Two new meta-analyses evaluate digital interventions for treating ADHD

Two New Meta-analyses Evaluate Digital Interventions for Treating ADHD

There is increasing interest in digital interventions to treat ADHD symptoms and to overcome deficits in executive functioning that are associated with this disorder. Executive functions such as working memory and cognitive speed originate in the frontal lobes of the brain, and guide voluntary goal-directed behavior.

There is increasing interest in digital interventions to treat ADHD symptoms and to overcome deficits in executive functioning that are associated with this disorder. Executive functions such as working memory and cognitive speed originate in the frontal lobes of the brain, and guide voluntary goal-directed behavior. They affect reading speed and accuracy, reading comprehension, attention, and impulse control, among other behaviors important to the ability to function in social, educational, and professional environments.

A Swedish study team based at Umeå University recently conducted a systematic search of the medical literature to explore the efficacy of computerized cognitive training (CCT) to improve executive functioning in adults with ADHD.

They included published randomized controlled trials (RCTs) involving adults 18 to 65 years old with a primary diagnosis of ADHD. The controls were participants with either a passive (wait-list) or active (modified simple training) intervention.

Nine RCTs with a combined total of 285 participants met inclusion criteria. Lumping together all cognitive outcome types, meta-analysis reported a small effect size improvement that was just barely statistically significant (p = .048, with p < .05 as the boundary).

However, when separated out by individual outcome types – executive functioning, cognitive speed, general short-term memory, or ADHD symptom severity – the meta-analyses found no improvements that reached statistical significance. 

Moreover, all RCTs except one were judged as high risk of bias.

While it is possible that additional studies enlarging the pool of participants could lead to statistical significance, all effect sizes were small to begin with, which is not encouraging.

The team concluded, “Considering the small positive effect in this meta-analysis for overall cognitive outcomes, together with the lack of evidence for far transfer, practitioners and individuals with ADHD should weigh the costs (resources and time) against the benefits of training.”

A South Korean study team recently concluded the first RCT-only meta-analysis of game-based digital therapeutics (DTx).

Combining 14 RCTs with a total of 1,183 participants, they found a small effect size improvement in parent-rated attention symptoms for game-based DTx interventions over controls. Nine RCTs combining 424 participants likewise found a small effect size improvement in teacher-rated attention symptoms. Between-study variation (heterogeneity) was negligible, and there was no indication of publication bias.

Combining five RCTs with a total of 256 participants, they reported small effect size improvements in both parent and teacher-rated hyperactivity/impulsivity symptoms. But they found no improvement in hyperactivity symptoms alone, whether evaluated by parents or teachers. Heterogeneity was in all instances negligible, with no sign of publication bias. 

The team then compared game-based DTx interventions with pharmaceutical treatment. 

ADHD medications outperformed game-based DTx interventions for improvement of attention symptoms in both parent (four RCTs with a total of 128 participants) and teacher (three RCTs with 92 participants) ratings, with small-to-medium effect sizes. Medications likewise prevailed in improving hyperactivity/impulsivity symptoms, whether rated by parents or teachers, with small-to-medium effect sizes.

The team concluded, “This study is the first direct and indirect meta-analysis to compare the efficacy of game-based DTx between control and medication according to the assessor in an RCT. In conclusion, game-based DTx had a more significant effect than the control. Additionally, between medication treatment versus DTx, medication was more effective.”

March 15, 2024

Meta-analysis finds methylphenidate, atomoxetine, and guanfacine use is associated with increased occurrence of headaches, but amphetamine use is not

Meta-analysis Finds Certain ADHD Meds Associated with Headaches, But Others Are Not

An international study team has just reported findings from a series of meta-analyses exploring associations between ADHD medications and headaches in children and adolescents.

An international study team has just reported findings from a series of meta-analyses exploring associations between ADHD medications and headaches in children and adolescents. 

First, to compare headache occurrence in individuals with ADHD to those without ADHD, the team performed a very large meta-analysis of twelve studies with over 2.7 million children and adolescents. Those with ADHD had twice the rate of headaches. 

There was no indication of publication bias, but there was considerable variation (heterogeneity) among studies, with crude odds ratios spanning from 0.9 to 3.37. Nevertheless, ten of the twelve studies pointed to higher odds among children and adolescents with ADHD. The four studies that controlled for age, sex, race, and other socioeconomic status variables reaffirmed the finding of a doubling of headache risk, this time with acceptable heterogeneity.

Three studies with a combined 7,755 participants found no difference in tension headaches, but five studies with over a quarter million persons found more than a doubling of the rate of migraine in children and adolescents with ADHD.

Next, the team performed meta-analyses of 58 randomized controlled trials (RCTs) of specific ADHD medications that met eligibility criteria for their systematic review. Because only a single eligible RCT apiece looked at bupropion and clonidine, these ADHD medications could not be included in the meta-analyses.

A meta-analysis of ten RCTs with a total of 2,672 participants found absolutely no association between use of amphetamines (including lisdexamphetamine) and headaches. Variation (heterogeneity) between studies was minimal, and there was no sign of publication bias.

A smaller meta-analysis of six RCTs with a combined 818 participants found a 24% increase in headaches among modafinil users, but it was not statistically significant, perhaps because of the much smaller combined sample size.

A meta-analysis of 17 RCTs with a total of 3,371 participants found a 33% increase in headache occurrence among methylphenidate users over placebo. Between-study variation (heterogeneity) was negligible, and there was absolutely no sign of publication bias. 

Similarly, a meta-analysis of 22 RCTs with a combined 3,857 participants reported a 29% increase in headache occurrence among atomoxetine users over placebo. Again, heterogeneity between studies was negligible, with absolutely no indication of publication bias.

Finally, a meta-analysis of eight RCTs with 1,956 participants found a 43% increase in headache occurrence among guanfacine users over placebo. Once again, with negligible heterogeneity and no indication of publication bias.

March 18, 2024

Meta-analysis finds improvements in executive functioning in children and adolescents from non-pharmacological treatments, but with methodological shortcomings

Meta-analysis Finds Improvements in Executive Functioning From Some Non-Pharmacological ADHD Treatments

ADHD is associated with impaired executive functioning. Executive functioning is a set of mental skills that include working memory, flexible thinking, and self-control. These are skills we use every day to learn, work, and manage daily life. Trouble with executive function can make it hard to focus, follow directions, and handle emotions.

ADHD is associated with impaired executive functioning. Executive functions are a set of mental skills that include working memory, flexible thinking, and self-control. These are skills we use every day to learn, work, and manage daily life. Trouble with executive function can make it hard to focus, follow directions, and handle emotions. 

A Chinese study team searched for studies on non-pharmacological treatments of children and adolescents with ADHD aged 5 to 18 years intended to improve their executive functioning. 

An initial methodological weakness was the decision to combine studies using formal ADHD diagnoses based on professional psychiatric manuals (DSM 3/4/5 and ICD 10/11) and studies relying on other methods such as parent reports.

This lack of rigor in identifying ADHD is surprising given that the team used studies that directly measured executive functioning through neurocognitive tasks, excluding those that relied on parent- or teacher-reported questionnaires. 

67 studies involving 74 training interventions met the criteria. Meta-analysis of all these studies, encompassing a total of 3,101 participants, suggested medium-to-large effect size improvements in executive functioning. There was evidence of publication bias, but trim-and-fill adjustment increased the estimated effect size to large.

Nevertheless, there were further methodological shortcomings:

  • The meta-analysis mixed studies of substantially different interventions: cognitive training, executive function-specific curriculum, game-based training, neurofeedback, mindfulness, and physical exercise.
  • There was tremendous variation (heterogeneity) between study outcomes. Such inconsistency casts doubt on the outcome unless subgroup analysis can explain it. 

In this case, subgroup analysis mostly failed to explain the heterogeneity, with a single exception. Meta-analysis of the 16 studies with 744 participants that explored executive function-specific curriculum found small-to-medium effect size improvements, with no heterogeneity. 

Unfortunately, the team did not perform a separate publication bias analysis on this subgroup, just as it failed to do so on any of the other subgroups.

By far the strongest evidence of benefit came from meta-analysis of the 17 studies with 558 participants evaluating physical exercise. Here the outcome pointed to very large effect size improvements in executive functioning. Yet once again, heterogeneity was extremely high. Breaking this down further between aerobic exercise and cognitively engaged physical exercise made no difference. Both types had the same very high effect size, with very wide heterogeneity. Again, there was no separate evaluation of publication bias on this group.

Meta-analyses of thirteen studies of neurofeedback combining 444 participants, and fifteen studies of cognitive training encompassing 727 participants, both pointed to just-short-of-large effect size improvements in executive function. Meta-analysis of twelve studies of game-based training with 598 participants indicated medium effect size gains. But again, in all three subgroups there was great variation between studies, and no analysis of publication bias.

While these meta-analyses are suggestive of efficacy, especially for physical exercise interventions, their methodological shortcomings mean we will have to await more rigorous meta-analyses to draw any more settled conclusions. Moreover, these meta-analyses did not evaluate the adequacy of the control groups used in the trials, which is a big shortcoming given prior work showing that the effect of non-pharmacologic treatments are very weak or non-existent when adequate controls are used.

March 13, 2024

Meta-Analysis: Is Neurofeedback A Viable Treatment For ADHD?

New meta-analysis of 17 RCTs finds no evidence of efficacy for neurofeedback treatment of ADHD

Neurofeedback, also known as EEG (electroencephalogram)biofeedback, is a treatment that seeks to alleviate symptoms of various neurological and mental health disorders, including ADHD. It does this through immediate feedback from a computer program that tracks a client's brainwave activity, then uses sound or visual signals to retrain these brain signals. This in principle enables patients to learn to regulate and improve their brain function and reduce symptoms.

An Iranian study team recently performed a systematic search of the peer-reviewed medical literature. It identified seventeen randomized-controlled trials (RCTs) of neurofeedback treatment for children and adolescents with ADHD that could be aggregated for meta-analysis.

A meta-analysis of twelve RCTs with a combined total of 740 youths looked at parent ratings of changes in hyperactivity/impulsivity symptoms, and separately of changes in inattention symptoms. In both instances, the net pooled effect centered on zero.

A meta-analysis of nine RCTs with a combined total of 787 youths examined teacher ratings. Once again, the pooled change hyperactivity/impulsivity symptoms centered on zero. For inattention symptoms, the teacher ratings centered on a tiny improvement, but it did not approach statistical significance. The 95% confidence interval stretched well into negative territory.

There was no sign of publication bias. Between-study heterogeneity, on the other hand, was high, with some small sample size RCTs pointing to reduced symptoms, and other small sample size RCTs pointing to increased symptoms. However, the RCTs with the larger sample sizes clustered close around zero effect size.

The authors concluded,"The results provide preliminary evidence that neurofeedback treatment is not an efficacious clinical method for ADHD."

March 23, 2022

Meta-analysis suggests ADHD contributes significantly to alcohol use disorder in persons with comorbid bipolar disorder

Meta-analysis suggests ADHD contributes significantly to alcohol use disorder in persons with comorbid bipolar disorder

Bipolar disorder is a severe mental illness that afflicts over one in fifty persons worldwide. About a quarter of those with bipolar disorder also has alcohol use disorder (AUD). This in turn complicates the treatment of their bipolar disorder. It exacerbates their symptoms, makes them more likely to be suicidal, and increases the risk of hospitalization.

More than one in five persons with bipolar disorder also have ADHD, which is likewise known to be correlated with AUD. To what extent does ADHD contribute to AUD in persons with comorbid bipolar disorder?

A European study team recently conducted a systematic search of the peer-reviewed medical literature to address that question. The team identified eleven studies with a combined total of 2,734 participants that could be aggregated to perform a meta-analysis.

They found that persons with comorbid ADHD and bipolar disorder were two and a half times more likely to be diagnosed with alcohol use disorder than persons with bipolar disorder but no ADHD.

Between-study heterogeneity was negligible, and there was no sign of publication bias.

The authors concluded, "At least a portion of the high rates of AUD in BD may, thereby, be related to comorbid ADHD. Longitudinal studies are needed to clarify the nature of this relationship."

March 13, 2022

Association Found: Maternal PCOS and ADHD

Meta-analysis with over 1.3 million participants finds clear association between maternal polycystic ovary syndrome and ADHD in offspring

Meta-analysis discovers clear link between mothers with PCOS and children with ADHD.

Polycystic ovary syndrome (PCOS) affects somewhere between 6 and 20% of women of reproductive age. Typical effects include:

·  failure to ovulate;

·      high levels of male hormones (androgens), which can lead to acne, seborrhea, hair loss on the scalp, increased body or facial hair, and infrequent or absent menstruation;

·       metabolic disruption, including obesity and insulin resistance.

In pregnancy, PCOS is also known to increase the chances of birth complications.

Previous studies have suggested a link between maternal PCOS and ADHD.

A team of Arabian (Saudi and United Arab Emirates) researchers conducted a systematic review of the peer-reviewed medical literature and were able to identify four studies with a total of 1,354,182 participants that could be combined into a meta-analysis.

The meta-analysis found that children born to mothers with PCOS were 43% more likely to develop ADHD. The 95% confidence interval stretched from 35% to 51%, indicating a highly reliable finding.

Moreover, there was between-study variation: They all produced essentially identical results. There was also no sign of publication bias.

"However,"the authors noted, "the reported results do not necessarily provide definitive findings of a causal inference due to the randomized study design. All the included studies were observational in design." With this caution, they could only conclude that "the results of this meta-analysis showed that there might be a link between maternal PCOS and the risk of developing ASD and ADHD in the offspring."

March 11, 2022

Meta-analysis: No Association Between Childhood Antibiotic Use and ADHD

Meta-analysis finds no association between childhood antibiotic exposure and ADHD

Childhood antibiotic use is not found to be associated with development of ADHD

A Chinese research team recently conducted a systematic search of the peer-reviewed medical journal literature for studies exploring the association between childhood antibiotic exposure and subsequent diagnosis of ADHD in youths 18 years and younger.

A meta-analysis of six studies with a combined total of over 1.5 million participants found that children exposed to antibiotics were 18% more likely to later be diagnosed with ADHD.

There was absolutely no indication of publication bias. Between-study heterogeneity, on the other hand, was extremely high.

With such large cohorts, one can often tease out whether an association is causal, or due to genetic and familial confounding, by looking at matched close relatives.

Three of the studies, with a combined total of well over half a million participants, also compared matched siblings.

Significantly, the meta-analysis among matched siblings found no association whatsoever between childhood exposure to antibiotics and subsequent ADHD. Between-study heterogeneity was virtually nonexistent.

The team concluded, "Our meta-analysis indicated that early-life antibiotic exposure was associated with a subsequent increased risk of ASD or ADHD. However, such association was not found in the sibling-matched analysis, indicating that genetic and familial confounding factors may largely explain the observed association."

March 3, 2022

Meta-Analysis Finds No Significant Benefit For ADHD Patients in tCDS

New meta-analysis finds no significant gains from transcranial direct stimulation (tCDS)

Based on current findings, repetitive transcranial magnetic stimulation cannot yet be officially recommended as an alternative neurotherapy for ADHD.

Noting that "despite a lack of solid evidence for their use, rTMS [repetitive transcranial magnetic stimulation]and tDCS [transcranial direct current stimulation] are already offered clinically and commercially in ADHD," and that a recent meta-analysis of ten tDCS studies found small but significant improvements in outcomes, but had several methodological shortcomings and did not include two studies reporting mostly null effects, a team of British neurologists performed a meta-analysis of all twelve sham-controlled, non-open-label, studies found in a comprehensive search of the peer-reviewed literature.

Ten of the twelve randomized-controlled trials used anodal stimulation of the dorsolateral prefrontal cortex, while the other two used anodal stimulation of the right inferior frontal cortex.

The trials explored several measures of cognition. The research team carried out a meta-analysis of all twelve trials, with a total of 232 participants, and found no significant improvement in attention scores from CDC, relative to sham stimulation. A second meta-analysis, of eleven trials with a total of 220 participants, assessed the efficacy of tDCS on improving inhibition scores, and again found no significant effect. A third meta-analysis, encompassing eight trials with a total of 124 participants, evaluated the efficacy of tDCS on improving processing speed scores, once again finding no significant effect.

The latter two meta-analyses approached the border of significance, prompting the authors to speculate that larger sample sizes could bring the results just over the threshold of significance. Even so, effect sizes would be small.

It is also possible that the trials focused on regions of the brain suboptimal for this objective, and thus the authors "cannot rule out the possibility that stimulation of other prefrontal regions (such as the right hemispheric inferior frontal cortex or dorsolateral prefrontal cortex or parietal regions), multiple session tDCS or tDCS in combination with cognitive training could improve clinically or cognitive functions in ADHD."

As to concerns about safety, on the other hand, "stimulation was well-tolerated overall."

The authors concluded that based on current evidence, tDCS of the dorsolateral prefrontal cortex cannot yet be recommended as an alternative Neurotherapy for ADHD.

February 15, 2022

Meta-analysis finds mindfulness-based interventions reduce ADHD symptoms in adults, but no better than active psychological controls

Meta-analysis finds mindfulness-based interventions reduce ADHD symptoms in adults, but no better than active psychological controls

According to Dexing Zhang et al., writing in the British Medical Bulletin, "Mindfulness is a moment-by-moment awareness of thoughts, feelings, bodily sensations, and surrounding environment. ... These practices can be formal (e.g. breathing, sitting, walking, body scan) or informal (e.g. mindfulness in everyday life).... Mindfulness is rooted in Buddhist traditions. However, it has become popular in recent years among various secular populations in healthcare, educational, and workplace settings: from pre-schoolchildren to older adults across the world." The two most widely adopted mindfulness-based interventions (MBIs) are mindfulness-based stress reduction and mindfulness-based cognitive therapy (Zhang, 2021).

An Italian research team recently conducted a comprehensive search of the peer-reviewed literature to identify studies exploring the efficacy of mindfulness-based treatments for ADHD. It found 31 studies that qualified for review, ten of which met the criteria for meta-analysis, with a total of 596 participants.

A meta-analysis of seven studies with a combined total of 489 participants found MBIs reduced ADHD symptoms with medium effect size and no sign of publication bias. When split into subgroups with and without active controls - in this case, psychoeducation and skills training groups -the outcomes diverged. In the three studies with non-active controls (187 participants), there was a large reduction in ADHD symptoms. In the four with active controls (302 participants), there was no significant difference.

A meta-analysis of ten studies with 596 participants found MBIs reduced inattention symptoms, with a medium-sized effect. Pooling the five studies without active controls (261 participants) produced a very large reduction in inattention symptoms. Once again, in the five studies with active controls (335 participants), there was no significant difference.

After adjusting for publication bias, a third meta-analysis of nine studies with 563 participants found no significant effect of MBIs hyperactivity symptoms. However, when limited to the five studies with-active controls (261 participants), it found a large reduction in hyperactivity symptoms.

After adjusting for publication bias, the fourth meta-analysis of four studies with a combined 243 participants found no significant improvement in executive function.

After adjusting for publication bias, a fifth meta-analysis combining six studies with 449 participants reported a moderate improvement in mindfulness skills. There was no significant improvement when looking only at the three studies with active controls (262 participants).

The team concluded that MBIs seemed to be effective in treating ADHD, but no more so than psychoeducation and skills training groups.

Yet they cautioned that the use of a waiting list for non-active controls muddies that conclusion: "It could be suggested that any intervention seems to have a significantly higher effect than WL [waiting list]in improving ADHD symptoms." This is a known hazard of using waiting lists as control groups (Cunningham, 2013).

Noting "the low general methodological quality," they stated, "From a clinical standpoint, according to the poor available evidence, we cannot conclude that MBIs are superior to other active [psychological] interventions in ameliorating all the considered outcomes, suggesting a role complementation and not as a replacement of the psychoeducation in the management of patients with ADHD, consistently with some guidelines' recommendations."

December 27, 2021

Can Neurofeedback Improve Executive Functions? New Meta-Analysis

New meta-analysis finds no improvement in executive functions from neurofeedback in children and adolescents

Meta-analysis show neurofeedback treatments resulted in no noticeable improvements in the working memory, response inhibition, or sustained attention of youth with ADHD.

Neurofeedback, also known as electroencephalogram (EEG) biofeedback, aims to help persons with ADHD train themselves to self-regulate patterns of brain activity associated with the disorder.

An example is theta-beta ratio frequency (TBR) training. Beta waves, with a frequency of 18 to 25 Hz, are associated with electrical activity when the brain is conscious or alert. Theta waves, with a frequency of 4 to 7 Hz, are associated with meditative, daydreaming, or drowsy states. In youths with ADHD, the theta to beta ratio tends to be elevated. TBR training seeks to reduce it.

Neurofeedback is often described as a promising emerging alternative or complement to pharmaceutical treatment. Previous meta-analyses have found neurofeedback can reduce symptoms of ADHD.

But what effect does it have on executive functions? A Thai research team based at Chiang Mai University conducted a comprehensive search of the peer-reviewed journal literature and identified ten studies with results suitable for meta-analysis.

A meta-analysis of all ten studies with a combined total, of 378 participants found no improvement whatsoever in response inhibition.

A second meta-analysis, of nine studies with a combined total of 349 participants, found no improvement in sustained attention.

Finally, a meta-analysis of three studies with a total of 121 participants likewise found no improvement in working memory.

In all three cases, there was no evidence of publication bias.

The authors concluded, "Results did not show the benefits of neurofeedback on executive functions assessed by neuropsychological tests."

December 15, 2021

What is Evidenced-Based Medicine?

What is Evidenced-Based Medicine?

With the growth of the Internet, we are flooded with information about attention deficit hyperactivity disorder from many sources, most of which aim to provide useful and compelling "facts" about the disorder.  But, for the cautious reader, separating fact from opinion can be difficult when writers have not spelled out how they have come to decide that the information they present is factual. 

My blog has several guidelines to reassure readers that the information they read about ADHD is up-to-date and dependable. They are as follows:

Nearly all the information presented is based on peer-reviewed publications in the scientific literature about ADHD. "Peer-reviewed" means that other scientists read the article and made suggestions for changes and approved that it was of sufficient quality for publication. I say "nearly all" because in some cases I've used books or other information published by colleagues who have a reputation for high-quality science.

When expressing certainty about putative facts, I am guided by the principles of evidence-based medicine, which recognizes that the degree to which we can be certain about the truth of scientific statements depends on several features of the scientific papers used to justify the statements, such as the number of studies available and the quality of the individual studies. For example, compare these two types of studies.  One study gives drug X to 10 ADHD patients and reported that 7 improved.  Another gave drug Y to 100 patients and a placebo to 100 other patients and used statistics to show that the rate of improvement was significantly greater in the drug-treated group. The second study is much better and much larger, so we should be more confident in its conclusions. The rules of evidence are fairly complex and can be viewed at the Oxford Center for Evidenced Based Medicine (OCEBM;http://www.cebm.net/).


The evidenced-based approach incorporates two types of information: a) the quality of the evidence and b) the magnitude of the treatment effect. The OCEBM levels of evidence quality are defined as follows (higher numbers are better:

  1. Mechanism-based reasoning.  For example, some data suggest that oxidative stress leads to ADHD, and we know that omega-3 fatty acids reduce oxidative stress. So there is a reasonable mechanism whereby omega-3 therapy might help ADHD people.
  2. Studies of one or a few people without a control group, or studies that compare treated patients to those that were not treated in the past.

Non-randomized, controlled studies.    In these studies, the treatment group is compared to a group that receives a placebo treatment, which is a fake treatment not expected to work.  

  1. Non-randomized means that the comparison might be confounded by having placed different types of patients in the treatment and control groups.
  2. A single randomized trial.  This type of study is not confounded.
  3. Systematic review and meta-analysis of randomized trials. This means that many randomized trials have been completed and someone has combined them to reach a more accurate conclusion.

It is possible to have high-quality evidence proving that a treatment works but the treatment might not work very well. So it is important to consider the magnitude of the treatment effect, also called the "effect size" by statisticians. For ADHD, it is easiest to think about ranking treatments on a ten-point scale. The stimulant medications have a quality rating of 5 and also have the strongest magnitude of effect, about 9 or 10.Omega-3 fatty acid supplementation 'works' with a quality rating of 5, but the score for the magnitude of the effect is only 2, so it doesn't work very well. We have to take into account patient or parent preferences, comorbid conditions, prior response to treatment, and other issues when choosing a treatment for a specific patient, but we can only use an evidence-based approach when deciding which treatments are well-supported as helpful for a disorder.

April 23, 2021