Measuring Preferences for Colorectal Cancer Screening: What are the Implications for Moving Forward?

Abstract

There is a growing interest in the application of preferences to inform healthcare planning and delivery. Clinical practice guidelines are encouraging incorporation of preferences in patient management choices in recognition that often no single approach is best.The objective of this focused review is to provide an overview of the current state of preference measurement for colorectal cancer screening (CRCS) and highlight the implications for health policy, CRCS program implementation, and further research.MEDLINE and EMBASE electronic databases were searched (1990-May 2009) for English-language literature examining patient preferences for CRCS, using conjoint analysis methods. We systematically extracted information on the study population, whether the choice sets were framed around specific CRCS tests or the overall program, the attributes and levels included, and, where available, the ordering of importance of the attributes and key study findings.Qualitative data synthesis of key differences and commonalities in the approaches and findings are presented. Six conjoint analysis studies of CRCS were identified. While 66-88% of respondents in the general population indicated they would choose CRCS, this is greater than observed rates of uptake (40-50%). All studies were administered in a sample of the general population at average risk of CRC, except one that included a sample of physicians. The studies varied in the attributes and levels they included. However, accuracy, whether expressed in the context of a CRCS test or program, was consistently identified as a statistically significant and important attribute. Other attributes included in the conjoint analysis studies included level of discomfort during the test, preparation for the test, the testing process, frequency of testing, frequency of complications, the process of follow up, and cost.Our results suggested that (i) a majority of people would choose to be screened for CRC, although actual CRCS participation rates suggest otherwise; (ii) patients have distinct preferences for CRCS tests that can be linked to CRC test attributes; and consequently, (iii) there is no single CRCS test that is preferred by everyone. In addition, although the specific approach, attributes, and levels in the studies varied, they consistently found that multiple factors are important from the patient's perspective and that preferences vary amongst subgroups. Consequently, careful consideration should be given to the design and implementation of a CRCS program based on a broader range of factors than the traditional outcomes such as mortality and incidence reduction. Attention should now be focused on how to use this information to inform health policy and develop CRCS programs that will increase screening uptake and adherence in the context of the health system in which the program will be provided. We propose a two-step process to designing and implementing a CRCS program based on evidence and preferences that informs patient choice.