Uncommon Haemophilus infections.

in which the incidence of the effect would be measured so that a careful benefit-risk assessment could be formed, taking account of information on other similar drugs.Resources giving rapid and accurate assessments on these lines would allow better decisions about products by both companies and regulatory authorities.Perhaps the most important conclusion of the Minster Lovell seminar was that an automated national data base linking prescriptions filled with hospital diagnoses in Britain would be extremely valuable in allowing the rapid study of drug effects and providing quantitative data.This would lessen public anxiety and restore much needed confidence.The Prescription Pricing Authority and the National Health Service -hospital system provide an opportunity for such a system of drug surveillance based on computerised linkage of prescriptions with hospital diagnoses.A start has been made in the Dundee record linkage study3 and the cimetidine postmarketing surveillance study.4Confidentiality should not be allowed to be an insuperable problem, and the Minster Lovell seminar also concluded that technical and organisational barriers could be overcome-provided that enough individuals were encouraged to make a commitment to research in pharmacoepidemiology.