Good agreement between self-report and centralized hospitalizations data for arthritis-related surgeries

J Clin Epidemiol. 2013 Oct;66(10):1128-34. doi: 10.1016/j.jclinepi.2013.04.012. Epub 2013 Jul 8.


Objectives: To examine the level of agreement between self-reported and hospital administration records of arthritis-related surgeries for two large samples of community-dwelling older women in Australia, born between 1921-1926 and 1946-1951.

Study design and setting: Self-report survey data from the Australian Longitudinal Study on Women's Health was linked to inpatient hospital data from the New South Wales Admitted Patient Data Collection. Levels of agreement were compared using Cohen's kappa, sensitivity, specificity, and positive and negative predictive values. Reasons for false positives were examined.

Results: This study found good agreement (kappa >0.70; sensitivity and specificity >0.80) between self-report and hospitalizations data for arthritis-related surgeries.

Conclusions: This study provides new evidence for good agreement between self-reported health survey data and administrative records of arthritis-related joint procedures, and supports the use of self-report surveys in epidemiological studies of joint procedures where administrative data are either not available or not readily accessible, or where more extensive contextual information is needed. The use of health survey data in conjunction with administrative data has an important role to play in public health planning and policy.

Keywords: Arthroplasty; Health surveys; Hip; Knee; Medical record linkage; Replacement; Self-report.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Aged
  • Arthritis / surgery*
  • Arthroplasty, Replacement, Hip / statistics & numerical data
  • Arthroplasty, Replacement, Knee / statistics & numerical data
  • Female
  • Health Surveys*
  • Hospital Records*
  • Hospitalization / statistics & numerical data*
  • Humans
  • Longitudinal Studies
  • Middle Aged
  • New South Wales
  • Reproducibility of Results
  • Self Report*