rss
J Epidemiol Community Health 50:86-93 doi:10.1136/jech.50.1.86
  • Research Article

Comparing survey data on functional disability: the impact of some methodological differences.

  1. H S Picavet,
  2. G A van den Bos
  1. Academic Medical Center, University of Amsterdam, Netherlands.

      Abstract

      STUDY OBJECTIVE: To examine the impact of some differences in survey methodology on the prevalence of functional disability in population based surveys of the elderly. DESIGN AND METHODS: Nine surveys of Dutch people aged 55 years and older were compared to investigate the differences in the methods of data collection (proxy questioning, yes/no; interview versus self administered questionnaire) and construction of the questionnaire (wording of introductory text, activities, and response categories). The effect of these differences on prevalences in three domains of functional disability--activities of daily living, mobility, and communication--were studied. Both univariate analyses and multivariate logistic regression were used to quantify the methodological influences. RESULTS: No effect of proxy questioning could be shown. Self administered questionnaires yielded higher prevalences of disability than interviewer administered questionnaire--in particular for mobility (odds ratio (OR) 1.4, 95% confidence interval (95% CI) 1.3, 1.6) and communication (OR = 1.7, 95% CI 1.5, 1.9), resulting in prevalence differences of 9 and 11 percentage points respectively. Seemingly minor differences in the structure and wording of the questionnaires resulted in major differences (up to 15.6 percentage points) in prevalence estimates of functional disability. These differences were associated with the severity level of the disability indicated by the wording of the questions. CONCLUSIONS: Differences in survey methodology have a substantial effect on the prevalence estimates of disability in the elderly. These differences should be taken into account when making international comparisons and studying time trends based on survey data.