Published utilities for the same health condition vary across studies. We investigated the factors related to utilities using the case of colorectal cancer for which we ran a meta-regression on the utilities of colorectal cancer derived from the relev...
Published utilities for the same health condition vary across studies. We investigated the factors related to utilities using the case of colorectal cancer for which we ran a meta-regression on the utilities of colorectal cancer derived from the relevant literature. We searched the literature published until December, 2010 in Medline, Science Direct, CINAHL, EMBASE, and KoreaMed from which 88 studies were selected. Among the 88 studies, we finally chose 16 studies from which 228 utility scores were retrieved to be included in the meta-regression. For each of the 228 utilities, information was recorded on its cancer stage, cancer type, cancer treatment, adverse reaction, remission, definition of the lower bound, definition of the upper bound, respondent, preference elicitation method, source of utility, and survey method. Fixed effect model was used to control for the correlations within the same study. We found that in the case of colorectal cancer, utilities were affected by metastatic stage, the type of colorectal cancer, adjuvant therapies, colostomy, preference elicitation method, and respondent, which suggests that in practice we should be careful in choosing utilities for economic evaluations and clinical decisions because several characteristics of health condition and utility measurement methods may affect the level of utilities.