The positive predictive value of hip fracture diagnoses and surgical procedure codes in the danish multidisciplinary hip fracture registry and the danish national patient registry

Aim: The health-care databases may be a valuable source for epidemiological research in hip fracture surgery, if the diagnoses are valid. We examined the validity of hip fracture diagnoses and surgical procedure codes in the Danish Multidisciplinary Hip Fracture Registry (DMHFR) and the Danish National Patient Registry (DNPR) by calculating the positive predictive value (PPV). Methods: We identified a random sample of 750 hip fracture patients registered in the DMHFR between 2014 and 2017. Diagnoses have been coded by the 10^th revision of the International Classification of Diseases, while procedures have been coded by the Nordic Medico-Statistical Committee classification in the DNPR and directly transferred to the DMHFR. Using the surgical procedure description from the medical record as gold standard, we estimated the PPV of the hip fracture diagnoses and surgical procedure codes in the DMHFR and the DNPR with 95% confidence interval (CIs). Results: The PPV was 90% (95% CI: 86%-93%) for fracture of the neck of femur, 92% (95% CI: 87%-95%) for trochanteric fracture, and 83% (95% CI: 78%-88%) for subtrochanteric fracture. Joining trochanteric and subtrochanteric fracture resulted in a PPV of 97% (95% CI: 95%-98%). Procedure codes had a PPV of 100% for primary prosthetic replacement and internal fixation with intramedullary nail, 96% (95% CI: 85%-99%) for internal fixation using screws alone, 91% (95% CI: 84%-96%) for internal fixation using plates and screws, and 89% (95% CI: 83%-94%) for internal fixation with other or combined methods. Stratifying by age group, gender, hospital type and calendar year of surgery showed similar results as the overall PPV estimates. Conclusion: Our findings indicate a high quality of the hip fracture diagnoses and corresponding procedure codes in the DMHFR and the DNPR, with a majority of PPVs above 90%. Thus, the DMHFR and the DNPR are a valuable data source on hip fracture for epidemiological research.