Cookie Preferences
By clicking, you agree to store cookies on your device to enhance navigation, analyze usage, and support marketing. More Info
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
August 26, 2024
Most previous studies of suicide and self-harm risk among persons with ADHD have focused on adolescents and adults. They’ve also tended to be cross-sectional, analyzing data from a population at a specific point in time.
An Australian study team took a different approach, conducting a before-and-after study through the birth cohort of the Longitudinal Study of Australian Children (LSAC), comprising 5,107 children who have been followed up every two years since birth.
The diagnosis of ADHD was based on parents reporting that their child had received a diagnosis of ADHD at or before age ten.
Suicide and self-harm were defined as children’s self-report at age 14 of any thought or attempt of suicide and self-harm respectively over the past year.
The team adjusted for the following confounders: socioeconomic status, birth weight, ADHD medication history, maternal education level, maternal age at birth, experience in bullying victimization at age 12, and depression score based on Short Mood and Feelings Questionnaire (SMFQ).
Of the 5,107 participants, 3,696 had all the valid data required for analysis and were included in the final cohort. Of these, 3.6% were diagnosed with ADHD by age 10.
With a diagnosis of ADHD at age 10 and all other factors held constant:
Both depression and exposure to bullying were statistically significant mediators for the relationship. Nevertheless, depression and exposure to bullying each accounted for well under 10% of the overall effect.
Neither socioeconomic status nor maternal factors had any significant mediating effect on outcomes.
The authors concluded, “This study provides compelling evidence that children diagnosed with ADHD at the age of 10 years face significantly elevated risks of experiencing suicidal thoughts, planning, or attempts, as well as self-harm, by the age of 14 years, which underscores the critical importance of recognizing and addressing these heightened risks in children with ADHD.”
Ping-I Lin, Weng Tong Wu, Enoch Kordjo Azasu, and Tsz Ying Wong, “Pathway from attention-deficit/hyperactivity disorder to suicide/self-harm,” Psychiatry Research (2024), 337:115936, https://doi.org/10.1016/j.psychres.2024.115936.
In the general population, most mothers experience mood disturbances right after childbirth, commonly known as postpartum blues, baby blues, or maternity blues. Yet only about one in six develop symptoms with a duration and magnitude that require treatment for depressive disorder, and one in ten for anxiety disorder.
To what extent does ADHD contribute to the risk of such disorders following childbirth? A Swedish study team used the country’s single-payer health insurance database and other national registers to conduct the first nationwide population study to explore this question.
They used the medical birth register to identify all 420,513 women above 15 years of age who gave birth to their first child, and all 352,534 who gave birth to their second child, between 2005 and 2013. They excluded miscarriages. They then looked for diagnoses of depression and/or anxiety disorders up to a year following childbirth.
In the study population, 3,515 mothers had been diagnosed with ADHD, and the other 769,532 had no such diagnosis.
Following childbirth, depression disorders were five times more prevalent among mothers with ADHD than among their non-ADHD peers. Excluding individuals with a prior history of depression made little difference, lowering the prevalence ratio to just under 5. Among women under 25, the prevalence ratio was still above 3, while for those 25 and older it was above 6.
Similarly, anxiety disorders were over five times more prevalent among mothers with ADHD than among their non-ADHD peers. Once again, excluding individuals with a prior history of depression made little difference, lowering the prevalence ratio to just under 5. Among women under 25, the prevalence ratio was still above 3, while for those 25 and older it was above 6.
The team cautioned, “There is a potential risk of surveillance bias as women diagnosed with ADHD are more likely to have repeated visits to psychiatric care and might have an enhanced likelihood of also being diagnosed with depression and anxiety disorders postpartum, compared to women without ADHD.”
Nevertheless, they concluded, “ADHD is an important risk factor for both depression and anxiety disorders in the postpartum period and should be considered in the post- pregnancy maternal care, regardless of sociodemographic factors and the presence of other psychiatric disorders. Parental education prior to conception, psychological surveillance during, and social support after childbirth should be provided to women diagnosed with ADHD.”
Suicide is one of the most feared outcomes of any psychiatric condition. Although its association with depression is well known, a small but growing research literature shows that ADHD is also a risk factor for suicidality. Suicide is difficult to study. Because it is relatively rare, large samples of patients are needed to make definitive statements.
Studies of suicide and ADHD must also consider the possibility that medications might elevate that risk. For example, the FDA placed a black box warning on atomoxetine because that ADHD medication had been shown to increase suicidal risk in youth. A recent study of 37,936 patients with ADHD now provides much insight into these issues (Chen, Q., Sjolander, A., Runeson, B., D'Onofrio, B. M., Lichtenstein, P. & Larsson, H. (2014). Drug treatment for attention-deficit/hyperactivity disorder and suicidal behavior: a register-based study. BMJ 348, g3769.). In Sweden, such large studies are possible because researchers have computerized medical registers that describe the disorders and treatments of all people in Sweden. Among 37,936 patients with ADHD, 7019 suicide attempts or completed suicides occurred during 150,721 person-years of follow-up. This indicates that, in any given year, the risk for a suicidal event is about 5%. For ADHD patients, the risk for a suicide event is about 30% greater than for non-ADHD patients. Among the ADHD patients who attempted or completed suicide, the risk was increased for those who had also been diagnosed with a mood disorder, conduct disorder, substance abuse, or borderline personality. This is not surprising; the most serious and complicated cases of ADHD are those that have the greatest risk for suicidal events. The effects of the medication were less clear. The risk for suicide events was greater for ADHD patients who had been treated with non-stimulant medication compared with those who had not been treated with non-stimulant medication. A similar comparison showed no effect of stimulant medications. This first analysis suffers from the fact that the probability of receiving medication increases with the severity of the disorder. To address this problem, the researchers limited the analyses to ADHD patients who had some medication treatment and then compared suicidal risk between periods of medication treatment and periods of no medication treatment. This analysis found no increased risk for suicide from non-stimulant medications and, more importantly, found that for patients treated with stimulants, the risk for suicide was lower when they were taking stimulant medications. This protective effect of stimulant medication provides further evidence of the long-term effects of stimulant medications, which have also been shown to lower the risks for traffic accidents, criminality, smoking, and other substance use disorders.
A Chinese research team performed two types of meta-analyses to compare the risk of suicide for ADHD patients taking ADHD medication as opposed to those not taking medication.
The first type of meta-analysis combined six large population studies with a total of over 4.7 million participants. These were located on three continents - Europe, Asia, and North America - and more specifically Sweden, England, Taiwan, and the United States.
The risk of suicide among those taking medication was found to be about a quarter less than for unmediated individuals, though the results were barely significant at the 95 percent confidence level (p = 0.49, just a sliver below the p = 0.5 cutoff point). There were no significant differences between males and females, except that looking only at males or females reduced sample size and made results non-significant.
Differentiating between patients receiving stimulant and non-stimulant medications produced divergent outcomes. A meta-analysis of four population studies covering almost 900,000 individuals found stimulant medications to be associated with a 28 percent reduced risk of suicide. On the other hand, a meta-analysis of three studies with over 62,000 individuals found no significant difference in suicide risk for non-stimulant medications. The benefit, therefore, seems limited to stimulant medication.
The second type of meta-analysis combined three within-individual studies with over 3.9 million persons in the United States, China, and Sweden. The risk of suicide among those taking medication was found to be almost a third less than for unmediated individuals, though the results were again barely significant at the 95 percent confidence level (p =0.49, just a sliver below the p = 0.5 cutoff point). Once again, there were no significant differences between males and females, except that looking only at males or females reduced the sample size and made results non-significant.
Differentiating between patients receiving stimulant and non-stimulant medications once again produced divergent outcomes. Meta-analysis of the same three studies found a 25 percent reduced risk of suicide among those taking stimulant medications. But as in the population studies, a meta-analysis of two studies with over 3.9 million persons found no reduction in risk among those taking non-stimulant medications.
A further meta-analysis of two studies with 3.9 million persons found no reduction in suicide risk among persons taking ADHD medications for 90 days or less, "revealing the importance of duration and adherence to medication in all individuals prescribed stimulants for ADHD."
The authors concluded, "exposure to non-stimulants is not associated with a higher risk of suicide attempts. However, a lower risk of suicide attempts was observed for stimulant drugs. However, the results must be interpreted with caution due to the evidence of heterogeneity ..."
Stimulant medications have long been considered the default first-line treatment for attention-deficit/hyperactivity disorder (ADHD). Clinical guidelines, prescribing practices, and public narratives all reinforce the idea that stimulants should be tried first, with non-stimulants reserved for cases where stimulants fail or are poorly tolerated.
I recently partnered with leading ADHD researcher Jeffrey Newcorn for a Nature Mental Health commentary on the subject. We argue that this hierarchy deserves reexamination. It is important to note that our position is not anti-stimulant. Rather, we call into question whether the evidence truly supports treating non-stimulants as secondary options, and we propose that both classes should be considered equal first-line treatments.
Stimulants have earned their reputation as the go-to drug of choice for ADHD. They are among the most effective medications in psychiatry, reliably reducing core ADHD symptoms and improving daily functioning when properly titrated and monitored. However, when stimulant and non-stimulant medications are compared more closely, the gap between them appears smaller than commonly assumed.
Meta-analyses often report slightly higher average response rates for stimulants, but head-to-head trials where patients are directly randomized to one medication versus another frequently find no statistically significant differences in symptom improvement or tolerability. Network meta-analyses similarly show that while some stimulant formulations have modest advantages, these differences are small and inconsistent, particularly in adults.
When translated into clinical terms, the advantage of stimulants becomes even more modest. Based on existing data, approximately eight patients would need to be treated with a stimulant rather than a non-stimulant for one additional person to experience a meaningful benefit. This corresponds to only a 56% probability that a given patient will respond better to a stimulant than to a non-stimulant. This difference is not what we would refer to as “clinically significant.”
One reason non-stimulants may appear less effective is the way efficacy is typically reported. Most comparisons rely on standardized mean differences, a method of averages that may mask heterogeneity of treatment effects. In reality, ADHD medications do not work uniformly across patients.
For example, evidence suggests that response to some non-stimulants, such as atomoxetine, is bimodal: this means that many patients respond extremely well, while others respond poorly, with few in between. When this happens, average effect sizes can obscure the fact that a substantial subgroup benefits just as much as they would from a stimulant. In other words, non-stimulants are not necessarily less effective across the board, but that they are simply different in who they help.
In our commentary, we also highlight structural issues in ADHD research. Stimulant trials are particularly vulnerable to unblinding, as their immediate and observable physiological effects can reveal treatment assignment, potentially inflating perceived efficacy. Non-stimulants, with slower onset and subtler effects, are less prone to this bias.
Additionally, many randomized trials exclude patients with common psychiatric comorbidities such as anxiety, depression, or substance-use disorders. Using co-diagnoses as exclusion criteria for clinical trials on ADHD medications is nonviable when considering the large number of ADHD patients who also have other diagnoses. Real-world data suggest that a large proportion of individuals with ADHD would not qualify for typical trials, limiting how well results generalize to everyday clinical practice.
Standard evaluations of medication tolerability focus on side effects experienced by patients, but this narrow lens misses broader societal consequences. Stimulants are Schedule II controlled substances, which introduces logistical barriers, regulatory burdens, supply vulnerabilities, and administrative strain for both patients and clinicians.
When used as directed, stimulant medications do not increase risk of substance-use disorders (and, in fact, tend to reduce these rates); however, as ADHD awareness has spread and stimulants are more widely prescribed, non-medical use of prescription stimulants has become more widespread, particularly among adolescents and young adults. Non-stimulants do not carry these risks.
Non-stimulants are not without drawbacks themselves, however. They typically take longer to work and have higher non-response rates, making them less suitable in situations where rapid results are essential. These limitations, however, do not justify relegating them to second-line status across the board.
This is a call for abandoning a one-size-fits-all approach. Instead, future guidelines should present stimulant and non-stimulant medications as equally valid starting points, clearly outlining trade-offs related to onset, efficacy, misuse risk, and practical burden.
The evidence already supports this shift. The remaining challenge is aligning clinical practice and policy with what the data, and patient-centered care, are increasingly telling us.
Today, most treatment guidelines recommend starting ADHD treatment with stimulant medications. These medicines often work quickly and can be very effective, but they do not help every child, and they can have bothersome side effects, such as appetite loss, sleep problems, or mood changes. Families also worry about long-term effects, the possibility of misuse or abuse, as well as the recent nationwide stimulant shortages. Non-stimulant medications are available, but they are usually used only after stimulants have not been effective.
This stimulant-first approach means that many patients who would respond well to a non-stimulant will end up on a stimulant medication anyway. This study addresses this issue by testing two different ways of starting medication treatment for school-age children with attention-deficit/hyperactivity disorder (ADHD). We want to know whether beginning with a non-stimulant medicine can work as well as the “stimulant-first” approach, which is currently used by most prescribers.
From this study, we hope to learn:
Our goal is to give families and clinicians clear, practical evidence to support a truly shared decision: “Given this specific child, should we start with a stimulant or a non-stimulant?”
Who will be in the study?
We will enroll about 1,000 children and adolescents, ages 6 to 16, who:
We will include children with common co-occurring conditions (such as anxiety, depression, learning or developmental disorders) so that the results reflect the “real-world” children seen in clinics, not just highly selected research volunteers.
How will the treatments be assigned?
This is a randomized comparative effectiveness trial, which means:
Parents and clinicians will know which type of medicine the child is taking, as in usual care. However, the experts who rate how much each child has improved using our main outcome measure will not be told which treatment strategy the child received. This helps keep their ratings unbiased.
What will participants be asked to do?
Each family will be followed for 12 months. We will collect information at:
At these times:
We will also track:
Data will be entered into a secure, HIPAA-compliant research database. Study staff at each site will work closely with families to make participation as convenient as possible, including offering flexible visit schedules and electronic options for completing forms when feasible.
How will we analyze the results?
Using standard statistical methods, we will:
All analyses will follow the “intention-to-treat” principle, meaning we compare children based on the strategy they were originally assigned to, even if their medication is later changed. This mirrors real-world decision-making: once you choose a starting strategy, what tends to happen over time?
Why is this study necessary now?
This study addresses a critical, timely gap in ADHD care:
In short, this study is needed now to move ADHD medication decisions beyond “one-size-fits-all.” By rigorously comparing stimulant-first and non-stimulant-first strategies in real-world settings, and by focusing on what matters most to children and families overall functioning, side effects, and long-term well-being, we aim to give patients, parents, and clinicians the information they need to choose the best starting treatment for each child.
This project was conceived by Professor Stephen V. Faraone, PhD (SUNY Upstate Medical University, Department of Psychiatry, Syracuse, NY) and Professor Jeffrey H. Newcorn, MD (Icahn School of Medicine at Mount Sinai, Department of Psychiatry, New York, NY). It will be conducted at nine sites across the USA.
EBI-ADHD:
If you live with ADHD, treat ADHD, or write about ADHD, you’ve probably run into the same problem: there’s a ton of research on treatments, but it’s scattered across hundreds of papers that don’t talk to each other. The EBI-ADHD website fixes that.
EBI-ADHD (Evidence-Based Interventions for ADHD) is a free, interactive platform that pulls together the best available research on how ADHD treatments work and how safe they are. It’s built for clinicians, people with ADHD and their families, and guideline developers who need clear, comparable information rather than a pile of PDFs. EBI-ADHD Database The site is powered by 200+ meta-analyses covering 50,000+ participants and more than 30 different interventions. These include medications, psychological therapies, brain-stimulation approaches, and lifestyle or “complementary” options.
The heart of the site is an interactive dashboard. You can:
The dashboard then shows an evidence matrix: a table where each cell is a specific treatment–outcome–time-point combination. Each cell tells you two things at a glance:
Clicking a cell opens more detail: effect sizes, the underlying meta-analysis, and how the certainty rating was decided.
EBI-ADHD is not just a curated list of papers. It’s built on a formal umbrella review of ADHD interventions, published in The BMJ in 2025. That review re-analyzed 221 meta-analyses using a standardized statistical pipeline and rating system.
The platform was co-created with 100+ clinicians and 100+ people with lived ADHD experience from around 30 countries and follows the broader U-REACH framework for turning complex evidence into accessible digital tools.
Why it Matters
ADHD is one of the most studied conditions in mental health, yet decisions in everyday practice are still often driven by habit, marketing, or selective reading of the literature. EBI-ADHD offers something different: a transparent, continuously updated map of what we actually know about ADHD treatments and how sure we are about it.
In short, it’s a tool to move conversations about ADHD care from “I heard this works” to “Here’s what the best current evidence shows, and let’s decide together what matters most for you.”
_logo%201.svg)
