OBJECTIVES: The objective of the study was to compare patients with hip fracture in the Swedish Hip Fracture Register (SHR) and the National Patient Register (NPR) between 2008 and 2017 regarding coverage, agreement, and representativeness.
STUDY DESIGN AND SETTING: The NPR and SHR were linked through the personal identity number assigned to all residents of Sweden. The proportion of matching records in both registers was calculated in total, for different fracture types, and for the first and recurrent fractures separately. Representativeness was estimated using logistic regression and survival analysis.
RESULTS: The agreement between fracture types in both registers was excellent, but our comparison with the SHR indicated that the NPR may overestimate the number of recurrent fractures in Sweden. The SHR covers more than 80% of all hip fractures in the NPR. Patients in the SHR were similar to those in the NPR with regard to many, but not all, characteristics and long-term survival. However, patients who die shortly after hip fracture were underreported to the SHR.
CONCLUSION: Both registers are valuable data sources for epidemiological research. Although neither register constitutes a gold standard, their excellent agreement suggests high-data quality. Nevertheless, both registers have some limitations that may be relevant depending on the research question under study.
Keywords: Completeness Hip fracture Register-based research Registers Representativeness Validation