Determinants of infant mortality for children of women prisoners: a longitudinal linked data study

Determinants of infant mortality for children of women prisoners: a longitudinal linked data study Background: There is limited information on the determinants of infant mortality outcomes for the children of women prisoners. This study aimed to explore determinants of infant mortality for Indigenous and non-Indigenous children, with a specific focus on maternal imprisonment during pregnancy as a risk factor. Methods: Using linked administrative data we obtained a longitudinal sample of 42,674 infants born in Western Australia between October 1985 and June 2013. Data were analysed by maternal contact with corrective services, including; (i) imprisonment during pregnancy, (ii) imprisonment before (but not during) pregnancy, (iii) imprisonment after birth, (iv) community-based correctional orders (but no imprisonment), and (v) no corrections record. Infant mortality rates were calculated. Univariate and multivariate log-binomial regression was undertaken to identify key demographic and pregnancy-related risk factors for infant mortality. Risk factor prevalence was calculated for infants by maternal corrections history. Results: 430 Indigenous and 116 non-Indigenous infants died aged 0–12 months. For singletons, infant mortality rates were highest in Indigenous infants with mothers imprisoned during pregnancy (32.1 per 1000) and non-Indigenous infants whose mothers were first imprisoned after birth (14.2 per 1000). For all Indigenous children, the strongest determinants of infant mortality were: abruptio placentae and other placental disorders (RR = 2.85; 95%CI 1.46–5.59; p =0. 002), maternal imprisonment during pregnancy (RR = 2.55; 95%CI 1.69–3.86; p < 0.001), and multiple gestation (RR = 2.29; 95% CI1.51–3.46; p < 0.001). Indigenous and non-Indigenous infants with mothers imprisoned at any time, and particularly before or during pregnancy, experienced higher prevalence of key pregnancy risk factors. Conclusions: This is the first comprehensive study of the determinants of infant mortality for children of women prisoners. Infants with any maternal corrections history, including community-based orders or imprisonment outside of pregnancy, had increased infant mortality. Indigenous infants whose mothers were imprisoned during pregnancy were at particular risk. There was a low incidence of infant death in the non-Indigenous sample which limited the investigation of the impact of the specific aspects of maternal corrections history on infant mortality. Non-Indigenous Infants whose mothers were imprisoned before or during pregnancy experienced higher prevalence of pregnancy risk factors than infants of mothers first imprisoned after birth. Keywords: Infant mortality, Women prisoners, Linked data, Australia * Correspondence: caitlin.dowell@mymail.unisa.edu.au Health Economics and Social Policy Group, Centre for Population Health Research, Sansom Institute, School of Health Sciences, University of South Australia, GPO Box 2471, Adelaide, SA 5001, Australia Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 2 of 16 Background Infant Death Syndrome [10], as well as increased paren- Women prisoners constitute a highly vulnerable popula- tal employment and maternal education [11]. Although tion which is exposed to multiple and complex risk fac- the overall rates of infant mortality declined between tors, including domestic violence, substance abuse, 1980 and 2001, the disparity between Indigenous and poverty, discrimination and mental illness, placing them non-Indigenous populations increased from a Relative and their children at risk of poor pregnancy and health Risk of 3.0 (95%CI 2.5–3.6) in 1980–84 to 4.4 (95%CI outcomes [1]. However, few reports, exist which have in- 3.5–5.5) in 1998–2001 [10]. For Indigenous infants, vestigated pregnancy outcomes for children of women postneonatal mortality was higher than neonatal mortal- prisoners [2, 3]. Of those studies that have investigated ity, a pattern that indicates the impact of socioeconomic pregnancy outcomes for infants, the main outcomes in- disadvantage and marginalisation on infant mortality vestigated include preterm delivery and low birth weight outcomes [10]. [3]. The few studies that report on infant death have The primary objective of this study was to explore deter- been limited by small numbers of events, in part due to minants of infant mortality for Indigenous and the size of the cohort sampled [3, 4]. non-Indigenous children, with a focus on whether maternal While the international literature on the pregnancy out- imprisonment during pregnancy is a risk factor for infant comes of prisoners is equivocal, a review of studies from mortality. The specific aims of the study were to determine: across the United States (US), United Kingdom and Eur- infant mortality rates for Indigenous and non-Indigenous ope found women imprisoned in pregnancy generally had infants of mothers with different corrections histories; the poorer maternal and infant outcomes than community key demographic and pregnancy-related risk factors which controls, and better outcomes than disadvantaged com- may contribute to infant mortality in Indigenous and munity controls [3]. It has been concluded that this may non-Indigenous populations; the importance of maternal indicate that imprisonment in pregnancy may be benefi- corrections history as a determinant of infant mortality cial for some pregnancy outcomes [3]. The only Australian after accounting for significant demographic and study on pregnancy outcomes of women prisoners found, pregnancy-related risk factors; and the prevalence of key however, that women imprisoned during pregnancy did risk factors for infant mortality between infants with differ- not have better perinatal outcomes than women impri- ent maternal corrections histories. soned at times other than pregnancy [4]. Thus the unique context of the justice systems and prisoner populations in Methods specific jurisdictions limit the transferability of findings Study design across jurisdictions in the absence of an understanding of We have used data from a large data linkage project the determinants of pregnancy outcomes for women to explore infant mortality within Indigenous and prisoners. non-Indigenous children of mothers who have been Studies in the US have found racial differences in the exposed to the corrections system at different times pregnancy outcomes of women prisoners [5]. In Australia, in relation to their pregnancy. We compared infant research on the pregnancy outcomes of Indigenous mortality outcomes for children whose mothers had; women prisoners is lacking despite their overrepresenta- (i) any period of imprisonment during pregnancy, (ii) tion in the prison population. In Western Australia, for imprisonment before (but not during) pregnancy, (iii) example, Indigenous peoples represent 4% of the general their first period of imprisonment after birth, (iv) population but 46% of the female prison population [6, 7]. community-based correctional orders (but no impris- This reflects the high levels of social and economic disad- onment), or (v) no corrections record at any time vantage and discrimination experienced by Indigenous over the study period. peoples in Australia [8]. Similarly, Indigenous peoples in Australia experience poorer pregnancy outcomes com- pared with non-Indigenous mothers, and while infant Conceptual framework mortality rates have been improving for both Indigenous Mosley and Chen’s[12] analytical framework for the study and non-Indigenous populations across time, the racial of child survival was adapted for the present study. The disparity remains [9, 10]. basis of this framework is that broader determinants ne- Across 1980 to 2001, infant mortality rates in Western cessarily act through biological pathways, or mechanisms, Australia declined for both Indigenous (25.0 in 1980–84 which impact before, during and after pregnancy on the to 16.1 in 1998–2001) and non-Indigenous infants (8.4 healthy development of the fetus and infant, and ultim- in 1980–84 to 3.7 in 1998–2001) [10]. Some important ately on infant mortality. There is evidence that adverse changes that occurred across the past 30-years include events experienced before and during pregnancy can im- improved transport for rural and remote pregnant pact on fetal development and result in increased risk of women, immunisation of infants, prevention of Sudden poor infant and childhood health outcomes [13–17]. Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 3 of 16 Our adaptation of this framework first groups together cause of death, diagnosis, or reason for health service key demographic factors, including birth year, sex, Indigen- contact, respectively. For HMDC records, we obtained ous status, socioeconomic status, and geographical remote- one code for the principal diagnosis of the episode of ness. The second grouping includes baseline pregnancy risk care, and up to four codes for external causes of factors, such as multiple gestation, birth spacing < episodes of care. Over the study period, the ICD 9th Re- 18 months, maternal age, and parity, which are largely vision with Clinical Modification (ICD-9-CM) [21] re- unmodifiable from the commencement of pregnancy [18]. lated to services contacts before July 1999, and service The third grouping includes key pregnancy complications contacts from that date used ICD 10th Revision with that might be a precursor of infant mortality, such as nutri- Australian Modification (ICD-10-AM) [22]. tional deficiencies, placental disorders, prematurity, and in- fection [19]. The last group includes other maternal factors Study population and exposures which are known risk factors for infant mor- The study population was drawn from a retrospective lon- tality or may indicate maternal vulnerability or household gitudinal cohort study of all liveborn children born in dysfunction, such as substance use or mental health related Western Australia from 1985 to 2011 whose biological service contacts, external causes of injury, and having other mother was imprisoned at least once within 18-years after children in contact with the child protection system. their birth. The cohort study population included a com- parison group of children whose mother had no record of Data sources imprisonment from their date of birth to their 18th birth- Data were obtained through the Western Australian day, which was identified through the same data sources Data Linkage System (WADLS). The WADLS uses as the cohort and matched 3:1 to cohort children on Indi- highly-accurate computerised, probabilistic matching genous status, age and sex. Data on second-generation with clerical review to create linkages within and be- children, born between 1998 and 2014 to the female tween administrative data collections across a range of members of the cohort and comparison group, were also Western Australian government agencies [20]. The obtained. Western Australian Data Linkage Branch conducted the Stillbirths (second-generation only) and infants with linkage and provided de-identified data extracts. Records chromosomal abnormalities, identified through HMDC were extracted from the Midwives Notifications System, and death records, were excluded (Fig. 1). Erroneous re- Birth Registrations, Death Registrations, Department of cords with multiple mothers or missing key information Justice, Hospital Morbidity Data System Collection such as birthdate were removed. The final study popula- (HMDC), Mental Health Information System (MHIS), tion was restricted to children born from October 1985 to and Department of Communities: Child Protection and June 2013 (inclusive) to ensure pregnancy exposure and Family Support (CPFS) data collections. These are all death data was available for all infants. statutory State-wide data collections with complete In total, there were 42,674 infants in the final study coverage. population, 37,469 from the first-generation (original co- The Birth Registration and Midwives Notifications Sys- hort and comparison group) and 5205 from the tem data provided social and demographic characteristics second-generation (children of the original cohort and of mothers and children at time of birth. Mortality data comparison group). Data from the birth and death regis- include all deaths registered in Western Australia. The trations, midwives notifications records, and CPFS data Department of Justice data collection includes all custodial were available for first- and second-generation children. records for offenders held in Western Australian prisons Only first-generation children had HMDC record data. and records for offenders on community-based correc- Mothers had corrections, HMDC, and MHIS data tional orders. Data excluded unsentenced individuals available. detained in police stations and courts, immigration deten- tion centres, and mental health facilities. The HMDC in- cludes all inpatient records for Western Australian public Definition of maternal corrections history and private hospitals and day surgeries. The MHIS in- The study population was categorised into: a) infants cludes presentations to all inpatient and public commu- whose mothers had a record of imprisonment at any nity mental health services. The CPFS data include all time (n = 7317 Indigenous; n = 3504 non-Indigenous); b) reports of concerns for child welfare made to the child infants whose mothers had community-based correc- protection system and the details of investigations, protec- tional orders but no record of imprisonment (n =5828 tion applications and orders as well as placements in Indigenous; n = 653 non-Indigenous); and c) infants out-of-home care. whose mothers had no record within any Department of The Death Registrations, HMDC, and MHIS use the Justice database (n = 12,817 Indigenous; n = 12,555 International Classification of Diseases (ICD), to classify non-Indigenous). Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 4 of 16 Fig. 1 Selection of the study population and classification by maternal corrections history Imprisonment records covered prison stays of any Community-based sentences may involve treatment or length of time, and included unsentenced remandees vocational programs, community service, and place re- detained before trial as well as sentenced prisoners. In- strictions on offenders. Breach of conditions while on fants of mothers who had a prison record at any time over community orders may result in imprisonment. Accord- the study period were further categorised based on the ingly, women with community-based correctional orders timing of their mother’s imprisonment in relation to their are sentenced offenders, but may differ to women given pregnancy. The first group included infants whose custodial prison sentences in terms of severity or fre- mothers had any record of imprisonment during preg- quency of their offending and other individual factors. nancy. The second group included infants whose mothers They are not exposed to the prison environment which had imprisonment records in the period before, but not generally places more stringent conditions on offenders during, pregnancy. The third group included infants and has different implications for them and their families. whose mothers first record of imprisonment only oc- The proportion of infants in the various maternal cor- curred after the child’s birth. These groupings are shown rections history sub-groups (Fig. 1) relate only to the in Fig. 1. Classifications were based on mother’sprisonre- study sample and do not reflect the prevalence of these ception dates and the infant’s birth date. groups across the whole Western Australian population. Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 5 of 16 Pregnancy and birth dates siblings which enabled parity and duration of birth spa- Child month and year of birth was provided by the Mid- cing to be determined. Multiple gestation pregnancies wives Notification System, or if missing from the Birth were also derived based on siblings having shared birth Registration data. As gestational age was not available, dates or through maternal or child HDMC data. Child-level pregnancy start date was calculated as being nine HDMC data were not obtained for the second-generation months before the first day of the child’s birth-month. children, however, as stillbirths were captured for the second-generation this assisted in identifying multiple ges- Definition of infant mortality tation pregnancies. Infant mortality was defined as the death of a live born child under one year of age [23]. Full date of death was Pregnancy complications provided in the death registration data, however, birth The separate and combined effects of key pregnancy data were available only for month and year of birth. Ac- complications identified from maternal hospital records cordingly, infant mortality was defined as death within were evaluated. Pregnancy complications included the 12-full months after birth. For example, for a child born in effects of infection-related hospitalisations, anaemia, dia- January 2000, death on or before 31 January 2001 would betes, hypertension, preeclampsia, eclampsia, abruptio be determined within the category of infant mortality. placentae, placenta previa, other placental disorders, pre- mature rupture of membranes and renal disorders dur- Demographic characteristics ing pregnancy on infant mortality (Additional file 1) The Birth Registration and Midwives Notifications Sys- [19]. Complications were excluded due to low incidence tem data provides social and demographic characteristics or non-significance (p > 0.05), including hospitalisations of mothers and children at time of birth, including sex, for anaemia, diabetes, hypertension, preeclampsia, and socioeconomic status and geographical remoteness. Sex eclampsia (Additional file 1). was taken primarily from Midwives Data, or if missing from Birth Data. Other maternal risk factors and exposures Area-based socio-economic status of infant’s place of Maternal substance use (including alcohol) and residence at time of birth was assigned using the poisoning-related service contacts during pregnancy Socio-economic Indexes for Areas (SEIFA) Index of were identified from HMDC and MHIS data. Maternal Relative Socio-economic Disadvantage [24]. The smallest hospital admissions for any injuries from external area of SEIFA reporting is Collectors District (CD) level, causes, excluding substance use, self-harm and which is approximately 250 households or less in rural poisoning-related contacts were identified during preg- areas. Missing CD scores were imputed with mean nancy. Maternal hospital admissions for mental and be- CD-score by postcode before using broader area scores havioural disorders, self-harm, and mental health service of SEFIA available for Statistical Local Area or Local presentations, both excluding substance use and Government Area [25]. poisoning-related contacts, were identified during preg- The Accessibility/Remoteness Index of Australia nancy from HMDC and MHIS data. Having an older (ARIA) [26] was used to classify the geographical re- sibling in contact with child protection services during moteness of infant’s place of residence at time of birth. the infant’s pregnancy was also identified using sibling’s ARIA is derived from the measure of place of residence child protection data. to populated locations and key services and classified as major cities, inner regional, outer regional, remote, and Statistical analyses very remote locations. For the current study major cities All analyses were conducted using Stata Version 14.0. and inner regional areas were combined given both have All analyses were stratified by Indigenous status of the greater accessibility of relevant services [26]. infants. Indigenous status for infants and mothers in the study Infant mortality rates (per 1000 population) were calcu- populations was ascertained from the Derived Indigen- lated for singleton infants by maternal corrections history. ous Status Flag variable generated by the WADLS using Prevalence of demographic and pregnancy-related risk fac- best-practice algorithms, which assess individuals’ Indi- tors were calculated for Indigenous and non-Indigenous genous status across multiple data collections to en- populations. Log-binomial regression was used to calculate hance accuracy [27]. the Relative Risk of infant death for all univariate and multi- variate analyses. Baseline pregnancy risk factors The strength of correlation between all variables of inter- Maternal birth date was determined using all available est was assessed using Chi-square tests with Cramer’sV data sources. Maternal age was calculated as the age of statistic. For variable pairs with a medium effect size (> 0.3) mother at time of birth. Birth date was available for all [28], one variable was excluded from further multivariate Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 6 of 16 analysis on the basis of the univariate Relative Risk and imprisoned during pregnancy had the highest rates of in- level of statistical significance of each variable with infant fant mortality (32.1 per 1000), compared to infants mortality. Multivariate regression was then conducted for whose mothers were imprisoned either only before preg- each grouping of variables (demographic factors, baseline nancy (22.1 per 1000) or after birth (23.6 per 1000) pregnancy risk factors, pregnancy complications, and (Table 1). other maternal factors and exposures) separately with In non-Indigenous infants, those whose mothers were infant mortality, and variables that were not statisti- imprisoned before (9.4 per 1000) or during (8.2 per cally significant (p < 0.05) were excluded from further 1000) pregnancy had apparently lower rates of infant analysis (Additional file 2). All remaining variables were en- mortality than those whose mothers were imprisoned tered with maternal corrections history into a full regres- for the first time after their birth (14.2 per 1000) or who sion model (Model 1). Variables were removed by key had community orders alone (12.6 per 1000). This differ- groups; other maternal factors and exposures (Model 2), ence was not significant and confidence intervals wide, pregnancy complications (Model 3), baseline pregnancy risk possibly due to the low numbers of infant deaths within factors (Model 4), leaving the combined effects of maternal the sample of non-Indigenous children with mothers corrections history and demographic factors with infant imprisoned before or during pregnancy. mortality. Goodness of model fit was assessed from the Akaike Information Criterion (AIC) value. There were in- Univariate analysis of infant mortality sufficient numbers of non-Indigenous infant deaths whose As shown in Table 2, abruptio placentae and other pla- mothers were imprisoned before or during pregnancy to cental disorders (excluding placenta previa) were associ- undertake multivariate regression for non-Indigenous ated with the highest risk of death for both Indigenous children. and non-Indigenous infants. Other important risk factors Prevalence of key demographic and pregnancy-related for Indigenous and non-Indigenous infants were low so- risk factors, as determined from the univariate and cioeconomic status, multiple gestation pregnancies, birth multivariate analyses, were calculated for each maternal spacing < 18-months, having an older sibling in contact corrections history grouping and by Indigenous status. with the child protection system during an infant’sgesta- tion, and maternal substance use or poisoning-related ser- Results vice contacts during pregnancy. In total, 546 infants in the sample died aged 0 to Maternal hospitalisations for premature rupture of 12 months, between October 1985 and June 2014. There membranes, or external causes of injury were important was a 2.39-fold risk of infant mortality for Indigenous in- risk factors for infant mortality only for Indigenous in- fants compared to non-Indigenous infants (95% CI: fants. Male sex was also only a risk factor for infants in 1.95–2.93, p < 0.001). the Indigenous subgroup. Young maternal age and geo- graphical location for those living in outer regional areas, Infant mortality by maternal corrections history were only significant risk factors in the non-Indigenous For singleton births, Indigenous infants whose mothers sample. were imprisoned at any time had a significantly higher Factors not strongly associated with infant death included risk of death than infants of mothers with community parity, infection related hospitalisations, or mental health corrections orders alone, or with no corrections history related service contact (not related to substance use) in (Table 1). Indigenous infants whose mothers were pregnancy, and hospital admissions for placenta previa. Table 1 Infant mortality for singleton births, by maternal corrections history and Indigenous status Indigenous infants Non-Indigenous infants Survived Died Survived Died a a nn IMR RR (95% CI) p-value nn IMR RR (95% CI) p-value Maternal corrections history Prison (before pregnancy ) 2742 62 22.1 1.95 (1.45–2.62) <.001 634 6 9.4 2.25 (0.97–5.22) 0.059 Prison (during pregnancy) 753 25 32.1 2.84 (1.87–4.31) <.001 243 < 5 8.2 1.96 (0.48–8.00) 0.349 Prison (after birth only) 3481 84 23.6 2.08 (1.59–2.72) <.001 2491 36 14.2 3.42 (2.24–5.23) <.001 Community corrections 5588 94 16.5 1.46 (1.13–1.89) 0.004 628 8 12.6 3.02 (1.44–6.33) 0.003 No corrections* 12,400 142 11.3 12,186 51 4.2 Note Excludes 591 Indigenous and 427 non-Indigenous multiples births Infant mortality rate: deaths under 12 months, per 1000 live births Prison before, but not during, pregnancy Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 7 of 16 Table 2 Risk factors for infant mortality, by Indigenous status Indigenous infants (n = 25,962) Non-Indigenous infants (n = 16,712) Survived Died Survived Died nn % RR (95% CI) p-value nn % RR (95% CI) p-value Infants 25,532 430 1.7 16,596 116 0.7 Demographic factors Sex Male 13,052 240 1.8 1.20 (0.99–1.45) 0.058 8586 64 0.7 1.14 (0.79–1.65) 0.468 Female* 12,438 190 1.5 7991 52 0.6 Maternal Indigenous status Indigenous 24,321 418 1.7 1.72 (0.97–3.05) 0.062 Non-Indigenous* 1211 12 1.0 16,596 116 0.7 Socioeconomic status Very low (0–5%) 6793 128 1.8 2.16 (1.38–3.39) 0.001 996 15 1.5 3.50 (1.88–6.54) <.001 Low (6–25%) 11,095 195 1.7 2.02 (1.30–3.13) 0.002 4461 50 1.1 2.62 (1.65–4.15) <.001 Medium (26–50%) 5054 83 1.6 1.89 (1.18–3.01) 0.008 4517 23 0.5 1.20 (0.69–2.07) 0.523 High (51–100%)* 2546 22 0.9 6584 28 0.4 Geographical remoteness Major cities/Inner regional* 10,331 160 1.5 13,574 88 0.6 Outer regional 4185 73 1.7 1.12 (0.85–1.48) 0.404 1811 22 1.2 1.86 (1.17–2.97) 0.009 Remote 4776 75 1.5 1.01 (0.77–1.33) 0.922 834 5 0.6 0.93 (0.38–2.27) 0.865 Very remote 6195 120 1.9 1.25 (0.98–1.58) 0.066 326 < 5 0.3 0.47 (0.07–3.40) 0.458 Baseline Pregnancy Risk Factors Multiple gestation Yes 568 23 3.9 2.43 (1.61–3.66) <.001 414 13 3.0 4.81 (2.73–8.50) <.001 No* 24,964 407 1.6 16,182 103 0.6 Birth spacing < 18 months 3011 82 2.7 1.74 (1.37–2.21) <.001 1112 19 1.7 2.70 (1.66–4.40) <.001 Firstborn/18 months+* 22,521 348 1.5 15,484 97 0.6 Maternal age 12–19 years 7628 122 1.6 0.91 (0.72–1.17) 0.471 1740 21 1.2 2.36 (1.41–3.94) 0.001 20–24 years 9420 158 1.6 0.96 (0.76–1.20) 0.713 3916 47 1.2 2.35 (1.57–3.52) <.001 25–34 years* 7533 132 1.7 9054 46 0.5 35 + years 951 18 1.9 1.08 (0.66–1.76) 0.761 1886 < 5 0.1 0.21 (0.05–0.86) 0.030 Parity Nulliparous 9035 126 1.4 0.82 (0.66–1.03) 0.093 7954 53 0.7 0.99 (0.67–1.45) 0.939 Parity 1–2* 10,847 184 1.7 7539 51 0.7 Parity 3+ 5650 120 2.1 1.25 (0.99–1.57) 0.058 1103 12 1.1 1.60 (0.86–2.99) 0.140 Pregnancy complications Abruptio placentae and other disorders Yes 122 8 6.2 3.77 (1.91–7.42) <.001 90 < 5 4.3 6.31 (2.38–16.8) <.001 No* 25,410 422 1.6 16,506 112 0.7 Placenta previa Yes 87 < 5 3.3 2.02 (0.66–6.17) 0.217 109 0 0 No* 25,445 427 1.7 16,487 116 0.7 Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 8 of 16 Table 2 Risk factors for infant mortality, by Indigenous status (Continued) Indigenous infants (n = 25,962) Non-Indigenous infants (n = 16,712) Survived Died Survived Died nn % RR (95% CI) p-value nn % RR (95% CI) p-value Premature rupture of membranes Yes 1200 34 2.8 1.72 (1.22–2.43) 0.002 445 5 1.1 1.63 (0.67–3.97) 0.284 No* 24,332 396 1.6 16,151 111 0.7 Infection related hospitalisation in pregnancy Yes 2041 34 1.6 0.99 (0.70–1.40) 0.947 388 < 5 0.5 0.73 (0.18–2.96) 0.664 No* 23,491 396 1.7 16,208 114 0.7 Other maternal factors/Exposures during pregnancy Substance use related service contact Yes 397 15 3.6 2.24 (1.35–3.72) 0.002 293 6 2.0 2.99 (1.33–6.76) 0.008 No* 25,135 415 1.6 16,303 110 0.7 Hospitalisation for external causes of injury Yes 1479 38 2.5 1.56 (1.12–2.17) 0.008 278 < 5 1.1 1.55 (0.50–4.86) 0.450 No* 24,053 392 1.6 16,318 113 0.7 Mental health related service contact Yes 768 12 1.5 0.93 (0.52–1.64) 0.794 427 5 1.2 1.70 (0.70–4.14) 0.244 No* 24,764 418 1.7 16,169 111 0.7 Sibling in contact with child protection Yes 852 30 3.4 2.13 (1.48–3.07) <.001 237 6 2.5 3.70 (1.64–8.33) 0.002 Firstborn/No* 24,680 400 1.6 16,359 110 0.7 *Reference category 42 Indigenous and 19 non-Indigenous infants missing sex 46 Indigenous and 38 non-Indigenous infants missing socioeconomic status 47 Indigenous and 51 non-Indigenous infants missing remoteness Excludes placenta previa Infection related hospitalisation in pregnancy Substance use (including alcohol) or poisoning related service contact (hospital or mental health service) Hospitalisation for external causes of injury in pregnancy (excludes poisoning) Mental health service contact in pregnancy (excludes substance use) (hospital or mental health service) Older sibling(s) in contact with child protection system in infant’s pregnancy Multivariate regression of indigenous infant mortality The best model fit was achieved with the inclusion of all Multivariate regression was only performed for Indigenous variables (i.e., Model 1). Abruptio placentae and other pla- children because of the low number of deaths in the sample cental disorders contributed the highest risk of infant death, of non-Indigenous infants with incarcerated mothers. followed by maternal imprisonment during pregnancy, and Variables for maternal Indigenous status, geographical multiple gestation pregnancy. The effect of birth year remoteness, maternal age, parity, placenta previa, remained stable across Models 1–4 and approximated to a infection-related hospitalisations, external injury-related 4% reduction in risk of infant mortality each year. hospitalisations, and mental health related service con- tacts in pregnancy (other than for substance use) were Prevalence of key demographic and pregnancy risk excluded based on their lack of association with infant factors by maternal corrections history mortality in our data (Additional file 2). The remaining Table 4 shows the prevalence of pregnancy risk factors variables were included and are shown in Model 1 selected from the previous univariate and multivariate (Table 3). In Models 2–4, variables were removed by analyses, for Indigenous and non-Indigenous infants re- group in order of; other maternal factors and exposures ported by their mothers’ contact with corrective services. (Model 2), pregnancy complications (Model 3), baseline Infants of mothers with any corrections history had, in pregnancy risk factors (Model 4), leaving the combined most instances, higher prevalence of maternal service con- effects of maternal corrections history and demographic tact in pregnancy for substance use (including poisoning), factors on infant mortality in Model 4. external causes of injury, mental health related service Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 9 of 16 Table 3 Regression model of infant mortality, Indigenous children Model 1 Model 2 Model 3 Model 4 RR (95% CI) p-value RR (95% CI) p-value RR (95% CI) p-value RR (95% CI) p-value Maternal corrections history Prison (before pregnancy*) 1.83 (1.36–2.47) <.001 1.99 (1.49–2.67) <.001 2.03 (1.51–2.71) <.001 2.10 (1.57–2.81) <.001 Prison (during pregnancy) 2.55 (1.69–3.86) <.001 2.96 (1.98–4.43) <.001 3.01 (2.01–4.51) <.001 3.10 (2.07–4.64) <.001 Prison (after birth only) 1.55 (1.18–2.03) 0.001 1.64 (1.25–2.14) <.001 1.67 (1.28–2.18) <.001 1.73 (1.32–2.25) <.001 Community-only 1.38 (1.07–1.77) 0.012 1.41 (1.10–1.81) 0.007 1.42 (1.11–1.83) 0.006 1.45 (1.13–1.87) 0.003 No corrections history ref. Demographic factors Birth year 0.96 (0.95–0.97) <.001 0.96 (0.95–0.98) <.001 0.96 (0.95–0.98) <.001 0.97 (0.95–0.98) <.001 Sex Male 1.21 (1.01–1.47) 0.044 1.22 (1.01–1.47) 0.039 1.22 (1.01–1.47) 0.038 1.22 (1.01–1.48) 0.036 Female ref. Socioeconomic status Very low (0–5%) 2.11 (1.35–3.31) 0.001 2.08 (1.33–3.27) 0.001 2.1 (1.34–3.30) 0.001 2.11 (1.34–3.31) 0.001 Low (6–25%) 1.88 (1.21–2.91) 0.005 1.86 (1.20–2.88) 0.006 1.87 (1.20–2.89) 0.005 1.91 (1.23–2.96) 0.004 Medium (26–50%) 1.75 (1.10–2.80) 0.018 1.74 (1.09–2.78) 0.02 1.75 (1.10–2.79) 0.019 1.78 (1.11–2.83) 0.016 High (51–100%) ref. Baseline pregnancy risk factors Multiple gestation Yes 2.29 (1.51–3.46) <.001 2.37 (1.57–3.59) <.001 2.58 (1.71–3.88) <.001 No ref. Birth spacing < 18 months 1.52 (1.19–1.93) 0.001 1.57 (1.24–2.00) <.001 1.60 (1.26–2.03) <.001 Firstborn/18 months+ ref. Pregnancy complications Abruptio placentae and other disorders Yes 2.85 (1.46–5.59) 0.002 2.92 (1.49–5.72) 0.002 No ref. Premature rupture of membranes Yes 1.66 (1.18–2.35) 0.004 1.67 (1.18–2.37) 0.003 No ref. Other maternal factors/exposures in pregnancy Substance use related service contact Yes 1.71 (1.02–2.87) 0.042 No ref. External injury related hospitalisation Yes 1.20 (0.86–1.68) 0.283 No ref. Sibling with Child Protection contact Yes 1.57 (1.07–2.31) 0.022 Firstborn/No ref. Observations 25,875 25,875 25,875 265,875 AIC 0.1640 0.1642 0.1646 0.1656 *Prison before, but not during, pregnancy Excludes placenta previa Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 10 of 16 Table 4 Prevalence of pregnancy risk factors for infant mortality, by maternal corrections exposure and Indigenous status Indigenous infants Non-Indigenous infants Prison Prison Prison Community correct. No correct. Prison Prison Prison Community correct No correct (before preg.) (during preg.) (after birth) (before preg.) (during preg.) (after birth) n = 2857 n = 802 n = 3658 n = 5828 n = 12,817 n =667 n = 249 n =2588 n = 653 n = 12,555 Very low (0–5%) SES % 22.9 22.7 29.1 29.8 25.7 13.3 13.4 12.8 12.3 3.8 RR 0.89 0.88 1.13 1.16 ref. 3.48 3.49 3.35 3.20 ref. p-value 0.002 0.057 <.001 <.001 <.001 <.001 <.001 <.001 Multiple gestation % 1.9 3.0 2.5 2.5 2.1 4.0 1.6 2.4 2.6 2.5 RR 0.86 1.39 1.18 1.17 ref. 1.60 0.63 0.93 1.03 ref. p-value 0.328 0.113 0.152 0.126 0.017 0.362 0.602 0.911 Birth spacing < 18mths % 17.9 11.2 14.1 12.8 9.6 12.7 8.0 9.8 10.4 5.6 RR 1.87 1.17 1.47 1.34 ref. 2.27 1.43 1.75 1.86 ref. p-value <.001 0.127 <.001 <.001 <.001 0.099 <.001 <.001 Maternal age < 20 yrs. % 15.3 28.6 41.7 33.1 28.3 9.3 10 25.9 26 6.6 RR 0.54 1.01 1.47 1.17 ref. 1.40 1.51 3.91 3.92 ref. p-value <.001 0.895 <.001 <.001 0.007 0.032 <.001 <.001 Placental disorders % 0.5 0.7 0.6 0.7 0.4 0.1 0.8 1.1 1.1 0.4 RR 1.17 1.92 1.54 1.72 ref. 0.34 1.83 2.56 2.45 ref. p-value 0.620 0.130 0.090 0.011 0.288 0.398 <.001 0.025 PROM % 6.8 5.5 5.1 5.2 4.0 5.5 3.2 3.7 3.2 2.3 RR 1.71 1.39 1.28 1.32 ref. 2.41 1.4 1.59 1.40 ref. p-value <.001 0.032 0.003 <.001 <.001 0.344 <.001 0.133 Substance use % 3.2 7.4 2.2 1.6 0.7 12.3 10.0 4.7 3.2 0.4 RR 4.64 10.59 3.15 2.27 ref. 30.87 25.21 11.74 8.08 ref. p-value <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 External injury Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 11 of 16 Table 4 Prevalence of pregnancy risk factors for infant mortality, by maternal corrections exposure and Indigenous status (Continued) Indigenous infants Non-Indigenous infants Prison Prison Prison Community correct. No correct. Prison Prison Prison Community correct No correct (before preg.) (during preg.) (after birth) (before preg.) (during preg.) (after birth) n = 2857 n = 802 n = 3658 n = 5828 n = 12,817 n =667 n = 249 n =2588 n = 653 n = 12,555 % 10.9 14.5 9.4 6.5 2.9 6.9 6.4 3.4 2.1 0.9 RR 3.78 5.02 3.26 2.25 ref. 7.46 6.95 3.72 2.32 ref. p-value <.001 <.001 <.001 <.001 <.001 <.001 <.001 0.003 Mental health contact % 6.7 8.4 3.6 3.1 1.6 13.5 10.8 5.0 5.1 1.2 RR 4.20 5.22 2.27 1.96 ref. 11.15 8.96 4.15 4.17 ref. p-value <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 Sibling in contact CP % 11.6 14.3 5.2 2.6 0.7 12.9 11.2 4.4 1.5 0.05 RR 15.68 19.35 6.97 3.50 ref. 269.80 235.30 91.36 32.04 ref. p-value <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 46 Indigenous and 38 non-Indigenous infants missing socioeconomic status Excludes placenta previa Premature rupture of membranes Substance use (including alcohol) or poisoning related service contact (Mental Health or Hospital) Hospitalisation for external causes of injury in pregnancy (excludes poisoning) Mental health related service contact in pregnancy (excludes substance use) (Mental Health or Hospital) Older sibling(s) in contact with child protection system in infant’s pregnancy Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 12 of 16 contacts, and sibling contact with the child protection sys- between infant mortality and maternal imprisonment in tem compared to infants whose mothers had no record pregnancy for Indigenous infants remained in the full with corrective services. Prevalence of these service con- model after adjusting for other important risk factors. tacts in pregnancy was highest for infants whose mothers The only determinant to have a greater association with were imprisoned during pregnancy, and higher for infants Indigenous infant mortality was abruptio placentae and whose mothers were imprisoned before pregnancy com- other placental disorders, a serious pregnancy complica- pared to those imprisoned for the first time after birth or tion. Imprisonment during pregnancy was a stronger de- who had community corrections orders for these service terminant of infant mortality than all other pregnancy contacts. complications and baseline pregnancy risk factors in- The proportion of infants born in areas of very low so- cluding multiple gestation pregnancies. Indigenous in- cioeconomic status was high for all groups of Indigenous fants whose mothers were imprisoned during pregnancy children (> 20%) and did not differ by maternal correc- also experienced the highest prevalence of maternal con- tions history. There was a higher proportion of low so- tact with services during pregnancy for substance use, cioeconomic status for non-Indigenous children with mental illness, and external injury. These findings clearly any maternal corrections record (12–13%) compared to demonstrate the significant vulnerability of Indigenous in- those with no maternal corrections history (4%). fants whose mothers are imprisoned during pregnancy. There was a higher prevalence of a birth spacing of It was not possible to determine the relationship be- less than 18-months for infants with any maternal cor- tween maternal imprisonment before and during preg- rections record compared to no maternal corrections nancy and the risk of infant mortality for non-Indigenous history, except the difference was not significant for infants due to the relatively small numbers of infants in those infants whose mothers were imprisoned during the sample populations whose mothers were imprisoned pregnancy. The prevalence of infants born to mothers before or during pregnancy. However, non-Indigenous in- aged less than 20 years was highest among infants whose fants whose mothers who were imprisoned before or dur- mothers were first imprisoned after birth or had ing pregnancy had a significantly higher prevalence of community-corrections orders. several pregnancy risk factors including maternal service For Indigenous children, maternal hospitalisation for contact in pregnancy for substance use, external injury, or abruptio placentae and other placental disorders (exclud- mental health issues, and having siblings in contact with ing placenta previa) was not different between infants by the child protection system, compared to non-Indigenous maternal corrections history, whereas prevalence of ma- infants whose mothers were first imprisoned after birth or ternal hospitalisation for premature rupture of membranes had community-based corrections orders alone. was higher where there had been any record of maternal Infant mortality is a marker of adversity which is contact with the corrections system. strongly linked to social and economic disadvantage [10]. For the non-Indigenous children, hospitalisation of the It is well-established that Indigenous children experience mother for abruptio placentae and other placental disor- higher rates of socioeconomic disadvantage and infant ders (excluding placenta previa) was higher where in- mortality than non-Indigenous children [10, 29]. Over a fants’ mothers were first imprisoned after birth or had quarter of our Indigenous subgroup, compared to only 6% community-correctional orders, and prevalence of hos- of the non-Indigenous sample, was born in the lowest 5% pitalisation for premature rupture of membranes was of areas by socioeconomic status. Socioeconomic disad- higher where infants’ mothers were imprisoned before vantage provides a broad measure of social determinants or after pregnancy. such as parental education, employment, disability, and overcrowding as well as a greater prevalence of health Discussion conditions and risk behaviours such as alcohol and sub- This is the first study to provide a comprehensive investi- stance use, domestic violence, and mental illness, which gation of infant mortality outcomes for children of women are related to infant mortality risk. For Indigenous peo- prisoners. Children of mothers with a history of contact ples, socioeconomic disadvantage is also associated with with corrective services, including community-based cor- experiences of racism and discrimination in service access rections orders and imprisonment before or after preg- and broader society, and in increased contact with health nancy, had increased rates of infant mortality. and criminal justice systems [30]. Within the Indigenous sample, rates of infant mortal- It is recognised that imprisonment likely acts as ity were highest for infants whose mothers were impri- both a proxy for socioeconomic disadvantage and for soned during pregnancy, when compared to similarly risk behaviours which are associated with imprison- disadvantaged mothers who were imprisoned at times ment, including increased substance use, injury and other than pregnancy or who had community-based cor- mental illness, as evidenced in our study by increased rectional orders. The strength of the relationship prevalence of service contacts related to these risks in Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 13 of 16 pregnancy for pregnant prisoners. Similarly, having a represents a positive pregnancy outcome [31, 32]. For sibling in contact with the child protection system example, Walker and colleagues [4] reported the average during pregnancy can be considered as a proxy for length of stay for sentenced women prisoners in New maternal vulnerability and socioeconomic disadvan- South Wales, Australia, is 196 days compared to 547 days tage. Whether maternal imprisonment during preg- in US prisons [4]. In our study of Western Australian nancy has an impact on birth outcomes, including children exposed to maternal incarceration before their infant mortality, over-and-above the effects of second birthday, almost half of all prison stays for either pre-existing disadvantage, is a key issue within the sentenced prisoners or unsentenced remandees were < international literature [3]. 2 weeks [33]. In this context the presumption that im- It has been proposed that there may be a possible pro- prisonment has a ‘dose-response’ effect on perinatal out- tective effect of imprisonment in pregnancy for birth comes [32], and that short-term imprisonment does not outcomes [3]. The proposed protective effect of impris- carry excess risks for the mother or her infant warrants onment in pregnancy on birth outcomes is thought to further research. be contributed to by a reduction of exposure to risk fac- Within Western Australia, all pregnant prisoners are tors such as domestic violence and substance use while provided with health care “commensurate with commu- in custody, and in improved nutrition and access to nity standards” [34]. However, health service provision antenatal care. Within our study, infants whose mothers varies between prisons [35–40], and regional prisons were imprisoned before pregnancy had a higher preva- face additional challenges such as in the transportation lence of risk factors during pregnancy related to mater- of prisoners to community health centres [41]. There is nal service contact for substance use, injury, and mental limited evidence available on the provision and impact illness, when compared to infants whose mothers were of antenatal care provided to pregnant prisoners in first imprisoned after birth. This finding suggests that Australia. In New South Wales, Walker and colleagues risk behaviours of this kind do occur concurrently with [4] found that women imprisoned during pregnancy imprisonment. However, the strength of our findings were more likely to initiate antenatal care after 20 weeks with respect to maternal imprisonment during preg- gestation than women with no record of imprisonment. nancy for Indigenous infants suggest there may be an However as many of these women were not imprisoned additional impact of imprisonment during pregnancy on for the duration of pregnancy, as was the case for infant mortality risk, at least in certain circumstances. women in the present study, it is possible antenatal care Similar to our findings, the general protective effect of was first initiated during imprisonment. imprisonment during pregnancy on birth outcomes that Imprisonment during pregnancy may have negative has been reported in the broader international literature impacts on the mental health and wellbeing of women, was not replicated in the only prior Australian study of particularly for Indigenous mothers, as a consequence of the effects of maternal imprisonment on pregnancy out- being separated from family and country during their comes [4]. While just over one-quarter of pregnant pris- pregnancy [42]. There is evidence that maternal stress in oners were Indigenous, Walker and colleagues [4] did pregnancy can impact on birth outcomes [43, 44]. Re- not investigate the outcomes for the Indigenous and search on offender health has demonstrated that the non-Indigenous populations separately. Our study has year following release from prison, particularly within clearly demonstrated that infant mortality is higher for the first month of release, is also a key risk period for an Western Australian Indigenous infants whose mothers offender’s own hospitalisation and mortality [45–48]. were imprisoned during pregnancy. Taken together, our Whether the release period also leads to increased risk study and that of Walker and colleagues [4] suggest that for offenders’ children, including the unborn children of there are different outcomes for the infants of Australian pregnant prisoners, has not been investigated. The re- women imprisoned during pregnancy than those reported sults from the present study highlight that this as an im- for other jurisdictions [3]. It is not yet clear, however, portant area of future research. whether this difference is restricted to the Indigenous Further research is needed to understand whether population as our findings were inconclusive, due to small there are particular characteristics of maternal im- numbers, with respect to non-Indigenous infants. prisonment during pregnancy that are associated The difference in outcomes for Australian pregnant with infant mortality, and whether the effect is prisoners compared to those within other criminal just- restricted to Indigenous populations. The study find- ice systems reported in the international literature [3], ings highlight the importance of separate consider- may relate to the longer periods of imprisonment experi- ation of Indigenous populations when investigating enced by women in other jurisdictions compared to outcomes for children of prisoners, as combining Australia, as an increased length of imprisonment in populations may mask important differences in pregnancy is associated with higher birth weight which outcomes. Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 14 of 16 Limitations mortality. This study provides the first detailed analysis of These results need to be considered within the context infant mortality outcomes for children whose mothers of the study’s limitations, for example, the power to de- were imprisoned in pregnancy. The study demonstrates tect relationships between infant mortality and maternal that there are higher rates of infant mortality for Indigen- imprisonment before or during pregnancy in the ous, compared to non-Indigenous, children of prisoners non-Indigenous sample was limited by the small num- and that within the Indigenous sample any maternal con- bers of non-Indigenous infant deaths reported within tact with the corrections system is associated with an in- those groups. crease in infant mortality. As gestational age was not available, pregnancy was taken Maternal imprisonment in pregnancy is an important de- to begin nine-months prior to month of birth for all infants. terminant of infant mortality for Indigenous children. Fur- Accordingly, some records of exposure to maternal impris- ther research is needed to determine what factors onment, and service contacts related to substance use, contribute to this increased risk of infant mortality, and mental health or child protection, that were attributed to whether particular groups of prisoners are more affected. having occurred during pregnancy may have occurred prior Due to the relatively low incidence of infant deaths within to pregnancy if gestation was shorter than nine-months. the non-Indigenous sample, it was not possible to deter- There is evidence, however, that adverse events within the mine the impact of maternal imprisonment on infant mor- preconception period (6–0 months before pregnancy) can tality in this sub-population. It was the case, however, that increase infant mortality risk [49, 50]. Therefore any re- non-Indigenous infants whose mothers were imprisoned cords of exposure misclassified as having occurred during before or during pregnancy experienced higher rates of pregnancy, may still have been expected to have an impact pregnancy risk factors, than infants whose mothers were on infant mortality risk. first imprisoned after birth or had community-based cor- Additionally, without gestational age we have not been rectional orders. This highlights the vulnerability of able to measure preterm birth (delivery before 37 weeks non-Indigenous and Indigenous pregnant prisoners, and of gestation), which is associated with infant mortality the importance of providing support services to address [51]. However, preterm birth shares many of the same pregnancy risk factors for women in contact with the cor- risk factors for infant mortality identified within the rections system. study including low socioeconomic status, maternal age, maternal stress, infections, and multiple gestation preg- Additional files nancies [51]. The study has provided the first evidence of an association between maternal incarceration in preg- Additional file 1: ICD codes. List of ICD codes used to define the study variables. (PDF 393 kb) nancy and infant mortality, further research is needed in- Additional file 2: Regression models of infant mortality by main group vestigate the impact of factors not able to be measured in for Indigenous children. Results of the multivariate regression conducted the current study, notably preterm birth, antenatal care for each main grouping of variables to eliminate non-significant variables and caesarean section rates. from including the final multivariate analysis. (PDF 330 kb) Administrative data alone cannot fully capture occur- rences of heavy drinking or substance use, mental ill- Abbreviations ness, or injuries resulting from domestic violence, within AIC: Akaike Information Criterion; ARIA: Accessibility/Remoteness Index of Australia; CPFS: Department of Communities: Child Protection and Family pregnancy [52]. In addition, the study only obtained the Support; HMDC: Hospital Morbidity Data System Collection; ICD: International primary diagnosis code (not co-diagnoses) for hospital Classification of Diseases; MHIS: Mental Health Information System; and mental health service records. Consequently, the as- SEIFA: Socio-economic Indexes for Areas; WADLS: Western Australian Data Linkage System sociations observed in our study with drinking or other substance use in pregnancy, mental illness or injuries are Acknowledgements likely under ascertained. We thank the Developmental Pathways in Western Australian Children Hospital data were not obtained for second-generation Project, the Western Australian Data Linkage Branch, the Western Australian children. While explanatory variables based on hospital Aboriginal Health Research Ethics Committee, the Western Australian Aboriginal and Torres Strait Islander community, the Western Australian data were primarily taken from maternal hospital re- Department of Justice, the Department of Communities: Child Protection cords (available for all mothers), there may have been and Family Support, the Western Australian Department of Health and the some missed cases of chromosomal abnormalities, and custodians of the Western Australian Birth and Death Registrations, and Midwives Notification System for the linkages, provision of data and ongoing substance use related service contacts in pregnancy, for ethical review of project outputs. the second-generation children. Material presented cannot be considered as either endorsed by the Western Australian Department of Justice or an expression of the policies or view of the Department, and any errors or omission or commission are the Conclusions responsibility of the Researchers. To date there have been few studies which have focussed Responsibility for interpretation of the data supplied through the on the impact of maternal imprisonment on infant Government of Western Australia, the Developmental Pathways in Western Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 15 of 16 Australian Children Project, and by any other data custodian, is the 4. Walker JR, Hilder L, Levy MH, Sullivan EA. Pregnancy, prison and perinatal Researchers’ alone. outcomes in new South Wales, Australia: a retrospective cohort study using linked health data. BMC Pregnancy Childb. 2014;14:214. 5. Howard DL, Strobino D, Sherman SG, Crum RM. Timing of incarceration Funding during pregnancy and birth outcomes: exploring racial differences. Matern CD was funded by an Australian Government Research Training Program Child Healt J. 2008;13:457–66. (RTP) Scholarship. 6. Australian Bureau of Statistics. Prisoners in Australia, 2016. Table 20. Canberra: Commonwealth of Australia; 2016. http://www.abs.gov.au/ Availability of data and materials AUSSTATS/abs@.nsf/DetailsPage/4517.02016?OpenDocument. Accessed 20 This study used linked routinely-collected administrative data from the West- Dec 2017 ern Australian Departments of Health, Justice, and Communities: Child Pro- 7. Australian Bureau of Statistics. Estimated resident Aboriginal and Torres tection and Family Support. While these data are accessible to approved Strait Islander and Non-Indigenous population, States and Territories - 30 researchers through the standard ethical and data application processes, they June 2011. Tables 1 and 3. Canberra: Commonwealth of Australia; 2011. are not publicly available and are maintained by the relevant government http://www.abs.gov.au/AUSSTATS/abs@.nsf/DetailsPage/3238.0.55. data custodians. As a result, while future researchers are able to apply dir- 001June%202011?OpenDocument. Accessed 20 Dec 2017 ectly for these data we are restricted under Statutory requirement from re- 8. Council of Australian Governments. Closing the Gap on Indigenous leasing any of our project data to a third party. This includes making the Disadvantage: The Challenge for Australia, February 2009. Canberra: study data available in a publicly available repository. Government of Australia; 2009. https://www.dss.gov.au/sites/default/files/ documents/05_2012/closing_the_gap.pdf. Accessed 20 Dec 2017 Authors’ contributions 9. Council of Australian Governments. Closing the Gap: Prime Minister’s Report CD had primary responsibility of the study design, conduct of the data 2016. Canberra: Government of Australia; 2016. https://www.pmc.gov.au/ analysis and interpretation, and drafting the initial and final versions of the sites/default/files/publications/closing_the_gap_report_2016.pdf. Accessed manuscript. LS, GM and DP contributed to the study design, interpretation of 20 Dec 2017 findings and drafts of the manuscript. GM contributed to the data analysis. 10. Freemantle CJ, Read AW, de Klerk AH, McAullay D, Anderson AP, Stanley FJ. All authors read and approved the final manuscript. Patterns, trends, and increasing disparities in mortality for aboriginal and non-aboriginal infants born in Western Australia, 1980-2001: population Ethics approval and consent to participate database study. Lancet. 2006;367:1758–66. Ethical approval for the project was received from the Department of Health 11. Hayes A, Weston R, Qu L, gray M. Families then and now: 1980-2010. Human Research Ethics Committee, the Western Australian Aboriginal Health Melbourne: Australian institute of family Studies; 2010. Ethics Committee, and the University of South Australia Human Research 12. Mosely HW, Chen LC. An analytical framework for the study of child survival Ethics Committee. in developing countries. Popul Dev Rev. 1984;10:25–45. The study uses data that is routinely collected by the Department of Justice 13. Barker DJP, Harding JE, Owens JA, Robinson JS. Fetal nutrition and and other Western Australian Government departments and services and cardiovascular disease in adult life. Lancet. 1993;341(8850):938–41. linked for research purposes by the Western Australian Data Linkage Branch. 14. Barker DJP. The origins of the developmental origins theory. J Intern Med. In cases such as this the National Health and Medical Research Council 2007;261(5):412–7. outlines in the National Statement on Conduct in Human Research the 15. Eriksson JG. Developmental origins of health and disease – from a small requirements for waiver of consent. Under the National Statement, a Human body size at birth to epigenetics. Ann Med. 2016;48(6):456–67. Research Ethics Council can grant a waiver of consent if it is satisfied certain 16. Halfon N, Larson K, Lu M, Tullis E, Russ S. Lifecourse health development: conditions have been met. The criteria for allowing a waiver listed under the past, present and future. Matern Child Healt J. 2014;18:344–65. National Statement apply to the project. Specifically as the project carries no 17. Heindel JJ, Vandenberg LN. Developmental origins of health and disease: a more than a low risk; it is not feasible to obtain consent due to the age and paradigm for understanding disease etiology and prevention. Curr Opin quantity of the data required; and measures are being taken to ensure the Pediatr. 2016;27(2):248–53. continued protection of the confidentiality of the data. 18. Kozuki N, Lee ACC, Silveira MF, Sania A, Vogel J, Adair L, Barros F, Caulfield LE, Christian P, Fawzi W, Humphrey J, Huybregts L, et al. The associations of parity Competing interests and maternal age with small-for-gestational-age, preterm, and neonatal and The authors declare that they have no competing interests. infant mortality: a meta-analysis. BMC Public Health. 2013;13(Suppl 3):S2. 19. Salihu HM, August EM, de la Cruz C, Mogos MF, Weldeselasse H, Alio AP. Infant mortality and the risk of small for gestational age in the subsequent pregnancy: a Publisher’sNote retrospective cohort study. Matern Child Healt J. 2013;17(6):1044–51. Springer Nature remains neutral with regard to jurisdictional claims in 20. Holman CD, Bass AJ, Rouse IL, Hobbs MS. Population based linkage of published maps and institutional affiliations. health records in Western Australia: development of a health services research linked database. Aust N Z J Public Health. 1999;23(5):453–9. Author details 21. National Coding Centre. The Australian version of the international th Health Economics and Social Policy Group, Centre for Population Health classification of diseases, 9 revision, clinical modification (ICD-9-CM). Research, Sansom Institute, School of Health Sciences, University of South Sydney: National Coding Centre; 1996. Australia, GPO Box 2471, Adelaide, SA 5001, Australia. Centre for Health 22. National Centre for Classification in Health. The international statistical th Services Research, School of Population and Global Health, University of classification of diseases and related health problems, 10 revision, nd Western Australia, Crawley, WA, Australia. Australian modification, ICD-10-AM Australian Coding Standards. 2 edition. Sydney: Faculty of Health Sciences, University of Sydney; 2000. Received: 9 January 2018 Accepted: 22 May 2018 23. Organisation for Economic Co-operation and Development. Glossary of Statistical Terms: Infant mortality rate. Organisation for economic co- operation and development. 2006. https://stats.oecd.org/glossary/detail. References asp?ID=1347. Accessed 20 Dec 2017. 1. Knight M, Plugge E. Risk factors for adverse perinatal outcomes in 24. Australian Bureau of Statistics. 2033.0.55.001 – Census of Population and imprisoned pregnant women: a systematic review. BMC Public Health. 2005; Housing: Socio-Economic Indexes for Areas (SEIFA), Australia – Data Only 5(111) https://doi.org/10.1186/1471-2458-5-111. 2006, Data Cube Only 2001 and Data Only 1996. Commonwealth of 2. Shaw J, Downe S, Kingdon C. Systematic mixed-methods review of Australia, Canberra. 2006. http://www.abs.gov.au/AUSSTATS/abs@.nsf/second interventions, outcomes and experiences for imprisoned pregnant women. +level+view?ReadForm&prodno=2033.0.55.001&viewtitle= J Adv Nurs. 2015;71(7):1451–63. Census%20of%20Population%20and%20Housing:%20%20Socio- 3. Knight M, Plugge E. The outcomes of pregnancy among imprisoned Economic%20Indexes%20for%20Areas%20(SEIFA),%20Australia%20- women: a systematic review. BJOG-Int J Obstet Gyn. 2005;112:1467–74. %20Data%20Cube%20only~2001~Previous~05/10/2006&&tabname= Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 16 of 16 Past%20Future%20Issues&prodno=2033.0.55.001&issue=2001&num=&view= J Prisoner Health. 2005;11(2):108–18. https://doi.org/10.1108/IJPH-10- &. Accessed 20 Dec 2017. 2013-0046. 25. Dowell CM, Preen DB, Segal L. Quantifying maternal incarceration: a whole- 48. Kinner SA, Preen DB, Kariminia A, Butler T, Andrews JY, Stoové M, Law population linked data study of western Australian children born 1985-2011. M. Counting the cost: estimating the number of deaths among recently Aust N Z J Public Health. 2017;41(2):151–7. released prisoners in Australia. Med J Australia. 2011;195(2):64–8. 49. Class QA, Khashan AS, Lichtenstein P, Långström N, D’Onofrio BM. Maternal 26. Australian Bureau of Statistics. Remoteness Structure. 2014. http://www. stress and infant mortality: the importance of the preconception period. abs.gov.au/websitedbs/d3310114.nsf/home/remoteness+structure. Psychol Sci. 2013;24(7):1309–16. Accessed 20 Dec 2017. 50. Class QA, Mortensen PB, Henriksen TB, Dalman C, D’Onofrio BM, 27. Christensen D, Davis G, Draper G, Mitrou F, McKeown S, Lawrence D, et al. Khashan AS. Preconception maternal bereavement and infant and Evidence for the use of an algorithm in resolving inconsistent and missing childhood mortality: a Danish population-based study. Psychosom Med. indigenous status in administrative data collections. Aust J Soc Issues. 2014; 2015;77(8):863–9. 49(4):423–43. 51. Frey HA, Klebanoff MA. The epidemiology, etiology, and costs of preterm 28. Cohen J. Statistical power analysis for the behavioral sciences, second birth. Seminars in Fetal & Neonatal Medicine. 2016;21(2):68–73. edition: Lawrence Erlbaum Associates, USA. ISBN 0-8058-0283-5; 1988. 52. O’Leary CM, Halliday J, Bartu A, D’Antoine H, Bower C. Alcohol-use disorders 29. Shah PS, Zao J, Al-Wassia H, Shah V. Pregnancy and neonatal outcomes of during and within one year of pregnancy: a population-based cohort study aboriginal women: a systematic review and meta-analysis. Women Health 1985-2006. BJOG-Int J Obstet Gyn. 2013;120(6):744–53. Iss. 2011;21(1):28–39. 30. Prandl KJ, Rooney R, Bishop BJ. Mental health of Australian aboriginal women during pregnancy: identifying the gaps. Arch Womens Ment Health. 2012;15:149–54. 31. Martin SL, Rieger RH, Kupper LL, Meyer RE, Qaqish BF. The effect of incarceration during pregnancy on birth outcomes. Public Health Rep. 1997; 112(4):340–6. 32. Howard DL, Strobino D, Sherman SG, Crum RM. Maternal incarceration during pregnancy and infant birthweight. Matern Child Healt J. 2011;15: 478–86. 33. Dowell CM, Mejia GC, Preen DB, Segal L. Maternal incarceration, child protection, and infant mortality: A descriptive study of infant children of women prisoners in Western Australia. Manuscript submitted for publication. 34. Department of Corrective Services. Policy Directive 74 – Management of Women in Custody. Perth: Government of Western Australia; 2011. 35. Office of the Inspector of Custodial Services. Report of an Announced Inspection of Boronia Pre-release Centre for Women; Report 79. Perth: Government of Western Australia; 2012. 36. Office of the Inspector of Custodial Services. Report of an Announced Inspection of Broome Regional Prison; Report 77. Perth: Government of Western Australia; 2012. 37. Office of the Inspector of Custodial Services. Report of an Announced Inspection of Bandyup Women’s Prison; Report 73. Perth: Government of Western Australia; 2011. 38. Office of the Inspector of Custodial Services. Report of an Announced Inspection of Eastern Goldfields Regional Prison; Report 72. Perth: Government of Western Australia; 2011. 39. Office of the Inspector of Custodial Services. Report of an Announced Inspection of Roebourne Regional Prison; Report 70. Perth: Government of Western Australia; 2011. 40. Office of the Inspector of Custodial Services. Report of an Announced Inspection of Greenough Regional Prison; Report 66. Perth: Government of Western Australia; 2010. 41. Geraghty S. Reaching out: caring for women prisoners in Western Australia. The Practising Midwife. 2015;18(1):26–8. 42. Mukherjee S, Pierre-Victor D, Bahelah R, Madhivanan P. Mental health issues among pregnant women in correctional facilities: a systematic review. Women Health. 2014;54(8):816–42. 43. Bussières E, Tarabulsy G, Pearson J, Tessier R, Forest J, Giguèrec Y. Maternal prenatal stress and infant birth weight and gestational age: a meta-analysis of prospective studies. Dev Rev. 2015;36:179–99. 44. Beijers R, Buitelaar JK, de Weerth C. Mechanisms underlying the effects of prenatal psychosocial stress on child outcomes: beyond the HPA axis. Eur Child Adolesc Psychiatry. 2014;23:943–56. 45. Alan J, Burmas M, Preen D, Pfaff J. Inpatient hospital use in the first year after release from prison: a western Australian population-based record linkage study. Aust N Z J Public Healt. 2011;35(3):264–9. 46. Kariminia A, Butler T, Jones J, Lew M. Increased mortality among indigenous persons during and after release from prison in new South Wales. Aust N Z J Public Healt. 2012;36(3):274–80. 47. Jama-Alol KA, Malacova E, Ferrante A, Alan J, Stewart L, Preen D. Influence of offence type and prior imprisonment on risk of death following release from prison: a whole-population linked data study. Int http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png BMC Pregnancy and Childbirth Springer Journals

Determinants of infant mortality for children of women prisoners: a longitudinal linked data study

Free
16 pages

Loading next page...
 
/lp/springer_journal/determinants-of-infant-mortality-for-children-of-women-prisoners-a-qF9x1hRjjj
Publisher
Springer Journals
Copyright
Copyright © 2018 by The Author(s).
Subject
Medicine & Public Health; Reproductive Medicine; Maternal and Child Health; Gynecology
eISSN
1471-2393
D.O.I.
10.1186/s12884-018-1840-z
Publisher site
See Article on Publisher Site

Abstract

Background: There is limited information on the determinants of infant mortality outcomes for the children of women prisoners. This study aimed to explore determinants of infant mortality for Indigenous and non-Indigenous children, with a specific focus on maternal imprisonment during pregnancy as a risk factor. Methods: Using linked administrative data we obtained a longitudinal sample of 42,674 infants born in Western Australia between October 1985 and June 2013. Data were analysed by maternal contact with corrective services, including; (i) imprisonment during pregnancy, (ii) imprisonment before (but not during) pregnancy, (iii) imprisonment after birth, (iv) community-based correctional orders (but no imprisonment), and (v) no corrections record. Infant mortality rates were calculated. Univariate and multivariate log-binomial regression was undertaken to identify key demographic and pregnancy-related risk factors for infant mortality. Risk factor prevalence was calculated for infants by maternal corrections history. Results: 430 Indigenous and 116 non-Indigenous infants died aged 0–12 months. For singletons, infant mortality rates were highest in Indigenous infants with mothers imprisoned during pregnancy (32.1 per 1000) and non-Indigenous infants whose mothers were first imprisoned after birth (14.2 per 1000). For all Indigenous children, the strongest determinants of infant mortality were: abruptio placentae and other placental disorders (RR = 2.85; 95%CI 1.46–5.59; p =0. 002), maternal imprisonment during pregnancy (RR = 2.55; 95%CI 1.69–3.86; p < 0.001), and multiple gestation (RR = 2.29; 95% CI1.51–3.46; p < 0.001). Indigenous and non-Indigenous infants with mothers imprisoned at any time, and particularly before or during pregnancy, experienced higher prevalence of key pregnancy risk factors. Conclusions: This is the first comprehensive study of the determinants of infant mortality for children of women prisoners. Infants with any maternal corrections history, including community-based orders or imprisonment outside of pregnancy, had increased infant mortality. Indigenous infants whose mothers were imprisoned during pregnancy were at particular risk. There was a low incidence of infant death in the non-Indigenous sample which limited the investigation of the impact of the specific aspects of maternal corrections history on infant mortality. Non-Indigenous Infants whose mothers were imprisoned before or during pregnancy experienced higher prevalence of pregnancy risk factors than infants of mothers first imprisoned after birth. Keywords: Infant mortality, Women prisoners, Linked data, Australia * Correspondence: caitlin.dowell@mymail.unisa.edu.au Health Economics and Social Policy Group, Centre for Population Health Research, Sansom Institute, School of Health Sciences, University of South Australia, GPO Box 2471, Adelaide, SA 5001, Australia Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 2 of 16 Background Infant Death Syndrome [10], as well as increased paren- Women prisoners constitute a highly vulnerable popula- tal employment and maternal education [11]. Although tion which is exposed to multiple and complex risk fac- the overall rates of infant mortality declined between tors, including domestic violence, substance abuse, 1980 and 2001, the disparity between Indigenous and poverty, discrimination and mental illness, placing them non-Indigenous populations increased from a Relative and their children at risk of poor pregnancy and health Risk of 3.0 (95%CI 2.5–3.6) in 1980–84 to 4.4 (95%CI outcomes [1]. However, few reports, exist which have in- 3.5–5.5) in 1998–2001 [10]. For Indigenous infants, vestigated pregnancy outcomes for children of women postneonatal mortality was higher than neonatal mortal- prisoners [2, 3]. Of those studies that have investigated ity, a pattern that indicates the impact of socioeconomic pregnancy outcomes for infants, the main outcomes in- disadvantage and marginalisation on infant mortality vestigated include preterm delivery and low birth weight outcomes [10]. [3]. The few studies that report on infant death have The primary objective of this study was to explore deter- been limited by small numbers of events, in part due to minants of infant mortality for Indigenous and the size of the cohort sampled [3, 4]. non-Indigenous children, with a focus on whether maternal While the international literature on the pregnancy out- imprisonment during pregnancy is a risk factor for infant comes of prisoners is equivocal, a review of studies from mortality. The specific aims of the study were to determine: across the United States (US), United Kingdom and Eur- infant mortality rates for Indigenous and non-Indigenous ope found women imprisoned in pregnancy generally had infants of mothers with different corrections histories; the poorer maternal and infant outcomes than community key demographic and pregnancy-related risk factors which controls, and better outcomes than disadvantaged com- may contribute to infant mortality in Indigenous and munity controls [3]. It has been concluded that this may non-Indigenous populations; the importance of maternal indicate that imprisonment in pregnancy may be benefi- corrections history as a determinant of infant mortality cial for some pregnancy outcomes [3]. The only Australian after accounting for significant demographic and study on pregnancy outcomes of women prisoners found, pregnancy-related risk factors; and the prevalence of key however, that women imprisoned during pregnancy did risk factors for infant mortality between infants with differ- not have better perinatal outcomes than women impri- ent maternal corrections histories. soned at times other than pregnancy [4]. Thus the unique context of the justice systems and prisoner populations in Methods specific jurisdictions limit the transferability of findings Study design across jurisdictions in the absence of an understanding of We have used data from a large data linkage project the determinants of pregnancy outcomes for women to explore infant mortality within Indigenous and prisoners. non-Indigenous children of mothers who have been Studies in the US have found racial differences in the exposed to the corrections system at different times pregnancy outcomes of women prisoners [5]. In Australia, in relation to their pregnancy. We compared infant research on the pregnancy outcomes of Indigenous mortality outcomes for children whose mothers had; women prisoners is lacking despite their overrepresenta- (i) any period of imprisonment during pregnancy, (ii) tion in the prison population. In Western Australia, for imprisonment before (but not during) pregnancy, (iii) example, Indigenous peoples represent 4% of the general their first period of imprisonment after birth, (iv) population but 46% of the female prison population [6, 7]. community-based correctional orders (but no impris- This reflects the high levels of social and economic disad- onment), or (v) no corrections record at any time vantage and discrimination experienced by Indigenous over the study period. peoples in Australia [8]. Similarly, Indigenous peoples in Australia experience poorer pregnancy outcomes com- pared with non-Indigenous mothers, and while infant Conceptual framework mortality rates have been improving for both Indigenous Mosley and Chen’s[12] analytical framework for the study and non-Indigenous populations across time, the racial of child survival was adapted for the present study. The disparity remains [9, 10]. basis of this framework is that broader determinants ne- Across 1980 to 2001, infant mortality rates in Western cessarily act through biological pathways, or mechanisms, Australia declined for both Indigenous (25.0 in 1980–84 which impact before, during and after pregnancy on the to 16.1 in 1998–2001) and non-Indigenous infants (8.4 healthy development of the fetus and infant, and ultim- in 1980–84 to 3.7 in 1998–2001) [10]. Some important ately on infant mortality. There is evidence that adverse changes that occurred across the past 30-years include events experienced before and during pregnancy can im- improved transport for rural and remote pregnant pact on fetal development and result in increased risk of women, immunisation of infants, prevention of Sudden poor infant and childhood health outcomes [13–17]. Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 3 of 16 Our adaptation of this framework first groups together cause of death, diagnosis, or reason for health service key demographic factors, including birth year, sex, Indigen- contact, respectively. For HMDC records, we obtained ous status, socioeconomic status, and geographical remote- one code for the principal diagnosis of the episode of ness. The second grouping includes baseline pregnancy risk care, and up to four codes for external causes of factors, such as multiple gestation, birth spacing < episodes of care. Over the study period, the ICD 9th Re- 18 months, maternal age, and parity, which are largely vision with Clinical Modification (ICD-9-CM) [21] re- unmodifiable from the commencement of pregnancy [18]. lated to services contacts before July 1999, and service The third grouping includes key pregnancy complications contacts from that date used ICD 10th Revision with that might be a precursor of infant mortality, such as nutri- Australian Modification (ICD-10-AM) [22]. tional deficiencies, placental disorders, prematurity, and in- fection [19]. The last group includes other maternal factors Study population and exposures which are known risk factors for infant mor- The study population was drawn from a retrospective lon- tality or may indicate maternal vulnerability or household gitudinal cohort study of all liveborn children born in dysfunction, such as substance use or mental health related Western Australia from 1985 to 2011 whose biological service contacts, external causes of injury, and having other mother was imprisoned at least once within 18-years after children in contact with the child protection system. their birth. The cohort study population included a com- parison group of children whose mother had no record of Data sources imprisonment from their date of birth to their 18th birth- Data were obtained through the Western Australian day, which was identified through the same data sources Data Linkage System (WADLS). The WADLS uses as the cohort and matched 3:1 to cohort children on Indi- highly-accurate computerised, probabilistic matching genous status, age and sex. Data on second-generation with clerical review to create linkages within and be- children, born between 1998 and 2014 to the female tween administrative data collections across a range of members of the cohort and comparison group, were also Western Australian government agencies [20]. The obtained. Western Australian Data Linkage Branch conducted the Stillbirths (second-generation only) and infants with linkage and provided de-identified data extracts. Records chromosomal abnormalities, identified through HMDC were extracted from the Midwives Notifications System, and death records, were excluded (Fig. 1). Erroneous re- Birth Registrations, Death Registrations, Department of cords with multiple mothers or missing key information Justice, Hospital Morbidity Data System Collection such as birthdate were removed. The final study popula- (HMDC), Mental Health Information System (MHIS), tion was restricted to children born from October 1985 to and Department of Communities: Child Protection and June 2013 (inclusive) to ensure pregnancy exposure and Family Support (CPFS) data collections. These are all death data was available for all infants. statutory State-wide data collections with complete In total, there were 42,674 infants in the final study coverage. population, 37,469 from the first-generation (original co- The Birth Registration and Midwives Notifications Sys- hort and comparison group) and 5205 from the tem data provided social and demographic characteristics second-generation (children of the original cohort and of mothers and children at time of birth. Mortality data comparison group). Data from the birth and death regis- include all deaths registered in Western Australia. The trations, midwives notifications records, and CPFS data Department of Justice data collection includes all custodial were available for first- and second-generation children. records for offenders held in Western Australian prisons Only first-generation children had HMDC record data. and records for offenders on community-based correc- Mothers had corrections, HMDC, and MHIS data tional orders. Data excluded unsentenced individuals available. detained in police stations and courts, immigration deten- tion centres, and mental health facilities. The HMDC in- cludes all inpatient records for Western Australian public Definition of maternal corrections history and private hospitals and day surgeries. The MHIS in- The study population was categorised into: a) infants cludes presentations to all inpatient and public commu- whose mothers had a record of imprisonment at any nity mental health services. The CPFS data include all time (n = 7317 Indigenous; n = 3504 non-Indigenous); b) reports of concerns for child welfare made to the child infants whose mothers had community-based correc- protection system and the details of investigations, protec- tional orders but no record of imprisonment (n =5828 tion applications and orders as well as placements in Indigenous; n = 653 non-Indigenous); and c) infants out-of-home care. whose mothers had no record within any Department of The Death Registrations, HMDC, and MHIS use the Justice database (n = 12,817 Indigenous; n = 12,555 International Classification of Diseases (ICD), to classify non-Indigenous). Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 4 of 16 Fig. 1 Selection of the study population and classification by maternal corrections history Imprisonment records covered prison stays of any Community-based sentences may involve treatment or length of time, and included unsentenced remandees vocational programs, community service, and place re- detained before trial as well as sentenced prisoners. In- strictions on offenders. Breach of conditions while on fants of mothers who had a prison record at any time over community orders may result in imprisonment. Accord- the study period were further categorised based on the ingly, women with community-based correctional orders timing of their mother’s imprisonment in relation to their are sentenced offenders, but may differ to women given pregnancy. The first group included infants whose custodial prison sentences in terms of severity or fre- mothers had any record of imprisonment during preg- quency of their offending and other individual factors. nancy. The second group included infants whose mothers They are not exposed to the prison environment which had imprisonment records in the period before, but not generally places more stringent conditions on offenders during, pregnancy. The third group included infants and has different implications for them and their families. whose mothers first record of imprisonment only oc- The proportion of infants in the various maternal cor- curred after the child’s birth. These groupings are shown rections history sub-groups (Fig. 1) relate only to the in Fig. 1. Classifications were based on mother’sprisonre- study sample and do not reflect the prevalence of these ception dates and the infant’s birth date. groups across the whole Western Australian population. Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 5 of 16 Pregnancy and birth dates siblings which enabled parity and duration of birth spa- Child month and year of birth was provided by the Mid- cing to be determined. Multiple gestation pregnancies wives Notification System, or if missing from the Birth were also derived based on siblings having shared birth Registration data. As gestational age was not available, dates or through maternal or child HDMC data. Child-level pregnancy start date was calculated as being nine HDMC data were not obtained for the second-generation months before the first day of the child’s birth-month. children, however, as stillbirths were captured for the second-generation this assisted in identifying multiple ges- Definition of infant mortality tation pregnancies. Infant mortality was defined as the death of a live born child under one year of age [23]. Full date of death was Pregnancy complications provided in the death registration data, however, birth The separate and combined effects of key pregnancy data were available only for month and year of birth. Ac- complications identified from maternal hospital records cordingly, infant mortality was defined as death within were evaluated. Pregnancy complications included the 12-full months after birth. For example, for a child born in effects of infection-related hospitalisations, anaemia, dia- January 2000, death on or before 31 January 2001 would betes, hypertension, preeclampsia, eclampsia, abruptio be determined within the category of infant mortality. placentae, placenta previa, other placental disorders, pre- mature rupture of membranes and renal disorders dur- Demographic characteristics ing pregnancy on infant mortality (Additional file 1) The Birth Registration and Midwives Notifications Sys- [19]. Complications were excluded due to low incidence tem data provides social and demographic characteristics or non-significance (p > 0.05), including hospitalisations of mothers and children at time of birth, including sex, for anaemia, diabetes, hypertension, preeclampsia, and socioeconomic status and geographical remoteness. Sex eclampsia (Additional file 1). was taken primarily from Midwives Data, or if missing from Birth Data. Other maternal risk factors and exposures Area-based socio-economic status of infant’s place of Maternal substance use (including alcohol) and residence at time of birth was assigned using the poisoning-related service contacts during pregnancy Socio-economic Indexes for Areas (SEIFA) Index of were identified from HMDC and MHIS data. Maternal Relative Socio-economic Disadvantage [24]. The smallest hospital admissions for any injuries from external area of SEIFA reporting is Collectors District (CD) level, causes, excluding substance use, self-harm and which is approximately 250 households or less in rural poisoning-related contacts were identified during preg- areas. Missing CD scores were imputed with mean nancy. Maternal hospital admissions for mental and be- CD-score by postcode before using broader area scores havioural disorders, self-harm, and mental health service of SEFIA available for Statistical Local Area or Local presentations, both excluding substance use and Government Area [25]. poisoning-related contacts, were identified during preg- The Accessibility/Remoteness Index of Australia nancy from HMDC and MHIS data. Having an older (ARIA) [26] was used to classify the geographical re- sibling in contact with child protection services during moteness of infant’s place of residence at time of birth. the infant’s pregnancy was also identified using sibling’s ARIA is derived from the measure of place of residence child protection data. to populated locations and key services and classified as major cities, inner regional, outer regional, remote, and Statistical analyses very remote locations. For the current study major cities All analyses were conducted using Stata Version 14.0. and inner regional areas were combined given both have All analyses were stratified by Indigenous status of the greater accessibility of relevant services [26]. infants. Indigenous status for infants and mothers in the study Infant mortality rates (per 1000 population) were calcu- populations was ascertained from the Derived Indigen- lated for singleton infants by maternal corrections history. ous Status Flag variable generated by the WADLS using Prevalence of demographic and pregnancy-related risk fac- best-practice algorithms, which assess individuals’ Indi- tors were calculated for Indigenous and non-Indigenous genous status across multiple data collections to en- populations. Log-binomial regression was used to calculate hance accuracy [27]. the Relative Risk of infant death for all univariate and multi- variate analyses. Baseline pregnancy risk factors The strength of correlation between all variables of inter- Maternal birth date was determined using all available est was assessed using Chi-square tests with Cramer’sV data sources. Maternal age was calculated as the age of statistic. For variable pairs with a medium effect size (> 0.3) mother at time of birth. Birth date was available for all [28], one variable was excluded from further multivariate Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 6 of 16 analysis on the basis of the univariate Relative Risk and imprisoned during pregnancy had the highest rates of in- level of statistical significance of each variable with infant fant mortality (32.1 per 1000), compared to infants mortality. Multivariate regression was then conducted for whose mothers were imprisoned either only before preg- each grouping of variables (demographic factors, baseline nancy (22.1 per 1000) or after birth (23.6 per 1000) pregnancy risk factors, pregnancy complications, and (Table 1). other maternal factors and exposures) separately with In non-Indigenous infants, those whose mothers were infant mortality, and variables that were not statisti- imprisoned before (9.4 per 1000) or during (8.2 per cally significant (p < 0.05) were excluded from further 1000) pregnancy had apparently lower rates of infant analysis (Additional file 2). All remaining variables were en- mortality than those whose mothers were imprisoned tered with maternal corrections history into a full regres- for the first time after their birth (14.2 per 1000) or who sion model (Model 1). Variables were removed by key had community orders alone (12.6 per 1000). This differ- groups; other maternal factors and exposures (Model 2), ence was not significant and confidence intervals wide, pregnancy complications (Model 3), baseline pregnancy risk possibly due to the low numbers of infant deaths within factors (Model 4), leaving the combined effects of maternal the sample of non-Indigenous children with mothers corrections history and demographic factors with infant imprisoned before or during pregnancy. mortality. Goodness of model fit was assessed from the Akaike Information Criterion (AIC) value. There were in- Univariate analysis of infant mortality sufficient numbers of non-Indigenous infant deaths whose As shown in Table 2, abruptio placentae and other pla- mothers were imprisoned before or during pregnancy to cental disorders (excluding placenta previa) were associ- undertake multivariate regression for non-Indigenous ated with the highest risk of death for both Indigenous children. and non-Indigenous infants. Other important risk factors Prevalence of key demographic and pregnancy-related for Indigenous and non-Indigenous infants were low so- risk factors, as determined from the univariate and cioeconomic status, multiple gestation pregnancies, birth multivariate analyses, were calculated for each maternal spacing < 18-months, having an older sibling in contact corrections history grouping and by Indigenous status. with the child protection system during an infant’sgesta- tion, and maternal substance use or poisoning-related ser- Results vice contacts during pregnancy. In total, 546 infants in the sample died aged 0 to Maternal hospitalisations for premature rupture of 12 months, between October 1985 and June 2014. There membranes, or external causes of injury were important was a 2.39-fold risk of infant mortality for Indigenous in- risk factors for infant mortality only for Indigenous in- fants compared to non-Indigenous infants (95% CI: fants. Male sex was also only a risk factor for infants in 1.95–2.93, p < 0.001). the Indigenous subgroup. Young maternal age and geo- graphical location for those living in outer regional areas, Infant mortality by maternal corrections history were only significant risk factors in the non-Indigenous For singleton births, Indigenous infants whose mothers sample. were imprisoned at any time had a significantly higher Factors not strongly associated with infant death included risk of death than infants of mothers with community parity, infection related hospitalisations, or mental health corrections orders alone, or with no corrections history related service contact (not related to substance use) in (Table 1). Indigenous infants whose mothers were pregnancy, and hospital admissions for placenta previa. Table 1 Infant mortality for singleton births, by maternal corrections history and Indigenous status Indigenous infants Non-Indigenous infants Survived Died Survived Died a a nn IMR RR (95% CI) p-value nn IMR RR (95% CI) p-value Maternal corrections history Prison (before pregnancy ) 2742 62 22.1 1.95 (1.45–2.62) <.001 634 6 9.4 2.25 (0.97–5.22) 0.059 Prison (during pregnancy) 753 25 32.1 2.84 (1.87–4.31) <.001 243 < 5 8.2 1.96 (0.48–8.00) 0.349 Prison (after birth only) 3481 84 23.6 2.08 (1.59–2.72) <.001 2491 36 14.2 3.42 (2.24–5.23) <.001 Community corrections 5588 94 16.5 1.46 (1.13–1.89) 0.004 628 8 12.6 3.02 (1.44–6.33) 0.003 No corrections* 12,400 142 11.3 12,186 51 4.2 Note Excludes 591 Indigenous and 427 non-Indigenous multiples births Infant mortality rate: deaths under 12 months, per 1000 live births Prison before, but not during, pregnancy Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 7 of 16 Table 2 Risk factors for infant mortality, by Indigenous status Indigenous infants (n = 25,962) Non-Indigenous infants (n = 16,712) Survived Died Survived Died nn % RR (95% CI) p-value nn % RR (95% CI) p-value Infants 25,532 430 1.7 16,596 116 0.7 Demographic factors Sex Male 13,052 240 1.8 1.20 (0.99–1.45) 0.058 8586 64 0.7 1.14 (0.79–1.65) 0.468 Female* 12,438 190 1.5 7991 52 0.6 Maternal Indigenous status Indigenous 24,321 418 1.7 1.72 (0.97–3.05) 0.062 Non-Indigenous* 1211 12 1.0 16,596 116 0.7 Socioeconomic status Very low (0–5%) 6793 128 1.8 2.16 (1.38–3.39) 0.001 996 15 1.5 3.50 (1.88–6.54) <.001 Low (6–25%) 11,095 195 1.7 2.02 (1.30–3.13) 0.002 4461 50 1.1 2.62 (1.65–4.15) <.001 Medium (26–50%) 5054 83 1.6 1.89 (1.18–3.01) 0.008 4517 23 0.5 1.20 (0.69–2.07) 0.523 High (51–100%)* 2546 22 0.9 6584 28 0.4 Geographical remoteness Major cities/Inner regional* 10,331 160 1.5 13,574 88 0.6 Outer regional 4185 73 1.7 1.12 (0.85–1.48) 0.404 1811 22 1.2 1.86 (1.17–2.97) 0.009 Remote 4776 75 1.5 1.01 (0.77–1.33) 0.922 834 5 0.6 0.93 (0.38–2.27) 0.865 Very remote 6195 120 1.9 1.25 (0.98–1.58) 0.066 326 < 5 0.3 0.47 (0.07–3.40) 0.458 Baseline Pregnancy Risk Factors Multiple gestation Yes 568 23 3.9 2.43 (1.61–3.66) <.001 414 13 3.0 4.81 (2.73–8.50) <.001 No* 24,964 407 1.6 16,182 103 0.6 Birth spacing < 18 months 3011 82 2.7 1.74 (1.37–2.21) <.001 1112 19 1.7 2.70 (1.66–4.40) <.001 Firstborn/18 months+* 22,521 348 1.5 15,484 97 0.6 Maternal age 12–19 years 7628 122 1.6 0.91 (0.72–1.17) 0.471 1740 21 1.2 2.36 (1.41–3.94) 0.001 20–24 years 9420 158 1.6 0.96 (0.76–1.20) 0.713 3916 47 1.2 2.35 (1.57–3.52) <.001 25–34 years* 7533 132 1.7 9054 46 0.5 35 + years 951 18 1.9 1.08 (0.66–1.76) 0.761 1886 < 5 0.1 0.21 (0.05–0.86) 0.030 Parity Nulliparous 9035 126 1.4 0.82 (0.66–1.03) 0.093 7954 53 0.7 0.99 (0.67–1.45) 0.939 Parity 1–2* 10,847 184 1.7 7539 51 0.7 Parity 3+ 5650 120 2.1 1.25 (0.99–1.57) 0.058 1103 12 1.1 1.60 (0.86–2.99) 0.140 Pregnancy complications Abruptio placentae and other disorders Yes 122 8 6.2 3.77 (1.91–7.42) <.001 90 < 5 4.3 6.31 (2.38–16.8) <.001 No* 25,410 422 1.6 16,506 112 0.7 Placenta previa Yes 87 < 5 3.3 2.02 (0.66–6.17) 0.217 109 0 0 No* 25,445 427 1.7 16,487 116 0.7 Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 8 of 16 Table 2 Risk factors for infant mortality, by Indigenous status (Continued) Indigenous infants (n = 25,962) Non-Indigenous infants (n = 16,712) Survived Died Survived Died nn % RR (95% CI) p-value nn % RR (95% CI) p-value Premature rupture of membranes Yes 1200 34 2.8 1.72 (1.22–2.43) 0.002 445 5 1.1 1.63 (0.67–3.97) 0.284 No* 24,332 396 1.6 16,151 111 0.7 Infection related hospitalisation in pregnancy Yes 2041 34 1.6 0.99 (0.70–1.40) 0.947 388 < 5 0.5 0.73 (0.18–2.96) 0.664 No* 23,491 396 1.7 16,208 114 0.7 Other maternal factors/Exposures during pregnancy Substance use related service contact Yes 397 15 3.6 2.24 (1.35–3.72) 0.002 293 6 2.0 2.99 (1.33–6.76) 0.008 No* 25,135 415 1.6 16,303 110 0.7 Hospitalisation for external causes of injury Yes 1479 38 2.5 1.56 (1.12–2.17) 0.008 278 < 5 1.1 1.55 (0.50–4.86) 0.450 No* 24,053 392 1.6 16,318 113 0.7 Mental health related service contact Yes 768 12 1.5 0.93 (0.52–1.64) 0.794 427 5 1.2 1.70 (0.70–4.14) 0.244 No* 24,764 418 1.7 16,169 111 0.7 Sibling in contact with child protection Yes 852 30 3.4 2.13 (1.48–3.07) <.001 237 6 2.5 3.70 (1.64–8.33) 0.002 Firstborn/No* 24,680 400 1.6 16,359 110 0.7 *Reference category 42 Indigenous and 19 non-Indigenous infants missing sex 46 Indigenous and 38 non-Indigenous infants missing socioeconomic status 47 Indigenous and 51 non-Indigenous infants missing remoteness Excludes placenta previa Infection related hospitalisation in pregnancy Substance use (including alcohol) or poisoning related service contact (hospital or mental health service) Hospitalisation for external causes of injury in pregnancy (excludes poisoning) Mental health service contact in pregnancy (excludes substance use) (hospital or mental health service) Older sibling(s) in contact with child protection system in infant’s pregnancy Multivariate regression of indigenous infant mortality The best model fit was achieved with the inclusion of all Multivariate regression was only performed for Indigenous variables (i.e., Model 1). Abruptio placentae and other pla- children because of the low number of deaths in the sample cental disorders contributed the highest risk of infant death, of non-Indigenous infants with incarcerated mothers. followed by maternal imprisonment during pregnancy, and Variables for maternal Indigenous status, geographical multiple gestation pregnancy. The effect of birth year remoteness, maternal age, parity, placenta previa, remained stable across Models 1–4 and approximated to a infection-related hospitalisations, external injury-related 4% reduction in risk of infant mortality each year. hospitalisations, and mental health related service con- tacts in pregnancy (other than for substance use) were Prevalence of key demographic and pregnancy risk excluded based on their lack of association with infant factors by maternal corrections history mortality in our data (Additional file 2). The remaining Table 4 shows the prevalence of pregnancy risk factors variables were included and are shown in Model 1 selected from the previous univariate and multivariate (Table 3). In Models 2–4, variables were removed by analyses, for Indigenous and non-Indigenous infants re- group in order of; other maternal factors and exposures ported by their mothers’ contact with corrective services. (Model 2), pregnancy complications (Model 3), baseline Infants of mothers with any corrections history had, in pregnancy risk factors (Model 4), leaving the combined most instances, higher prevalence of maternal service con- effects of maternal corrections history and demographic tact in pregnancy for substance use (including poisoning), factors on infant mortality in Model 4. external causes of injury, mental health related service Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 9 of 16 Table 3 Regression model of infant mortality, Indigenous children Model 1 Model 2 Model 3 Model 4 RR (95% CI) p-value RR (95% CI) p-value RR (95% CI) p-value RR (95% CI) p-value Maternal corrections history Prison (before pregnancy*) 1.83 (1.36–2.47) <.001 1.99 (1.49–2.67) <.001 2.03 (1.51–2.71) <.001 2.10 (1.57–2.81) <.001 Prison (during pregnancy) 2.55 (1.69–3.86) <.001 2.96 (1.98–4.43) <.001 3.01 (2.01–4.51) <.001 3.10 (2.07–4.64) <.001 Prison (after birth only) 1.55 (1.18–2.03) 0.001 1.64 (1.25–2.14) <.001 1.67 (1.28–2.18) <.001 1.73 (1.32–2.25) <.001 Community-only 1.38 (1.07–1.77) 0.012 1.41 (1.10–1.81) 0.007 1.42 (1.11–1.83) 0.006 1.45 (1.13–1.87) 0.003 No corrections history ref. Demographic factors Birth year 0.96 (0.95–0.97) <.001 0.96 (0.95–0.98) <.001 0.96 (0.95–0.98) <.001 0.97 (0.95–0.98) <.001 Sex Male 1.21 (1.01–1.47) 0.044 1.22 (1.01–1.47) 0.039 1.22 (1.01–1.47) 0.038 1.22 (1.01–1.48) 0.036 Female ref. Socioeconomic status Very low (0–5%) 2.11 (1.35–3.31) 0.001 2.08 (1.33–3.27) 0.001 2.1 (1.34–3.30) 0.001 2.11 (1.34–3.31) 0.001 Low (6–25%) 1.88 (1.21–2.91) 0.005 1.86 (1.20–2.88) 0.006 1.87 (1.20–2.89) 0.005 1.91 (1.23–2.96) 0.004 Medium (26–50%) 1.75 (1.10–2.80) 0.018 1.74 (1.09–2.78) 0.02 1.75 (1.10–2.79) 0.019 1.78 (1.11–2.83) 0.016 High (51–100%) ref. Baseline pregnancy risk factors Multiple gestation Yes 2.29 (1.51–3.46) <.001 2.37 (1.57–3.59) <.001 2.58 (1.71–3.88) <.001 No ref. Birth spacing < 18 months 1.52 (1.19–1.93) 0.001 1.57 (1.24–2.00) <.001 1.60 (1.26–2.03) <.001 Firstborn/18 months+ ref. Pregnancy complications Abruptio placentae and other disorders Yes 2.85 (1.46–5.59) 0.002 2.92 (1.49–5.72) 0.002 No ref. Premature rupture of membranes Yes 1.66 (1.18–2.35) 0.004 1.67 (1.18–2.37) 0.003 No ref. Other maternal factors/exposures in pregnancy Substance use related service contact Yes 1.71 (1.02–2.87) 0.042 No ref. External injury related hospitalisation Yes 1.20 (0.86–1.68) 0.283 No ref. Sibling with Child Protection contact Yes 1.57 (1.07–2.31) 0.022 Firstborn/No ref. Observations 25,875 25,875 25,875 265,875 AIC 0.1640 0.1642 0.1646 0.1656 *Prison before, but not during, pregnancy Excludes placenta previa Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 10 of 16 Table 4 Prevalence of pregnancy risk factors for infant mortality, by maternal corrections exposure and Indigenous status Indigenous infants Non-Indigenous infants Prison Prison Prison Community correct. No correct. Prison Prison Prison Community correct No correct (before preg.) (during preg.) (after birth) (before preg.) (during preg.) (after birth) n = 2857 n = 802 n = 3658 n = 5828 n = 12,817 n =667 n = 249 n =2588 n = 653 n = 12,555 Very low (0–5%) SES % 22.9 22.7 29.1 29.8 25.7 13.3 13.4 12.8 12.3 3.8 RR 0.89 0.88 1.13 1.16 ref. 3.48 3.49 3.35 3.20 ref. p-value 0.002 0.057 <.001 <.001 <.001 <.001 <.001 <.001 Multiple gestation % 1.9 3.0 2.5 2.5 2.1 4.0 1.6 2.4 2.6 2.5 RR 0.86 1.39 1.18 1.17 ref. 1.60 0.63 0.93 1.03 ref. p-value 0.328 0.113 0.152 0.126 0.017 0.362 0.602 0.911 Birth spacing < 18mths % 17.9 11.2 14.1 12.8 9.6 12.7 8.0 9.8 10.4 5.6 RR 1.87 1.17 1.47 1.34 ref. 2.27 1.43 1.75 1.86 ref. p-value <.001 0.127 <.001 <.001 <.001 0.099 <.001 <.001 Maternal age < 20 yrs. % 15.3 28.6 41.7 33.1 28.3 9.3 10 25.9 26 6.6 RR 0.54 1.01 1.47 1.17 ref. 1.40 1.51 3.91 3.92 ref. p-value <.001 0.895 <.001 <.001 0.007 0.032 <.001 <.001 Placental disorders % 0.5 0.7 0.6 0.7 0.4 0.1 0.8 1.1 1.1 0.4 RR 1.17 1.92 1.54 1.72 ref. 0.34 1.83 2.56 2.45 ref. p-value 0.620 0.130 0.090 0.011 0.288 0.398 <.001 0.025 PROM % 6.8 5.5 5.1 5.2 4.0 5.5 3.2 3.7 3.2 2.3 RR 1.71 1.39 1.28 1.32 ref. 2.41 1.4 1.59 1.40 ref. p-value <.001 0.032 0.003 <.001 <.001 0.344 <.001 0.133 Substance use % 3.2 7.4 2.2 1.6 0.7 12.3 10.0 4.7 3.2 0.4 RR 4.64 10.59 3.15 2.27 ref. 30.87 25.21 11.74 8.08 ref. p-value <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 External injury Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 11 of 16 Table 4 Prevalence of pregnancy risk factors for infant mortality, by maternal corrections exposure and Indigenous status (Continued) Indigenous infants Non-Indigenous infants Prison Prison Prison Community correct. No correct. Prison Prison Prison Community correct No correct (before preg.) (during preg.) (after birth) (before preg.) (during preg.) (after birth) n = 2857 n = 802 n = 3658 n = 5828 n = 12,817 n =667 n = 249 n =2588 n = 653 n = 12,555 % 10.9 14.5 9.4 6.5 2.9 6.9 6.4 3.4 2.1 0.9 RR 3.78 5.02 3.26 2.25 ref. 7.46 6.95 3.72 2.32 ref. p-value <.001 <.001 <.001 <.001 <.001 <.001 <.001 0.003 Mental health contact % 6.7 8.4 3.6 3.1 1.6 13.5 10.8 5.0 5.1 1.2 RR 4.20 5.22 2.27 1.96 ref. 11.15 8.96 4.15 4.17 ref. p-value <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 Sibling in contact CP % 11.6 14.3 5.2 2.6 0.7 12.9 11.2 4.4 1.5 0.05 RR 15.68 19.35 6.97 3.50 ref. 269.80 235.30 91.36 32.04 ref. p-value <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 46 Indigenous and 38 non-Indigenous infants missing socioeconomic status Excludes placenta previa Premature rupture of membranes Substance use (including alcohol) or poisoning related service contact (Mental Health or Hospital) Hospitalisation for external causes of injury in pregnancy (excludes poisoning) Mental health related service contact in pregnancy (excludes substance use) (Mental Health or Hospital) Older sibling(s) in contact with child protection system in infant’s pregnancy Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 12 of 16 contacts, and sibling contact with the child protection sys- between infant mortality and maternal imprisonment in tem compared to infants whose mothers had no record pregnancy for Indigenous infants remained in the full with corrective services. Prevalence of these service con- model after adjusting for other important risk factors. tacts in pregnancy was highest for infants whose mothers The only determinant to have a greater association with were imprisoned during pregnancy, and higher for infants Indigenous infant mortality was abruptio placentae and whose mothers were imprisoned before pregnancy com- other placental disorders, a serious pregnancy complica- pared to those imprisoned for the first time after birth or tion. Imprisonment during pregnancy was a stronger de- who had community corrections orders for these service terminant of infant mortality than all other pregnancy contacts. complications and baseline pregnancy risk factors in- The proportion of infants born in areas of very low so- cluding multiple gestation pregnancies. Indigenous in- cioeconomic status was high for all groups of Indigenous fants whose mothers were imprisoned during pregnancy children (> 20%) and did not differ by maternal correc- also experienced the highest prevalence of maternal con- tions history. There was a higher proportion of low so- tact with services during pregnancy for substance use, cioeconomic status for non-Indigenous children with mental illness, and external injury. These findings clearly any maternal corrections record (12–13%) compared to demonstrate the significant vulnerability of Indigenous in- those with no maternal corrections history (4%). fants whose mothers are imprisoned during pregnancy. There was a higher prevalence of a birth spacing of It was not possible to determine the relationship be- less than 18-months for infants with any maternal cor- tween maternal imprisonment before and during preg- rections record compared to no maternal corrections nancy and the risk of infant mortality for non-Indigenous history, except the difference was not significant for infants due to the relatively small numbers of infants in those infants whose mothers were imprisoned during the sample populations whose mothers were imprisoned pregnancy. The prevalence of infants born to mothers before or during pregnancy. However, non-Indigenous in- aged less than 20 years was highest among infants whose fants whose mothers who were imprisoned before or dur- mothers were first imprisoned after birth or had ing pregnancy had a significantly higher prevalence of community-corrections orders. several pregnancy risk factors including maternal service For Indigenous children, maternal hospitalisation for contact in pregnancy for substance use, external injury, or abruptio placentae and other placental disorders (exclud- mental health issues, and having siblings in contact with ing placenta previa) was not different between infants by the child protection system, compared to non-Indigenous maternal corrections history, whereas prevalence of ma- infants whose mothers were first imprisoned after birth or ternal hospitalisation for premature rupture of membranes had community-based corrections orders alone. was higher where there had been any record of maternal Infant mortality is a marker of adversity which is contact with the corrections system. strongly linked to social and economic disadvantage [10]. For the non-Indigenous children, hospitalisation of the It is well-established that Indigenous children experience mother for abruptio placentae and other placental disor- higher rates of socioeconomic disadvantage and infant ders (excluding placenta previa) was higher where in- mortality than non-Indigenous children [10, 29]. Over a fants’ mothers were first imprisoned after birth or had quarter of our Indigenous subgroup, compared to only 6% community-correctional orders, and prevalence of hos- of the non-Indigenous sample, was born in the lowest 5% pitalisation for premature rupture of membranes was of areas by socioeconomic status. Socioeconomic disad- higher where infants’ mothers were imprisoned before vantage provides a broad measure of social determinants or after pregnancy. such as parental education, employment, disability, and overcrowding as well as a greater prevalence of health Discussion conditions and risk behaviours such as alcohol and sub- This is the first study to provide a comprehensive investi- stance use, domestic violence, and mental illness, which gation of infant mortality outcomes for children of women are related to infant mortality risk. For Indigenous peo- prisoners. Children of mothers with a history of contact ples, socioeconomic disadvantage is also associated with with corrective services, including community-based cor- experiences of racism and discrimination in service access rections orders and imprisonment before or after preg- and broader society, and in increased contact with health nancy, had increased rates of infant mortality. and criminal justice systems [30]. Within the Indigenous sample, rates of infant mortal- It is recognised that imprisonment likely acts as ity were highest for infants whose mothers were impri- both a proxy for socioeconomic disadvantage and for soned during pregnancy, when compared to similarly risk behaviours which are associated with imprison- disadvantaged mothers who were imprisoned at times ment, including increased substance use, injury and other than pregnancy or who had community-based cor- mental illness, as evidenced in our study by increased rectional orders. The strength of the relationship prevalence of service contacts related to these risks in Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 13 of 16 pregnancy for pregnant prisoners. Similarly, having a represents a positive pregnancy outcome [31, 32]. For sibling in contact with the child protection system example, Walker and colleagues [4] reported the average during pregnancy can be considered as a proxy for length of stay for sentenced women prisoners in New maternal vulnerability and socioeconomic disadvan- South Wales, Australia, is 196 days compared to 547 days tage. Whether maternal imprisonment during preg- in US prisons [4]. In our study of Western Australian nancy has an impact on birth outcomes, including children exposed to maternal incarceration before their infant mortality, over-and-above the effects of second birthday, almost half of all prison stays for either pre-existing disadvantage, is a key issue within the sentenced prisoners or unsentenced remandees were < international literature [3]. 2 weeks [33]. In this context the presumption that im- It has been proposed that there may be a possible pro- prisonment has a ‘dose-response’ effect on perinatal out- tective effect of imprisonment in pregnancy for birth comes [32], and that short-term imprisonment does not outcomes [3]. The proposed protective effect of impris- carry excess risks for the mother or her infant warrants onment in pregnancy on birth outcomes is thought to further research. be contributed to by a reduction of exposure to risk fac- Within Western Australia, all pregnant prisoners are tors such as domestic violence and substance use while provided with health care “commensurate with commu- in custody, and in improved nutrition and access to nity standards” [34]. However, health service provision antenatal care. Within our study, infants whose mothers varies between prisons [35–40], and regional prisons were imprisoned before pregnancy had a higher preva- face additional challenges such as in the transportation lence of risk factors during pregnancy related to mater- of prisoners to community health centres [41]. There is nal service contact for substance use, injury, and mental limited evidence available on the provision and impact illness, when compared to infants whose mothers were of antenatal care provided to pregnant prisoners in first imprisoned after birth. This finding suggests that Australia. In New South Wales, Walker and colleagues risk behaviours of this kind do occur concurrently with [4] found that women imprisoned during pregnancy imprisonment. However, the strength of our findings were more likely to initiate antenatal care after 20 weeks with respect to maternal imprisonment during preg- gestation than women with no record of imprisonment. nancy for Indigenous infants suggest there may be an However as many of these women were not imprisoned additional impact of imprisonment during pregnancy on for the duration of pregnancy, as was the case for infant mortality risk, at least in certain circumstances. women in the present study, it is possible antenatal care Similar to our findings, the general protective effect of was first initiated during imprisonment. imprisonment during pregnancy on birth outcomes that Imprisonment during pregnancy may have negative has been reported in the broader international literature impacts on the mental health and wellbeing of women, was not replicated in the only prior Australian study of particularly for Indigenous mothers, as a consequence of the effects of maternal imprisonment on pregnancy out- being separated from family and country during their comes [4]. While just over one-quarter of pregnant pris- pregnancy [42]. There is evidence that maternal stress in oners were Indigenous, Walker and colleagues [4] did pregnancy can impact on birth outcomes [43, 44]. Re- not investigate the outcomes for the Indigenous and search on offender health has demonstrated that the non-Indigenous populations separately. Our study has year following release from prison, particularly within clearly demonstrated that infant mortality is higher for the first month of release, is also a key risk period for an Western Australian Indigenous infants whose mothers offender’s own hospitalisation and mortality [45–48]. were imprisoned during pregnancy. Taken together, our Whether the release period also leads to increased risk study and that of Walker and colleagues [4] suggest that for offenders’ children, including the unborn children of there are different outcomes for the infants of Australian pregnant prisoners, has not been investigated. The re- women imprisoned during pregnancy than those reported sults from the present study highlight that this as an im- for other jurisdictions [3]. It is not yet clear, however, portant area of future research. whether this difference is restricted to the Indigenous Further research is needed to understand whether population as our findings were inconclusive, due to small there are particular characteristics of maternal im- numbers, with respect to non-Indigenous infants. prisonment during pregnancy that are associated The difference in outcomes for Australian pregnant with infant mortality, and whether the effect is prisoners compared to those within other criminal just- restricted to Indigenous populations. The study find- ice systems reported in the international literature [3], ings highlight the importance of separate consider- may relate to the longer periods of imprisonment experi- ation of Indigenous populations when investigating enced by women in other jurisdictions compared to outcomes for children of prisoners, as combining Australia, as an increased length of imprisonment in populations may mask important differences in pregnancy is associated with higher birth weight which outcomes. Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 14 of 16 Limitations mortality. This study provides the first detailed analysis of These results need to be considered within the context infant mortality outcomes for children whose mothers of the study’s limitations, for example, the power to de- were imprisoned in pregnancy. The study demonstrates tect relationships between infant mortality and maternal that there are higher rates of infant mortality for Indigen- imprisonment before or during pregnancy in the ous, compared to non-Indigenous, children of prisoners non-Indigenous sample was limited by the small num- and that within the Indigenous sample any maternal con- bers of non-Indigenous infant deaths reported within tact with the corrections system is associated with an in- those groups. crease in infant mortality. As gestational age was not available, pregnancy was taken Maternal imprisonment in pregnancy is an important de- to begin nine-months prior to month of birth for all infants. terminant of infant mortality for Indigenous children. Fur- Accordingly, some records of exposure to maternal impris- ther research is needed to determine what factors onment, and service contacts related to substance use, contribute to this increased risk of infant mortality, and mental health or child protection, that were attributed to whether particular groups of prisoners are more affected. having occurred during pregnancy may have occurred prior Due to the relatively low incidence of infant deaths within to pregnancy if gestation was shorter than nine-months. the non-Indigenous sample, it was not possible to deter- There is evidence, however, that adverse events within the mine the impact of maternal imprisonment on infant mor- preconception period (6–0 months before pregnancy) can tality in this sub-population. It was the case, however, that increase infant mortality risk [49, 50]. Therefore any re- non-Indigenous infants whose mothers were imprisoned cords of exposure misclassified as having occurred during before or during pregnancy experienced higher rates of pregnancy, may still have been expected to have an impact pregnancy risk factors, than infants whose mothers were on infant mortality risk. first imprisoned after birth or had community-based cor- Additionally, without gestational age we have not been rectional orders. This highlights the vulnerability of able to measure preterm birth (delivery before 37 weeks non-Indigenous and Indigenous pregnant prisoners, and of gestation), which is associated with infant mortality the importance of providing support services to address [51]. However, preterm birth shares many of the same pregnancy risk factors for women in contact with the cor- risk factors for infant mortality identified within the rections system. study including low socioeconomic status, maternal age, maternal stress, infections, and multiple gestation preg- Additional files nancies [51]. The study has provided the first evidence of an association between maternal incarceration in preg- Additional file 1: ICD codes. List of ICD codes used to define the study variables. (PDF 393 kb) nancy and infant mortality, further research is needed in- Additional file 2: Regression models of infant mortality by main group vestigate the impact of factors not able to be measured in for Indigenous children. Results of the multivariate regression conducted the current study, notably preterm birth, antenatal care for each main grouping of variables to eliminate non-significant variables and caesarean section rates. from including the final multivariate analysis. (PDF 330 kb) Administrative data alone cannot fully capture occur- rences of heavy drinking or substance use, mental ill- Abbreviations ness, or injuries resulting from domestic violence, within AIC: Akaike Information Criterion; ARIA: Accessibility/Remoteness Index of Australia; CPFS: Department of Communities: Child Protection and Family pregnancy [52]. In addition, the study only obtained the Support; HMDC: Hospital Morbidity Data System Collection; ICD: International primary diagnosis code (not co-diagnoses) for hospital Classification of Diseases; MHIS: Mental Health Information System; and mental health service records. Consequently, the as- SEIFA: Socio-economic Indexes for Areas; WADLS: Western Australian Data Linkage System sociations observed in our study with drinking or other substance use in pregnancy, mental illness or injuries are Acknowledgements likely under ascertained. We thank the Developmental Pathways in Western Australian Children Hospital data were not obtained for second-generation Project, the Western Australian Data Linkage Branch, the Western Australian children. While explanatory variables based on hospital Aboriginal Health Research Ethics Committee, the Western Australian Aboriginal and Torres Strait Islander community, the Western Australian data were primarily taken from maternal hospital re- Department of Justice, the Department of Communities: Child Protection cords (available for all mothers), there may have been and Family Support, the Western Australian Department of Health and the some missed cases of chromosomal abnormalities, and custodians of the Western Australian Birth and Death Registrations, and Midwives Notification System for the linkages, provision of data and ongoing substance use related service contacts in pregnancy, for ethical review of project outputs. the second-generation children. Material presented cannot be considered as either endorsed by the Western Australian Department of Justice or an expression of the policies or view of the Department, and any errors or omission or commission are the Conclusions responsibility of the Researchers. To date there have been few studies which have focussed Responsibility for interpretation of the data supplied through the on the impact of maternal imprisonment on infant Government of Western Australia, the Developmental Pathways in Western Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 15 of 16 Australian Children Project, and by any other data custodian, is the 4. Walker JR, Hilder L, Levy MH, Sullivan EA. Pregnancy, prison and perinatal Researchers’ alone. outcomes in new South Wales, Australia: a retrospective cohort study using linked health data. BMC Pregnancy Childb. 2014;14:214. 5. Howard DL, Strobino D, Sherman SG, Crum RM. Timing of incarceration Funding during pregnancy and birth outcomes: exploring racial differences. Matern CD was funded by an Australian Government Research Training Program Child Healt J. 2008;13:457–66. (RTP) Scholarship. 6. Australian Bureau of Statistics. Prisoners in Australia, 2016. Table 20. Canberra: Commonwealth of Australia; 2016. http://www.abs.gov.au/ Availability of data and materials AUSSTATS/abs@.nsf/DetailsPage/4517.02016?OpenDocument. Accessed 20 This study used linked routinely-collected administrative data from the West- Dec 2017 ern Australian Departments of Health, Justice, and Communities: Child Pro- 7. Australian Bureau of Statistics. Estimated resident Aboriginal and Torres tection and Family Support. While these data are accessible to approved Strait Islander and Non-Indigenous population, States and Territories - 30 researchers through the standard ethical and data application processes, they June 2011. Tables 1 and 3. Canberra: Commonwealth of Australia; 2011. are not publicly available and are maintained by the relevant government http://www.abs.gov.au/AUSSTATS/abs@.nsf/DetailsPage/3238.0.55. data custodians. As a result, while future researchers are able to apply dir- 001June%202011?OpenDocument. Accessed 20 Dec 2017 ectly for these data we are restricted under Statutory requirement from re- 8. Council of Australian Governments. Closing the Gap on Indigenous leasing any of our project data to a third party. This includes making the Disadvantage: The Challenge for Australia, February 2009. Canberra: study data available in a publicly available repository. Government of Australia; 2009. https://www.dss.gov.au/sites/default/files/ documents/05_2012/closing_the_gap.pdf. Accessed 20 Dec 2017 Authors’ contributions 9. Council of Australian Governments. Closing the Gap: Prime Minister’s Report CD had primary responsibility of the study design, conduct of the data 2016. Canberra: Government of Australia; 2016. https://www.pmc.gov.au/ analysis and interpretation, and drafting the initial and final versions of the sites/default/files/publications/closing_the_gap_report_2016.pdf. Accessed manuscript. LS, GM and DP contributed to the study design, interpretation of 20 Dec 2017 findings and drafts of the manuscript. GM contributed to the data analysis. 10. Freemantle CJ, Read AW, de Klerk AH, McAullay D, Anderson AP, Stanley FJ. All authors read and approved the final manuscript. Patterns, trends, and increasing disparities in mortality for aboriginal and non-aboriginal infants born in Western Australia, 1980-2001: population Ethics approval and consent to participate database study. Lancet. 2006;367:1758–66. Ethical approval for the project was received from the Department of Health 11. Hayes A, Weston R, Qu L, gray M. Families then and now: 1980-2010. Human Research Ethics Committee, the Western Australian Aboriginal Health Melbourne: Australian institute of family Studies; 2010. Ethics Committee, and the University of South Australia Human Research 12. Mosely HW, Chen LC. An analytical framework for the study of child survival Ethics Committee. in developing countries. Popul Dev Rev. 1984;10:25–45. The study uses data that is routinely collected by the Department of Justice 13. Barker DJP, Harding JE, Owens JA, Robinson JS. Fetal nutrition and and other Western Australian Government departments and services and cardiovascular disease in adult life. Lancet. 1993;341(8850):938–41. linked for research purposes by the Western Australian Data Linkage Branch. 14. Barker DJP. The origins of the developmental origins theory. J Intern Med. In cases such as this the National Health and Medical Research Council 2007;261(5):412–7. outlines in the National Statement on Conduct in Human Research the 15. Eriksson JG. Developmental origins of health and disease – from a small requirements for waiver of consent. Under the National Statement, a Human body size at birth to epigenetics. Ann Med. 2016;48(6):456–67. Research Ethics Council can grant a waiver of consent if it is satisfied certain 16. Halfon N, Larson K, Lu M, Tullis E, Russ S. Lifecourse health development: conditions have been met. The criteria for allowing a waiver listed under the past, present and future. Matern Child Healt J. 2014;18:344–65. National Statement apply to the project. Specifically as the project carries no 17. Heindel JJ, Vandenberg LN. Developmental origins of health and disease: a more than a low risk; it is not feasible to obtain consent due to the age and paradigm for understanding disease etiology and prevention. Curr Opin quantity of the data required; and measures are being taken to ensure the Pediatr. 2016;27(2):248–53. continued protection of the confidentiality of the data. 18. Kozuki N, Lee ACC, Silveira MF, Sania A, Vogel J, Adair L, Barros F, Caulfield LE, Christian P, Fawzi W, Humphrey J, Huybregts L, et al. The associations of parity Competing interests and maternal age with small-for-gestational-age, preterm, and neonatal and The authors declare that they have no competing interests. infant mortality: a meta-analysis. BMC Public Health. 2013;13(Suppl 3):S2. 19. Salihu HM, August EM, de la Cruz C, Mogos MF, Weldeselasse H, Alio AP. Infant mortality and the risk of small for gestational age in the subsequent pregnancy: a Publisher’sNote retrospective cohort study. Matern Child Healt J. 2013;17(6):1044–51. Springer Nature remains neutral with regard to jurisdictional claims in 20. Holman CD, Bass AJ, Rouse IL, Hobbs MS. Population based linkage of published maps and institutional affiliations. health records in Western Australia: development of a health services research linked database. Aust N Z J Public Health. 1999;23(5):453–9. Author details 21. National Coding Centre. The Australian version of the international th Health Economics and Social Policy Group, Centre for Population Health classification of diseases, 9 revision, clinical modification (ICD-9-CM). Research, Sansom Institute, School of Health Sciences, University of South Sydney: National Coding Centre; 1996. Australia, GPO Box 2471, Adelaide, SA 5001, Australia. Centre for Health 22. National Centre for Classification in Health. The international statistical th Services Research, School of Population and Global Health, University of classification of diseases and related health problems, 10 revision, nd Western Australia, Crawley, WA, Australia. Australian modification, ICD-10-AM Australian Coding Standards. 2 edition. Sydney: Faculty of Health Sciences, University of Sydney; 2000. Received: 9 January 2018 Accepted: 22 May 2018 23. Organisation for Economic Co-operation and Development. Glossary of Statistical Terms: Infant mortality rate. Organisation for economic co- operation and development. 2006. https://stats.oecd.org/glossary/detail. References asp?ID=1347. Accessed 20 Dec 2017. 1. Knight M, Plugge E. Risk factors for adverse perinatal outcomes in 24. Australian Bureau of Statistics. 2033.0.55.001 – Census of Population and imprisoned pregnant women: a systematic review. BMC Public Health. 2005; Housing: Socio-Economic Indexes for Areas (SEIFA), Australia – Data Only 5(111) https://doi.org/10.1186/1471-2458-5-111. 2006, Data Cube Only 2001 and Data Only 1996. Commonwealth of 2. Shaw J, Downe S, Kingdon C. Systematic mixed-methods review of Australia, Canberra. 2006. http://www.abs.gov.au/AUSSTATS/abs@.nsf/second interventions, outcomes and experiences for imprisoned pregnant women. +level+view?ReadForm&prodno=2033.0.55.001&viewtitle= J Adv Nurs. 2015;71(7):1451–63. Census%20of%20Population%20and%20Housing:%20%20Socio- 3. Knight M, Plugge E. The outcomes of pregnancy among imprisoned Economic%20Indexes%20for%20Areas%20(SEIFA),%20Australia%20- women: a systematic review. BJOG-Int J Obstet Gyn. 2005;112:1467–74. %20Data%20Cube%20only~2001~Previous~05/10/2006&&tabname= Dowell et al. BMC Pregnancy and Childbirth (2018) 18:202 Page 16 of 16 Past%20Future%20Issues&prodno=2033.0.55.001&issue=2001&num=&view= J Prisoner Health. 2005;11(2):108–18. https://doi.org/10.1108/IJPH-10- &. Accessed 20 Dec 2017. 2013-0046. 25. Dowell CM, Preen DB, Segal L. Quantifying maternal incarceration: a whole- 48. Kinner SA, Preen DB, Kariminia A, Butler T, Andrews JY, Stoové M, Law population linked data study of western Australian children born 1985-2011. M. Counting the cost: estimating the number of deaths among recently Aust N Z J Public Health. 2017;41(2):151–7. released prisoners in Australia. Med J Australia. 2011;195(2):64–8. 49. Class QA, Khashan AS, Lichtenstein P, Långström N, D’Onofrio BM. Maternal 26. Australian Bureau of Statistics. Remoteness Structure. 2014. http://www. stress and infant mortality: the importance of the preconception period. abs.gov.au/websitedbs/d3310114.nsf/home/remoteness+structure. Psychol Sci. 2013;24(7):1309–16. Accessed 20 Dec 2017. 50. Class QA, Mortensen PB, Henriksen TB, Dalman C, D’Onofrio BM, 27. Christensen D, Davis G, Draper G, Mitrou F, McKeown S, Lawrence D, et al. Khashan AS. Preconception maternal bereavement and infant and Evidence for the use of an algorithm in resolving inconsistent and missing childhood mortality: a Danish population-based study. Psychosom Med. indigenous status in administrative data collections. Aust J Soc Issues. 2014; 2015;77(8):863–9. 49(4):423–43. 51. Frey HA, Klebanoff MA. The epidemiology, etiology, and costs of preterm 28. Cohen J. Statistical power analysis for the behavioral sciences, second birth. Seminars in Fetal & Neonatal Medicine. 2016;21(2):68–73. edition: Lawrence Erlbaum Associates, USA. ISBN 0-8058-0283-5; 1988. 52. O’Leary CM, Halliday J, Bartu A, D’Antoine H, Bower C. Alcohol-use disorders 29. Shah PS, Zao J, Al-Wassia H, Shah V. Pregnancy and neonatal outcomes of during and within one year of pregnancy: a population-based cohort study aboriginal women: a systematic review and meta-analysis. Women Health 1985-2006. BJOG-Int J Obstet Gyn. 2013;120(6):744–53. Iss. 2011;21(1):28–39. 30. Prandl KJ, Rooney R, Bishop BJ. Mental health of Australian aboriginal women during pregnancy: identifying the gaps. Arch Womens Ment Health. 2012;15:149–54. 31. Martin SL, Rieger RH, Kupper LL, Meyer RE, Qaqish BF. The effect of incarceration during pregnancy on birth outcomes. Public Health Rep. 1997; 112(4):340–6. 32. Howard DL, Strobino D, Sherman SG, Crum RM. Maternal incarceration during pregnancy and infant birthweight. Matern Child Healt J. 2011;15: 478–86. 33. Dowell CM, Mejia GC, Preen DB, Segal L. Maternal incarceration, child protection, and infant mortality: A descriptive study of infant children of women prisoners in Western Australia. Manuscript submitted for publication. 34. Department of Corrective Services. Policy Directive 74 – Management of Women in Custody. Perth: Government of Western Australia; 2011. 35. Office of the Inspector of Custodial Services. Report of an Announced Inspection of Boronia Pre-release Centre for Women; Report 79. Perth: Government of Western Australia; 2012. 36. Office of the Inspector of Custodial Services. Report of an Announced Inspection of Broome Regional Prison; Report 77. Perth: Government of Western Australia; 2012. 37. Office of the Inspector of Custodial Services. Report of an Announced Inspection of Bandyup Women’s Prison; Report 73. Perth: Government of Western Australia; 2011. 38. Office of the Inspector of Custodial Services. Report of an Announced Inspection of Eastern Goldfields Regional Prison; Report 72. Perth: Government of Western Australia; 2011. 39. Office of the Inspector of Custodial Services. Report of an Announced Inspection of Roebourne Regional Prison; Report 70. Perth: Government of Western Australia; 2011. 40. Office of the Inspector of Custodial Services. Report of an Announced Inspection of Greenough Regional Prison; Report 66. Perth: Government of Western Australia; 2010. 41. Geraghty S. Reaching out: caring for women prisoners in Western Australia. The Practising Midwife. 2015;18(1):26–8. 42. Mukherjee S, Pierre-Victor D, Bahelah R, Madhivanan P. Mental health issues among pregnant women in correctional facilities: a systematic review. Women Health. 2014;54(8):816–42. 43. Bussières E, Tarabulsy G, Pearson J, Tessier R, Forest J, Giguèrec Y. Maternal prenatal stress and infant birth weight and gestational age: a meta-analysis of prospective studies. Dev Rev. 2015;36:179–99. 44. Beijers R, Buitelaar JK, de Weerth C. Mechanisms underlying the effects of prenatal psychosocial stress on child outcomes: beyond the HPA axis. Eur Child Adolesc Psychiatry. 2014;23:943–56. 45. Alan J, Burmas M, Preen D, Pfaff J. Inpatient hospital use in the first year after release from prison: a western Australian population-based record linkage study. Aust N Z J Public Healt. 2011;35(3):264–9. 46. Kariminia A, Butler T, Jones J, Lew M. Increased mortality among indigenous persons during and after release from prison in new South Wales. Aust N Z J Public Healt. 2012;36(3):274–80. 47. Jama-Alol KA, Malacova E, Ferrante A, Alan J, Stewart L, Preen D. Influence of offence type and prior imprisonment on risk of death following release from prison: a whole-population linked data study. Int

Journal

BMC Pregnancy and ChildbirthSpringer Journals

Published: Jun 1, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off