Reliability of International Classification of Diseases, Ninth Edition, Codes to Detect Morbid Obesity in Patients Undergoing Total Hip Arthroplasty

Abstract

BackgroundAlthough the impact of coding errors, with respect to obesity, has been previously reported, it is unclear whether morbid obesity is prone to similar coding inaccuracies. Therefore, the purpose of this study was to evaluate the reliability of coding for morbid obesity in patients who underwent total hip arthroplasty (THA). MethodsA total of 10,475 primary THAs performed at a single institution from 2004 to 2014 were identified. The presence of International Classification of Diseases, ninth edition diagnosis codes denoting any grade of obesity or morbid obesity during the admission was noted. The sensitivity of coding was evaluated along with the effect of morbid obesity (defined by body mass index or coding) on complications within 90 days of THA. ResultsThe sensitivity of obesity coding was 28.3%, while that of morbid obesity was 27.9% (area under the curve: 0.63 vs 0.63, P = .765). Among the 882 surgeries performed in morbidly obese patients, a code for any obesity was present in 467 (53%) surgeries, but only 53% (246) of these patients had a code specific for morbid obesity, while 47% (221) had a code for obesity not specifying morbid obesity. Nevertheless, the effects of morbid obesity on complications were similar regardless of how it was defined (coding or body mass index). ConclusionAlthough morbidly obese patients are likely to be easily identified as obese using codes, these patients may not be receiving a specific code. Researchers and clinicians should be aware that coding errors are prevalent even among higher grades of obesity in patients undergoing THA which may lead to suboptimal reimbursements and affect the results of studies using codes.