Estimating the effect of changes in body mass index on health state preferences

To investigate the effect of changes in body mass index (BMI) on health state preferences (HSP).

Multiple regression analysis on pooled data from a clinical study, predicting final HSP as a function of changes in BMI and initial HSP, controlling for age and gender. Subgroup analyses according to clinically relevant subgroups were performed.

Primary care practice sites.

402 obese patients of varying disease severity.

BMI was operationalised as (weight in kilograms)/ (height in metres)(2). HSP were measured directly as visual analogue scale (VAS) scores, and converted to time trade-off (TTO) scores.

A one unit decrease in BMI over a 1-year period was associated with a 0.017 gain in utility units (utils). This estimate is comparable to utility gains observed for other widely used treatments (e.g. revascularisation for intermittent claudication, renal transplantation). It varies little over the observed range of VAS to TTO conversion values, and is a conservative estimate compared with using the unconverted VAS scores. If attempts to use weight reduction treatment in only the patients who successfully meet strict weight reduction targets are successful, the gain in HSP experienced by such patients may exceed this estimate. Patients with BMI values >or=28 with hyperglycaemia appear to have the greatest gain in HSP for a given change in BMI over 1 year.

The effect of weight reduction on HSP can be significant, at least in the short term. Estimates of HSP changes presented herein may be useful in economic evaluations of weight reduction treatments.