Measles in rural Bangladesh: issues of validation and age distribution.

In the Third World, scientists have described thoroughly the pattern of measles in Africa. In Asia, however, the epidemiology of measles has been poorly described. In 1989, a measles surveillance system was initiated in Matlab, rural Bangladesh in order to monitor measles incidence in the area. The population (100,000), which has relatively low immunization coverage, has an excellent demographic surveillance system which allows accurate follow-up. The system works through trained community health workers who visit all households every fortnight and report possible cases of measles. Medical doctors review a proportion of index cases and cases in infants soon after the appearance of the rash. Of the 4673 cases reported in all age groups, 18% were visited by medics. Confirmation of diagnosis was related to the age at onset and overreporting by community health workers is documented. The incidence of measles was estimated at 43% of children < 5 years old. Of cases < 5 years of age, 14% occurred below the age at vaccination (incidence rate = 5.5%). A strong seasonal pattern is reported. Case-fatality rates are low. This is the first report which shows the effect of age on reliability of reported measles. It shows the misclassification likely to occur when reporting measles in large-scale surveys, particularly in infants. It also shows that the pattern of measles in this community resembles the pattern reported for African communities and calls for a review of immunization strategies with vaccination below 9 months of age.