National perinatal audit, a feasible initiative for the Netherlands!? A validation study

Objective. To explore the feasibility of a national perinatal audit organization. Design. Validation study. Setting. Three regions in the Netherlands. Population. 228 cases of perinatal mortality. Methods. Narratives of perinatal mortality cases were assessed by a panel of representatives of all perinatal care provider groups. 123 cases were assessed twice. Consensus was defined as 75% agreement. For the chance corrected agreement Cohen's kappa statistic was used. Main outcome measures. Consensus and the chance corrected agreement on three cause of death classifications. The presence or absence of substandard factors (SSF) with the care provider, the organization of care and the relation of the SSF with perinatal death. Results. Consensus rates and chance corrected agreement for three cause of death classifications ranged from 92 to 96% and kappa 0.87 to kappa 0.93 (very good agreement), with comparable confidence intervals and similar values in the validation subset of 123 cases. On the presence of SSF at the level of the care provider consensus and chance corrected agreement was 68% and kappa 0.53 (moderate), with comparable values in the subset of 123 cases. Consensus for the relation between SSF at the level of the care provider and perinatal death was 81.4% and kappa 0.68 (good). Conclusion. Perinatal audit on a national level with relatively large audit groups with many different care providers is feasible.