Abstract
Background
Having a preterm (<37 weeks' gestation) birth may increase a woman's risk of early mortality. Aboriginal and Torres Strait Islander (hereafter Aboriginal) women have higher preterm birth and mortality rates compared with other Australian women.
Objectives
We investigated whether a history of having a preterm birth was associated with early mortality in women and whether these associations differed by Aboriginal status.
Methods
This retrospective cohort study used population‐based perinatal records of women who had a singleton birth between 1980 and 2015 in Western Australia linked to Death Registry data until June 2018. The primary and secondary outcomes were all‐cause and cause‐specific mortality respectively. After stratification by Aboriginal status, rate differences were calculated, and Cox proportional hazard regression was used to estimate adjusted hazard ratios (HR) and 95% confidence intervals (CI) for all‐cause and cause‐specific mortality.
Results
There were 20,244 Aboriginal mothers (1349 deaths) and 457,357 non‐Aboriginal mothers (7646 deaths) with 8.6 million person‐years of follow‐up. The all‐cause mortality rates for Aboriginal mothers who had preterm births and term births were 529.5 and 344.0 (rate difference 185.5, 95% CI 135.5, 238.5) per 100,000 person‐years respectively. Among non‐Aboriginal mothers, the corresponding figures were 125.5 and 88.6 (rate difference 37.0, 95% CI 29.4, 44.9) per 100,000 person‐years. The HR for all‐cause mortality for Aboriginal and non‐Aboriginal mothers associated with preterm birth were 1.48 (95% CI 1.32, 1.66) and 1.35 (95% CI 1.26, 1.44), respectively, compared with term birth. Compared with mothers who had term births, mothers of preterm births had higher relative risks of mortality from diabetes, cardiovascular, digestive and external causes.
Conclusions
Both Aboriginal and non‐Aboriginal women who had a preterm birth had a moderately increased risk of mortality up to 38 years after the birth, reinforcing the importance of primary prevention and ongoing screening.
Keywords: Aboriginal, birth, data linkage, mortality, preterm, Western Australia, women
Synopsis.
Study question
The study aimed to investigate whether having a preterm birth was associated with early mortality in Aboriginal and non‐Aboriginal women in Western Australia.
What is already known
Having a preterm birth has been associated with maternal all‐cause and cause‐specific mortality in later life, but there is a paucity of data from minority populations, many of whom who have higher rates of preterm birth.
What this study adds
Having a preterm birth was associated with an increased risk of early mortality, particularly related to diabetes, cardiovascular and digestive conditions, and external causes, compared with those who had a term birth. While the level of relative risk appeared similar among Aboriginal and non‐Aboriginal women, the absolute risks of mortality were distinctly higher among Aboriginal women.
1. BACKGROUND
Preterm birth (<37 weeks) is a universal public health issue, with a global incidence of around 11%. 1 In Australia, the preterm birth rate rose from 6.8% in 1991 2 to 8.6% in 2019, 3 which, at least in more recent years, may be related to an increase in clinician‐initiated preterm births. 4 In addition to adverse consequences for the offspring, 5 preterm birth may be associated with long‐term health risks for the mother. Having a preterm birth has been associated with maternal early all‐cause 6 , 7 , 8 , 9 , 10 and cause‐specific mortality (cardiovascular, 7 , 8 , 9 , 10 , 11 cancer, 6 , 7 diabetes, 7 respiratory, 7 non‐cardiovascular 10 and alcohol‐related 6 ) and a doubling of the risk of combined fatal and non‐fatal cardiovascular disease (CVD). 12 The biological mechanisms behind these associations between having a preterm birth and early mortality are unknown but could include shared contributary factors or pathophysiology. In response to the increasing evidence, the American Heart Association (AHA) called for the introduction of strategies to reduce later cardiometabolic risk among women with adverse pregnancy outcomes, including preterm birth. 13
To date, most studies of having a preterm birth and early mortality among women have been conducted in the Scandinavian countries 6 , 7 , 9 , 10 , 11 where preterm birth rates are low (6%–7%). 14 The AHA has called for more research into the relationship between adverse pregnancy outcomes and CVD in more diverse populations given the paucity of data from ethnic and minority populations, 13 many of whom have higher rates of preterm birth. 1 In Australia, for example, Aboriginal and Torres Strait Islander (hereafter ‘Aboriginal’) women have both distinctly higher rates of preterm birth (nearly 14%) compared with other Australian women (8%) 15 and higher mortality rates across the lifespan, most notably in the 35–44 years age group with rates 4.5 times higher than other Australian women. While a high proportion of the mortality gap can be explained by CVD, diabetes, respiratory diseases and cancer, 16 Aboriginal women are also more likely to die at an earlier age and from external causes of death (e.g. accidents, assaults and suicides) than non‐Aboriginal women. 17 Therefore, any association between having a preterm birth and mortality could add a disproportionate burden to an already vulnerable population.
We investigated whether a history of having a preterm birth is associated with subsequent mortality and whether these associations differed by Aboriginal status. In addition, we investigated cause‐specific mortality and whether the association changed by time since the birth, type or number of preterm births.
2. MATERIAL AND METHODS
2.1. Cohort selection
This study followed the REporting of studies Conducted using Observational Routinely collected health Data (RECORD) guidelines 18 (see checklist, Table S1). This was a population‐based retrospective cohort study, which included all mothers who had a singleton birth in Western Australia between 1980 and 2015, linked to Death Registry data until June 2018 and stratified by Aboriginal status. For each mother, the index birth was the pregnancy with the shortest gestation irrespective of birth order and was chosen after exclusion of multiple births, pregnancy terminations for foetal anomalies and births with missing data for gestation and maternal age (Figure 1).
FIGURE 1.
Study flow chart for investigation of association between preterm birth and maternal deaths in Western Australia 1980–2015.
2.2. Data sources
The study data were obtained from population health data sets including those routinely linked by the data linkage team of the Western Australian Department of Health (Midwives Notification System (MNS), Hospital Morbidity Data Collection (HMDC), Birth, Cancer and Death Registers) as well as the Western Australian Register of Developmental Anomalies, Western Australian Electoral Roll and the Derived Aboriginal and Torres Islander Status Flag. After linkage using probabilistic matching techniques 19 by the data linkage team, these data were provided with identifying fields removed. The MNS contained information about all births from 20 weeks' gestation while the HMDC had information about inpatient episodes. The Western Australian Register of Developmental Anomalies contained information about pregnancy terminations for foetal anomalies which allowed their exclusion while the Western Australian Electoral Roll contained information about when registered voters had left the State. Registration is compulsory for all eligible Australian citizens and information about Electoral Roll status was available for 79.2% of Aboriginal and 82.5% of non‐Aboriginal mothers. The Derived Aboriginal and Torres Islander Status Flag was a variable created by applying a validated algorithm 20 across a range of administrative data sets to produce a single indicator of Aboriginal status to improve the quality of information about Aboriginal status.
2.2.1. Aboriginal status
Indigenous status of the mother was identified using the Derived Aboriginal and Torres Islander Status Flag and mothers were classified as Aboriginal or non‐Aboriginal.
2.2.2. Exposures
Prior to 1998, gestation was calculated using the birth and last menstrual period (LMP) dates if LMP was certain, otherwise by clinical signs (fundal height measurement or neonatal assessment). 21 Since 1998, midwives have recorded the basis of the gestational estimate 22 (ultrasound dating (73%), LMP/clinical signs (25%) and unknown for 2%). 22 , 23 Completed weeks of gestation were categorised as 20–27, 28–31, 32–36 and ≥37 weeks. Preterm births (<37 weeks) were further classified as spontaneous (spontaneous onset of labour or vaginal or caesarean births with pre‐labour rupture of membranes) or clinician initiated (induced labour or pre‐labour caesarean with intact membranes). The total number of singleton preterm births per mother was categorised as 0, 1, 2 and ≥3.
2.2.3. Outcomes
Mortality was identified in the Western Australian Death Registry. Causes of Death, coded using the International Classification of Diseases (ICD) were grouped using Australian Institute of Health and Welfare categories 24 , 25 (Table S2). Cause‐specific mortality was examined in groups with more than 30 deaths among mothers with a preterm birth. Mothers were followed from the index birth date until death, the date they left the State as recorded on the Electoral Roll, or until 30 June 2018, whichever came first.
2.2.4. Covariates
Maternal age at the first and index births were categorised as <20, 20–24, 25–29, 30–34 and ≥35 years. The following were categorised as maternal age at end of follow‐up (≤35, 35–44, 45–54 and ≥55 years), parity in the index pregnancy (0, 1, 2 or ≥3) and after the last birth record in MNS (1, 2, 3 or ≥4), and year of the index birth (1980–1989, 1990–1999, 2000–2009 and 2010–2015). Socio‐economic status (SES) was categorised using quintiles of a composite, area‐based measure of the birth residence 26 based on the Western Australian distribution of the Index of Relative Socio‐Economic Disadvantage Small Area 1 values, generated for the Census closest to the birth year. Areas of remoteness of birth residence, based on the five categories of the Accessibility/Remoteness Index of Australia, 27 were collapsed into two groups (Very remote/remote and Perth metropolitan/inner or outer regional). Smoking during pregnancy (since 1998) was categorised as an indicator (yes/no). Appropriateness of fetal growth was categorised as small‐, appropriate‐ or large for gestational age (<10th, 10–90th and >90th centiles, respectively), based on Australian national birthweight centiles for singletons by gestational age and sex. 28
Pre‐birth malignant cancer diagnoses were identified in the Cancer Register. Women were classified as having a mental health‐related hospitalisation 29 or other chronic condition (cardiovascular, chronic kidney, asthma/chronic respiratory or autoimmune conditions, thyroid disorders, pre‐existing diabetes, human immunodeficiency virus, cystic fibrosis, thalassemia/sickle cell disorders and other diseases of the blood and blood forming organs; see Table S3 for ICD codes) if they had a related diagnosis in a HMDC record between a year prior to conception and birth or were noted in the MNS (pre‐existing diabetes, essential hypertension, asthma). The choice of chronic conditions was based on previous Australian 30 and US 31 comorbidity indices. Women were classified as having the following pregnancy complications: gestational diabetes, preeclampsia/gestational hypertension, or placenta praevia/ placental abruption/ other antepartum haemorrhage if noted in the MNS or were listed as a diagnosis in the HMDC during the index pregnancy (see Table S3 for ICD codes). All diagnostic codes were converted into ICD‐9‐AM and ICD‐10‐AM and modified to reflect coding changes with the assistance of the Western Australian Clinical Coding Authority.
2.2.5. Statistical analyses
Unless stated, analyses were stratified by Aboriginal status. Medians and interquartile ranges (IQR) were calculated for key maternal ages and follow‐up time. Absolute and relative measures of association between gestation and all‐cause and cause‐specific mortality were estimated and rate differences calculated. Associations were examined over the maximum possible follow‐up period (0–38 years). The Kaplan–Meier method was used to estimate the survival functions while hazard ratios (HR) and 95% confidence intervals (CI) were estimated using Cox proportional hazards regression, with maternal age as the time‐scale. 32 Potential confounders were selected based on a priori knowledge and availability of variables in the data set and refined with the development of a Directed Acyclic Graph (Figure S1). The final model included birth year (continuous), maternal age at first birth, parity in index pregnancy, SES, remoteness of residence, gestational diabetes, preeclampsia/gestational hypertension, any antepartum haemorrhage, pre‐birth cancer diagnosis, mental health‐related hospitalisation and chronic disease. The reference group was term births (37–41 weeks). Further models were run to assess whether the association varied by the type or number of preterm births.
2.2.6. Missing data
If the mother's first birth record was not in the MNS (births prior to 1980 or not in Western Australia), her age at her first birth was missing (21%) and information about smoking during pregnancy was only collected from 1998, so 43.1% of smoking data were not collected. The MICE (Multivariate Imputation by Chained Equations) imputation method 33 was used with 50 iterations to impute the missing data for maternal age at the first birth as well as missing SES (3.4%) and remoteness of residence (3%) (See Appendix S1 and Table S4 for more details). Because there was a decline in smoking rates among Australian women between 1980 and 1998 34 (the years of missing data), it was not appropriate to impute these data, so smoking was not included as a covariate in the main regression models. After imputation, the final multivariable models included >99% of the cohort.
2.2.7. Sensitivity analyses
As a woman's risk of having a preterm birth increases with each birth, 35 the analyses were repeated using the first birth in the MNS for each mother (regardless of parity at the time), instead of the shortest gestation. Analyses were repeated excluding mothers missing age at the first birth and smoking during pregnancy in turn. The latter analyses were run with and without smoking as a covariate. Although the sample contained only singleton births, 8572 (1.8%) of these mothers had also had a multiple birth history, so the analyses were repeated excluding them.
Analyses were conducted using SAS version 9.4 (SAS Institute Inc, Cary, NC, USA) and RStudio version 2021.09.0. 36 Cell counts with less than 10 were not tabulated to ensure confidentiality.
2.3. Ethics approval
Ethics approvals were obtained from the Western Australian Department of Health Human Ethics Research Committee (201610.01) and the Western Australian Aboriginal Health Ethics Committee (797).
3. RESULTS
The total study size included 487,291 mothers who had a birth between 1980 and 2015. After exclusion of births missing key data (1.1%), multiple births (2.8%) and pregnancy terminations for foetal anomaly (0.1%), the final sample consisted of 477,601 mothers who had a singleton birth (8,605,490 person‐years of follow‐up), including 20,244 (4.2%) Aboriginal mothers (Figure 1). The maximum follow‐up was 38 years with a median (IQR) follow‐up of 15 6, 25 and 17 7, 27 years among Aboriginal and non‐Aboriginal mothers, respectively. There were 1349 deaths (6.7%) among Aboriginal mothers with a median age at death of 43 (IQR 34, 51) years and 7646 deaths (1.7%) among non‐Aboriginal mothers, with a median age at death of 49 (IQR 40, 57) years. There were distinct differences in most demographic and other characteristics by Aboriginal status (Table 1). Notably, Aboriginal mothers were younger at their first birth, had a higher final parity, and were more likely to live in a low SES or remote/very remote area, smoke during pregnancy, have a mental health hospitalisation or other chronic condition and have a preterm birth (27.3%) than non‐Aboriginal mothers (10.7%). In both population groups, mothers of preterm births were more likely to live in a low SES area, smoke, have a pregnancy complication, mental health hospitalisation or other chronic condition than those with a term birth.
TABLE 1.
Demographic and other characteristics of mothers with and without a preterm birth in Western Australia 1980–2015, stratified by Aboriginal status
| Aboriginal mothers | Non‐Aboriginal mothers | |||||
|---|---|---|---|---|---|---|
| Term (n = 14,716) | Preterm (n = 5528) | Total (n = 20,244) | Term (n = 408,511) | Preterm (n = 48,846) | Total (n = 457,357) | |
| n (%) | n (%) | n (%) | n (%) | n (%) | n (%) | |
| Maternal age (years) at first birth | ||||||
| <20 | 6949 (47.2) | 3168 (57.3) | 10,117 (50.0) | 26,797 (6.6) | 5116 (10.5) | 31,913 (7.0) |
| 20–24 | 3728 (25.3) | 1145 (20.7) | 4873 (24.1) | 79,260 (19.4) | 10,526 (21.5) | 89,786 (19.6) |
| 25–29 | 1018 (6.9) | 262 (4.7) | 1280 (6.3) | 110,420 (27.0) | 12,842 (26.3) | 123,262 (27.0) |
| 30–34 | 356 (2.4) | 92 (1.7) | 448 (2.2) | 74,325 (18.2) | 8901 (18.2) | 83,226 (18.2) |
| ≥35 | 122 (0.8) | 46 (0.8) | 168 (0.8) | 28,516 (7.0) | 3766 (7.7) | 32,282 (7.1) |
| Missing a | 2543 (17.3) | 815 (14.7) | 3358 (16.6) | 89,193 (21.8) | 7695 (15.8) | 96,888 (21.2) |
| Parity in first pregnancy in MNS | ||||||
| 0 | 12,173 (82.7) | 4713 (85.3) | 16,886 (83.4) | 319,318 (78.2) | 41,151 (84.2) | 360,469 (78.8) |
| 1 | 1070 (6.9) | 395 (7.1) | 1415 (7.0) | 51,430 (12.6) | 4369 (8.9) | 55,799 (12.2) |
| 2 | 672 (4.6) | 191 (3.5) | 863 (4.3) | 25,604 (6.3) | 2108 (4.3) | 27,712 (6.1) |
| ≥3 | 850 (5.8) | 228 (4.1) | 1078 (5.3) | 12,062 (3.0) | 1203 (2.5) | 13,265 (2.9) |
| Maternal age (years) at end of follow‐up | ||||||
| 14–34 | 5982 (40.6) | 1719 (31.1) | 7701 (38.0) | 65,527 (16.0) | 6898 (14.1) | 72,425 (15.8) |
| 35–44 | 3551 (24.1) | 1613 (29.2) | 5164 (25.5) | 110,589 (27.1) | 13,857 (28.4) | 124,446 (27.2) |
| 45–54 | 3098 (21.1) | 1492 (27.0) | 4590 (22.7) | 112,230 (27.5) | 14,591 (29.9) | 126,821 (27.7) |
| ≥55 | 2085 (14.2) | 704 (12.7) | 2789 (13.8) | 120,165 (29.4) | 13,500 (27.6) | 133,665 (29.2) |
| Parity after last birth record in MNS | ||||||
| 1 | 3654 (24.8) | 623 (11.3) | 4277 (21.1) | 99,381 (24.3) | 8790 (18.0) | 108,171 (23.7) |
| 2 | 3468 (23.6) | 877 (15.9) | 4345 (21.5) | 173,251 (42.4) | 18,385 (37.6) | 191,636 (41.9) |
| 3 | 2862 (19.4) | 1057 (19.1) | 3919 (19.4) | 90,910 (22.3) | 12,567 (25.7) | 103,477 (22.6) |
| ≥4 | 4731 (32.1) | 2970 (53.7) | 7701 (38.0) | 44,872 (11.0) | 9089 (18.6) | 53,961 (11.8) |
| Index birth | ||||||
| Gestation (completed weeks) | ||||||
| 20–27 | 676 (12.2) | 676 (3.3) | 4007 (8.2) | 4007 (0.9) | ||
| 28–31 | 665 (12.0) | 665 (3.3) | 4515 (9.2) | 4515 (1.0) | ||
| 32–36 | 4187 (75.7) | 4187 (20.7) | 40,324 (82.6) | 40,324 (8.8) | ||
| 37–38 | 7555 (51.3) | 7555 (37.3) | 166,302 (40.7) | 166,302 (36.4) | ||
| 39–41 | 7026 (47.7) | 7026 (34.7) | 238,904 (58.5) | 238,904 (52.2) | ||
| >41 | 135 (0.9) | 135 (0.7) | 3305 (0.8) | 3305 (0.7) | ||
| Birth year group | ||||||
| 1980–1989 | 3349 (22.8) | 1310 (23.7) | 4659 (23.0) | 103,549 (25.3) | 11,414 (23.4) | 114,963 (25.1) |
| 1990–1999 | 3064 (20.8) | 1384 (25.0) | 4448 (22.0) | 96,062 (23.5) | 12,507 (25.6) | 108,569 (23.7) |
| 2000–2009 | 4081 (27.7) | 1669 (30.2) | 5750 (28.4) | 106,719 (26.1) | 14,028 (28.7) | 120,747 (26.4) |
| 2010–2015 | 4222 (28.7) | 1165 (21.1) | 5387 (26.6) | 102,181 (25.0) | 10,897 (22.3) | 113,078 (24.7) |
| Maternal age (years) at index birth | ||||||
| <20 | 4601 (31.3) | 1593 (28.8) | 6194 (30.6) | 17,030 (4.2) | 2657 (5.4) | 19,687 (4.3) |
| 20–24 | 5128 (34.8) | 1817 (32.9) | 6945 (34.3) | 71,472 (17.5) | 8680 (17.8) | 80,152 (17.5) |
| 25–29 | 2922 (19.9) | 1121 (20.3) | 4043 (20.0) | 132,098 (32.3) | 14,803 (30.3) | 146,901 (32.1) |
| 30–34 | 1395 (9.5) | 632 (11.4) | 2027 (10.0) | 121,990 (29.9) | 13,921 (28.5) | 135,911 (29.7) |
| ≥35 | 670 (4.6) | 365 (6.6) | 1035 (5.1) | 65,921 (16.1) | 8785 (18.0) | 74,706 (16.3) |
| Parity at index birth | ||||||
| 0 | 7204 (49.0) | 1909 (34.5) | 9113 (45.0) | 208,537 (51.0) | 24,514 (50.2) | 233,051 (51.0) |
| 1 | 3518 (23.9) | 1366 (24.7) | 4884 (24.1) | 128,417 (31.4) | 13,527 (27.7) | 141,944 (31.0) |
| 2 | 1932 (13.1) | 862 (15.6) | 2794 (13.8) | 49,607 (12.1) | 6575 (13.5) | 56,182 (12.3) |
| ≥3 | 2061 (14.0) | 1390 (25.1) | 7701 (38.0) | 21,853 (5.3) | 4215 (8.6) | 26,068 (5.7) |
| SES area birth residence quintiles | ||||||
| 1 (Most disadvantaged) | 8551 (58.1) | 3534 (63.9) | 12,085 (59.7) | 75,067 (18.4) | 10,184 (20.8) | 85,251 (18.6) |
| 2 | 2983 (20.3) | 1051 (19.0) | 4034 (19.9) | 83,376 (20.4) | 10,388 (21.3) | 93,764 (20.5) |
| 3 | 1642 (11.2) | 518 (9.4) | 2160 (10.7) | 81,778 (20.0) | 9622 (19.7) | 91,410 (20.0) |
| 4 | 840 (5.7) | 253(4.6) | 1093 (5.4) | 80,186 (19.6) | 9296 (19.0) | 89,482 (19.6) |
| 5 (Least disadvantaged) | 330 (2.2) | 77 (1.4) | 407 (2.0) | 73,874 (18.1) | 7978 (16.3) | 81,852 (17.9) |
| Missing b | 370 (2.5) | 95 (1.7) | 465 (2.3) | 14,220 (3.5) | 1378 (2.8) | 15,598 (3.4) |
| Remote or very remote birth residence | 7069 (48.0) | 2692 (48.7) | 9761 (48.2) | 40,186 (9.8) | 4612 (9.4) | 44,798 (9.8) |
| Missing b | 530 (3.6) | 250 (4.5) | 780 (3.9) | 11,980 (2.9) | 1425 (2.9) | 13,405 (2.9) |
| Smoker during pregnancy | ||||||
| No | 5092 (34.6) | 1460 (26.4) | 6552 (32.4) | 203,804 (49.9) | 23,161 (47.4) | 226,965 (49.6) |
| Yes | 3952 (26.9) | 1690 (30.6) | 5642 (27.9) | 27,175 (6.7) | 4573 (9.4) | 31,748 (6.9) |
| Missing c | 5650 (38.4) | 2347 (42.5) | 7997 (39.5) | 176,952 (43.3) | 20,917 (42.8) | 197,869 (43.3) |
| Appropriateness of foetal growth d | ||||||
| Small for gestational age | 2234 (15.2) | 636 (11.5) | 2870 (14.2) | 36,169 (8.9) | 4773 (9.8) | 40,942 (9.0) |
| Appropriate for gestational age | 11,253 (76.5) | 3989 (72.2) | 15,242 (75.3) | 33,3019 (81.5) | 36,771 (75.3) | 369,790 (80.9) |
| Large for gestational age | 1225 (8.3) | 898 (16.2) | 2123 (10.5) | 39,257 (9.6) | 7281 (14.9) | 46,538 (10.2) |
| Pregnancy complications | ||||||
| Pre‐eclampsia/gestational hypertension | 1487 (10.1) | 867 (15.7) | 2354 (11.6) | 36,660 (9.0) | 8815 (18.0) | 45,475 (9.9) |
| Gestational diabetes | 772 (5.2) | 325 (5.9) | 1097 (5.4) | 1, 774 (4.4) | 2715 (5.6) | 20,489 (4.5) |
| Placenta praevia, placental abruption or other ante partum haemorrhage | 480 (3.3) | 727 (13.2) | 1207 (6.0) | 17,522 (4.3) | 8945 (18.3) | 26,467 (5.8) |
| Existing conditions | ||||||
| Pre‐birth cancer diagnosis | 19 (0.1) | 10 (0.2) | 29 (0.1) | 1351 (0.3) | 265 (0.5) | 1616 (0.4) |
| Mental health‐related hospitalisation e | 958 (6.5) | 582 (10.5) | 1540 (7.6) | 7577 (1.9) | 1969 (4.0) | 9546 (2.1) |
| Other chronic condition f | 1593 (10.8) | 808 (14.6) | 2401 (11.9) | 33,400 (8.2) | 5820 (11.9) | 39,220 (8.6) |
| Stillbirth | 47 (0.3) | 428 (7.7) | 475 (2.3) | 410 (0.1) | 3078 (6.3) | 3488 (0.8) |
| Type of censoring | ||||||
| Left State g | 228 (1.5) | 45 (0.8) | 273 (1.3) | 6543 (1.6) | 1050 (2.1) | 10,137 (2.2) |
| Deceased | 838 (5.7) | 511 (9.2) | 1349 (6.7) | 9087 (2.2) | 1103 (2.3) | 7646 (1.7) |
| Censored on 30 June 2018 | 13,651 (92.8) | 4972 (89.9) | 18,623 (92.0) | 392,886 (96.2) | 46,694 (95.6) | 439,580 (96.1) |
| Median follow‐up year (interquartile range) | 14 (7, 26) | 16 (9, 26) | 15 (7, 26) | 17 (8, 28) | 17 (9, 28) | 17 (8, 28) |
Abbreviations: MNS: Midwives Notification System, SES: socio‐economic status.
If there was no birth record in the MNS for mother's first birth, the age was classified as unknown (births prior to 1980 or not in Western Australia).
Missing values only given if ≥1% for any of the subgroups.
Data about maternal smoking only collected from 1998.
These were assigned using Australian sex‐specific national birthweight for age gestation centiles for singletons. 19
Any diagnoses in the Hospital Morbidity Data Collection from 1 year prior to conception until the birth hospitalisation (see Table S3).
Chronic conditions were defined as asthma, pre‐existing diabetes or hypertension recorded in the Midwives Notification System or any of the following diagnoses in the Hospital Morbidity Data Collection from 1 year prior to conception until the birth hospitalisation: cardiovascular, chronic kidney, asthma/chronic respiratory or autoimmune conditions, thyroid disorders, pre‐existing diabetes, human immunodeficiency virus, cystic fibrosis, thalassemia/sickle cell disorders and other diseases of the blood and blood forming organs (see Table S3).
Based on Electoral Roll registrations (Restricted to Australian citizens who had registered for voting ~79% of Aboriginal mothers and 82% of non‐Aboriginal mothers).
3.1. All‐cause mortality
The Kaplan–Meier survival curves show that Aboriginal mothers died younger than non‐Aboriginal mothers (Figure 2). Similarly, mothers who had a preterm died younger than mothers who had only term births, with the difference more pronounced among Aboriginal women. The absolute mortality rates were lowest for term birth and highest in one of the three preterm groups for both Aboriginal and non‐Aboriginal mothers—with higher mortality rates among Aboriginal mothers resulting in at least a four‐fold excess risk (Table 2).
FIGURE 2.
Survival probabilities of women stratified by Aboriginal status and birth group (term/preterm).
TABLE 2.
Association between gestation and maternal all‐cause mortality by Aboriginal status Western Australia 1980–2015 (Follow‐up time 0–38 years)
| Aboriginal mothers (n = 20,244) | Non‐Aboriginal mothers (n = 457,357) | |||||||
|---|---|---|---|---|---|---|---|---|
| Gestation (completed weeks) | Person‐ years | Deaths (Rate a ) | Rate difference (95% CI) | Hazard ratio (95% CI) b | Person‐ years | Deaths (Rate a ) | Rate difference (95% CI) | Hazard ratio (95% CI) b |
| ≥37 | 243, 595 | 838 (344.0) | 0.0 (Reference) | 1.00 (Reference) | 7,386,735 | 6543 (88.6) | 0.0 (Reference) | 1.00 (Reference) |
| All preterm births | ||||||||
| <37 | 96, 505 | 511 (529.5) | 185.5 (135.5, 238.3) | 1.48 (1.32, 1.66) | 878,655 | 1103 (125.5) | 37.0 (29.4, 44.9) | 1.35 (1.26, 1.44) |
| 20–27 | 11, 365 | 63 (554.3) | 210.3 (87.2, 366.2) | 1.62 (1.23, 2.12) | 74,182 | 121 (163.1) | 74.5 (47.9, 106.3) | 1.65 (1.37, 1.98) |
| 28–31 | 11, 000 | 76 (690.9) | 346.9 (206.3, 521.3) | 1.88 (1.47, 2.41) | 85,609 | 118 (137.8) | 49.3 (26.4, 76.5) | 1.36 (1.13, 1.64) |
| 32–36 | 74, 140 | 372 (501.8) | 157.7 (103.7, 215.8) | 1.41 (1.24, 1.60) | 718,864 | 864 (120.2) | 31.6 (23.6, 40.2) | 1.31 (1.22, 1.41) |
| Spontaneous preterm births c | ||||||||
| <37 | 76, 610 | 373 (486.9) | 142.9 (90.2, 199.4) | 1.45 (1.27, 1.64) | 600,746 | 709 (118.0) | 29.4 (20.8, 38.7) | 1.29 (1.20, 1.40) |
| 20–27 | 9443 | 50 (529.5) | 185.5 (55.6, 354.8) | 1.51 (1.12, 2.05) | 53,854 | 83 (154.1) | 65.5 (35.7, 102.5) | 1.49 (1.19, 1.86) |
| 28–31 | 8042 | 46 (572.0) | 228.0 (83.1, 419.4) | 1.71 (1.26, 2.32) | 53,514 | 66 (123.3) | 34.8 (8.3, 68.4) | 1.23 (0.96, 1.57) |
| 32–36 | 59,125 | 277 (468.5) | 124.5 (67.3, 187.0) | 1.40 (1.22, 1.61) | 493,378 | 560 (113.5) | 24.9 (15.7, 34.9) | 1.27 (1.17, 1.39) |
| Initiated preterm births c | ||||||||
| <37 | 19,804 | 138 (696.8) | 352.8 (243.5, 480.6) | 1.62 (1.33, 1.97) | 277,274 | 394 (142.1) | 53.5 (40.0, 68.4) | 1.47 (1.32, 1.63) |
| 20–27 | 1844 | 13 (705.0) | 361.0 (67.5, 859.0) | 2.15 (1.23, 3.76) | 19,734 | 38 (192.6) | 104.0 (51.7, 175.6) | 2.11 (1.53, 2.91) |
| 28–31 | 2945 | 30 (1018.7) | 674.7 (369.5, 1107.1) | 2.25 (1.53, 3.32) | 32,095 | 52 (162.0) | 73.4 (34.9, 123.9) | 1.56 (1.19, 2.06) |
| 32–36 | 15,015 | 95 (632.7) | 288.7 (171.4, 430.5) | 1.45 (1.15, 1.82) | 225,445 | 304 (134.8) | 46.3 (31.8, 62.4) | 1.39 (1.24, 1.57) |
Mortality rate per 100,000 person‐years.
Maternal age used as the Cox model time scale and adjusted for birth year, maternal age at first birth, parity at index pregnancy, socio‐economic status quintile, remote or very remote birth residence, gestational diabetes, preeclampsia/gestational hypertension, any ante‐partum haemorrhage, pre‐birth cancer diagnosis, mental health‐related hospitalisation or chronic disease from 12 months prior to conception to birth.
Spontaneous preterm births were defined as births with spontaneous onset or labour or vaginal births with pre‐labour rupture of membranes. Initiated births were those whose labour was induced or had a pre‐labour caesarean section with intact membranes.
The adjusted HRs (aHR) for mortality associated with preterm birth compared with term birth were 1.48 (95% CI 1.32, 1.66) for Aboriginal and 1.35 (95% CI 1.26, 1.44) for non‐Aboriginal mothers, while the corresponding rate differences were 185.5 (95% CI 135.5, 238.3) and 37.0 (95% CI 29.4, 44.9) per 100,000 person‐years, respectively (Table 2). While the absolute rates were higher in mothers following a clinician‐initiated than a spontaneous preterm birth, the hazard ratios were too imprecise to detect a difference. (Table 2).
3.2. Cause‐specific mortality
Of the 8995 deaths, 7050 (78.4%) were classified into a cause‐specific category. The most common causes of death were cancer (41.2% of total deaths), external causes of death (16.6%), circulatory (11.0%) and digestive (3.0% deaths) systems. Compared to mothers who had a term birth, mothers of preterm birth had higher relative risks of death from most of the causes investigated, with the largest effect sizes recorded for diabetes and diseases of the circulatory system and digestive system. For Aboriginal mothers, there was a more than 10‐fold excess risk for diseases of the circulatory and digestive systems, diabetes and external causes (Table 3). Further breakdowns of some of the causes of death (e.g., cancer and external causes) are provided in Table S5.
TABLE 3.
Association between pregnancy length and maternal cause‐specific mortality by Aboriginal status Western Australia 1980–2015
| Aboriginal mothers | Non‐Aboriginal mothers | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Cause of death | Pregnancy length | Person‐ years | Deaths (Rate a ) | Rate difference (95% CI) | Hazard ratio (95% CI) b | Person‐ years | Deaths (Rate a ) | Rate difference (95% CI) | Hazard ratio (95% CI) b |
| Cancer | |||||||||
| Preterm | 88,736 | 43 (48.5) | −9.8 (−26.3, 9.2) | 0.87 (0.61, 1.25) | 866,926 | 421 (48.6) | 6.1 (1.5, 11.2) | 1.18 (1.06, 1.31) | |
| Term | 231,583 | 135 (58.3) | 0.0 (Reference) | 1.00 (Reference) | 7,321,331 | 3107 (42.4) | 0.0 (Reference) | 1.00 (Reference) | |
| External causes | |||||||||
| Preterm | 89,089 | 109 (122.3) | 39.1 (14.7, 66.6) | 1.31 (1.02, 1.67) | 861,966 | 183 (21.2) | 7.4 (4.4, 10.8) | 1.28 (1.08, 1.50) | |
| Term | 230,748 | 192 (83.2) | 0.0 (Reference) | 1.00 (Reference) | 7,274,710 | 1005 (13.8) | 0.0 (Reference) | 1.00 (Reference) | |
| Circulatory system | |||||||||
| Preterm | 89,930 | 117 (130.1) | 64.9 (40.7, 92.5) | 1.92 (1.49, 2.47) | 861,998 | 126 (14.6) | 6.5 (4, 9.3) | 1.66 (1.36, 2.03) | |
| Term | 231,757 | 151 (65.2) | 0.0 (Reference) | 1.00 (Reference) | 7,273,518 | 592 (8.1) | 0.0 (Reference) | 1.00 (Reference) | |
| Diabetes | |||||||||
| Preterm | 88,828 | 42 (47.3) | 29.9 (16.1, 47.1) | 2.44 (1.54, 3.87) | 860,458 | 33 (3.8) | 2.9 (1.7, 4.4) | 3.99 (2.59, 6.15) | |
| Term | 229,701 | 40 (17.4) | 0.0 (Reference) | 1.00 (Reference) | 7,263,844 | 70 (1.0) | 0.0 (Reference) | 1.00 (Reference) | |
| Digestive system | |||||||||
| Preterm | 88,633 | 37 (41.5) | 16.5 (2.9, 33.3) | 1.67 (1.09, 2.56) | 860,163 | 28 (3.3) | 1.2 (0.2, 2.7) | 1.43 (0.94, 2.19) | |
| Term | 229,746 | 58 (25.2) | 0.0 (Reference) | 1.00 (Reference) | 7,264,994 | 148 (2.0) | 0.0 (Reference) | 1.00 (Reference) | |
| Respiratory system | |||||||||
| Preterm | 88,227 | 18 (20.4) | 7.3 (−2.0, 19.8) | 1.48 (0.81, 2.71) | 860,436 | 40 (4.6) | 1.9(0.6, 3.6) | 1.52 (1.06, 2.16) | |
| Term | 229,208 | 30 (13.1) | 0.0 (Reference) | 1.00 (Reference) | 7,266,135 | 198 (2.7) | 0.0 (Reference) | 1.00 (Reference) | |
| Nervous system | |||||||||
| Preterm | 88,122 | 10 (11.3) | 5.2 (−1.4, 15.1) | 1.71 (0.73, 4.00) | 860,168 | 23 (2.7) | 0.7 (−0.3, 2.1) | 1.42 (0.90, 2.25) | |
| Term | 228,988 | 14 (6.1) | 0.0 (Reference) | 1.00 (Reference) | 7,265,038 | 143 (2.0) | 0.0 (Reference) | 1.00 (Reference) | |
Mortality rate per 100,000 person‐years.
Maternal age used as the Cox model time scale and adjusted for birth year, maternal age at first birth, parity at index pregnancy, Socio‐economic status quintile, remote or very remote birth residence, gestational diabetes, preeclampsia/gestational hypertension, any ante‐partum haemorrhage, pre‐birth cancer diagnosis, mental health‐related hospitalisation or chronic disease from 12 months prior to conception to birth.
3.3. Mortality by number of preterm births
The relative risks of mortality increased by the number of preterm births (Figure 3).
FIGURE 3.
Adjusted hazard ratiosa (95% confidence intervals) for mortality among women by the number of preterm births. aMaternal age used as the Cox model time scale and adjusted for birth year, maternal age at first birth, parity at index pregnancy, socio‐economic status quintile, remote or very remote birth residence, gestational diabetes, preeclampsia/gestational hypertension, any ante‐partum haemorrhage, pre‐birth cancer diagnosis, mental health‐related hospitalisation, or chronic disease from 12 months prior to conception to birth.
3.4. Sensitivity analyses
When the analyses were repeated using the first birth in the MNS for each mother, the proportion with a preterm birth decreased from 27.3% to 13.4% and from 10.7% to 7.1% among Aboriginal and non‐Aboriginal women respectively. However, the mortality risk results remained similar to the main findings as they did when mothers with unknown age at the first birth or any multiple births were excluded (Table S6). When the analyses were restricted to those with smoking data (births from 1998), the addition of smoking to the model made little difference to the aHR (Table S7).
4. COMMENT
4.1. Principal findings
We found that both Aboriginal and non‐Aboriginal women who had a history of a singleton preterm birth in Western Australia between 1980 and 2015 had an increased risk of early mortality compared with those who had a term birth. The level of risk decreased with each additional week of gestation and increased with each additional preterm birth. In both populations, mothers who had a preterm birth had an increased risk of death related to diabetes, cardiovascular and digestive conditions and external causes. The absolute mortality risk associated with a preterm birth was distinctly higher among Aboriginal than non‐Aboriginal mothers. While the level of relative risk generally appeared to be similar for Aboriginal and non‐Aboriginal mothers, this can only be confirmed in studies with larger sample sizes and longer follow‐up to give more precise estimates among Aboriginal women.
4.2. Strengths of the study
This was a population‐based study which included linkages to hospital admissions and the cancer register so we could identify potential risk factors present before the birth.
4.3. Limitations of the data
While we featured one of the longest follow‐up periods for a study of this nature, the median age of the cohort at the end of follow‐up was 39 and 47 years for Aboriginal and non‐Aboriginal women—well short of the average life expectancy for Aboriginal (76 years) and non‐Aboriginal women (83 years) 37 so we could not identify if any excess risk continued later in life. Only information about the main cause of death was available with no data about contributing causes. We also only had information about maternal smoking during pregnancy from 1998. Although smoking made no difference in the sensitivity analyses restricted to a maximum of 20 years follow‐up, this may have been related to the age at the end of follow‐up being younger than when smoking‐related deaths occur, for example the mean age at death from lung cancer in Australian women is 71 years. 38 We also did not have information about substance misuse which is higher in Aboriginal women 39 , 40 , 41 and contributes to preterm birth (drugs and alcohol) 42 and deaths (alcohol). 43
4.4. Interpretation
Only limited epidemiological studies have investigated the relationship between having a preterm birth and subsequent early mortality in mothers, 6 , 7 , 8 , 9 , 10 , 11 , 44 , 45 , 46 , 47 with the majority conducted in Scandinavia, 6 , 7 , 9 , 10 , 11 , 47 and none focusing on any minority populations. Despite methodological differences and follow‐up duration, four 6 , 7 , 8 , 10 of the five studies which investigated overall mortality, 6 , 7 , 8 , 9 , 10 also reported a moderately increased risk with aHR estimates ranging from 1.2 (95% CI 1.1, 1.3) in Sweden after up to 25 years of follow‐up 6 to 1.7 (95% 1.5, 1.8) in Denmark with up to 30 years of follow‐up. 10 Using 39–41 weeks' gestation as the reference, the latest Swedish study by Crump et al. 7 with up to 44 years of follow‐up, also found that the risk was highest at very early gestations (22–27 weeks), and with recurrent preterm births. Unlike us, they found that the risk was higher following initiated preterm births than spontaneous ones, 7 which could be due to a larger sample size. However, indications for initiated preterm births (e.g. ischaemic placental disease) can also predispose to spontaneous births. 48
An umbrella review of literature until 2019 6 , 7 , 8 , 9 , 10 , 11 , 44 , 45 , 46 further confirms the relationship between preterm birth and cardiovascular mortality 12 calculating the risk for cardiovascular mortality following a preterm birth to be 1.9 (95% CI 1.8, 2.0) which is similar to estimates from both our study and the more recent study by Crump et al. 7 Like us, Crump et al. 7 reported that diabetes mortality was associated with preterm birth, although our study extends the cause‐specific outcomes to digestive disorders and external causes of death.
The biological mechanisms underlying the association between having a preterm birth and early mortality are unknown, but there are several theories and likely to be more than one true mechanism. Firstly, it has been said that pregnancy may unmask a predisposition to disease in a diverse range of systems, particularly cardiometabolic disease but also thrombosis, liver, renal, thyroid and pituitary disorders. 49 Further, shared pathophysiology may contribute both to pregnancy complications including preterm birth and to subsequent causes of mortality, such as, inflammation, which is a key contributor to both spontaneous 50 and initiated 51 , 52 , 53 preterm birth and is also implicated in cardiovascular disease. 54 Whatever the mechanisms, it is likely that social deprivation can accentuate the level of risk as it associated with increased risks of both early death 55 and preterm birth. 56 In particular, a high proportion of poor birth outcomes among Aboriginal women have been attributed to social, economic and environmental factors (poverty, lack of access to appropriate care, smoking, alcohol, substance abuse, violence). 42 Importantly, these inequalities are intrinsically linked to the legacies of colonisation in Australia and the associated distress of ongoing marginalisation, discrimination, dispossession and exclusion experienced by Aboriginal peoples over generations. 57 The literature is now replete with examples highlighting the direct and indirect effects of racial discrimination on social and economic circumstances and myriad health outcomes. 58
Having a preterm birth can impact a woman's later health. Because of the baby's health needs, these women often have frequent contact with the health system which makes them an easily identifiable target for monitoring and primary prevention, both in the postpartum period and beyond. This fits well with the concept of the postpartum period being treated as the ‘fourth trimester’ with the woman receiving ongoing care, rather than at a single point of contact. 59 Our findings add to the existing evidence behind the American Heart Association's call for more primary prevention of cardiovascular disease following adverse pregnancy outcomes, 14 which could be extended to other chronic conditions such as diabetes and those of the digestive system. Policies related to improving the socio‐economic circumstances of women, particularly Aboriginal women, and their families, has the potential to reduce the risk of both preterm birth and early mortality among women.
5. CONCLUSIONS
Both Aboriginal and non‐Aboriginal women in Western Australia who had a preterm birth had a moderately increased risk of early mortality up to 38 years. These findings highlight the importance of primary prevention, screening for and management of chronic conditions such as cardiovascular disease and diabetes during and after pregnancy and into middle‐age.
AUTHOR CONTRIBUTIONS
HDB, CG, NAS, RM, CCJS and MS conceived and designed the study. CCJS obtained access to the data. HDB managed the data with assistance from AAA. HDB performed the analyses, did the initial data interpretation, and drafted the initial manuscript. All authors made substantial contributions to theanalysis plan and data interpretation and critically revised the manuscript.
FUNDING STATEMENT
This work was supported by funding from an Australian National Health and Medical Research Council (NHMRC) Project Grant (GNT1127265) which funded HDB and CCJS. The funding source did not participate in the design or conduct of the study, collection, data management, analysis or interpretation or any part of the manuscript preparation.
CONFLICT OF INTEREST
None declared.
Supporting information
Figure S1
Appendix S1
ACKNOWLEDGEMENTS
Within Western Australia, the term Aboriginal is used in preference to Aboriginal and Torres Strait Islander, in recognition that Aboriginal people are the original inhabitants of Western Australia. No disrespect is intended to our Torres Strait Islander colleagues and communities. We would like to acknowledge the Data Linkage team (Western Australian Government Department of Health), the custodians of the Midwives' Notification System, Hospital Morbidity Data Collection, Birth, Death and Cancer Registries, Western Australian Electoral Roll and Western Australian Register of Developmental Anomalies for providing data for this project. We acknowledge the staff at the WA Clinical Coding Authority (WA Department of Health), who advised about the International Classification of Diseases 9‐AM and 10‐AM classifications. We also wish to thank the Australian Co‐ordinating Registry, the Registries of Births, Deaths and Marriages, the Coroners, the National Coronial Information System and the Victorian Department of Justice and Community Safety for enabling COD URF data to be used for this publication. Open access publishing facilitated by Curtin University, as part of the Wiley ‐ Curtin University agreement via the Council of Australian University Librarians. [Correction added on 23 November, after first online publication: CAUL funding statement has been added.]
Bailey HD, Gray C, Adane AA, et al. Early mortality among aboriginal and non‐aboriginal women who had a preterm birth in Western Australia: A population‐based cohort study. Paediatr Perinat Epidemiol. 2023;37:31‐44. doi: 10.1111/ppe.12929
DATA AVAILABILITY STATEMENT
The data underlying this article were provided by the Data Linkage team of the Western Australian Department of Health and cannot be publicly shared by the authors. Researchers wanting to access these data will need to apply to the Data Linkage team of the Western Australian Department of Health (https://www.datalinkage‐wa.org.au/).
REFERENCES
- 1. Blencowe H, Cousens S, Oestergaard MZ, et al. National, regional, and worldwide estimates of preterm birth rates in the year 2010 with time trends since 1990 for selected countries: a systematic analysis and implications. Lancet. 2012;379:2162‐2172. [DOI] [PubMed] [Google Scholar]
- 2. Lancaster P, Huang J, Pedisich E. Australia's Mothers and Babies 1991. Sydney Australian Institute of Health and Welfare, National Pertinatal Statistics Unit; 1994. [Google Scholar]
- 3. Australian Institute of Health and Welfare . Australia's Mothers and Babies: Summary. Australian Institute of Health and Welfare; 2021. [Google Scholar]
- 4. Burger RJ, Temmink JD, Wertaschnigg D, et al. Trends in singleton preterm birth in Victoria, 2007 to 2017: a consecutive cross‐sectional study. Acta Obstet Gynecol Scand. 2021;100:1230‐1238. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Vogel JP, Chawanpaiboon S, Moller AB, Watananirun K, Bonet M, Lumbiganon P. The global epidemiology of preterm birth. Best Pract Res Clin Obstet Gynaecol. 2018;52:3‐12. [DOI] [PubMed] [Google Scholar]
- 6. Smith GD, Sterne J, Tynelius P, Lawlor DA, Rasmussen F. Birth weight of offspring and subsequent cardiovascular mortality of the parents. Epidemiology. 2005;16:563‐569. [DOI] [PubMed] [Google Scholar]
- 7. Crump C, Sundquist J, Sundquist K. Preterm delivery and long term mortality in women: national cohort and co‐sibling study. BMJ. 2020;370:m2533. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Smith GC, Pell JP, Walsh D. Pregnancy complications and maternal risk of ischaemic heart disease: a retrospective cohort study of 129,290 births. Lancet. 2001;357:2002‐2006. [DOI] [PubMed] [Google Scholar]
- 9. Smith GD, Whitley E, Gissler M, Hemminki E. Birth dimensions of offspring, premature birth, and the mortality of mothers. Lancet. 2000;356:2066‐2067. [DOI] [PubMed] [Google Scholar]
- 10. Lykke JA, Langhoff‐Roos J, Lockwood CJ, Triche EW, Paidas MJ. Mortality of mothers from cardiovascular and non‐cardiovascular causes following pregnancy complications in first delivery. Paediatr Perinat Epidemiol. 2010;24:323‐330. [DOI] [PubMed] [Google Scholar]
- 11. Rich‐Edwards JW, Klungsoyr K, Wilcox AJ, Skjaerven R. Duration of pregnancy, even at term, predicts long‐term risk of coronary heart disease and stroke mortality in women: a population‐based study. Am J Obstet Gynecol. 2015;213:e511‐e518. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Okoth K, Chandan JS, Marshall T, et al. Association between the reproductive health of young women and cardiovascular disease in later life: umbrella review. BMJ. 2020;371:m3502. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Parikh NI, Gonzalez JM, Anderson CAM, et al. Adverse pregnancy outcomes and cardiovascular disease risk: unique opportunities for cardiovascular disease prevention in women: a scientific statement from the American Heart Association. Circulation. 2021;143:e902‐e916. [DOI] [PubMed] [Google Scholar]
- 14. Chawanpaiboon S, Vogel JP, Moller AB, et al. Global, regional, and national estimates of levels of preterm birth in 2014: a systematic review and modelling analysis. Lancet Glob Health. 2019;7:e37‐e46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Australian Institute of Health and Welfare . Pregnancy and Birth Outcomes for Aboriginal and Torres Strait Islander Women 2016–2018. AIHW; 2021. [Google Scholar]
- 16. Australian Institute of Health and Welfare . The Health and Welfare of Australia's Aboriginal and Torres Strait Islander Peoples 2015. AIHW; 2015. [Google Scholar]
- 17. Fairthorne J, Walker R, de Klerk N, Shepherd C. Early mortality from external causes in aboriginal mothers: a retrospective cohort study. BMC Public Health. 2016;16:461. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Benchimol EI, Smeeth L, Guttmann A, et al. The REporting of studies conducted using observational routinely‐collected health data (RECORD) statement. PLoS Med. 2015;12:e1001885. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Holman CDJ, Bass AJ, Rouse IL, Hobbs MST. Population‐based linkage of health records in Western Australia: development of a health services research linked database. Aust N Z J Public Health. 1999;23(5):453‐459. [DOI] [PubMed] [Google Scholar]
- 20. Christensen D, Davis G, Draper G, et al. Evidence for the use of an algorithm in resolving inconsistent and missing indigenous status in administrative data collections. Aust J Social Issues. 2014;49:423‐443. [Google Scholar]
- 21. Blair E, Liu Y, Cosgrove P. Choosing the best estimate of gestational age from routinely collected population‐based perinatal data. Paediatr Perinat Epidemiol. 2004;18:270‐276. [DOI] [PubMed] [Google Scholar]
- 22. Maternal and Child Health Unit . Guidelines for Midwives Notification of Case Attended. Maternal and Child Health Unit; 2015. [Google Scholar]
- 23. Hutchinson M, Joyce A, Peirce A. Western Australia's Mothers and Babies, 2015: 33st Annual Report of the Western Australian Midwives' Notification System. Department of Health Western Australia; 2019. [Google Scholar]
- 24. Australian Institute of Health and Welfare . Mortality over the Twentieth Century in Australia: Trends and Patterns in Major Causes of Death. AIHW; 2006. [Google Scholar]
- 25. Australian Institute of Health and Welfare . General Record of Incidence of Mortality (GRIM) Books, 2019. AIHW; 2021. [Google Scholar]
- 26. Australian Bureau of Statistics . Socio‐Economic Indexes for Areas (SEIFA). Australian Bureau of Statistics; 2011. [Google Scholar]
- 27. Australian Bureau of Statistics . Australian Statistical Geography Standard (ASGS): Volume 5–Remoteness Structure 2016; (Catalogue No 1270.0.55.005).
- 28. Dobbins TA, Sullivan EA, Roberts CL, Simpson JM. Australian national birthweight percentiles by sex and gestational age, 1998–2007. Med J Aust. 2012;197:291‐294. [DOI] [PubMed] [Google Scholar]
- 29. Lima F, Shepherd C, Wong J, O'Donnell M, Marriott R. Trends in mental health related contacts among mothers of aboriginal children in Western Australia (1990–2013): a linked data population‐based cohort study of over 40,000 children. BMJ Open. 2019;9:e027733. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Chen JS, Roberts CL, Simpson JM, Ford JB. Use of hospitalisation history (lookback) to determine prevalence of chronic diseases: impact on modelling of risk factors for haemorrhage in pregnancy. BMC Med Res Methodol. 2011;11:68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Leonard SA, Kennedy CJ, Carmichael SL, Lyell DJ, Main EK. An expanded obstetric comorbidity scoring system for predicting severe maternal morbidity. Obstet Gynecol. 2020;136(3):440‐449. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Lamarca R, Alonso J, Gomez G, Munoz A. Left‐truncated data with age as time scale: an alternative for survival analysis in the elderly population. J Gerontol A Biol Sci Med Sci. 1998;53:M337‐M343. [DOI] [PubMed] [Google Scholar]
- 33. van Buuren S, Groothuis‐Oudshoorn K. Mice: multivariate imputation by chained equations in R. J Stat Softw. 2011;45:1‐67. [Google Scholar]
- 34. White V, Hill D, Siahpush M, Bobevski I. How has the prevalence of cigarette smoking changed among Australian adults? Trends in smoking prevalence between 1980 and 2001. Tob Control. 2003;12(Suppl 2):ii67‐ii74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Adane AA, Shepherd CCJ, Farrant BM, White SW, Bailey HD. Patterns of recurrent preterm birth in Western Australia: a 36‐year state‐wide population‐based study. Aust N Z J Obstet Gynaecol. 2022;62:494‐499. [DOI] [PubMed] [Google Scholar]
- 36. RStudio Team . RStudio: Integrated Development Environment for R. RStudio, PBC; 2021. [Google Scholar]
- 37. Australian Institute of Health and Welfare . Deaths in Australia. Australian Institute of Health and Welfare; 2020. [Google Scholar]
- 38. Australian Institute of Health and Welfare . Lung Cancer in Australia. Australian Institute of Health and Welfare; 2011. [Google Scholar]
- 39. Kennare R, Heard A, Chan A. Substance use during pregnancy: risk factors and obstetric and perinatal outcomes in South Australia. Aust N Z J Obstet Gynaecol. 2005;45:220‐225. [DOI] [PubMed] [Google Scholar]
- 40. Burns L, Mattick RP, Cooke M. The use of record linkage to examine illicit drug use in pregnancy. Addiction. 2006;101:873‐882. [DOI] [PubMed] [Google Scholar]
- 41. O'Leary CM, Halliday J, Bartu A, D'Antoine H, Bower C. Alcohol‐use disorders during and within one year of pregnancy: a population‐based cohort study 1985–2006. BJOG. 2013;120:744‐753. [DOI] [PubMed] [Google Scholar]
- 42. Gibberd AJ, Simpson JM, Jones J, Williams R, Stanley F, Eades SJ. A large proportion of poor birth outcomes among aboriginal Western Australians are attributable to smoking, alcohol and substance misuse, and assault. BMC Pregnancy Childbirth. 2019;19:110. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Australian Government, Australian Institute of Health and Welfare, National Indigenous Australian Agency . Tier 2‐ Determinents of Health 2.16 Risky Alcohol Behaviour. Australian Government, Australian Institute of Health and Welfare, National Indigenous Australian Agency; 2021. [updated 30 Apr 2021; cited 2021 19 Nov]. [Google Scholar]
- 44. Nardi O, Zureik M, Courbon D, Ducimetiere P, Clavel‐Chapelon F. Preterm delivery of a first child and subsequent mothers' risk of ischaemic heart disease: a nested case‐control study. Eur J Cardiovasc Prev Rehabil. 2006;13:281‐283. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Hastie CE, Smith GC, Mackay DF, Pell JP. Maternal risk of ischaemic heart disease following elective and spontaneous pre‐term delivery: retrospective cohort study of 750 350 singleton pregnancies. Int J Epidemiol. 2011;40:914‐919. [DOI] [PubMed] [Google Scholar]
- 46. Cirillo PM, Cohn BA. Pregnancy complications and cardiovascular disease death: 50‐year follow‐up of the child health and development studies pregnancy cohort. Circulation. 2015;132:1234‐1242. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Catov JM, Wu CS, Olsen J, Sutton‐Tyrrell K, Li J, Nohr EA. Early or recurrent preterm birth and maternal cardiovascular disease risk. Ann Epidemiol. 2010;20:604‐609. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48. Ananth CV, Vintzileos AM. Epidemiology of preterm birth and its clinical subtypes. J Matern Fetal Neonatal Med. 2006;19:773‐782. [DOI] [PubMed] [Google Scholar]
- 49. Williams D. Pregnancy: a stress test for life. Curr Opin Obstet Gynecol. 2003;15:465‐471. [DOI] [PubMed] [Google Scholar]
- 50. Cappelletti M, Della Bella S, Ferrazzi E, Mavilio D, Divanovic S. Inflammation and preterm birth. J Leukoc Biol. 2016;99:67‐78. [DOI] [PubMed] [Google Scholar]
- 51. Ananth CV, Getahun D, Peltier MR, Smulian JC. Placental abruption in term and preterm gestations: evidence for heterogeneity in clinical pathways. Obstet Gynecol. 2006;107:785‐792. [DOI] [PubMed] [Google Scholar]
- 52. Madan I, Romero R, Kusanovic JP, et al. The frequency and clinical significance of intra‐amniotic infection and/or inflammation in women with placenta previa and vaginal bleeding: an unexpected observation. J Perinat Med. 2010;38:275‐279. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53. Tenorio MB, Ferreira RC, Moura FA, Bueno NB, de Oliveira ACM, Goulart MOF. Cross‐talk between oxidative stress and inflammation in preeclampsia. Oxidative Med Cell Longev. 2019;2019:8238727. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54. Libby P. Inflammation and cardiovascular disease mechanisms. Am J Clin Nutr. 2006;83:456S‐460S. [DOI] [PubMed] [Google Scholar]
- 55. Adler NE, Stewart J. Health disparities across the lifespan: meaning, methods, and mechanisms. Ann N Y Acad Sci. 2010;1186:5‐23. [DOI] [PubMed] [Google Scholar]
- 56. Kramer MS, Seguin L, Lydon J, Goulet L. Socio‐economic disparities in pregnancy outcome: why do the poor fare so poorly? Paediatr Perinat Epidemiol. 2000;14:194‐210. [DOI] [PubMed] [Google Scholar]
- 57. Shepherd CC, Li J, Zubrick SR. Social gradients in the health of indigenous Australians. Am J Public Health. 2012;102:107‐117. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 58. Paradies Y, Ben J, Denson N, et al. Racism as a determinant of health: a systematic review and meta‐analysis. PLoS One. 2015;10:e0138511. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59. Optimizing postpartum care . ACOG Committee Opinion No. 736. American College of Obstetricians and Gynecologists. Obstet Gynecol. 2018;131:e140‐e150. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1
Appendix S1
Data Availability Statement
The data underlying this article were provided by the Data Linkage team of the Western Australian Department of Health and cannot be publicly shared by the authors. Researchers wanting to access these data will need to apply to the Data Linkage team of the Western Australian Department of Health (https://www.datalinkage‐wa.org.au/).