Answering questions about the risks and benefits of ADHD medications through better research design
Can improved research designs answer questions that arise between physicians and patients, such as a better understanding of the risks and benefits of prescription drugs for Attention Deficit Hyperactivity Disorder (ADHD)? Brian M. D’Onofrio, PhD, presented translational epidemiology in a talk titled The Risks and Benefits of ADHD: A Pharmacoepidemiologic Perspecitive to Answer These Questions at the 2018 Annual Meeting of the American Professional Society of ADHD and Related Disorders (APSARD).
Can improved research designs answer questions that arise between physicians and patients, such as a better understanding of the risks and benefits of prescription drugs for Attention Deficit Hyperactivity Disorder (ADHD)?
Brian M. D’Onofrio, PhD, director of clinical education in the Department of Psychological and Brain Sciences at Indiana University, presented translational epidemiology in a lecture titled “The Risks and Benefits of ADHD: A pharmacoepidemiologic perspective ”to answer these questions at the 2018 annual meeting of the American Professional Society of ADHD and Related Disorders (APSARD).
Making the audience laugh, he joked that putting “translational” in front of the word “epidemiology” is not a way of trying to convince the National Institutes of Health to give you money, but rather, its purpose. is to convince people that the use of big, large-scale data could be used to immediately answer clinical questions and put certain questions aside.
In the case of ADHD medications, D’Onofrio used the example of mainstream media headlines questioning possible associations between stimulant medications and substance abuse, as well as various randomized controlled trials (RCTs) published in scholarly journals.
RCTs have limits, he said, and cannot control everything. If there is an exposure associated with a result, any alternative explanation should be discarded, as the statistical covariates are not perfect and there may be a measurement error.
Better research design can account for occasional inferences, he said. For example, when studying side effects, is the drug the cause of the problem or is this the underlying condition for which a patient is given the drug? To try to get that answer, D’Onofrio looked at what happened to patients in the month leading up to an emergency room (ED) visit for substance use issues. Were they on or off their medication?
In one RCT, men with ADHD were 3.2% more likely to go to the emergency room for substance abuse problems compared to 1.2% of men without ADHD, and women with ADHD were 2.6 % more likely than 0.8% without. D’Onofrio said his research showed that when a male patient received his ADHD medication in the previous month compared to what he had not received, he had a 35% chance less than going to the emergency room for a drug addiction problem.
D’Onofrio used a large-scale data set from people with private employer insurance from 2005 to 2014. It included around 146 million people. In 2013, 51% were women and 78% had drug coverage. Since this was a private employment insurance data set, individuals were covered for different durations. From there, he withdrew inpatient, outpatient, and prescription claims for nearly 3 million ADHD patients.
He performed an “intra-individual analysis” where each person serves as their own control. The risks were compared from the moment when the same individual takes or not his drugs. The analysis took into account all stable factors and adjusted for time varying covariates to help rule out dynamic confusion.
From there, D’Onofrio looked at the odds ratio of the risk of a serious outcome, such as a trip to the emergency department for an addiction problem, in the month a patient was taking ADHD medication, compared to when the patient was arrested.
For each concurrent intra-individual comparison examined by D’Onofrio, there was a decreased risk (an odds ratio of less than 1) of going to the emergency room for patients who fell into these categories:
- No other psychiatric medication
- No psychotherapy
- Incident diagnosis cohort
- First events only
- Medication period extended by one month
- Stimulant drug only
- Ambulance, hospitalization or emergency events
- Drug selective serotonin reuptake inhibitor
In addition, the associations were held 2 years later. The same associations were also held when reviewing emergency room visits for motor vehicle accidents by people taking ADHD medication.
D’Onofrio said the results should help allay concerns about rare, but serious concerns about ADHD medications, and said there could be clinical issues with not prescribing ADHD medications. .