ACM responds to NZ study into maternity care, warns researchers not to confuse association with causation

A new study into maternity care in NZ has been gaining a large amount of media coverage in the past few days, with some sensationalist reporting of the results

However, ACM has reviewed the research, and identified some significant methodological flaws. 
With thanks to Deborah Davis, Ank de Jonge and Jane Sandall.

Our response:  

The conclusion drawn from the study carried out by Wernham, Gurney, Stanley, Ellison-Loschmann and Safati is a classic example of research establishing association, not causation. 
Ann Kinnear, CEO of the Australian College of Midwives, said:

“We have abundant high-level evidence which demonstrates the safety and efficacy of midwifery-led models of care. We know that health outcomes for women and babies are improved under continuity of care models, but health outcomes are adversely affected by a lack of access to quality maternity care, which this study demonstrates.”

The study has several key methodological flaws which go some way to explaining the difference in findings when compared to the highly regarded Cochrane Review: firstly, the Wernham et al study compares women based on lead maternity carer when the pregnancy was registered, but does not include data on what care model was in place throughout pregnancy, at the start of labour, or during the postnatal period. 

Deborah Davis, Professor of Midwifery, University of Canberra and previously ACM Vice President, said:

“Anyone with an understanding of the New Zealand model of maternity care will know that this is not the case and that midwives as “lead maternity carers” will often provide care in collaboration with obstetricians, acting as “case managers” for women with highly complex clinical and social situations." 

“The outcomes of this study therefore cannot with any veracity, be associated with midwife led care.” 

Secondly, the study also excludes data about any pregnancy of less than 37 weeks duration; and thirdly, does not include randomisation of the results. 
The available evidence includes a systematic review of randomised control trials in the Cochrane library (level 1 evidence). This study is a retrospective cohort study which is level III-3 evidence. There is a very good reason why randomisation is so important in studies that aim to show (or insinuate) that A causes B.

A further concern is that the authors have combined GP- and obstetrician lead maternity carers under the category of “medical-led care”. This obfuscates any differences or similarities between medical and midwifery led care at the primary level, and differences that might arise from care providers at the primary level (GP and midwife) and tertiary level (obstetricians).

The study clearly shows the disparity in clientele between medical and midwifery led models, with midwives more likely to provide care to more disadvantaged women. The analysis adjusts for this disparity though it is not possible to remove the effect of this from the results, especially when working with an existing database and a limited set of data fields. 

This is why randomisation is such a powerful tool in studies that aim to determine the effect of an intervention (like midwifery led care). With randomisation we can be assured that we are comparing like with like; that the women in both of the models of care are alike. One suspects in this study they are not at all alike despite adjusting for a number of confounding variables.

It is useful to think about what other factors (besides model of care) might influence the clinical outcomes. Women with an obstetrician lead maternity carer are more likely to receive continuity of obstetric care which is something that women accessing primary maternity carers (GPs and midwives) who require obstetric input do not usually receive and they may also have more ready access to obstetric care when it is required. 
Caroline Homer, President, ACM, said:

“While there are differences between the models of care available in Australia and New Zealand, we are confident that the model of midwifery-led woman-centred care that we strive to deliver to all women, no matter where they are, is best practice. We urge all researchers to similarly adhere to best practice by employing consistency, from the use of random control trials to standardised datasets.”      

Internationally renowned researchers will attend the the Normal Labour and Birth Conference, co-hosted by ACM, WSU, which takes place in Sydney on the 10th October, to discuss guidelines for creating consistency in reporting research into birth, and the development of a core outcome dataset for studies looking at outcomes of maternity services. That include outcomes important to women, to assess quality and safety of services and health and wellbeing outcomes, both clinical physical and psycho-social.

Caroline Homer, ACM President said:

“New Zealand has an excellent model of midwifery led care and ACM fully supports this as an option for all women.”


Improving Research into Models of Maternity Care to Inform Decision Making

A Comparison of Midwife-Led and Medical-Led Models of Care and Their Relationship to Adverse Fetal and Neonatal Outcomes: A Retrospective Cohort Study in New Zealand

The Guardian: Births led by midwives rather than doctors linked to greater risks – NZ study