Objectives: To describe the processes involved in the selection of cases and controls, and the validation of case diagnosis, in a large case-control study on the association between myocardial infarction (MI) and oral contraceptives.
Design: Community case-control study in England, Scotland and Wales.
Subjects: Cases were women aged 16-44 with a diagnosis of MI. Controls were age and general practice matched women without a diagnosis of MI.
Results: The proportion of interviewed cases per number eligible was 19% (95% CI 12-26%) lower in England and Wales than in Scotland. The kappa score for agreement between three cardiologists on diagnostic categories of the potential cases was 0.52 overall ('moderate' agreement). For the 35 cases resubmitted for categorization, the kappa scores for agreement between the first and second classification for the two national committee cardiologists were 0.32 and 0.23 ('fair' agreement). Age matching of controls was better in Scotland than England and Wales, the proportion of the first choice controls being 14% (95% CI 8-19%) higher in Scotland. The mean time to complete a cluster of interviews (one case and four controls) was 36 days in Scotland and 63 days in England and Wales.
Conclusions: Very close age-matching was achieved overall, although the process was more efficient in Scotland than in England and Wales. Some suggestions for improvement in the case validation process are made.
Copyright 1999 John Wiley & Sons, Ltd.