Comparing properties of audit data and routinely collected register data in case of performance assessment of hip fracture treatment in Finland

• Published in: Methods of information in medicine Vol. 46; no. 5; p. 558
• Main Authors: Sund, R, Nurmi-Lüthje, I, Lüthje, P, Tanninen, S, Narinen, A, Keskimäki, I
• Format: Journal Article
• Language: English
• Published: Germany 2007
• Subjects: Finland; Health Care Surveys; Hip Fractures - diagnosis; Hip Fractures - therapy; Humans; Medical Audit; Prospective Studies; Registries; Sensitivity and Specificity; Finland
• ISSN: 0026-1270
• DOI: 10.1160/ME0382

To compare prospective audit data and secondary administrative register data in the production of performance assessment information in the case of hip fracture treatment, and to cross-validate the quality of information. First, a conceptual model for the performance assessment of hip fracture treatment was defined. This model was then utilized in comparisons between the prospective audit data concerning 106 consecutive hip fracture patients from the Kuusankoski Regional Hospital and corresponding register data from the Finnish Health Care Register and the Causes of Death Register. We examined the completeness of registration of patients and also the accuracy and degree of completeness of the registered data. Observed differences were checked against the medical records. Register data lack clinical detail, but outperform prospective data in the recording of inpatient care history. Completeness of the register data is very good. The accuracy of easily measurable variables in the register is at least 95%. The agreement between register and audit data was 86.3% for detailed hip fracture diagnosis. Polyserial correlation between the functional dependency variables was 0.68. Register and audit data have certain limitations and problems, but both seem to be suitable for the performance assessment of hip fracture treatment. To improve the feasibility of the administrative register data, the voluntary input of additional hip fracture event data to the register should be made possible. Standardized instructions that guide the use of available register classifications in a sensible way would improve the quality of data.