Dealing with suspicions of malformation frequency increase. Experience with the Swedish register of congenital malformations.

Strategies in dealing with data obtained from malformation monitoring based on experiences with the Swedish monitoring systems, operating since 1965 (Register of Congenital Malformations) and 1973 (Medical Birth Register) are discussed. The importance of checking data that have sounded an alarm is stressed. Experience has shown that false alarms due to artefacts, such as changed diagnostic routines, changed reporting or registration of malformations, or random fluctuations, comprise most suspected changes in incidence. If a true increase in malformation frequency or a local cluster is observed, a hint of possible aetiological factors can be obtained from studies of maternal age distribution, seasonal viriation, geographical distribution, etc. The last step of the analysis consists of a case-control or a cohort study, aimed at revealing a specific teratogen. The importance of locating limited research resources to well-defined problems using high quality data is stressed.