Methods used in prevalence studies of disrespect and abuse during facility based childbirth: lessons learned

BACKGROUND: Several recent studies have attempted to measure the prevalence of disrespect and abuse (D&A) of women during childbirth in health facilities. Variations in reported prevalence may be associated with differences in study instruments and data collection methods. This systematic review and comparative analysis of methods aims to aggregate and present lessons learned from published studies that quantified the prevalence of Disrespect and Abuse (D&A) during childbirth. METHODS: We conducted a systematic review of the literature in accordance with PRISMA (Preferred Reporting Items for Systematic Review and Meta-Analysis) guidelines. Five papers met criteria and were included for analysis. We developed an analytical framework depicting the basic elements of epidemiological methodology in prevalence studies and a table of common types of systematic error associated with each of them. We performed a head-to-head comparison of study methods for all five papers. Using these tools, an independent reviewer provided an analysis of the potential for systematic error in the reported prevalence estimates. RESULTS: Sampling techniques, eligibility criteria, categories of D&A selected for study, operational definitions of D&A, summary measures of D&A, and the mode, timing, and setting of data collection all varied in the five studies included in the review. These variations present opportunities for the introduction of biases - in particular selection, courtesy, and recall bias - and challenge the ability to draw comparisons across the studies' results. CONCLUSION: Our review underscores the need for caution in interpreting or comparing previously reported prevalence estimates of D&A during facility-based childbirth. The lack of standardized definitions, instruments, and study methods used to date in studies designed to quantify D&A in childbirth facilities introduced the potential for systematic error in reported prevalence estimates, and affected their generalizability