Development and evaluation of an algorithm to identify users of Prolia® during the early postmarketing period using health insurance claims data. (23rd July 2014)

• Record Type: Journal Article
• Authors: Hoffman, Veena; Xue, Fei; Gardstein, Betsey; Skerry, Kathleen; Critchlow, Cathy W.; Enger, Cheryl
• Is Part Of: Pharmacoepidemiology and drug safety. Volume 23:Number 9(2014:Sep.)
• Page Start: 993
• Page End: 998
• Publication Date: 2014-07-23
• Subjects: Pharmacoepidemiology -- Periodicals; Chemotherapy -- Periodicals; Epidemiology -- Periodicals; 615.705
• DOI: 10.1002/pds.3680
• Languages: English
• ISSNs: 1053-8569

Abstract

<abstract abstract-type="main"> <title>ABSTRACT</title> <sec id="pds3680-sec-0001" sec-type="section"> <title>Purpose</title> <p>The goal of this study is to develop and validate an algorithm to identify Prolia<sup>®</sup> users within a health insurance claims database.</p> </sec> <sec id="pds3680-sec-0002" sec-type="section"> <title>Methods</title> <p>Patients with a denosumab‐specific or nonspecific administration claim during the early period of Prolia availability in the USA (June 1, 2010 to March 31, 2012) were classified as definite, probable, possible, and nonusers of Prolia using an algorithm consisting of nine different components based on claims patterns consistent with Prolia use. Medical record review confirmed a sample of definite, probable, and possible users and the positive predictive value (PPV) was estimated.</p> </sec> <sec id="pds3680-sec-0003" sec-type="section"> <title>Results</title> <p>The PPV of the claims‐based algorithm components varied (17.8–95.8%). Requiring claims for a bone or cartilage disorder or osteoporotic fracture after excluding claims for cancer prior to a denosumab‐specific administration code gave the highest PPV (95.8%), followed by requiring a Prolia National Drug Code on the same claim as a denosumab‐specific or nonspecific administration code (88.2%). Among the 87 confirmed Prolia users, osteoporosis diagnoses were seen more frequently in the medical record than in claims (83% vs 62%).</p> </sec> <sec id="pds3680-sec-0004" sec-type="section"> <title>Conclusions</title> <p>Prolia users are most accurately identified with administration code claims in conjunction with claims for Prolia National Drug Code and bone disorder treatment and diagnosis codes. Osteoporosis diagnoses may be under‐recorded in claims data. The algorithm may require reassessment as uptake for more recently approved indications increases. Copyright © 2014 John Wiley & Sons, Ltd.</p> </sec> </abstract>