Background: The main measure to generate utility data for economic evaluations is the EQ-5D, but no study has tested whether or how to map from palliative care measures to the EQ-5D.
Aims: To assess the level of conceptual overlap between palliative outcomes and the EQ-5D, and the feasibility of mapping between them to obtain utilities for the Palliative care Outcome Scale.
Design: A cross-sectional secondary analysis of data from three studies.
Setting/participants: Patients receiving palliative care and bereaved relatives, recruited from three tertiary National Health Service hospitals in South London.
Methods: The overlap between both measures was assessed using principal component analysis. The Palliative care Outcome Scale was mapped onto the EQ-5D using three regression models.
Results: Spearman's correlations between both instruments were low (mean rho = 0.11). The principal component analysis showed the Palliative care Outcome Scale is associated with only two EQ-5D dimensions (pain; and anxiety/depression). No Palliative care Outcome Scale items loaded onto the mobility, self-care and usual activities dimensions of the EQ-5D. The mapping models performed poorly at predicting utilities from Palliative care Outcome Scale data (mean absolute error >0.3 and R2 <0.10). Hence, none of the models can be recommended as acceptable for calculating utilities from Palliative care Outcome Scale responses.
Conclusion: Differences between the Palliative care Outcome Scale and the EQ-5D do not undermine the qualities of either instrument when used for their own purposes. However, due to conceptual differences, the EQ-5D does not capture some of the concerns measured by the Palliative care Outcome Scale, and therefore, mapping onto the EQ-5D is unlikely to provide an appropriate basis for estimating utilities for conducting economic evaluations in palliative care studies.
Keywords: EQ-5D; health-state utility; modelling; palliative care outcome scale; patient-reported outcomes.