OBJECTIVES:

Within the health economics literature, there is debate about the appropriateness of using patient versus general population samples in decisions about the allocation of health resources. An assumption behind this debate, is the existence of fundamental differences between patient and general population samples. In this study, potential differences between these two samples are explored through a discrete choice experiment (DCE) used to understand what features are important to include in an assessment for chemotherapy induced peripheral neuropathy (CIPN). This study also investigated whether providing additional information to the general population sample could mitigate any differences between a general population and patient sample.

METHODS:

In the survey, respondents completed 8 choice sets each. In each choice set, respondents were asked to choose between two hypothetical assessment tools. A patient and general population sample were recruited. The general population was split into two arms, with Arm 1 receiving the same introductory information as the patient sample. The Arm 2 received extra information in the form of moving GIFs and a short video about CIPN. The scaled-multinomial logit model was used to test for differences between Arms 1 and 2 and to compare Arms 1 and 2 to the patient sample.

RESULTS:

The general population sample consisted of 167 respondents in Arm 1 and 168 respondents in Arm 2. There were 117 respondents in the patient sample. The extra information in Arm 2 improved the general population sample’s understanding of the DCE without affecting preferences. Significant scale differences were detected between the general population arms and the patient sample.

CONCLUSIONS:

Although providing extra information to a general population sample was found to improve understanding of the choice sets without affecting choices made, it was not enough to mitigate scale differences with a patient sample. This has implications on evidence used in health decision making.