A midwifery critical analysis of: A retrospective cohort study of the association between midwifery experience and perinatal mortality (Lawton et al., 2015)

Abstract

Background: New Zealand has a unique model of maternity care, with midwives as the predominant carers. A recently published retrospective cohort study sought to compare the experience levels of Lead Maternity Carer midwives with rates of perinatal mortality. The paper claimed that an association was found between first year of practice midwives and increased perinatal mortality. However, the study design was seriously flawed and the data do not support the conclusions that were drawn. Aim: To systematically critique the study using the principles of epidemiological research. Methods: The study design and methodology were critically analysed in terms of the three potential sources of error that occur in observational studies: bias (measurement error), confounding and chance. The potential impacts on results were considered. Findings: Four areas of misclassification bias were identified, which introduced error to the data from the design stage. Important confounders were not distinguished and therefore neither excluded in the design stage nor adjusted for in the analysis. These flaws invalidate the results. The hypothesis test for the outcome of interest indicated an extremely high likelihood that the apparent association between midwives in their first year of practice and increased perinatal mortality was due to chance. Conclusion: Because of the major methodological flaws in this research, no conclusions can be drawn from the results. This paper highlights the importance of involving members of the profession being studied in the research team to improve quality by providing crucial insight into the sector and identifying potential sources of error.