L. M. Funk1,2, Y. Shan1, C. I. Voils3,4, J. Kloke5, L. Hanrahan6 1University Of Wisconsin,Surgery,Madison, WI, USA 2William S. Middleton (Madison) VA,Surgery,Madison, 53792, USA 3Duke University Medical Center,Medicine,Durham, NC, USA 4Durham VA,Medicine,Durham, NC, USA 5University Of Wisconsin,Biostatistics & Medical Informatics,Madison, WI, USA 6University Of Wisconsin,Family Medicine,Madison, WI, USA
Introduction: The measurement of population-level obesity rates is important for informing policy and targeting treatment. The gold standard method of estimating obesity rates in the U.S. is the National Health and Nutrition Examination Survey (NHANES). Given that NHANES requires household visits for height and weight measurement, NHANES samples < 0.1% of the adult population and does not target state- or and health system-level measurement. The objective of this study was to assess the feasibility of using body mass index (BMI) data from the electronic health record (HER) in a large health system to assess rates of overweight and obesity. To explore the possibility of selection bias in EHR data, we also compared overweight and obesity rates in the EHR to national NHANES estimates.
Methods: Using outpatient data from 42 clinics, we studied 388,762 patients who had at least one primary care visit in 2011-2012. We compared crude and adjusted overweight and obesity rates by age category and ethnicity between EHR patients and NHANES participants. Adjusted rates of overweight (BMI>25.0-29.9) and obesity (class I: BMI 30.0-34.4; class II: 35.0-39.9; and class III: >40) were calculated in a two-step process. The first step accounted for missing BMI data using inverse probability weighting via a multivariable logistic regression, while the second included a post-stratification correction to adjust the EHR population to a nationally representative sample.
Results: 59.6% (n=192,039) of patients in the EHR had at least one BMI value in the dataset. 70.0% (95% CI 69.8-70.2) of adults were overweight or obese, while 17.0% (95% CI 16.8.-17.1) had class II or III obesity. Adjusted rates of obesity for EHR patients were 37.3% (95% CI 37.1-37.5) compared to 35.1% (95% CI 32.3-38.1) for NHANES patients. Adjusted class III obesity rates were 7.4% (95% CI 7.3-7.5) and 6.4% (95% CI 5.2-7.7) for EHR and NHANES participants, respectively. Among the 16 obesity class and ethnicity (White, Black, Hispanic, Other) strata that were compared between EHR and NHANES patients, 14 (87.5%) contained similar obesity estimates (i.e. overlapping 95% CIs; Figure).
Conclusions: Obesity estimates from the analysis of electronic health records were largely similar to national estimates generated by NHANES. The electronic health record may be an ideal tool for identifying and targeting patients with obesity for implementation of public health and/or individual level interventions, such as behavioral, medical and/or surgical treatment.
