Abstract
Surveillance, monitoring, or registration of coronary heart disease (CHD) in populations large enough to follow trends involve the collection and coding of data provided by routine medical and medico-legal sources. The data may be inadequate, missing from file, or in conflict, but because of the need for completeness, all cases must be used. Registration of a chronic disease in the form of episodes causes some logical problems, as does the need to allocate an underlying cause of death in fatal cases. The problems of extracting and coding the relevant information are discussed with reference to specific items and the results of an international quality control exercise are used to demonstrate how items differ in their consistency of coding. Possible causes of spurious trends are discussed, including inaccurate population data.