Elsevier

Health Policy

Volume 124, Issue 12, December 2020, Pages 1395-1402
Health Policy

Birth models of care and intervention rates: The impact of birth centres

https://doi.org/10.1016/j.healthpol.2020.10.001Get rights and content

Highlights

  • Improving maternal autonomy and intervention rates are policy priorities in birth care.

  • Birth centres offer a non-medicalised approach to childbirth led by midwives.

  • However, women choosing birth centres differ considerably from those in hospitals.

  • Accounting for selection effects, birth centres lower birth intervention rates substantially.

  • Importantly, the effects are increasing over time.

Abstract

Birth centres offer a midwifery-led model of care which supports a non-medicalised approach to childbirth. They are often reported as having low rates of birth intervention, however the precise impact is obscured because less disadvantaged mothers with less complex pregnancies, and who prefer and often select little intervention, are more likely to choose a birth centre. In this paper, we use a methodology that purges the impact of these selection effects and provides a causal interpretation of the impact of birth centres on intervention outcomes. Using administrative birth data on over 364,000 births in Australia’s most populous state between 2001 and 2012, we implement an instrumental variables framework to address confounding factors influencing choice of birth setting. We find that giving birth in a birth centre results in significantly lower probabilities of intervention, and that critically, this impact has been increasing over time. Our estimates are larger than those in existing studies, reflecting our newer data, diverging intervention rates across birth settings, and our accounting for important selection effects. The results emphasise the greater role of birth centres in delivering on policy priorities which include greater maternal autonomy, lower intervention rates, and lower health system costs.

Introduction

Recent policy initiatives have drawn attention to variation in birth intervention rates, particularly in high-income countries, and prioritised reducing interventions in the absence of clinical indications [1,2]. Studies have found that the complexity of a woman’s pregnancy does not explain the substantial variation, and there is significant concern that birth care is tending towards ‘too much, too soon’ without corresponding benefit in outcomes for mother or baby [[3], [4], [5]]. Not only does such variation raise concern about appropriate care, but it also contributes to an inappropriate use of health resources. Studies have found that birth interventions increase the cost of childbirth by between 10 and 67 percent [[6], [7], [8]]. These proportionate costs are not trivial and add significantly to the national health budget in public health systems. In Australia, spontaneous birth and caesarean section of a singleton infant are the two most common reasons for an overnight public hospital admission, with labour and childbirth care comprising 19.5 percent of total public hospital expenditures [9].

In this study, we focus on the impact of birth setting on intervention outcomes, specifically the impact of giving birth in a birth centre. Birth centres are midwifery-led, separate spaces (in Australia they are typically within a hospital) designed to provide a home-like setting for low-risk women. Birth centres are characterised by a commitment to the ‘normality of pregnancy and birth’, and have arisen as a response to both the ‘medicalisation’ of the birth process and homebirth safety concerns [10]. Globally, there is substantial variation in where women choose to give birth. In high income countries, between 0.5 percent (in the United States) and 10 percent (in The Netherlands and New Zealand) of women choose birth centres [11]. These differences are attributable to variation in the status, scope and regulation of the role of the midwife, as well as other institutional factors such as funding and the level of integration across maternity care options [[11], [12], [13]]. In Australia, while the vast majority of women labour and give birth in a obstetric unit based within a hospital (97 % in 2015) [14], within our sample of low-risk women for whom birth centres represent a feasible choice, around 7.5 percent chose a birth centre.

The choice of birth setting for pregnant women is a complex one. Studies have shown that not only is the complexity of the pregnancy a factor, but the woman’s previous experience, knowledge of available services, her attitudes towards childbirth and risk, the opinions of her peers, the culture of the health service as well as provider factors contribute to her decision [[15], [16], [17]]. These factors contribute to observed and unobserved differences between the women giving birth in birth centres and in obstetric units, and make it empirically challenging to measure the impact of giving birth in a birth centre. Two systematic reviews [11,18] of studies examining care in midwifery-led settings versus standard maternity care found significantly lower odds of intervention in birth centres, and women were nearly three times as likely to have a normal vaginal birth. The studies in these reviews are challenged in two main ways. First, the randomised control trial (RCT) studies (including two Australian studies, [19,20]) are now quite dated making them problematic given the trends revealed below in Fig. 2. They are also based on small samples, with estimated effects which vary substantially, and may lack precision and generalisability [21]. Importantly, it is doubtful that the women who choose to be randomised to their birth care are a random sample of pregnant women – indeed over three-quarters of invitees refused to participate in the only Australian RCT with this information available [22]. Second, existing observational cohort studies control only for a limited set of observed maternal attributes [23,24], and are unable to account for selection effects. The aim of this study is to provide high-quality estimates of the impact of birth centres on intervention outcomes. We use a large, population-based dataset to implement an instrumental variables framework, which overcomes the selection effects affecting both randomisation (of participation) in RCT studies, and those associated with the observed and unobserved traits of women choosing birth centres in cohort studies.

Australia has a mixed public-private system of healthcare. All women receive universal maternity care provided in public hospitals, and giving birth in a traditional obstetric unit in a public hospital remains the default option for women after confirming their pregnancy with their GP. However, after this initial visit with a GP or with their hospital, some women choose a birth centre model of care, or alternatively pay for private maternity care. All birth centres in NSW are publicly funded and exist within a health service both organisationally and financially. Women in these centres who require higher level care are transferred to an obstetric hospital birth unit following the initial management of the emergency.

This study focuses on births in New South Wales (NSW). NSW is Australia’s most populous state, with a population of around 7.9 million people and a geographically vast area of over 800,000 square kilometres (over 310,000 square miles). Births in NSW account for over a third of all births in Australia. All birth centres are public facilities, and most are co-located with their public hospital counterparts; two facilities in our study are stand-alone facilities (i.e. geographically separated from an obstetric unit but part of a general hospital).

Section snippets

Study design

We present a retrospective cohort study using a population-based administrative dataset. We implement an instrumental variables framework to address the issue of selection effects, using a distance-to-facility-based instrument. Our analytical sample comprises all births to women with low-risk pregnancies in the years 2001–2012.

We focus on a sample of low-risk mothers to minimise the impact of unobserved pregnancy complications on intervention outcomes. The low-risk sample of women was defined

Results

We first present descriptive data on the sociodemographic profile of the women in our sample. Table 1 shows that women choosing a birth centre are on average older, more likely to be married or in a de facto relationship, live in a relatively affluent area, and hold private health insurance. Women with a European, North American, or Australian/Oceania background were more likely to choose a birth centre than women born in Asia or Africa. These characteristics may contribute to selection effects

Discussion

We have used a population-based dataset to show that, after accounting for selection effects which include clinical need and maternal preferences, birth centres have a substantial impact on lowering intervention rates (between 2 and 22 percentage points), and that this impact has been growing over time. The reasons for these differences reflect how birth centres are organised and deliver a different model of care from traditional obstetric units. A birth centre is a separate space, often

Conclusions

There has been strong international policy interest in variation in intervention rates and the role of birth setting. Australian and international policy initiatives have prioritised ensuring women have greater choice in birth place, and reducing unnecessary interventions [1,2]. While birth centres have long been assumed to reduce intervention rates, their impact has been obscured by selection effects attributable to the women who choose them, who are typically older, more educated and in

Declaration of Competing Interest

The authors declare that they have no conflicting interests.

Acknowledgements

Financial support for this study was provided in part by a grant from the National Health and Medical Research Council [grant APP1103015].

References (41)

  • M. McClellan et al.

    The marginal cost-effectiveness of medical technology: A panel instrumental-variables approach

    Journal of Econometrics

    (1997)
  • L. Dubay et al.

    The impact of malpractice fears on cesarean section rates

    Journal of Health Economics

    (1999)
  • S.K. Tracy et al.

    Costing the cascade: estimating the cost of increased obstetric intervention in childbirth using population data

    BJOG: An International Journal of Obstetrics & Gynaecology

    (2003)
  • AHMAC

    “National maternity services plan.” Australian health ministers advisory council

    (2011)
  • World Health Organization

    WHO recommendations: intrapartum care for a positive childbirth experience

    (2018)
  • Y.Y. Lee et al.

    Unexplained variation in hospital caesarean section rates

    Medical Journal of Australia

    (2013)
  • L. Schroeder et al.

    Birthplace cost-effectiveness analysis of planned place of birth: individual level analysis. Birthplace in England research programme. Final report part 5

    (2011)
  • Australian Institute of Health and Welfare

    Admitted patient care 2015-16: Australian hospital statistics

    (2017)
  • P.J. Laws et al.

    Characteristics and practices of birth centres in Australia

    Australian and New Zealand Journal of Obstetrics and Gynaecology

    (2009)
  • C. Benoit et al.

    Understanding the social organisation of maternity care systems: midwifery as a touchstone

    Sociology of Health & Illness

    (2005)
  • Cited by (0)

    View full text