Work

Osteoarthritis case identification in the million veteran program cohort: Comparison of diagnostic codes versus clinical notes

Year: 2026

Type: article

Abstract: Objective: This retrospective electronic medical record (EMR)-based cohort study compared osteoarthritis (OA) case identification using diagnostic codes versus clinical notes in the Million Veteran Program (MVP), leveraging the United States (US) Department of Veterans Affairs (VA) Centralized Interactive Phenomics Resource (CIPHER). Design: A random sample of 320 Veterans was evaluated for general OA using ICD-9-CM/ICD-10-CM codes from a published algorithm (PMID:29559693). OA cases required two OA codes at least 30 days apart. Controls had no OA or related codes. Chart abstractions identified OA and joint replacement documentation. A subset of 220 Veterans was reviewed for site-specific OA and joint replacements. Agreement was assessed using Cohen's kappa and Mcnemar's test. Results: Among 320 Veterans, 73 % had chart-documented OA. Agreement for general OA was moderate (kappa = 0.53). In the subset of 220 Veterans, joint-specific agreement was substantial for knee (kappa = 0.63) and hip (kappa = 0.58) but lower for spine (kappa = 0.16), hand (kappa = 0.34), and thumb OA (kappa = 0.12). Agreement was high for hip (kappa = 0.86) and knee replacements (kappa = 0.69). McNemar's test indicated significant asymmetry for general OA, hand OA, and thumb OA, with more instances documented in charts, suggesting under-coding of milder OA phenotypes. No significant asymmetry was observed for knee and hip OA, indicating stronger alignment between chart and coded data. Conclusions: Agreement between diagnostic codes and chart-documented OA varies by joint site, highlighting the need for refined phenotyping algorithms and standardized diagnostic protocols to improve OA detection in Veterans.

Source: Osteoarthritis and Cartilage Open

DOI: 10.1016/j.ocarto.2026.100758

Authors: Chelsea T. Nguyen, Joshua S. Richman, Joe W. Chiles, Jasvinder A. Singh, Merry-Lynn N. McDonald

Institutions: University of Alabama at Birmingham, Birmingham VA Medical Center, UAB Medicine, Michael E. DeBakey VA Medical Center

Language: English

Topic: Medical Coding and Health Information

Subfield: Health Information Management

Field: Health Professions

Domain: Health Sciences

SDG: Good health and well-being

Open Access status: gold

Estimated APC paid: $1,610