A patient-centred approach to estimate total annual healthcare cost by body mass index in the UK Counterweight programme

Abstract

Previous studies, based on relative risks for certain secondary diseases, have shown greater healthcare costs in higher body mass index (BMI) categories. The present study quantifies the relationship between BMI and total healthcare expenditure, with the patient as the unit of analysis. Analyses of cross-sectional data, collected over 18-months in 2002-2003, from 3324 randomly selected patients, in 65 general practices across UK. Healthcare costs estimated from primary care, outpatient, accident/emergency and hospitalisation attendances, weighted by unit costs taken from standard sources. In univariate analyses, significant associations (P<0.05) were found between total healthcare expenditure and all dependent variables (women>men, drinker<non-drinkers, smokers>non-smokers, and increasing with greater physical activity, age and BMI. In multivariate analysis, age, sex, BMI, smoking and alcohol consumption remained significantly associated with healthcare cost, and together explained just 9% of the variance in healthcare expenditure. Adjusted total annual healthcare cost was £16 (95% CI £11-£21) higher per unit BMI. All cost categories were significantly (P<0.003) higher for those with BMI >40 compared with BMI <20 kg m(-2): prescription drugs (men: £390 versus £16; women: £211 versus £73), hospitalisation (men: £72 versus £0; women: £243 versus £107), primary care (men: £191 versus £69; women: £268 versus £153) and outpatient care (£234 versus £107 women only). Annual healthcare expenditure rose a mean of £16 per unit greater BMI, doubling between BMI 20-40 kg m(-2). This gradient may be an underestimate if the lower-BMI patients with heights and weights recorded had other costly diseases.