Introduction

The aim of antenatal (prenatal) care is to provide effective and appropriate screening, preventive or treatment interventions [1] so as to maximise the health of all women during and after pregnancy, and their infants. In Australia, approximately 3.5% of all mothers are Aboriginal and/or Torres Strait Islander (Indigenous) [2]. The number and proportion of Indigenous mothers varies by State and Territory, with the highest proportion occurring in the Northern Territory (39% in 2002) and the highest number of Indigenous mothers reported in Queensland (8,212 women in 2002) [2]. Depending on the State or Territory, the type of community Australian Indigenous women reside in may also vary, ranging from major capital cities or large regional coastal towns to smaller communities in rural and remote areas.

Indigenous women suffer disproportionately high rates of adverse pregnancy outcomes, relative to other Australian women, as is clearly illustrated by the consistently higher rates of perinatal mortality (21.4 per 1,000 perinatal deaths for births to Indigenous mothers compared with 9.6 per 1,000 perinatal deaths for births to non-Indigenous mothers), preterm birth (13% compared with 7%) and low birth weight babies (12.9% compared with 6%) Footnote 1 [2]. The causes of such excessive rates of perinatal morbidity and mortality experienced by Indigenous women are likely multifactorial and complex. High rates of smoking [3], adolescent pregnancies and grand multiparity [4], often compounded by disadvantage due to socio-economic status and location, and the related factors of dispossession and alienation may all contribute.

Indigenous women access antenatal care differently from other Australian women. Most Indigenous women (98%), including those residing in rural and urban areas, give birth in a hospital setting [4]. They are however, more likely to present later in pregnancy for their first antenatal visit and consequently have fewer antenatal care visits [5–7]. Achieving equity in access and utilization of antenatal services is vital in order to address the clear health disparities Indigenous women face. Barriers to accessing antenatal care and other primary health care may be present at many levels; some may relate to the organization or availability of services and socioeconomic factors, however care that is inappropriate to the cultural context is often highlighted as a major barrier to accessing mainstream services [8]. Attempts to address some of these barriers have occurred with the increasing provision of primary health care for Indigenous people by Aboriginal community controlled health services (ACCHS), many of which offer antenatal care. ACCHS are distinct from mainstream services as they are initiated and managed by Boards which include elected members from the local Aboriginal community. These types of services provide a means of access to culturally appropriate and comprehensive primary health care.

The aim of this paper is to review evaluations of changes to services or changes in the delivery of antenatal care for Australian Indigenous women and the impact on utilization and quality of care, birth outcomes and maternal views about care.

Methods

Studies were considered for inclusion in this review if they were published evaluations of either an antenatal care program or an explicit change in the provision of maternity services specifically developed to address the health needs of pregnant Australian Indigenous women. Studies were excluded if they were: not specifically about antenatal care, for example family planning, sexual health, breast feeding or infant care interventions; commentary or discussion about the need for better service provision; reports of descriptive epidemiology about patterns of care utilization or disparities in birth outcomes; descriptions of Indigenous cultural practices surrounding pregnancy and giving birth; or changes in service provision or interventions that were not specifically for Australian Indigenous women. Where care programs were identified for Indigenous women, studies were excluded if there was no formal evaluation undertaken or reported. Decisions about inclusion or exclusion of studies in the review were not based on quality, either of the intervention or the evaluation.

The pre-specified outcomes of interest in the review included measures of care utilization (for example gestational age at first antenatal visit, number of antenatal visits, antenatal hospitalisation, and other measured indicators of use of care) and any of the following health outcomes: perinatal mortality, maternal mortality, pre-eclampsia, anaemia (antenatal or postpartum), treated urinary tract infection including pyelonephritis (requiring antibiotic treatment or hospital admission), preterm birth (< 37 weeks’ gestation), low birth weight (birth weight less than 10th percentile, or birth weight < 2,500 g), birth weight (mean or median) and measures of maternal alcohol, tobacco and other drug use in pregnancy. These health outcomes were chosen as they reflect outcomes most likely to be influenced by provision of antenatal care [1] and also represent health outcomes that are more commonly experienced by Australian Indigenous women compared with other Australian women.

Other outcomes of interest included measures of quality of care: care consistent with antenatal clinical practice guidelines; models or programs that included interventions where there is good evidence of benefit; adherence to treatment interventions; women’s views about care and satisfaction with care; and measures of cost and cost-effectiveness.

The following electronic databases were searched: PubMed (1950-February 2006); Nursing & allied health (CINAHL)-CD (1982-December 2005); Cochrane Controlled Trials Register (CCTR) (The Cochrane Library, Issue 1, 2006) [9]; and the following databases accessed through INFORMIT Footnote 2: ATSIROM (Aboriginal and Torres Strait Islander Online—a collection of bibliographies related to Indigenous affairs including the Aboriginal and Torres Strait Islander Health Bibliography, 1900-February 2006); Australian Medical Index (1968-February 2006); Australian Public Affairs Information Service-Health (1978-February 2006); and RURAL (Rural and remote health database, 1966-February 2006).

The search strategy differed slightly for each of the databases searched, to account for different items in each database (see Fig. 1). Manual searches were also performed of the publication sections on the websites of the Australian Government’s Office of Aboriginal and Torres Strait Islander Health, the Australian Government’s Department of Health and Ageing and State and Territory Government health departments in January 2006. Approval from the relevant Human Research Ethics Committee was not required for this project as it involved a review of published literature only.

Fig. 1
figure 1

Terms used in the literature search

Full size image

Results

Twenty-nine publications referring to fourteen distinct antenatal care programs or interventions were identified [6, 10–37].

Descriptions of each care program or intervention are provided in Table 1. Eight publications were identified from PubMed and a further 21 from the other electronic databases and a search of the bibliographies listed in the publications. Four of the 14 programs were excluded from further consideration in this review for the following reasons: no evaluation report was detected in the time frame of the review [35, 36] or the publications identified reported process indicators only and no health outcomes [30, 31, 37] or provided only a description of the program [32].

Table 1 Description of programs
Full size table

Not surprisingly, of the 10 care programs included in the review, most were linked in some way to an ACCHS often in partnership with mainstream health services. The majority of programs identified were based on a midwifery model of care including primary health care principles. The design of the program evaluation varied widely across publications. It was not possible to compile information from each study to produce a single estimate of effect for individual outcomes, due to the diversity in study design and control groups; including before and after community interventions, the use of historical control groups, comparisons made with other Indigenous women in the area, comparisons with non-Indigenous women in the area and the use of routinely collected State or Territory-level Indigenous and non-Indigenous statistics. Many programs had multiple comparison groups. Information in this review is therefore reported separately for each program including the crude data used in each comparison (where reported). For comparisons where there was a statistically significant difference, an estimate of the magnitude of effect (the absolute difference) and the level of statistical significance are reported.

Care utilization and quality of antenatal care

The number of women utilising the care programs overall and as a proportion of Indigenous women in the various catchment areas varied widely, as did the type of care utilization measures reported (Table 2). The largest reported number of women seen during one care program evaluation period was 280 women and 456 births over 4 years [17]. Statistically significant increases in the number of women initiating antenatal care in the first trimester were reported by two programs [6, 12]; typically increases in the order of 10% were seen. However, only one program [12] reported a significantly earlier gestational age at first antenatal visit for women attending the care program. Not unrelated to the timing of initiation of care, increases in the mean or median number of antenatal visits over pregnancy were reported by two programs [12, 17]. By contrast, one program reported an overall decline in antenatal visits pre versus post-intervention (8.9 visits, standard deviation (SD) 4.3; to 7.3 visits, SD 3.8) [6]. A reduction in the number of women with ‘inadequate care’ (variously defined) was reported by two programs [17, 29].

Table 2 Measures of care utilization and quality of care
Full size table

Measures of quality of care were poorly reported, with only three care programs reporting on quantifiable aspects of quality of care. An increase in the proportion of women attending for routine antenatal tests was reported for one care program [12] (94% compared with women attending two other hospital-based services reporting 71%, P = 0.01 and 84%, P = 0.02). For another program, graphical representations were provided showing increases in recorded care planning, smoking cessation advice, antenatal education and changes in the screening activity over the study period [17]. Another publication reported an assessment of the quality of care related to the specific intervention, an ultrasound training program [10], which indicated that after training, Aboriginal health workers were able to correctly identify most obstetric complications detectable on ultrasound, with the exception of identifying a gestational age sac < 12 weeks’ gestation.

Health outcomes

The most commonly reported health outcomes were perinatal deaths or the perinatal mortality rate, preterm birth, mean birth weight and the proportion of low birth weight babies (< 2,500 g) (Table 3). Of the five programs reporting perinatal deaths or the perinatal mortality rate, three demonstrated no significant difference in the comparisons [12, 17, 20], one [29] reported six fewer deaths in the 6 months after the intervention but no further information was provided, and one program [19] reported a reduction of 36.6 per 1,000 perinatal deaths (95% Confidence Interval (CI) around the difference −60.3 to −12.8) in the perinatal mortality rate (from 45.2/1000 to 8.6/1000) after introduction of the program. Improved perinatal survival was reported by one care program (100% vs. 98%, P = 0.05) [12].

Table 3 Health outcomes
Full size table

Of the four programs that reported preterm birth, one [17] reported a reduction in preterm birth when women attending the care program were compared with historical controls (8.7% vs. 16.7%, P = 0.044), compared with other Aboriginal women giving birth in the same time period (8.7% vs. 14.3%, P = 0.002) and when compared with births to Queensland Indigenous women overall (8.7% vs. 12.3%, P < 0.05). One other program [6] reported a reduction in preterm birth in communities pre and post-intervention (22.4% to 13.1%, P = 0.005). No other differences were seen in the rate of preterm birth [12, 22].

Mean birth weight was reported for four care programs, and increases in mean birth weight over the study periods were reported for at least one of the comparisons undertaken in all of these programs [6, 11, 17, 19]. The magnitude of the birth weight increases ranged from 103 to 196 g. The proportion of low birth weight (< 2,500 g) or small for gestational age babies was reported by five care programs [6, 12, 17, 19, 20], of which, two reported reductions in communities pre and post-intervention (15.3% to 10.9%, P = 0.0148; 14.2% to 8.1%, P = 0.01) [6, 19]. For the remaining programs, no significant differences in the proportion of low birth weight babies were reported.

Other reported health outcomes relevant to this review included the proportion of women diagnosed with pregnancy-induced hypertension (PIH), which was greater for women attending one of the care programs compared with other Indigenous women not attending the program [12] (11% vs. 5.1%, P = 0.03). This may however, be an indication of increased quality of care leading to increased detection of PIH. Another program reported no difference in the rate of detected and treated urinary tract infections but an increase was seen in the proportion of women found to be anaemic at any time in pregnancy (26.3% to 41.3%, P = 0.001) in communities pre and post-intervention [6]. The cause of the increased anaemia outcome is unclear. There were an increased proportion of women with missing data about anaemia in the post-intervention phase, however, as with PIH, improved quality of care may lead to increased detection of anaemia.

Self-reported smoking, use of alcohol or other drugs were reported by one care program only [12]. In this program, no significant differences were observed in the proportion of women who consumed alcohol during pregnancy compared with Indigenous women having care at two other hospital-based services, or in the proportion of women who smoked during pregnancy. Women attending the care program were however, more likely to smoke more than 10 cigarettes per day compared with other Indigenous women attending one of the other hospital-based services (79% vs. 60%, P = 0.01).

Women’s views about care

The methodologies used to collect women’s views about care varied widely, as did the format in which they were reported, for the four care programs reporting on women’s experiences or perceptions of care and/or community discussions about the care programs [11, 12, 14, 27]. Table 4 describes both the positive and negative views expressed about services, as they relate to the individual care programs. Not surprisingly, many of the reported comments were positive reflections about care, with negative comments most often related to waiting times or staffing/organizational issues.

Table 4 Women’s views about the care programs
Full size table

Measures of cost

Only one program directly reported information about a costing analysis [12]. The analysis took into account the running costs of the program minus the cost savings to other centres occurring through women utilising the specific care program services. Downstream health sector costs and savings, described as the difference between health resources consumed by clients compared with other Indigenous women (non-clients) giving birth at the same hospital as the program clients were also taken into account. Annual net health sector costs were estimated as AUD $74,414 ($1,772 per patient), with a downstream health sector saving of $507 per client compared with non-client. This cost saving was due primarily to the shorter postnatal length of stay in hospital seen for clients of the care program.

Discussion

Despite extensive searching we were only able to find published evaluations for 10 antenatal care programs for Australian Indigenous women. This is almost certainly a reflection of a lack of evaluations rather than a lack of activity in the area of antenatal care (e.g. funding for programs for Australian Indigenous women). Of those programs that had been evaluated and published, we found modest increases in indicators of antenatal care utilization, most notably increases in the proportion of women accessing antenatal care in the first trimester. For birth outcomes, benefits were seen in some but not all care programs for the risk of preterm birth and low birth weight babies, in addition to increases in mean birth weight over the time period of the evaluations. Despite these benefits being of potential clinical importance, there was a lack of consistency in these and other outcomes across all care programs, making it difficult to determine the overall effect of care programs, and ultimately to make recommendations about care. Importantly, none of the programs reported any detriments to the health of the women or their infants.

The lack of consistency in findings related to birth and other outcomes is not surprising for a number of reasons. Comparing the findings of individual care programs was made difficult by the diversity in comparison groups used in the evaluations. Many used a historical control group, which can lead to difficulties in distinguishing the impact of an intervention from underlying trends in the outcome over time. Other groups compared outcomes for other Indigenous women giving birth in the same time period but not attending the care programs. These kinds of comparisons are limited by women “self-selecting” or choosing to attend the intervention care programs, and it is likely that these women differ systematically from women not attending the care programs. This potential bias may be manifested by women in the care programs having a higher risk profile in pregnancy, or conversely higher risk women may instead opt to have care at a tertiary level service. Indeed, the interventions in this review were not randomly allocated to communities; therefore it is difficult to accurately assess the effects of the interventions, without considering the baseline health risk factors in the populations assessed as well as other changes in health services that may have occurred over the same time period. Other care programs used routinely collected pregnancy statistics (required for government reporting purposes) to demonstrate trends in birth outcomes occurring in the geographical location of the care program. While the use of such information may be necessary due to a lack of availability or quality of data from the individual health services, it is difficult to demonstrate causality, particularly when there are other health services available in the same location.

Our review has a number of limitations which must be considered, but perhaps the most important is the potential for publication bias. None of the care programs identified in this review demonstrated any real detriments to women’s health for those women attending the care programs. It remains possible that care programs that may have resulted in no overall benefit or even harmful effects on the health of women and their infants may not have been reported or published and therefore evaded detection for our review. Other care programs may not have been detected for this review due to a lack of dissemination of findings to the wider public (including research and policy audiences). This may occur particularly where evaluations are not undertaken by researchers or for the purpose of research.

We are also certain that there are other un-published Australian programs that have implemented strategies to increase Indigenous women’s utilization of antenatal care but which have not been formally evaluated, either at all or beyond the reporting of performance indicators required by some funding agencies. In trying to reduce the potential for publication bias, we searched various sources of ‘grey’ literature (i.e. information that is not commercially published and not accessible through conventional search engines); however it remains possible that potentially relevant sources of information have not been identified. For example, at the time of completion of this review we became aware of another literature review of relevance to this topic [38], which included the majority of antenatal care programs reported in our review. This Australian Government report was not available in the academic literature and there were delays in adding the report to the relevant website, leading to delays in dissemination of findings to the wider public.

Almost all of the care programs included in this review reported perinatal mortality trends. While the very nature of mortality rates make them a “hard outcome” (i.e. relatively unlikely to be influenced by measurement and other errors), none of the evaluations had sufficient statistical power to be able to show clinically important differences in perinatal mortality or other pregnancy and birth outcomes. With the exception of one antenatal care program (which incidentally did not report birth outcomes) [27], none of the care programs reported a power calculation for the study evaluation. It is unrealistic to expect to show a significant change in perinatal mortality from care programs that might see on average, 50 new antenatal patients per year, as was typical in the care programs identified in this review. Only with the collection and reporting of good quality longitudinal data about these and other care programs, will clinically relevant differences in perinatal outcomes be able to be demonstrated. Unfortunately, good quality longitudinal data about Australian Indigenous mothers and their babies is not yet available even on a national level [4], due in part to difficulties in the accurate ascertainment of Indigenous status of both mothers and fathers.

In Australia, maternity health services (including antenatal care) are provided by obstetricians, midwives, general practitioners, nurses and Aboriginal health workers, working within hospital obstetric units, birthing centres, community controlled health services, government clinics and private practitioner rooms. Despite the diversity in antenatal care providers available in Australia, there are no national guidelines regarding the provision of antenatal care. As a result, variations in the content of antenatal visits and recommended antenatal screening procedures exist across Australia [39], which can contribute to inequalities in care and outcomes.

Factors affecting the provision of high quality antenatal care to Indigenous women are likely to be different in rural and remote settings compared with urban areas. In remote areas, challenges exist around the delivery of health services to small discrete communities often in sparsely populated areas; however improved access to appropriate health services has occurred with the initiation of local Aboriginal community controlled health services. In urban areas, Indigenous people may access a range of providers of primary health care services, not just community controlled health services [40], which highlights the need to ensure mainstream service providers are responsive to the needs of Indigenous people, and can work effectively in a cross-cultural environment.

It is important to take a long-term view when assessing the impact of antenatal care programs developed specifically for Indigenous women. There are clear disparities in health outcomes faced by Indigenous women compared with other Australian women. These are not limited to pregnancy and childbirth, and it may not be possible to rectify such vast disparities in the short term. Given that the programs often represent a partnership between ACCHS and mainstream services, reaching the women in most need of these care programs may require addressing some of the long-held issues surrounding mistrust of mainstream services. In addition, there is a limited workforce available for these programs, many of which were undertaken outside of major cities or even in remote areas. Few programs however, addressed issues surrounding the long-term sustainability of the program in their evaluation.

The difficulties in comparing the care programs in this review highlight the tensions often seen in health services research between good research practice that is grounded in consistency and comparability of findings versus adequate and appropriate provision of care that responds to local needs and situations. Indeed, rigorous care evaluations often require significant time, expertise and financial investment which simply may not be feasible within already under-resourced health services. It is therefore not surprising that many of the evaluations in this review were limited by the quality and availability of data. For some comparisons there was incomplete capture of data, and in some instances conclusions were drawn about programs that were not explicitly supported by the evidence reported in the publications. These findings highlight the need for provision of adequate resources to ensure that health care evaluations are undertaken with appropriate rigour and reproducibility, and have the ability to produce meaningful findings that could be generalized to other populations. In the meantime, however, the limited high quality data demonstrated in this review should not impede the implementation and support of antenatal care programs for Australian Indigenous women, when there are clear inequalities in women’s access to appropriate care.

Women’s views about care form an important part of assessments of health care interventions. In this review, only four care programs formally reported on women’s experiences or views of care, and not surprisingly, most of the discussions about the care programs were positive. The use of female staff, supportive staff attitudes and the continuity of care approach used in the services were commonly “liked” aspects of care. The importance of women’s satisfaction with care must not be underestimated, and it may be that changes in women’s views about services (in contrast to quantifiable health outcomes) are the primary short term outcomes. This may be particularly true for organizations that see only small numbers of women per year. There may also be other outcomes that cannot be easily quantified, for example women “re-claiming cultural pride” or the “empowerment of women”. Statements such as these were often expressed in the publications about care programs, although not substantiated quantitatively. Women’s satisfaction with care may be an important part of an assessment of the opportunity-costs of a care program, and this highlights the need for the collection and reporting of qualitative aspects of care in addition to quantitative measures of health outcomes.

Conclusion

Indigenous people in Australia, the United States, Canada and New Zealand share a common history of dispossession of land and culture, and common experiences of disadvantage, marginalization and disproportionately high rates of adverse health outcomes compared with non-Indigenous people. There are however, no greater disparities in health outcomes than those seen for Australian Indigenous men, women and children relative to other Australians [41, 42]. In this review, the importance of providing care that addresses the needs of Australian Indigenous women is not disputed. However, assessing the impact of the antenatal care programs that have been developed and evaluated to date was problematic, particularly due to the short term focus of many of the evaluations. While some benefits were seen in relation to increased utilization of care in early pregnancy, improved birth outcomes and women’s views about care, it was difficult to compare the individual care programs in a meaningful way. Any comparisons made were hampered by the uneven quality and availability of the data reported from the care programs.

While the antenatal care programs reported in this review may be culture specific, the mechanisms of service delivery may be transferable to other Indigenous populations, but only if the evaluations are well conducted and widely disseminated. This review has highlighted the need to implement systems that will enable long term high quality data collection about the health services, and the health of women and infants utilizing the services.