LAUSR

253 Quality of administrative data in measuring hemorrhage-related morbidity Dena Goffman, Matthew Oberhardt, David Vawdrey, Alan Kessler, Robert Green, Jean Ju Sheen, Mary D’Alton, Alexander Friedman Columbia University Irving Medical Center, New York, NY, New York Presbyterian Hospital, New York, NY, Weill Cornell Medicine, New York, NY OBJECTIVE: Postpartum hemorrhage (PPH) is a leading cause of severe maternal morbidity (SMM). SMM related to PPH such as disseminated intravascular coagulation (DIC), transfusion, and acute renal failure (ARF) can be readily measured with billing diagnosis and procedure codes. Given that the validity of these codes relative to chart data is unknown, the purpose of this study was to assess hemorrhage-related morbidity codes in an obstetric population. STUDY DESIGN: Delivery hospitalizations between 7/2014 and 7/2017 in three hospitals were analyzed. Data was obtained by querying the electronic medical record (EMR) including laboratory values and blood bank documentation. We evaluated whether billing codes (ICD-9 and ICD-10) used by the CDC to measure SMM aligned with EMR documentation for (i) blood transfusion, (ii) ARF, and (iii) DIC. The accuracy of blood transfusion diagnosis codes was assessed based on whether blood bank documentation demonstrated transfusion. The accuracy of ARF diagnosis codes was assessed based on whether laboratory values demonstrated serum creatinine 1.2mg/dL. The accuracy of DIC diagnosis codes was assessed based on whether laboratory values demonstrated serum fibrinogen at the following cutoffs: 200, 150, and 100 mg/dL. Test characteristics were calculated including sensitivity, specificity, positive predictive value (PPV), and negative predictive values with corresponding 95% confidence intervals. RESULTS: 35,518 deliveries were analyzed. 786 women underwent transfusion, 168 had serum creatinine 1.2mg/dL, and 99, 40, and 16 had fibrinogen 200, 150, and 100 mg/dL, respectively. Transfusion codes were 65.4% sensitive and 99.9% specific with a 91.3% PPV (Table). DIC codes were 56.3% sensitive, 99.6% specific with a PPV of 6.3% for fibrinogen 100mg/dL, 35.0% sensitive 99.6% specific with a PPVof 9.8% for a fibrinogen 150mg/dL, and 22.2% sensitive, 99.7% specific with a PPV of 15.4% for fibrinogen 200mg/dL. ARF codes were 33.3% sensitive with a PPV of 62.9% and specificity of 99.7%. CONCLUSION: ICD 9 and 10 codes were of poor to moderate quality in measuring diagnoses commonly used to assess PPH-related morbidity. Many PPH-related outcomes can be readily queried from the EMR. For hospitals analyzing quality and safety outcomes related to PPH, data directly ascertained from the clinical record is more valid than administrative codes.