Abstract
BACKGROUND: Global policymakers have proposed strengthening midwifery regulation to improve access to and quality of care provided by midwives, thereby enhancing maternal healthcare delivery and outcomes. However, quantifying 'midwifery regulatory environments' as a construct across countries has been difficult, limiting our ability to evaluate relationships between regulatory environments and key outcomes and hindering actionable steps toward improvement. The Global Midwives' Associations map survey includes data across five domains of regulation (overarching regulatory policy and legislation; education and qualification; licensure; registration/re-licensure; and scope and conduct of practice). We aimed to use these data to develop a composite index that represents the midwifery regulatory environment in the countries that participated in the survey. METHODS: To develop our composite Midwifery Regulatory Environment (MRE) Index, we analyzed data from 115 countries in the Global Midwives' Associations map survey. We identified five different possible scoring characterizations for thirteen regulatory items. Four characterizations used continuous or categorical cumulative scoring and one used multiple individual components scoring. We compared these characterizations using Clarke's test and descriptive model fit metrics to identify the best fit and performance for three outcomes: maternal mortality ratio, low birthweight prevalence, and stillbirth rate. RESULTS: The Aggregated Domain Scoring method, which assigns one point for each of the five essential regulatory domains with activity (possible score range: 0-5), was the best fit and performing characterization for maternal mortality ratio and stillbirth outcomes. The Any-or-None Scoring method, which assigns one point per survey item with regulatory activity (possible score range: 0-13), best fit low birthweight prevalence. CONCLUSIONS: Our study demonstrates that developing composite characterizations of complex constructs, as exemplified by MRE Index development, can enhance the usability of existing global health datasets. Additionally, it highlights how employing model fit prediction provides a transparent, replicable, and accessible approach for identifying the optimal characterization of the construct based on a specific outcome. Specifically, we found that different characterizations for the MRE Index are preferred for different maternal health outcomes. The MRE Index we have developed stands as a valuable tool for future research exploring relationships between midwifery regulation and maternal health outcomes.