The Biometry and Field Studies Branch is the Statistical Coordinating Center for the ongoing clinical trial of behavioral and cognitive side effects of phenobarbital for prevention of febrile seizure recurrence. The accrual of patients was completed, with 367 patients on study. Follow-up with psychometric testing was completed in July, 1988. Statistical analyses and a report regarding the primary results of Binet IQ and recurrence of febrile seizures have been completed, and submitted for publication. A large proportion of subjects missed Binet IQ test visits. The missing IQ data were determined not to be missing at random. We developed a statistical analysis method that reduces the magnitude of bias due to the nonrandom missingness. Issues surrounding analysis by """"""""intended treatment"""""""" versus analysis by """"""""actual treatment' have been studied. We have shown analysis by actual treatment changes analysis results. We have found that the effective age range of the tests for measuring language development and maturing of play behavior does not extend beyond three years of age. A method for determining adequacy of final sample sizes in the presence of a substantial amount of incomplete data was developed and applied to the assessment of the power of the statistical analysis of recurrence rates of seizures on the different treatment arms. Data management, statistical analysis and manuscript preparation activities will continue into FY 1990. The final master file and analysis files are in place.
