Objective: To evaluate the reliability of direct observation for measuring intrapartum care and compare this method with clinical audits using objective criteria based on patients' medical charts.
Design: Cross-sectional study, data collected by two independent evaluators.
Setting: Hospital in Dakar, Senegal.
Participants: Thirty consecutive intrapartum care episodes provided by midwives and the corresponding medical charts. Outcome Measure The presence or absence of each of twelve criteria selected on the basis of national and international norms for monitoring of labour and delivery (six criteria) and the immediate postpartum period (six criteria).
Results: For direct observation, the labour and delivery mean quality scores ranged from 5.34 to 5.77. In contrast, for the chart-based method, the scores ranged from 0.32 to 0.45. For postpartum care evaluated only with direct observation, the scores were also high (5.21-5.65). For direct observation, inter-evaluator agreement was high: kappa coefficients varied from 0.78 to 0.93 depending on the criterion (total score ICC = 0.74). For the chart-based method, inter-evaluator agreement was also high: 0.66 to 1 (total score ICC = 0.72). Comparison of the two methods showed strong differences by items and subscores.
Conclusion: Using direct observation, the quality of obstetric care was high for both the monitoring of labour and delivery and postpartum care. Both measurement instruments showed high reliability. The chart-based method underestimated the quality of care because of poor medical record documentation. Medical-record-based measurement may not be appropriate for the evaluation of the quality of obstetric care in Senegal and other low-income settings.
Keywords: Senegal; obstetric care; quality; reliability.