Arthritis Research Campaign National Primary Care Centre, Keele University, Keele, Staffordshire, UK.
Abstract
BACKGROUND: Epidemiological surveys often rely on self-report as a measure of morbidity in a population. However, these data can also be extracted from primary care records.
OBJECTIVE: To compare morbidity estimates based on self-report with those obtained from primary care records.
METHODS: A cross-sectional survey and accompanying medical record review were carried out in all consenting adults aged > or =50 years in three general practices in North Staffordshire, UK. Self-reported morbidity was compared with computerized general practice consultation records for the 2 years prior to the survey.
RESULTS: Of the 7878 survey responders, 5889 consented to medical record review. Agreement between self-reported and consultation data was excellent for diabetes. Agreement between the two sources of data was lower for hypertension, heart problems, chest problems and eyesight problems. It was poor for deafness and falls.
CONCLUSIONS: Self-report and consultation data provide comparable estimates of the prevalence of specific diagnoses such as diabetes. For other conditions, self-report and consultation records provide different measures of prevalence, and the choice of measure will depend on the morbidity being studied.