White Rose University Consortium logo
University of Leeds logo University of Sheffield logo York University logo

Estimating health state utility values for comorbid health conditions using SF-6D data

Ara, R. and Brazier, J. (2010) Estimating health state utility values for comorbid health conditions using SF-6D data. (Unpublished)

Full text available as:
[img] Text
WHS_combining_data.pdf

Download (212Kb)

Abstract

BACKGROUND: When health state utility values for comorbid health conditions are not available, analysts frequently use data from cohorts with single health conditions to estimate proxy scores. The methods used can produce very different results and there is currently no consensus on which is the most appropriate approach. OBJECTIVE: The objective of the current study was to assess the accuracy of five different methods that have been used to estimated HSUVs for comorbid health conditions. METHOD: Data collected during five Welsh Health Surveys (WHS) were subgrouped by health status. Mean SF-6D scores from cohorts with a particular health condition were used to estimate mean SF-6D scores for cohorts with two comorbid health conditions using:the additive, multiplicative, and minimum methods, and the adjusted decrement estimator. A linear model was obtained by regressing mean HSUV from subgroups with single health conditions onto mean HSUVs from subgroups with combined health conditions. RESULTS: The pooled WHS data provided 64,437 cases with SF-6D scores. When subgrouped by self reported health condition(s), 32 groups (n>30) were identified with comorbid health conditions. The mean SF-6D for these subgroups ranged from 0.4648 to 0.6068. The linear model produced the most accurate HSUVs for the combined health conditions with 88% of values accurate to within the minimum important difference for the SF-6D. The additive method underestimated the actual SF-6D scores and produced some substantial errors in the estimated values. The minimum method overestimated all mean SF-6D scores but was more accurate when estimating higher values. The multiplicative and ADE methods both underestimated the majority of the actual SF-6D scores. However, both methods both performed better when estimating SF-6D scores smaller than 0.50 with 43% and 86% of estimated HSUVs accurate to within the MID for the multiplicative and ADE respectively. This study makes an important contribution to the existing evidence as it is the first to compare five different methods on SF-6D data. Although the range in actual HSUVs was relatively small, the data covered the lower end of the index while the majority of previous research has involved actual HSUVs covering the upper end of possible ranges. While the linear model gave the most accurate results in our data, additional research is required to develop and validate the model.

Item Type: Article
Keywords: health state utility values, comorbidities, quality of life, SF-6D
Institution: The University of Sheffield
Academic Units: The University of Sheffield > Faculty of Medicine, Dentistry and Health (Sheffield) > School of Health and Related Research (Sheffield) > Health Economics and Decision Science > HEDS Discussion Paper Series
Depositing User: Mrs Liz Metham
Date Deposited: 21 Jul 2010 13:12
Last Modified: 04 Jun 2014 17:15
Status: Unpublished
Refereed: No
URI: http://eprints.whiterose.ac.uk/id/eprint/11049

Actions (login required)

View Item View Item