Comparing Treatment Options for Uterine Fibroids: An EHR-based Algorithm for Registry Recruitment [12H

Abstract

INTRODUCTION: COMPARE-UF is a national registry of women with uterine fibroids (UF) that will examine the effectiveness of different fibroid treatments on patient-centered outcomes. Electronic health records (EHR) may be used for efficient identification of eligible women within healthcare systems. An EHR-based algorithm was developed to identify women with symptomatic UF for recruitment into the PCORI-supported COMPARE-UF registry. METHODS: An iterative process was undertaken with a goal to maximize the positive predictive value (PPV) of the final algorithm. For the first algorithm, women were required to be age 18-54, have imaging confirming UFs or a diagnosis code for UF and no history of hysterectomy. The second algorithm required both imaging and at least 2 diagnoses codes for UF. The third algorithm required at least 2 diagnosis codes for UF on separate dates, and excluded patients who had UF detected during a prenatal or emergency department visits. Randomly selected charts were reviewed for each algorithm. RESULTS: The first algorithm identified 4,342 patients, with a PPV of 47% (95% CI: 39-56%) after 150 charts were reviewed with ultrasound evidence of UFs in 139 (93%). The second algorithm yielded 1,174 patients, with a PPV of 65% (95% CI: 50-79%) after reviewing 51 charts. The third algorithm yielded 465 patients, with a PPV of 76% (95% CI: 71-81%) after 300 chart reviews. CONCLUSION: An EHR algorithm can identify patients with UFs. However symptomatic UF cases can be challenging to distinguish from asymptomatic cases and require additional steps to confirm eligibility.