Prevalence variations in psychotropic treatment of children

This study was undertaken to clarify several aspects of the estimation of prevalence of three commonly use pediatric psychotropic agents, namely, methylphenidate, desipramine, and imipramine. The study aims are threefold: (1) to show the variability of drug prevalence by comparing estimates from three data sources; (2) to show the misleading impression that can be created by reporting drug prevalence estimates based on counts of prescriptions rather than persons; (3) to show the utility of gender-by-age-specific prevalence of drug use as a marker for diagnosis. Two data sources that yield population-based prescription estimates were available: 1991 Medicaid administrative claims data for prescriptions from a mid-Atlantic state and 1991 prescription records of the northwest region of Kaiser Permanente, a staff-model health maintenance organization (HMO). Another source of data consists of the 1991 National Ambulatory Medical Care Survey, which records medication information reported during physician office visits. Data analysis consists of quantitative estimates of (1) drug prevalence from each source; (2) the ratio of prescription claims to persons; and (3) the proportion of drug use according to age and gender. Methylphenidate and desipramine prevalence had a twofold greater use among state Medicaid enrollees compared with HMO enrollees. Average claims-to-person ratios of 5:1 suggest better accuracy using persons with medication rather than prescription counts. Gender-by-age-specific prevalence rates showed that 75% of the drug use for desipramine among those less than 15 years old was found among males, whereas 75% of the desipramine use among those 15 or older was found among females, suggesting its use for the treatment of attention deficit-hyperactivity disorder among young males and for depression among older females. The variability of community physician decision making in pediatric psychopharmacology is better understood by observing drug prevalence rates from different settings. National sampling efforts should be undertaken to verify regional and setting-specific prevalence findings and to learn the reasons for their fluctuation.