Assessing disability weights based on the responses of 30,660 people from four European countries
Haagsma, Juanita A
Maertens de Noordhout, Charline
Havelaar, Arie H
Kretzschmar, Mirjam E
MetadataShow full item record
CitationHaagsma, Juanita A, Charline Maertens de Noordhout, Suzanne Polinder, Theo Vos, Arie H Havelaar, Alessandro Cassini, Brecht Devleesschauwer, Mirjam E Kretzschmar, Niko Speybroeck, and Joshua A Salomon. 2015. “Assessing disability weights based on the responses of 30,660 people from four European countries.” Population Health Metrics 13 (1): 10. doi:10.1186/s12963-015-0042-4. http://dx.doi.org/10.1186/s12963-015-0042-4.
AbstractBackground: In calculations of burden of disease using disability-adjusted life years, disability weights are needed to quantify health losses relating to non-fatal outcomes, expressed as years lived with disability. In 2012 a new set of global disability weights was published for the Global Burden of Disease 2010 (GBD 2010) study. That study suggested that comparative assessments of different health outcomes are broadly similar across settings, but the significance of this conclusion has been debated. The aim of the present study was to estimate disability weights for Europe for a set of 255 health states, including 43 new health states, by replicating the GBD 2010 Disability Weights Measurement study among representative population samples from four European countries. Methods: For the assessment of disability weights for Europe we applied the GBD 2010 disability weights measurement approach in web-based sample surveys in Hungary, Italy, Netherlands, and Sweden. The survey included paired comparisons (PC) and population health equivalence questions (PHE) formulated as discrete choices. Probit regression analysis was used to estimate cardinal values from PC responses. To locate results onto the 0-to-1 disability weight scale, we assessed the feasibility of using the GBD 2010 scaling approach based on PHE questions, as well as an alternative approach using non-parametric regression. Results: In total, 30,660 respondents participated in the survey. Comparison of the probit regression results from the PC responses for each country indicated high linear correlations between countries. The PHE data had high levels of measurement error in these general population samples, which compromises the ability to infer ratio-scaled values from discrete choice responses. Using the non-parametric regression approach as an alternative rescaling procedure, the set of disability weights were bounded by distance vision mild impairment and anemia with the lowest weight (0.004) and severe multiple sclerosis with the highest weight (0.677). Conclusions: PC assessments of health outcomes in this study resulted in estimates that were highly correlated across four European countries. Assessment of the feasibility of rescaling based on a discrete choice formulation of the PHE question indicated that this approach may not be suitable for use in a web-based survey of the general population. Electronic supplementary material The online version of this article (doi:10.1186/s12963-015-0042-4) contains supplementary material, which is available to authorized users.
Citable link to this pagehttp://nrs.harvard.edu/urn-3:HUL.InstRepos:24983992
- SPH Scholarly Articles