Validation of the accuracy of postpartum haemorrhage data in the ObstetriX database : a retrospective cohort study

Background: Data related to postpartum haemorrhage (PPH) are important clinical parameters which can be applied to all places of birth, and their recording can be missed by busy clinicians providing critical care to women. We compared the accuracy of electronic ObstetriX records to the paper-based medical records of the women who sustained PPH. Methods: In this retrospective cohort study over a period of one month, 363 electronic records were compared to the paper-based medical records. The volume of blood loss for each patient and interventions for PPH were compared across birth unit, operating theatre and postpartum ward. The kappa statistic for agreement between the two types of recording methods was calculated. Results: There was substantial agreement between the ObstetriX records and medical records for the volume of blood loss at birth (kappa = 0.74), but poor agreement between records for the cumulative total volume of blood loss (kappa = 0.18). More women who experienced PPH and delivered in the operating theatre had errors in their ObstetriX records compared to women who had PPH with births in the birth unit (50% vs 16%; n = 73, P = 0.005). Interventions for PPH were found to be poorly recorded in ObstetriX, with 84% (n = 64/76) of women who experienced PPH having none of the interventions they received recorded. Conclusions: The ObstetriX database was not a generally reliable source of data relating to PPH. However, some data were recorded reliably, in particular, the volume of blood loss at birth.