Between-register and within-register variability in event-finding and rates in the MONICA-France Project.

The 1985 and 1986 coronary event data from the three French MONICA Registers in Haute-Garonne (Toulouse), Bas-Rhin (Strasbourg) and the Urban Community of Lille (UCL) are used to compare the sources of information, the distribution of events in different diagnostic categories, and their possible effects on event rates. All three registers follow the MONICA protocol, and use the "cold pursuit" method, with data abstracted retrospectively from medical files or from the doctors establishing the causes of death. The health system and medical background are similar in the three centers. The distribution of the sources of information (in-hospital and out-of-hospital) and the distribution of events by diagnostic categories are different among centres. The possible effects of these differences on rates of definite myocardial infarction or possible coronary death are estimated between registers and within registers during a 2-year period. Cross-sectional comparisons are questionable, and possible false trends in event rates require special attention.