Background
Research using large administrative databases has substantially increased in recent years. Accuracy with which comorbidities are represented in these databases has been questioned. The purpose of this study was to evaluate the extent of errors in obesity coding and its impact on arthroplasty research.