OBJECTIVES: Scope tests are used in discrete choice experiment (DCE) studies to examine whether survey respondents are recoding quantitative attribute levels from the absolute numeric values to qualitative levels, such as “low” or “high.” A sample “fails” a between-subject scope test when the utility difference, or importance, associated with two numeric attribute levels varies with the range of numeric levels in the experimental design. This study, aimed to systematically review the healthcare DCE literature, was conducted to determine how many DCE studies conducted scope tests; describe the design and statistical analysis of the test data; and understand how researchers interpret and use the results of scope tests.
METHODS: A literature review using a form of “pearl growing” was conducted to identify healthcare DCE studies using scope tests. A compilation of existing systematic reviews of DCEs in healthcare was used as an original source. Key terms were initially searched in the full texts of all articles covered by existing reviews to identify “initial pearls” to be grown.
RESULTS: The original source comprised 628 healthcare DCE studies published between 1990 and 2018. Searching for key terms yielded eight relevant results including two empirical examples of scope tests. Searching the reference lists and citations of the eight initial pearls yielded a further four empirical examples in subsequent waves. The final review comprised six empirical studies. Authors appear to be using scope tests for cost (n=3; 50%) and risk attributes (n=3; 50%). The scope test was reported as successfully passed in two studies, and failed/unreported in four studies.
CONCLUSIONS: There exists little literature on scope tests in DCE studies of preferences for health and healthcare and few papers reporting scope test failures. Consequently, there are few examples in the literature demonstrating how scope test failures may be interpreted and their implications for preference measurement.
