Algorithm for identifying various second- and third-line chemotherapy regimens in elderly U.S. Medicare patients with metastatic colon cancer.

Abstract

553 Background: Patients with metastatic colon cancer (mCC) often receive multiple lines of chemotherapy treatment (TX) in response to disease progression or toxicities. A claims-based algorithm that identifies TX lines can provide information on “real world” clinical practice patterns that may not be captured by clinical trials. Methods: Our claims-based algorithm was applied to SEER-Medicare data of elderly mCC patients diagnosed in ‘03-‘07 and followed through ‘09. The algorithm included 17 clinical rules for identifying the beginning and end TX lines. The face validity of the algorithm was assessed by 1) examining the output against a TX map for a random sample of patients; 2) evaluating the overall results; and 3) conducting a sensitivity analysis, which evaluated the variability in the number of detected TX lines as a function of key algorithm parameters. Results: Of 7,951 mCC patients, 3,266 (41%) received TX; 1,440 (18% of all, 44% of TX) and 274 (3% of all, 8% of TX) received 2nd and 3rd line TX, respectively. Fewer than 1% of treated patients had a 4th TX line. The utilization patterns in terms of number and type of TX lines were robust to changes in the algorithm parameters. OX±BEV (45%), 5FU/LV±BEV (33%) and IRI±BEV (16%) were the three most common initial TXs. 2nd line TX most commonly consisted of IRI±BIOLOGIC (62%) and OX±BIOLOGIC (26%), but 6% of patients received only BIOLOGICS. CETUX (19%), PANIT (15%), IRI alone (17%) and OX alone (12%) were the most common 3rd line TXs. OX to IRI (49%), IRI to OX (14%), 5FU/LV to OX (12%), and 5FU/LV to IRI (12%) were the most frequent TX progressions for those with 2nd line TX. 5FU/LV to OX to IRI (26%), OX to IRI to BIOLOGICS alone (25%), 5FU/LV to IRI to OX (14%) and IRI to OX to BIOLOGICS alone (6%) were the most frequent TX progressions for those with 3rd line TX. Conclusions: Our claims-based algorithm suggests that during 2003-2009 relatively few elderly mCC patients received 2nd and 3rd line TX. Sensitivity analysis confirmed the robustness of the algorithm. Future observational studies should address the “real world” benefits and risks of 2nd and 3rd line TX among elderly mCC patients.