Reducing overuse of colony-stimulating factors without compromising patient safety: Evidence from a decision-support enabled utilization management in patients with lung cancer receiving chemotherapy.

Abstract

52 Background: Recent studies suggest significant overuse of colony-stimulating factors (CSFs) for the primary prevention of febrile neutropenia (FN) in patients with cancer receiving chemotherapy. Overtreatment with CSF increases costs without evidence of clinical benefit. We examined the impact of a decision-support tool to promote risk-appropriate CSF use in lung cancer. Methods: A retrospective cohort study design was used to analyze commercial claims data on lung cancer patients. The decision-support tool was implemented at participating oncology practices in staggered phases across 14 U.S states (from July 1, 2014 - November 1, 2014). Study population included adults with lung cancer who initiated chemotherapy from July 1, 2014 through March 30, 2015. Patients were assigned to case and control cohorts according to whether or not they resided in service areas where the tool had been implemented. Patients were followed up to 6 months after initiating chemotherapy in the pre- and post-implementation periods. The primary outcomes: CSF use and FN incidence rates were compared using difference-in-differences (DID) models; adjusting for baseline FN risk factors with generalized estimating equations. Results: The final study population included 3,470 patients (case: 1,857 and control: 1,613; overall mean age (SD), years: 65 (10)). There were no meaningful differences in FN risk factors at baseline between the cohorts. In adjusted results, CSF use decreased from 48.4% to 35.6% in the case compared to a change from 43.2% to 44.4% in control cohort (DID: -8.7% (95% CI: -14.65% to -2.67%), p - < 0.001) in the pre- and post-implementation periods. The rates of FN were consistent for both cohorts in each time period, with no statistical difference in trend for the case (2.8% to 4.3%) as compared with the control (3.1% to 5.1%) cohort (DID: -0.2% (95% CI: -0.63% to -0.33%), p – 0.953). Conclusions: These findings demonstrate the potential for decision-support tools to improve evidence-based guideline adherence and promote risk-appropriate CSF use in patients with lung cancer without adversely impacting patient safety.