The Validity of Using Administrative Claims Data in Total Joint Arthroplasty Outcomes Research

• Published in: The Journal of arthroplasty 2010-09, Vol.25 (6), p.58-61
• Main Authors: Bozic, Kevin J., MD, MBA, Chiu, Vanessa W., MPH, Takemoto, Steven K., PhD, Greenbaum, Jordan N., MD, MBA, Smith, Thomas M., DO, Jerabek, Seth A., MD, Berry, Daniel J., MD
• Format: Article
• Language: English
• Subjects: administrative claims; Arthroplasty, Replacement, Hip - classification; Arthroplasty, Replacement, Knee - classification; Clinical Coding - standards; health services research; Hospital Records - standards; Humans; ICD-9-CM; Orthopedics; Osteolysis - classification; Osteolysis - diagnosis; Outcome Assessment (Health Care); Periprosthetic Fractures - classification; Periprosthetic Fractures - diagnosis; Prosthesis Failure; Prosthesis-Related Infections - classification; Prosthesis-Related Infections - diagnosis; Reoperation - classification; Reproducibility of Results; Retrospective Studies; total joint arthroplasty
• ISSN: 0883-5403; 1532-8406
• DOI: 10.1016/j.arth.2010.04.006

Abstract The purpose of this study was to evaluate concordance between administrative and clinical diagnosis and procedure codes for revision total joint arthroplasty (TJA). Concordance between administrative and clinical records was determined for 764 consecutive revision TJA procedures from 4 hospitals. For revision total hip arthroplasty, concordance between clinical diagnoses and administrative claims was very good for dislocation, mechanical loosening, and periprosthetic joint infection (all κ > 0.6), but considerably lower for prosthetic implant failure/breakage and other mechanical complication (both κ < 0.25). Similarly, for revision total knee arthroplasty diagnoses, concordance was very good for periprosthetic fracture, periprosthetic joint infection, mechanical loosening, and osteolysis (all κ > 0.60), but much lower for implant failure/breakage and other mechanical complication (both κ < 0.24). Concordance for TJA-specific procedure codes was very good only for revision total knee arthroplasty patellar component revisions and tibial insert exchange procedures. Total (all-component) revisions were overcoded for hips (00.70) and undercoded for knees (00.80). Improved clinical documentation and continued education are needed to enhance the value of these codes.