Local and national electronic databases in Norway demonstrate a varying degree of validity

C M Lofthus1, I Cappelen, E K Osnes

  • 1Department of Medicine, Centre of Endocrinology, Aker University Hospital, N-0514 Oslo, Norway. c.m.lofthus@ioks.uio.no

Insights

Electronic databases for hip fractures show significant inaccuracies. Local registers underestimated by 46% and overestimated by up to 19%, while the national database overestimated by 19%.

Area of Science:

  • Health Informatics
  • Epidemiology
  • Medical Record Analysis

Background:

  • Electronic databases are crucial for epidemiological studies.
  • Accurate data is essential for reliable health research and patient care.

Purpose of the Study:

  • To assess the validity of local and national electronic databases for hip fracture data.
  • To compare electronic database accuracy against medical records as the gold standard.

Main Methods:

  • Identified hospital admissions for hip fractures (ICD 9-code 820.X) over one year.
  • Utilized hospital discharge registers and the Norwegian Patient Register.
  • Verified diagnoses by retrieving and reviewing medical records and operating theater logs.

Main Results:

  • One local electronic register underestimated hip fractures by 46%.
  • Two other local registers overestimated hip fractures by 17% and 19%.
  • The national electronic database showed an overall overestimation of 19%.

Conclusions:

  • The validity of electronic databases for hip fracture data is questionable.
  • Findings suggest potential inaccuracies in electronic health records impacting epidemiological research.
  • Further validation of electronic health data is recommended.
Abstract