Condition-specific Measure was More Responsive than Generic Measure in Colorectal Cancer: All but Social Domains

Abstract

Objective: To examine the responsiveness of generic and condition-specific instruments based on the anchor of self-reported level of global change in patients with colorectal cancer (CRC). Study Design and Setting: Three hundred thirty-three patients with CRC were surveyed at two assessments at baseline and follow-up at 6 months from September 2009 to July 2010 using the Short Form-12 Health Survey version 2 (SF-12v2) and Functional Assessment of Cancer Therapy-Colorectal (FACT-C) measures. The responsiveness of the two measures was evaluated using standardized effect size, standardized response mean, responsiveness statistic, and receiver operating characteristic (ROC) curve analysis. Results: In worsened group, internal responsiveness of detecting negative changes was satisfactory for most subscales of FACT-C and SF-12v2. The FACT-C subscales were significantly more responsive to positive changes detection than the SF-12v2 subscales in improved group. Physical well-being subscale, Trial Outcome Index (TOI), and total score of FACT-C were more externally responsive to ROC curve analysis. The FACT-C measure was generally more responsive to changes in health status compared with SF-12v2 measure. Conclusion: TOI and total score of FACT-C were the most responsive among subscales of condition-specific measure, which were more responsive than all generic subscales with the exception of social domain. Complementary use of condition-specific and generic instruments to evaluate the health-related quality of life of CRC patients is encouraged.