Mapping analyses to estimate health utilities based on responses to the OM8-30 otitis media questionnaire
Dakin H., Petrou S., Haggard M., Benge S., Williamson I.
To investigate the statistical relationship between the OM8-30 health-related quality of life measure for children with otitis media with effusion (OME) and measures of health utility (Health Utilities Index [HUI] Mark 3 and Mark 2) and to develop models to estimate HUI3 and HUI2 health utilities from OM8-30 scores. A placebo-controlled, randomised trial (GNOME) evaluating intranasal mometasone in 217 children with OME provided concurrent responses to OM8-30 and HUI at three time points. Ordinary least squares (OLS), generalised linear models and two-step regression analyses were used to predict HUI3 and HUI2 utilities based on OM8-30 facet and domain scores. OLS models including all nine OM8-30 facets with or without predicted hearing level (HL) produced the best predictions of HUI3 utilities (mean absolute error: 0.134 with HL and 0.132 without; R (2): 0.63 with HL and 0.596 without). An OLS model predicting HUI3 utilities based on the two OM8-30 domain scores, reported hearing difficulties, predicted HL, age and sex also produced accurate predictions. Regression equations predicting HUI3 and HUI2 utilities based on OM8-30 facet and domain scores have been developed. These provide an empirical basis for estimating quality-adjusted life years (QALYs) for interventions in children with OME.