Kristin M Sheffield1, Lee Bowman1, David M Smith2, Li Li1, Lisa M Hess1, Leslie B Montejano2, Tina M Willson2, Amy J Davidoff3. 1. Global Patient Outcomes & Real World Evidence, Eli Lilly & Company, Lilly Corporate Center, Indianapolis, IN 46285, USA. 2. Outcomes Research, Truven Health Analytics, an IBM Company, 100 Phoenix Drive, Ann Arbor, MI 48108, USA. 3. Department of Health Policy and Management, Yale School of Public Health, 60 College Street, PO Box 208034, New Haven, CT 06510, USA.
Abstract
AIM: To develop a claims-based prediction model of poor performance status (PS) in commercially insured and Medicare supplemental beneficiaries with cancer. PATIENTS & METHODS: Retrospective analysis was conducted of electronic medical records (EMR) from community oncology practices linked to MarketScan claims. Multivariable logistic regression predicted PS scores from the EMR using claims-based diagnostic and procedure codes. RESULTS: The study included 8442 patients diagnosed with cancer from 2007 to 2015. Overall, 8.1% of patients had poor EMR-based PS. Bootstrapping results from the final model showed sensitivity and specificity of approximately 75% with a predicted probability cutpoint = 0.078, c-statistic = 0.821 and pseudo-R2 = 0.25. CONCLUSION: Patients with poor PS can be identified in claims data. This prediction model enables future studies evaluating cancer treatments and outcomes to account for PS.
AIM: To develop a claims-based prediction model of poor performance status (PS) in commercially insured and Medicare supplemental beneficiaries with cancer. PATIENTS & METHODS: Retrospective analysis was conducted of electronic medical records (EMR) from community oncology practices linked to MarketScan claims. Multivariable logistic regression predicted PS scores from the EMR using claims-based diagnostic and procedure codes. RESULTS: The study included 8442 patients diagnosed with cancer from 2007 to 2015. Overall, 8.1% of patients had poor EMR-based PS. Bootstrapping results from the final model showed sensitivity and specificity of approximately 75% with a predicted probability cutpoint = 0.078, c-statistic = 0.821 and pseudo-R2 = 0.25. CONCLUSION:Patients with poor PS can be identified in claims data. This prediction model enables future studies evaluating cancer treatments and outcomes to account for PS.
Authors: Lisa M Hess; David Smith; Zhanglin L Cui; Leslie Montejano; Astra M Liepa; William Schelman; Lee Bowman Journal: J Drug Assess Date: 2020-12-16