1696. The Black Box of Using Health Claims-Based Analyses to Estimate UTI Prevalence in Community-Dwelling Women

BACKGROUND: Urinary tract infections (UTIs) are common in women but most epidemiology studies occurred in specialized settings (university health clinics) or used outdated methods (random digit dialing). Currently, women receive UTI care in systems with electronic health records (EHR), thus documenting care of a wider female demographic in real-world settings. We estimated the prevalence of acute, uncomplicated UTIs in community-dwelling women in a health claims database using various operational definitions of UTI. METHODS: We conducted a retrospective analysis of claims data from the OptumLabs® Data Warehouse (OLDW), a de-identified claims and clinical information repository for privately insured and Medicare Advantage enrollees in a large, private US health plan. Non-pregnant female patients ≥ 15 years of age with two years of continuous enrollment between 2007-2015 and a visit encounter in an outpatient office, urgent care, or emergency department were included. Women with lower urinary tract disease/abnormalities, neurological disease, urological treatment, procedures or urinary catheter use, cancer or HIV treatment were excluded. Decision rules for identifying UTIs were derived using one or more combinations of: relevant ICD-9 codes, UTI symptom diagnosis codes, positive urine test results, and/or antibiotic prescription recorded in the EHR and claims. Prevalence rates were calculated for each decision rule. RESULTS: Of the 7,337,700 females in the claims database, 947,041 (12.97%) had an index UTI diagnosis or symptoms and met eligibility criteria. The table below illustrates prevalence rates according to each decision rule. As shown, applying decision rules based on common UTI definitions resulted in large differences in prevalence rates. Table [Image: see text] CONCLUSION: Using common definitions for UTI to analyze claims data from an insurer of large proportions of the US, we obtained significantly different prevalence rates. This study highlights major limitations in using EHR and claims data for UTI quality initiatives such as tracking of practices associated with antimicrobial stewardship and lends credibility to proposals to track these infections as a reportable disease. DISCLOSURES: All Authors: No reported disclosures