Abstract

For two years, the following records were linked for 10 453 people: (1) basic attributes; (2) details of prescriptions; and (3) information about illnesses recorded by general practitioners (GPs) in an "E' book. Analyses were performed to reveal association between drugs and diagnoses. Although the "E' book has certain disadvantages for drug monitoring, the methods proved to be capable of detecting adverse effects of drugs. Unfortunately the number of practitioners using "E' books would be too small for detection of most serious hazards such as the induction of cancer. Hence it is concluded that the first priority should be to establish a record linkage scheme covering hospital admissions, obstetric deliveries, and deaths.