US) and United Kingdom (UK) have compared the maternity outcomes of women in prison with those of population control groups [8‐10], and disadvantaged control groups [11, 12]. Findings are equivocal, but a systematic review concluded that imprisoned women were more likely to experience poorer maternity outcomes than women in the general population, yet less likely to experience poorer outcomes than other disadvantaged women [13]. Studies have suggested that imprisonment provides safety from abusive relationships, better nutrition, improved access to health care, and cessation of drug and alcohol use [8‐12]. If this is the case, improvements in outcomes for imprisoned women in NSW may be less evident, since they are imprisoned for shorter periods than their US counterparts. The mean length of stay in full-time custody for sentenced female prisoners in NSW is 196 days (or 300 days for remand to sentenced custody) [14], compared with 18 months (547 days) in US state prisons [15].

The aim of this study is to determine whether women imprisoned during pregnancy have improved perinatal outcomes, compared with women in the general community, and women who share similar socio-demographic characteristics including a history of incarceration (at a time other than during their pregnancy).

In this study, imprisonment during pregnancy did not improve maternal and perinatal outcomes for similarly disadvantaged women. Unlike previous studies [13], which found that imprisoned pregnant women had better perinatal outcomes than disadvantaged controls, our study found no evidence of any significant benefits for women or neonates. We attribute this to using controls that were themselves prisoners during the same time period, but not during pregnancy. Pregnant prisoners in our study were at the extreme end of the continuum of social disadvantage (Table 1) and were 12.6 times more likely to be of Aboriginal and Torres Strait Islander origin than women in the community. Furthermore, 78.5% had either a mental health and/or drug and alcohol disorder; we concluded that these morbidities, coupled with very high rates of smoking during pregnancy, primarily accounted directly or indirectly for their significantly higher risk of poorer maternal and perinatal outcomes. The prison control group shared these characteristics, and our study demonstrates that interventions received whilst pregnant in prison were not successful in mitigating poorer outcomes.

Low birthweight was investigated in the only other study to use incarcerated women as a control group [11]. This US study, which also compared incarcerated pregnant women (n = 168) with women imprisoned at a time other than during their pregnancy, found that pregnant prisoners were less likely than prison controls to have low birthweight babies (adjusted OR 0.54, 95% CI 0.32-0.93). The same study concluded that following adjustment for social disadvantage, pregnant prisoners were not significantly more or less likely to have low birthweight babies than women in the community. Our study adjusted primarily for co-morbidities and found that there was no significant difference between the prison pregnancy and prison control groups (adjusted RR 0.93, 0.69-1.24), but that pregnant prisoners were significantly more likely than community controls to have a low birthweight baby (adjusted RR: 1.71, 95% CI 1.28-2.30). The US study also found that birthweight increased with number of weeks of pregnancy in prison [11]. The ‘high-volume, short-term’ nature of women’s imprisonment in Australia [21] may have diluted our findings. Shorter lengths of stay of women in Australian prisons may provide insufficient opportunity to improve birthweight. Mental health, including substance use disorders, was highly significant in all of our regression models, but was not included in the aforementioned study [11]. It is well established that poor maternal mental health is a risk factor for poorer neonatal and infant outcomes, and this risk increases during pregnancy and the postnatal period [22]. It is evident that consideration of maternal mental health is essential to any study of the maternity outcomes of prisoners, and indeed to any intervention with this group.

The literature is equivocal on incarceration and pre-term birth, with a small US cohort study of 69 prisoners who received prenatal care from a single medical centre finding no difference from community controls. In contrast, two studies found pregnant prisoners were more likely to have pre-term births than community controls [8, 9]. Whilst this finding held true in our study, there was no statistically significant difference between the prison pregnancies and prison controls. Apart from these two studies, no others have used disadvantaged controls to investigate pre-term birth.

We found a higher rate of neonatal hospital admission and length of stay in SCN/NICUs for babies born to pregnant prisoners, when compared to community controls. This persisted after controlling for known risk factors. This finding conflicts with an earlier US study that found no difference between prisoners and hospital controls, and reflects the potential burden of mental health and/or drug and alcohol disorders found among prisoners [9]. We found no significant difference when comparing prison pregnancies with prison controls.

Our study found that birthing prisoners had rates of Caesarean section and induction of labour that were comparable with the community, whilst former pregnant prisoners and prison controls tended to be more likely to have a spontaneous vaginal birth. The post-release period is one of high risk, when contact with health services is likely to be erratic [2]. It is possible that women in the latter two groups presented at a more advanced stage of labour. Imprisoned women and their clinicians, given the challenges of birthing spontaneously in a prison setting, may favour increased intervention and ‘control’ over birth. Data on place of birth would assist in interpreting these findings.

Our study found no evidence of a statistically significant difference in the likelihood of fetal death between the prison or control groups. However, the number of deaths was small; and of the two fetal deaths in the prison pregnancy group, neither was to a birthing prisoner. A UK study suggested that there might be a reduced fetal mortality rate among imprisoned women, compared with births in families where at least one parent was under the Probation Service; however, a more carefully controlled study was recommended [10]. This study provided the only evidence for the subsequent systematic review that suggested that imprisoned women might be less likely to have a stillbirth than similarly disadvantaged controls [13]. This was not our finding and no other studies have examined this outcome.

Internationally, there remains wide variation in prison maternity care. Despite the availability and proximity of prison medical facilities, women in our study who were pregnant in prison were more than twice as likely to initiate antenatal care after 20 weeks’ gestation compared with prison controls. One US study reported that prenatal education, childbirth instruction, and general health education classes were mandatory for pregnant prisoners [9]. A 2006 report by the UK’s Maternity Alliance found ‘a high level of variation in the quality of antenatal care between prisons… The good practice found in some establishments was not replicated in others’ [23], and a report by the UK National Childbirth Trust (NCT) noted that prison routines and poor communication interfered significantly with pregnant women’s participation in antenatal classes, during a pilot study at one of England’s largest women’s prisons [24].

This study has a number of strengths. It is population based and uses women with a history of incarceration as a control group to investigate a range of important maternal and perinatal outcomes. Additionally, the study design incorporated linkage of data on mental health and drug and alcohol disorders allowing us to control for these highly significant confounding factors. However, for perinatal deaths and maternal morbidity, the small population of incarcerated pregnant women likely contributed to the lack of statistically significant findings. Information about socio-demographic factors that have a well-established correlation with both maternity outcomes and female imprisonment, including maternal education, employment, and marital status as well as health risk factors such as sexually transmitted infections, blood-borne viruses, previous pregnancies and maternal weight, would assist with the interpretation of findings. More robust data about incarceration histories, police and Juvenile Justice involvement would enhance our understanding of the interaction between health and social factors in maternity outcome for these women, and help to signal opportunities for intervention. Data about legal status (remand or sentenced) would indicate the type of accommodation and level of access to services women may have had whilst imprisoned. Data on the timing of onset of antenatal care would assist in understanding the nature of the care women had received. Further detailed analyses will be undertaken to determine whether the number of days’ incarceration during pregnancy, and gestation at incarceration, are related to differences in outcome for women and babies.

The notion that prisons are “therapunitive” environments is not supported by empirical evidence [25]. Prison health services are ill equipped to mediate the cumulative disadvantages, both health and social, to which their clients are exposed throughout the life-course. Post-release, most return to ‘liminal, marginal spaces’, which provide the context for re-incarceration [2].