An Internet "Value of Health" panel: recruitment, participation and compliance

OBJECTIVES To recruit a panel of members of the public to provide preferences in response to the needs of economic evaluators over the course of a year.

METHODS A sample of members of the UK general public was recruited in a stratified random sample from the electoral roll and familiarised with the standard gamble method of preference elicitation using an internet based tool. Recruitment (proportion of people approached who were trained), participation (defined as the proportion of people trained who provided any preferences) and compliance (defined as the proportion of preference tasks which were completed) were described. The influence of covariates on these outcomes were investigated using univariate and multivariate analyses.

RESULTS A panel of 112 people was recruited. The eventual panel reflected national demographics to some extent, but recruitment from areas of high socioeconomic deprivation and among ethnic minority communities was low. 23% of people who were approached (n= 5,320) responded to the invitation to take part in the study, and 24% of respondents (n=1,215) were willing to participate. However, eventual recruitment rates, following training, were low (2.1% of those approached), although significantly higher in Exeter than other cities. 18 sets of health state descriptions were presented to the panel over 14 months. 74% of panel members praticipated in at least one valuation task. Socioeconomic and marital status were significantly associated with participation. Compliance varied from 3% to 100%, with the average per set of health state descriptions being 41%. Compliance was higher in retired people but otherwise no significant predictors were identified.

CONCLUSIONS It is feasible to recruit and train a panel of members of the general public to express preferences on a wide range of health states using the internet in response to the needs of analysts. In order to provide a sample which reflects the demographics of the general public, and capitalise on the increasing opportunities for the use of the internet in this field, over-sampling in areas of high socioeconomic deprivation and among ethnic minority communities is necessary.