///2012 Abstract Details
2012 Abstract Details2019-08-02T19:38:42-06:00

The incidence of ICD-coded postpartum hemorrhage may significantly underestimate the true number of cases

Abstract Number: T-9
Abstract Type: Original Research

Abelina D'Costa MD1 ; Rachel Martin MD2; Naveed Siddiqui MD3; Kristi Downey MSc4; Gareth Seaward MD5; Jose C.A. Carvalho MD6

Introduction: The incidence of postpartum hemorrhage (PPH) is reportedly increasing in North America, primarily due to uterine atony(1,2), but no convincing explanation for such an increase has been offered. It is possible that these increasing numbers may reflect inconsistency in the diagnostic criteria, coding and reporting of PPH(2). The objective of this study was to determine whether the ICD-coded PPH cases at our institution actually reflect the true incidence of PPH as per current diagnostic criteria.

Methods: This was a retrospective case-control study. We reviewed patient records from June 29, 2009 to June 28, 2010. The cases were identified by Medical Records and included all patients with ICD codes 072.002, 072.102, 072.202, 072.302. The controls were the 3 deliveries subsequent to each case, irrespective of the mode of delivery. All cases underwent a full chart review. All controls underwent an electronic review of their hemoglobin levels on post partum day 1. Patients with hemoglobin levels < 100 g/L had their blood loss calculated based on the difference between pre- and postpartum hemoglobin; those showing > 500mL for vaginal delivery and > 1000mL for CS were considered positive controls and underwent a full chart review. All cases and positive controls were confronted with the following diagnostic criteria: calculated blood loss > 500 mL for vaginal delivery or > 1000 mL for CS; use of 2 or more uterotonics; hemodynamic instability; blood transfusion. Patient demographics, ante- and intrapartum risk factors for PPH, estimated and calculated blood loss, pharmacological and other interventions, and complications were recorded for all cases and positive controls.

Results: There were 6,552 deliveries in the study period. Two hundred and forty five (245) patients were coded as PPH, of which 47 (19.2%) had no features in their medical records to suggest PPH; only 198 cases were accurately coded. Out of the 735 controls, 81 (11%) should have been coded as PPH based on the diagnostic criteria. The incidence of PPH at our institution based on ICD coding was 245/6552 (3.7%). Assuming that the incidence of undiagnosed/uncoded PPH (11%) would prevail in the entire cohort of patients that were not diagnosed/coded as PPH (6,552-245=6,307) the true incidence of PPH at our hospital could be as high as 198+693/6552 (13.6%). The incidence of blood transfusion was 18/6552 (0.3%).

Discussion: If the current diagnostic criteria for PPH were strictly followed, the incidence of ICD coded PPH should be substantially higher. Our results suggest that before we assume that an increase in the incidence of PPH is occurring, there is a need to refine the diagnostic criteria for PPH, and most importantly, define strategies to capture all PPH cases and code them correctly. The increase in PPH may solely reflect changes in the diagnosis, coding and reporting of this condition.

References: 1) Am J Obstet Gynecol 2010; 202:353; 2) BJOG 2007; 114:751

SOAP 2012