AHRQ patient safety indicators: time to include hemorrhage and infection during childbirth

Abstract

Background: Many Agency for Healthcare Research and Quality (AHRQ) Patient Safety Indicators (PSIs) partially or completely exclude pregnant women. Both postoperative hemorrhage or hematoma (PSI 9; hemorrhage), and Postoperative Sepsis (PSI 13; infection) appear to be adaptable to pregnancy hospitalizations.

Methods: Using the 2009 California Patient Discharge Dataset (N [total] = 508,842), the hemorrhage and infection PSIs were examined for their potential to include pregnant women in gynecological, antepartum, postpartum, and delivery subpopulations. The statewide and hospital-level indicator rates were calculated using hierarchical models adjusted for case mix.

Results: Only the delivery population had sufficient cases for hospital-level analysis. Both PSIs required major changes to the technical specifications because of pregnancy-specific codes and coding practices. Nevertheless, these revised indicators identified substantial morbidity that varied widely across hospitals. The hemorrhage indicator rate was 2.50% (95% confidence interval [CI], 2.45-2.54) for all deliveries, compared with 0.26% (95% CI, 0.25-0.27) in the AHRQ population and 0.18% (95% CI, 0.15-0.21) for nonpregnant women of reproductive age. Adjusted hospital rates averaged 2.52%, with a midquartile range of 1.16% to 3.09% Although infection rates were lower for all deliveries than for the AHRQ population (0.18% versus 1.20%), they were highly associated with cesarean versus vaginal birth (0.43% versus 0.05%) and ranged from 0% to 1.15% across hospitals.

Conclusions: Although codes and coding practices for pregnancy hospitalizations differ from those used for nonpregnant adults, hospital-level measures of childbirth-associated hemorrhage and infection are feasible, vary widely, and demonstrate considerable opportunity for improvement.