ACUTE ENCEPHALOPATHY AND SEIZURE RATES IN CHILDREN UNDER AGE 2 YEARS IN OREGON AND WASHINGTON-STATE

A surveillance network was established in Oregon and Washington state to identify cases of certain neurologic illnesses during a one-year period (August 1, 1987 to July 31, 1988) among children 1-24 months of age. Reported here are the observed case occurrence rates and findings from capture-recapture analyses used to predict rates that would have been observed had ascertainment been complete. The network consisted of a hospital component involving 98% of all eligible facilities, a provider component involving 93% of all eligible pediatricians and neurologists, and a research staff component of record abstractors. Of 3,876 potential cases reported, 626 met the study criteria. The observed rates for the primary diagnostic groups were: encephalopathies, 16 per 100,000 children; infantile spasms, 10 per 100,000 children; afebrile seizures, 159 per 100,000 children; and complex febrile seizures, 101 per 100,000 children. Of the 626 qualifying cases, 41% were reported by two or more surveillance components. Capture-recapture analyses with log-linear modeling to control for source dependence suggested 80% of all study cases were detected. Comparable percentages for the four illness groups were: encephalopathies, 82%; infantile spasms, 94%; afebrile seizures, 69%; and complex febrile seizures, 91%. The predicted rate for afebrile seizures, corrected for under-ascertainment, exceeded the upper 95% confidence interval bound around the observed rate. For all other conditions, predicted rates fell within the 95% confidence intervals around the observed rates. These findings suggest capture-recapture analyses should be applied to the full sample of cases and to relevant disease substrata.