Development and Validation of a Case Identification Algorithm for Hand Trauma Patients Using Health Administrative Data and the Epidemiology of Hand Trauma in a Universal Healthcare System
Author(s)
Chloe Wong, MD1, Karen Tu, MD, MSc2, Alejandro Hernandez, MSc3, David Urbach, MD, MSc, FRCSC4, Christopher Witiw, MD, MSc, FRCSC5, Bettina Hansen, MSc, PhD6, Alice Ko, BSc7, Pascale Tsai, MSc8, Heather Baltzer, MD, MSc, FRCSC, FACS1.
1Department of Surgery, Division of Plastic, Reconstructive & Aesthetic Surgery, University of Toronto, Toronto, ON, Canada, 2Department of Family and Community Medicine, University of Toronto, Toronto, ON, Canada, 3ICES, Toronto, ON, Canada, 4Department of Surgery, Division of General Surgery, University of Toronto, Toronto, ON, Canada, 5Department of Surgery, Division of Neurosurgery, University of Toronto, Toronto, ON, Canada, 6Biostatistics, Erasmus University Medical Center, Rotterdam, Netherlands, 7Cumming School of Medicine, University of Calgary, Calgary, AB, Canada, 8Michael DeGroote School of Medicine, McMaster University, Hamilton, ON, Canada.
1Department of Surgery, Division of Plastic, Reconstructive & Aesthetic Surgery, University of Toronto, Toronto, ON, Canada, 2Department of Family and Community Medicine, University of Toronto, Toronto, ON, Canada, 3ICES, Toronto, ON, Canada, 4Department of Surgery, Division of General Surgery, University of Toronto, Toronto, ON, Canada, 5Department of Surgery, Division of Neurosurgery, University of Toronto, Toronto, ON, Canada, 6Biostatistics, Erasmus University Medical Center, Rotterdam, Netherlands, 7Cumming School of Medicine, University of Calgary, Calgary, AB, Canada, 8Michael DeGroote School of Medicine, McMaster University, Hamilton, ON, Canada.
OBJECTIVES: Our primary objectives were (1) to develop and validate an administrative data algorithm for identifying hand trauma using clinical diagnoses documented in medical records as the reference standard and (2) to estimate the incidence of hand trauma in a universal public healthcare system from 1993 to 2023 using a validated algorithm.
METHODS: This retrospective study linked a random sample of patients from a hand trauma center in Ontario, Canada, to health administrative data. Combinations of physician fee-for-service procedure codes, diagnostic billing codes, hospitalization records, and emergency room records were tested. Using the optimal algorithm, age- and sex-standardized incidence rates were calculated by sex and age group from 1993 to 2023.
RESULTS: Among 301 patients included in the study, 147 (49%) had hand trauma. The most common injury was fracture or dislocation of the phalanges, carpus, radius, or ulna (57%), followed by tendon (12%), ligamentous (8%), crush (5%), and spaghetti wrist (4%) injuries. The optimal algorithm for identifying hand trauma cases in Ontario’s health administrative data was: “(2 specific physician diagnostic billing codes in 1 year with at least 1 billed by any hand trauma specialist) OR (1 specific or general physician diagnostic billing code and 1 fee-for-service procedure code in 1 year with at least 1 billed by any hand trauma specialist).” This algorithm had a sensitivity of 73.8% (95% CI 66.6-81.0), specificity of 80.1% (95% CI 73.8-86.5), PPV of 78.1% (95% CI 71.2-85.0), and NPV of 76.1% (95% CI 69.5-82.7). The standardized incidence of hand trauma increased from 384 to 530 per 100,000, with the greatest rise among males and individuals aged 0-19 and 80+.
CONCLUSIONS: Our algorithm identified hand trauma cases using health administrative data, revealing a rising burden. These findings can support improved surveillance, resource allocation, and care delivery.
METHODS: This retrospective study linked a random sample of patients from a hand trauma center in Ontario, Canada, to health administrative data. Combinations of physician fee-for-service procedure codes, diagnostic billing codes, hospitalization records, and emergency room records were tested. Using the optimal algorithm, age- and sex-standardized incidence rates were calculated by sex and age group from 1993 to 2023.
RESULTS: Among 301 patients included in the study, 147 (49%) had hand trauma. The most common injury was fracture or dislocation of the phalanges, carpus, radius, or ulna (57%), followed by tendon (12%), ligamentous (8%), crush (5%), and spaghetti wrist (4%) injuries. The optimal algorithm for identifying hand trauma cases in Ontario’s health administrative data was: “(2 specific physician diagnostic billing codes in 1 year with at least 1 billed by any hand trauma specialist) OR (1 specific or general physician diagnostic billing code and 1 fee-for-service procedure code in 1 year with at least 1 billed by any hand trauma specialist).” This algorithm had a sensitivity of 73.8% (95% CI 66.6-81.0), specificity of 80.1% (95% CI 73.8-86.5), PPV of 78.1% (95% CI 71.2-85.0), and NPV of 76.1% (95% CI 69.5-82.7). The standardized incidence of hand trauma increased from 384 to 530 per 100,000, with the greatest rise among males and individuals aged 0-19 and 80+.
CONCLUSIONS: Our algorithm identified hand trauma cases using health administrative data, revealing a rising burden. These findings can support improved surveillance, resource allocation, and care delivery.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
EPH59
Topic
Epidemiology & Public Health
Topic Subcategory
Disease Classification & Coding
Disease
Injury & Trauma, No Additional Disease & Conditions/Specialized Treatment Areas