Compliance with adjuvant capecitabine in patients with stage II and III colon cancer: comparison of administrative versus medical record data

We aimed to examine the frequency of treatment delays as well as the reasons and appropriateness of such delays in early stage colon cancer patients receiving adjuvant capecitabine by comparing data from pharmacy dispensing versus medical records. Patients diagnosed with stage II or III colon cancer from 2008 to 2012 and who received at least two cycle of adjuvant capecitabine were reviewed for treatment delays. Data from pharmacy dispensing and patient medical records were compared. Multivariate regression models were constructed to identify predictors of treatment delays. A total of 697 patients were analyzed: median age was 70 years (IQR 30–89), 394 (57%) were men, 598 (86%) reported Eastern Cooperative Oncology Group 0/1, and 191 (27%) had stage II disease. In this study cohort, 396 (57%) patients experienced at least 1 treatment delay during their adjuvant treatment. Upon medical record review, half of treatment delays identified using pharmacy administrative data were actually attributable to side effects, of which over 90% were considered clinically appropriate for patients to withhold rather than to continue the drug. The most prevalent side effects were hand‐foot syndrome and diarrhea which occurred in 176 (44%) and 67 (17%) patients, respectively. Multivariate analysis revealed a statistically significant association between stage and inappropriate treatment delays whereby patients with stage II disease were more likely to experience drug noncompliance (OR 1.79, 95% CI: 1.27–2.53, P < 0.001) than those with stage III disease. Compliance with adjuvant capecitabine was reasonable. Adherence ascertained from pharmacy administrative data differs significantly from that obtained from medical records.