Estimation of body weight in hospitalized patients

To examine the bias and precision of different methods of estimating body mass and height in hospitalized adult patients.

Patients were enrolled at the Alfred and Caulfield hospitals, Melbourne, Australia following verbal consent. Estimates were made using the Lorenz formula (that utilizes height, waist and hip circumference), the Crandell formula (that utilizes height and arm circumference) and visual estimation of weight based on the average results obtained by two pharmacy interns. Statistical error was calculated as the ratio of estimated to actual weight; bias was assessed as the mean error and precision as the proportion of estimates within 10 and 20% of measured weight and standard deviation of the error.

In a 5-week period July to August 2010, 198 patients were enrolled. The median age was 64 years (range 19-91) and 52% were female. Thirty-four (17%) patients were obese (BMI >30 kg/m(2)) and 8 (4%) were underweight (BMI <18 kg/m(2)). With the Lorenz formula an estimate within 10% was obtained for 56% of patients; with the Crandell formula prediction was poor. Documentation of body weight in notes and patient self-reporting were both accurate. Seventy-two patients (43%) were prescribed one or more drugs for which dosing potentially should be adjusted for body weight.

In adult hospitalized patients, the estimation of body weight by anthropomorphic measures is not accurate. This supports the need for equipment to be made widely available to accurately weigh patients directly in hospital, including in unconscious and immobile patients.