Healthcare providers (HCPs) should identify overweight and obesity for surveillance and intervention purposes. This study aimed (a) to assess the ability of HCPs to estimate body mass index (BMI) by visual inspection and (b) to determine how systematically BMI is measured in an outpatient setting. A mixed group of 219 HCPs and 33 medical students were asked to estimate the BMI of five photographed people with varying degrees of obesity. A systematic survey of hospital charts took place over one week in a stratified random sample of 26 out-patient clinics in a teaching hospital. Each chart was surveyed for recording of height, weight, BMI and waist circumference at that clinic, or if not measured then, during the preceding year. The study group underestimated the BMI of obese people, with greater discrepancies as obesity increased: an actual BMI of 32 kg/m² was underestimated by 9%, while one of 72 kg/m² was underestimated by 42%. Of 515 hospital charts identified, weight was recorded for 33.4% of patients, height for 4%, waist circumference for 1.4% and BMI for 0.4%. In the 10 clinics that measured weight during the data collection period, the percentage of patients weighed was 11.8 - 88.9%. However, clinics were reasonably well supplied with equipment: 100% had weighing scales, 76.9% had a measuring tape, 7.7% had a BMI chart and all but one had a stadiometer.

Conclusion HCPs neither recognise obesity reliably nor routinely record BMI. Systematic measurement strategies are required to tackle the global obesity epidemic.