Self-perceived vs actual and desired weight and body mass index in adult ambulatory general internal medicine patients: A cross sectional study

Background: No study has compared patients' self-reported heights and weights (and resultant self-reported body mass indexes [BMIs]) with their actual heights, weights, and BMIs; their self-perceived BMI categories; and their desired weights and BMIs and determined rates of clinicians' documented diagnoses of overweight and obesity in affected patients in a single patient group. The objectives of this study were to make these comparisons, determine patient factors associated with accurate self-perceived BMI categorization, and determine the frequency of clinicians' documented diagnoses of overweight and obesity in affected patients. Results: A total of 508 consecutive adult general internal medicine outpatients (257 women, 251 men; mean age, 62.9 ± 14.9 years) seen at Mayo Clinic in Rochester, Minnesota, between November 9 and 20, 2009, completed a questionnaire in which they reported their heights, weights, self-perceived BMI categories ("underweight," "about right," "overweight," or "obese"), and desired weights. These self-reported data were compared to actual heights, actual weights, and actual BMI categories (measured after the questionnaire was completed). Overall, 70% of the patients were overweight or obese. The average self-reported weight was significantly lower than the average actual weight (80.3 ± 20.1 kg vs 81.9 ± 21.1 kg; P <.001). The average self-reported BMI was significantly lower than the average actual BMI (27.6 ± 5.7 kg/m² vs 28.3 ± 6.1 kg/m²; P <.001). Overall, 32% of patients had obesity; however, only 6% perceived they were obese. Accuracy of self-perceived BMI category decreased with higher actual BMI category (P <.001 for trend). Female sex, higher education level, smoking status, and lower BMI were associated with higher accuracy of self-perceived BMI category. Desired weight loss increased with higher self-perceived and actual BMI categories (P <.001 for trends). Of the 165 patients who actually were obese, only 40 (24%) had obesity documented as a diagnosis in their medical records by their clinicians. Statistical tests used were the paired t test, the Pearson Χ2 test, the Cochrane-Armitage trend test, the Wald test of marginal homogeneity, analysis of variance, and univariate and multivariate logistic regression. Conclusions: Many obese patients inaccurately perceive their BMI categories; accuracy decreases with increasing BMI. Clinicians should inform patients of their BMIs and prescribe treatment plans for those with overweight and obesity.