APPLICATION OF AN ALGORITHM FOR DEFINING RETROSPECTIVE COHORTS OF COLORECTAL CANCER (CRC) PATIENTS TREATED WITH DIFFERENT FIRST-LINE CHEMOTHERAPY REGIMENS +/- BEVACIZUMAB TO ADMINISTRATIVE CLAIMS DATA
Author(s)
Cisternas MG1, Mrad R2, Noe LL1, Miller DP1, Guyan C11Ovation Research Group, Highland Park, IL, USA; 2 F. Hoffmann-La Roche Ltd, Basel, Switzerland
Presentation Documents
OBJECTIVES: To apply a previously described algorithm for defining cohorts of CRC patients treated with first-line chemotherapy agents +/- the targeted therapy bevacizumab to an administrative health insurance claims database. METHODS: Claims records for 717 patients newly diagnosed for CRC in 2003 or 2004 who initiated chemotherapy during 2004 were extracted from a large U.S. healthcare claims database. By applying a previously defined algorithm (see “Development of an algorithm for the identification and classification of colorectal cancer (CRC) patients according to first-line chemotherapy +/- bevacizumab using administrative claims records”) these patients were classified according to five, mutually exclusive, first-line chemotherapy regimens +/- bevacizumab. RESULTS: Of the 717 patients identified, 709 (99%) could be assigned to one of the first-line chemotherapy categories: oxaliplatin (25%); irinotecan+bolus 5-FU/LV (2%); irinotecan+infusional 5-FU/LV (1%); irinotecan+5-FU/LV (bolus vs. infusion not distinguishable) (5%); and 5-FU/LV or capecitabine alone (66%). Each category was further subdivided according to whether bevacizumab was administered during the first month of chemotherapy. Of the 97 patients with a J9999 (“antineoplastic drug not elsewhere classified”) HCPCS claim in 2004, 12 had only one such claim and therefore bevacizumab vs. cetuxmab could not be identified; all but 2 of the remaining had their J9999 claims identified definitively. Seventy patients incurred a J9999 claim within their first-line therapy: assignment to bevacizumab, cetuximab, and unknown was 63, 1, and 6, respectively. Eighty patients (11%) received second-line therapy. CONCLUSIONS: First-line chemotherapies for CRC can be identified in health insurance claims data through a careful examination of CPT, HCPCS, and revenue center codes and the intervals between them. However, distinguishing bolus vs. infusion regimens is challenging due to inconsistent coding of ambulatory pump and IV push claims. Newly approved agents billed under a "not otherwise classified" code can be distinguished if their costs or intervals of administration differ substantially.
Conference/Value in Health Info
2005-11, ISPOR Europe 2005, Florence, Italy
Value in Health, Vol. 8, No.6 (November/December 2005)
Code
PCN39
Topic
Methodological & Statistical Research, Real World Data & Information Systems
Topic Subcategory
Health & Insurance Records Systems, Modeling and simulation
Disease
Oncology