2018 Conference Presentation
Objective: Stated preference studies are used in health research to understand the relative value that people place on different health-related quality of life (QoL) states captured by multi-attribute measures to make them suitable for cost-effectiveness analysis. Preferences may differ due to differences in care systems, cultures, beliefs or politics. Hence, when valuing multi-attribute measures country-specific preferences tend to be used.
Few studies, however, have directly compared differences in preferences for health-related QoL states across countries and none have investigated differences in social care-related quality of life (SCRQoL) states, such as those described by the Adult Social Care Outcomes Toolkit (ASCOT-SCT4) – a measure more suitable for assessing outcomes of long-term care users. The current study aimed to compare preferences for different SCRQoL states in three European countries (Austria, England and Finland) using Best-Worst Scaling (BWS) to elicit preferences.
Data and Methods: Data was collected from a representative sample of the general adult population in Austria (n=1,000), England (n=1,001) and Finland (n=1,008). Participants completed an online survey, which included a BWS experiment. The BWS experiment used an orthogonal main effects plan to select tasks. This produced an experiment consisting of 32 tasks, which were blocked into four segments (8 tasks per participant). The online survey also includes questions to capture participant’s socio-demographic, socio-economic characteristics and information about the quality of life.
A pooled multinomial logit regression (MNL) model was used to assess any differences in preferences between the three countries. Covariates describing observable population characteristics were also introduced into the model to test and control for differences in preferences for attributes and attribute levels between the three countries. The models also controlled for unobservable differences in variance (scale) across the countries, using a scale heterogeneity MNL model.
Results and conclusions: Estimation results from the MNL models will be presented and discussed. We will highlight whether there are significant differences in preferences for SCRQoL states across the three countries when controlling for observed sample differences in socio-economic and demographic characteristics, as well as shared unobservable heterogeneity. These findings and their implications for long-term care research, including cost-effectiveness analysis, will also be explored and discussed.