ADHD treatment—supply or demand?
A. Winterstein
DOI: https://doi.org/10.1002/pds.3192
2012-04-01
Pharmacoepidemiology and Drug Safety
Abstract:The steady rise in the prevalence of pediatric mental health diagnoses and treatment in the past 2 decades has resulted in great concern and an ongoing debate about its causes. At the top of the list of psychiatric diagnoses remains attention deficit hyperactivity disorder (ADHD) with staggering estimates of 10%1 and, depending on the specific socio-demographic stratum, up to 20% of all school-aged children receiving treatment for ADHD.2 Determinants of diagnoses or treatment include a variety of patient, provider, and environmental factors, raising the question whether ADHD is over-diagnosed in some groups or under-diagnosed in others.3,4 Bruckner et al. in this issue of PDS report the findings of an ecologic study examining the relationship between changes in medical supply (i.e., the number of physicians per capita) and ADHD drug utilization per youth diagnosed with ADHD.5 During their study period of 2001–2003, the authors report a 33% increase in ADHD prescriptions measured as total mg of methylphenidate equivalents dispensed across more than 2000 US counties. Simultaneous trends include increases in the number of all physicians and pediatricians per capita, the proportion of the population with public insurance or of Hispanic ethnicity, and presence of state school accountability measures. The density of psychiatrists and the student-to-teacher ratio appears to drop slightly across all counties. In their multivariate analysis of these changes on the level of individual counties, the authors find significant associations between increases in ADHD treatment utilization per county-specific ADHD prevalence and increases in physician and psychiatrist density, a decline in private insurance penetration, a decrease in individualized school education programs, an increase in per capita income, and a decrease of the non-Hispanic Black population. The findings are intuitive and consistent with previous reports with respect to effects of racial/ethnic characteristics or the available educational support structure.4,6 Likewise, associations between healthcare supply and use have been reported previously, and the findings at hand solidify concerns about treatment decisions that are not motivated by clinical indication but by market pressure. The importance of these findings notwithstanding, it is important to consider some methodological limitations. Bruckner and his colleagues assembled an impressive array of data sources that quantify stimulant sales, ADHDprevalence, physician density, healthmaintenance organization penetration, certain school parameters, and socio-demographic characteristics on the level of more than 2000 US counties. Although the idea and analytical execution are commendable, it is important to note that this is an ecologic study with inherent limitations in (causal) inference. It is further important to note that key factors, which consistently predict ADHD treatment, have been omitted (on either the patient or county level). These include, for example, presence of mental comorbidities, concomitant use of other psychotropic medications, as well as age. Importantly, in the absence of county-specific prevalence data for all three study years, the authors assumed a constant ADHD prevalence and report changes in ADHD drug volume (in total mg methylphenidate equivalents) per fixed number of youths diagnosed with ADHD per county. Following this rationale, study conclusions suggest that an increased physician density results in more ADHD drug volume, deriving from either a larger proportion of children with ADHD who use drug *Correspondence to: A. G. Winterstein, Pharmaceutical Outcomes and Policy, College of Pharmacy, Epidemiology, Colleges of Medicine and Public Health and Health Professions, University of Florida, PO Box 100496 Gainesville, FL 32610, USA. E-mail: almut@cop.ufl.edu