Epidemiologic Issues in ADHD
April 14, 1999
Benjamin Lahey, Ph.D., University of Chicago
Andrew Rowland, Ph.D., National Institute for Environmental Health Sciences
Prevalence estimates of Attention-Deficit/Hyperactivity Disorder (ADHD) have historically varied over time with estimates ranging from as low as under 1% to as high as nearly 20% in school-age children. There is no systematic monitoring of ADHD, no gold standard for epidemiologic research of the disorder, and minimal population-based epidemiologic research of ADHD in the United States. Consequently, even rough estimates of the prevalence provide no meaningful indication of the level of burden this disorder poses in our society. Unfortunately, current research endeavors provide insufficient information to explore, even roughly, basic descriptive epidemiologic questions such as how the disorder may vary by race, ethnicity, socio-economic status, and age. Currently there is not a known etiology for the disorder; therefore, epidemiologic research is imperative. However, there is a paucity of such rigorous science in the ADHD field. The U.S. Census Bureau estimates that there will be 51 million children between the ages of 5-17 by the year 2000. With prevalence estimates of ADHD ranging higher and higher, a public health perspective needs to be applied to this disorder.
In an effort to delineate the issues in the epidemiologic research of ADHD, the Division of Birth Defects, Child Development, and Disability and Health (proposed)/NCEH hosted a one-day meeting with two researchers in the epidemiologic study of ADHD and the NCEH ADHD work group which is comprised of DDB staff. Prior to the meeting, an outline of the specific epidemiologic questions Division of Birth Defects, Child Development, and Disability and Health (proposed) had identified was provided to all participants. The questions and the discussion which ensued resulted in the following points during the one-day meeting:
Please note that all answers are the summarized sentiments of the participants invited to this meeting and not those of the CDC. Statements herein are not, in any way, to be interpreted as promissary for inclusion in any ADHD research agenda setting nor are they ADHD policy statements by the Centers for Disease Control and Prevention, National Center for Environmental Health. This meeting was exploratory in nature and the results are provided here in an attempt to share the most information with the public.
Q1. What are the most important surveillance and epidemiologic research questions/needs in ADHD that are currently not being addressed by other researchers (or not rigorously addressed)?
- The etiology of ADHD is unknown and risk factors, aside from genetic predispositions, are under-studied, especially at the population level.
- There have been few longitudinal population-based studies, and, particularly in the U.S., there is a strong inverse relationship between the adequacy of the sample and the adequacy of the measures.
- Epidemiology that relates to risk factors is very incomplete. While there is clear evidence in the differences in gender and age, little is known about risk factors for ADHD. What are the risk factors and how do they vary by subtype, gender, age, race, SES, etc?
- The efficacy of pharmacological treatments in the general population has not been studied well, and we know little about the long-term benefits of such treatments from childhood into adulthood.
- Both maternal cigarette smoking and alcohol abuse and their association to ADHD are important research topics.
- The primarily inattentive subtype of ADHD is a group of individuals with consistent long-term problems in school settings but little epidemiologic work exists on this population.
- More research/discussion is needed about measurement of impairment.
- Under-diagnosis/some misdiagnosis may be occurring.
- Diagnosis depends on histories provided by parents and teachers; how are these reports shaped by social factors?
Q2. What types of epidemiologic study designs are recommended considering the limitations in previous ADHD risk factor studies? What are promising etiologies to focus on–genetic, environmental toxins, other environmental exposures, pre- or postnatal factors?
- Large longitudinal follow-up studies of ADHD children which can provide information on prognosis by subtype and comorbidity; information on how prevalence varies by age, race, gender, and SES; as well as information on diagnosis and treatment patterns in community settings.
- Intervention effectiveness trials at the population-based level to ascertain whether current practices in the general population effectively identify and treat ADHD individuals, and whether community interventions improve diagnosis, treatment, and outcome?
- Longitudinal follow-up to evaluate the long-term efficacy and safety of stimulant medication and to identify preventable risk factors, including exposures during pregnancy, childhood lead exposure, and risk factors not yet identified.
- More research on the relationship between very low birth weight and ADHD, as well as on possible risk factors, such as perinatal complications, which have not been measured as well as they need to be, and other potential risk factors such as lead, maternal cigarette smoking (genetically informative design), and urbanicity.
- Future studies of risk factors for ADHD should: examine (and control for) genetic influences, consider mediators and moderators, assess common comorbid psychopathology, assess cognitive ability, and measure impairment.
- Research on both maternal cigarette smoking and alcohol abuse and their association to ADHD were identified as promising areas of investigation. Such research should be done in a genetically informative manner, so that genetic and environmental factors can be teased apart.
- Research endeavors that include, at a minimum, peer functioning, academic functioning, and accidental injuries in their data collection.
Q3. What is the feasibility of conducting an ongoing, population-based surveillance program for ADHD in a public health setting, as opposed to a research setting?
- Passive surveillance of the disorder is not possible. Case ascertainment using an active surveillance system through school record abstracting is also not feasible.
- ADHD identification is not amenable to surveillance at present; other research designs must be employed.
- However, it is very possible to do some population-based surveying for ADHD using something like NHANES. A questionnaire could be designed that parents could fill out in 5 to 20 minutes.
Q4. If ongoing, population-based surveillance or research of ADHD were to be designed, how would the methodologic issues be resolved? Specifically consider case definition, target population, case ascertainment, and data collection items.
- A nationally representative household study of ADHD would provide extraordinarily important basic descriptive epidemiologic information that we currently do not have.
- Research should be population-based and might work best through household contact rather than the more common practice of engaging school systems for large-scale research.
- Case ascertainment should be employed using the DSM-IV criteria in any checklists (an interesting comparison of ADHD identification rates would be possible if a study identified cases using both DSM and ICD-10 criteria).
- Thorough case ascertainment would include ratings from both parents and teachers and a measure of impairment (note: do not use CBCL for impairment or ADHD diagnosis). Impairment is a crucial piece to making the diagnosis.
Q5. Besides surveillance methodologies, what other avenues of research in ADHD could be explored to address the critical research needs earlier identified and discussed?
- We need an active study of a cohort of 100,000 children. This is not a job for passive surveillance. We need to know where and when the drugs work and do not work.
- We also need larger longitudinal follow-up studies of ADHD children, which would include prognosis by subtype and co-morbidity; in addition, we should work toward an understanding of the developmental trajectories of the various subtypes of ADHD.
- Both long-term efficacy and safety of stimulant medications need to be addressed.
- CDC could facilitate such research by organizing collaboration between skilled public health/epidemiologic researchers and psychological researchers. There may be ways to bring together experts to answer some of the questions raised by studies lacking scientific rigor.
- CDC could also bring researchers together to agree on a standard set of case ascertainment methods for ADHD research across disciplines and agencies. Such standardization would help provide a way to utilize all the research being done by various groups.
“We cannot understand a disorder until we can describe it, and we cannot fully describe ADHD. There has never been a good epidemiologic study of ADHD in U.S., using good measures.”
Benjamin Lahey, Ph.D.