Comparing 15D valuation studies in Norway and Finland - Challenges when combining information from severel valuation tasks

Abstract

Background: The 15D is a generic preference-based health-related quality-of-life instrument developed in Finland. Values for the 15D instrument are estimated by combining responses to three distinct valuation tasks. The impact of how these tasks are combined is relatively unexplored. Objectives: To compare 15D valuation studies conducted in Norway and Finland in terms of scores assigned in the valuation tasks and resulting value algorithms, and to discuss the contributions of each task and the algorithm estimation procedure to observed differences. Methods: Norwegian and Finnish scores from the three valuation tasks were compared using independent samples t tests and Lin concordance correlation coefficients. Covariance between tasks was assessed using Pearson product-moment correlations. Norwegian and Finnish value algorithms were compared using concordance correlation coefficients, total ranges, and ranges for individual dimensions. Observed differences were assessed using minimal important difference. Results: Mean scores in the main valuation task were strikingly similar between the two countries, whereas the final value algorithms were less similar. The largest differences between Norway and Finland were observed for depression, vision, and mental function. Conclusions: 15D algorithms are a product of combining scores from three valuation tasks by use of methods involving multiplication. This procedure used to combine scores from the three tasks by multiplication serves to amplify variance from each task. From relatively similar responses in Norway and Finland, diverging value algorithms are created. We propose to simplify the 15D algorithm estimation procedure by using only one of the valuation tasks.