Exploring variation in the performance of planned birth: A mixed method study.

Objective: Variation in practice in relation to indications and timing for both induction of labour (IOL) and planned caesarean section (CS) clearly exists. However, the extent of this variation, and how this variation is explained by clinicians remains unclear. The aim of this study was to map the variation in IOL and planned CS at eight Australian hospitals, and understand why variation occurs from the perspective of clinicians at these hospitals. Our ultimate aim was to identify opportunities for improvement as evidenced by hospital data, clinician experiences, and feedback.

Design: A two-phased mixed method study using sequential explanatory study design. The first phase consisted of an analysis of routinely collected patient data to map variation between hospitals. The second phase consisted of focus groups with clinicians to gain their perspectives on the reasons for variation.

Setting And Participants: Patient data consisted of routine data from 19,073 women giving birth at eight Sydney hospitals between November 2017 and October 2018. Focus groups were attended by a total of 61 medical staff and 121 midwives.

Results: Hospital data analysis found substantial variation, before and after adjustment for case-mix, in rates of both IOL (adjusted rates 27.6%-42%) and planned CS (adjusted rate 15.4%-22.6%). Planned CS by gestation also showed variation, although after restricting analysis to term (≥37 weeks gestation) births, variation was reduced. At focus groups, five main themes explaining variation emerged: local guidelines, policies and procedures (inconsistency and ambiguity); uncertainty of the evidence/what is best practice (contradictory research and different interpretations of evidence); clinician preferences, beliefs and values; the culture of the unit; and organisational influences (access to specialised clinics, theatre time).

Key Conclusions: Considerable variation in IOL and planned CS, even after case-mix adjustment, was found in this sample of Australian hospitals. Engagement with hospital clinicians identified likely sources of this variation and enabled clinicians at each hospital to consider appropriate local responses to address variation, such as more detailed review of their planned birth cases.

Implications For Practice: At a macro level, measures to reduce unwarranted variation should initially focus on consistent national guidelines, while supporting equitable access to operating theatres for optimal CS timing, and shared decision-making training to reduce influence of clinician preference.