April 12, 2022

ADHD is underdiagnosed, to varying degrees, among adults of different ethnicities, ages, and education levels in the U.S.

A cohort study looked at over five million adults and over 850,000 children between the ages of five and eleven who received care at Kaiser Permanente Northern California during the ten-year period from the beginning of 2007 through the end of 2016. At any given time, KPNC serves roughly four million persons. It is representative of the population of the region, except for the highest and lowest income strata.

ADHD Diagnosis Rates:

  • Adults: Diagnosis rates rose from 0.43% in 2007 to 0.96% in 2016.
  • Children: Diagnosis rates went up from 2.96% to 3.74%, nearly four times higher than in adults.

Diagnosis Rates by Ethnicity:

  • Non-Hispanic whites had the highest adult diagnosis rates, increasing from 0.67% to 1.42%.
  • American Indian/Alaska Native (AIAN): Rates grew from 0.56% to 1.14%.
  • Black and Hispanic adults had similar rates: Black adults increased from 0.22% to 0.69%, and Hispanic adults rose from 0.25% to 0.65%.
  • Asian adults had the lowest rates (0.11% to 0.35%), followed by Native Hawaiian/Pacific Islanders (0.11% to 0.39%).

ADHD Diagnosis and Age:

The likelihood of being diagnosed with ADHD dropped sharply with age.

(When compared to 18-24-year-olds):

  • 25-34-year-olds were 16% less likely.
  • 35-44-year-olds were 33% less likely.
  • 45-54-year-olds were less than half as likely.
  • 55-64-year-olds were less than a quarter as likely.
  • Adults over 65 were about 5% as likely.

This matches findings from other studies showing that ADHD diagnoses become less common with age.

Other Factors:

  • Adults with higher education levels were twice as likely to be diagnosed as those with less education.
  • Household income had little effect on diagnosis rates.
  • Women were slightly less likely to be diagnosed than men.

ADHD and Comorbidity:

Adults with ADHD were more likely to have other mental health conditions:

  • Eating disorders: 5 times more likely.
  • Bipolar disorder or depression: Over 4 times more likely.
  • Anxiety: More than twice as likely.
  • Substance abuse: Slightly more likely.

Key Findings:
  1. Rising ADHD Diagnosis Rates: The increase in diagnoses may be due to better recognition of ADHD by doctors and greater public awareness during the study period.
  2. Differences by Ethnicity: The differences in diagnosis rates by ethnicity could be related to access to healthcare, cultural attitudes toward mental health, or even attempts to obtain ADHD medications for non-medical reasons, which may be more common among white patients.
Conclusion:

The authors speculate that rising rates of diagnosis “could reflect increasing recognition of ADHD in adults by physicians and other clinicians as well as growing public awareness of ADHD during the decade under study.” Turning to the notable differences by ethnicity, they note, “Racial/ethnic differences could also reflect differential rates of treatment-seeking or access to care. … Racial/ethnic background is known to play an important role in opinions on mental health services, health care utilization, and physician preferences. In addition, rates of diagnosis- seeking to obtain stimulant medication for non-medical use may be more common among white vs nonwhite patients.” They conclude, “greater consideration must be placed on cultural influences on health care seeking and delivery, along with an increased understanding of the various social, psychological, and biological differences among races/ethnicities as well as culturally sensitive approaches to identify and treat ADHD in the total population.”

The study highlights that many cases of adult ADHD go undiagnosed. Research shows about 3% of adults worldwide have ADHD, but this study found that less than 1% are diagnosed by doctors. This points to the need for better training for clinicians to recognize, diagnose, and treat ADHD in adults. It also emphasizes the importance of understanding cultural factors that affect how people seek and receive care.

Winston Chung, MD, MS; Sheng-Fang Jiang, MS; Diana Paksarian, MPH, PhD; Aki Nikolaidis, PhD; F. Xavier Castellanos, MD; Kathleen R. Merikangas, PhD; Michael P. Milham, MD, PhD, “Trends in the Prevalence and Incidence of Attention-Deficit/Hyperactivity Disorder Among Adults and Children of Different Racial and Ethnic Groups,” JAMA Network Open (2019) 2(11): e1914344. DOI:10.1521/adhd.2019.27.4.8.

Related posts

Variations in Diagnosis

Variations in Diagnosis

A cohort study looked at over five million adults, and over 850,000 children between the ages of five and eleven, who received care at Kaiser Permanente Northern California during the ten-year period from the beginning of 2007 through the end of 2016. At any given time, KPNC serves roughly four million persons. It is representative of the population of the region, except for the highest and lowest income strata.

Among adults rates of ADHD diagnosis rose from 0.43% to 0.96%. Among children the diagnosis rates rose from 2.96% to 3.74%, ending up almost four times as high as for adults.

Non-Hispanic whites had the highest adult rates throughout, increasing from 0.67% in 2007 to 1.42% in 2016. American Indian or Alaska Native (AIAN) had the second highest rates, rising from 0.56% to 1.14%. Blacks and Hispanics had roughly comparable rates of diagnosis, the former rising from 0.22% to 0.69%, the latter from 0.25% to 0.65%. The lowest rates were among Asians (rising from 0.11% to 0.35%) and Native Hawaiian or other Pacific Islanders (increasing from 0.11% to 0.39%).

Odds of diagnosis dropped steeply with age among adults. Relative to 18-24-year-olds, 25-34-year-olds were 1/6th less likely to be diagnosed with ADHD, 35-44-year-olds 1/3rd less likely, 45-54-year-olds less than half as likely, 55-64-year-olds less than a quarter as likely, and those over 65 about a twentieth as likely. This is consistent with other studies reporting and age dependent decline in the diagnosis.

Adults with the highest levels of education were twice as likely to be diagnosed as those with the lowest levels. But variations in median household income had almost no effect. Women were marginally less likely to be diagnosed than men.

ADHD is associated with some other psychiatric disorders. Compared with normally developing adults, and adjusted for confounders, those with ADHD were five times as likely to have an eating disorder, over four times as likely to be diagnosed with bipolar disorder or depression, more than twice as likely to suffer from anxiety, but only slightly more likely to abuse drugs or alcohol.

The authors speculate that rising rates of diagnosis could reflect increasing recognition of ADHD in adults by physicians and other clinicians as well as growing public awareness of ADHD during the decade under study. Turning to the strong differences among ethnicities, they note, Racial/ethnic differences could also reflect differential rates of treatment seeking or access to care. Racial/ethnic background is known to play an important role in opinions on mental health services, health care utilization, and physician preferences. In addition, rates of diagnosis- seeking to obtain stimulant medication for nonmedical use may be more common among white vs nonwhite patients. They conclude, greater consideration must be placed on cultural influences on health care seeking and delivery, along with an increased understanding of the various social, psychological, and biological differences among races/ethnicities as well as culturally sensitive approaches to identify and treat ADHD in the total population.

But the main take home message of this work is that most cases of ADHD in adults are not being diagnosed by clinicians. We know from population studies, worldwide, that about three percent of adults suffer from the disorder. This study found that less than 1 percent are diagnosed by their doctors. Clearly, more education is needed to teach clinicians how to identify, diagnose and treat ADHD in adults.

December 18, 2023

Inequities in ADHD diagnosis in the United States

Inequities in ADHD diagnosis in the United States

A transcontinental study team (California, Texas, Florida) used a nationally representative sample – the 2018 National Survey of Children’s Health – to query 26,205 caregivers of youth aged 3 to 17 years old to explore inequities in ADHD diagnosis.  

With increasing accessibility of the internet in the U.S., more than 80% of adults now search for health information online. Recognizing that search engine data could help clarify patterns of inequity, the team also consulted Google Trends.

The team noted at the outset that “[d]ocumenting the true prevalence of ADHD remains challenging in light of problems of overdiagnosis (e.g., following quick screening rather than full evaluation incorporating multi-informant and multi-method data given limited resources) and underdiagnosis (e.g., reflecting inequities in healthcare and education systems).” Underdiagnosis is known to be influenced by lack or inadequacy of health insurance, inadequate public health funding, stigma, sociocultural expectations in some ethnic groups, and structural racism, among other factors.

After controlling for poverty status, highest education in household, child’s sex, and child’s age, the team reported that Black youth were a quarter (22%) less likely to receive ADHD diagnoses than their white peers. Latino/Hispanic youth were a third (32%) less likely and Asian youth three-quarters (73%) less likely to receive ADHD diagnoses than their white peers.

The team also found that state-level online search interest in ADHD was positively associated with ADHD diagnoses, after controlling for race/ethnicity, poverty status, highest education in household, child’s sex, and child’s age. However, the odds ratio was low (1.01), “suggesting the need for additional evaluation.” Furthermore, “There was no interaction between individual-level racial/ethnic background and state-level information-seeking patterns. … the state-level online information-seeking variation did not affect the odds that youth of color would have a current ADHD diagnosis over and above other included characteristics.” 

That could be due in part to the gap in high-speed broadband access between Black and Hispanic in contrast to white populations, but that would not explain the even larger gaps in diagnosis for Asian youth, who tend to come from more prosperous backgrounds.

The team concluded, “Persistent racial/ethnic inequities warrant systematic changes in policy and clinical care that can attend to the needs of underserved communities. The digital divide adds complexity to persistent racial/ethnic and socioeconomic inequities in ADHD diagnosis …”

Using Video Analysis and Machine Learning in ADHD Diagnosis

NEWS TUESDAY: Machine Learning and The Possible Future of Diagnosing ADHD

Typically, clinicians rely on both subjective and objective observations, patient interviews and questionnaires, as well as reports from family and (in the case of children) parents and teachers, in order to diagnose ADHD. 

A group of researchers are aiming to find a diagnostic test that is purely objective and utilizes recent technological advancements. The method they developed involves analyzing videos of children in outpatient settings, focusing on their movements. The study included 96 children, half of whom had ADHD and half who did not.

How It Works

  1. Video Recording: Children were recorded during their outpatient visits.
  2. Skeleton Detection: Using a tool called OpenPose, the researchers detected and tracked the children's skeletons (essentially a map of their body's movements) in the videos.
  3. Movement Analysis: The researchers analyzed these movements, looking at 11 different movement features. They specifically focused on the angles of different body parts and how much they moved.
  4. Machine Learning: Six different machine learning models were used to see which movement features could best distinguish between children with ADHD and those without.

Key Findings

  • Movement Differences: Children with ADHD showed significantly more movement in all the features analyzed compared to children without ADHD.
  • Thigh Angle: The angle of the thigh was the most telling feature. On average, children with ADHD had a thigh angle of about 157.89 degrees, while those without ADHD had an angle of 15.37 degrees.
  • High Accuracy: Using thigh angle alone, the model could diagnose ADHD with 91.03% accuracy. It was very sensitive (90.25%) and specific (91.86%), meaning it correctly identified most children with ADHD and correctly recognized most children without it.

This new method could potentially provide a more objective way to diagnose ADHD, reducing the reliance on subjective observations and reports. It can help doctors make more accurate diagnoses, ensuring that those who need help get it and that those who don't aren't misdiagnosed.

May 28, 2024

Here’s What the Wall Street Journal Got Wrong about the Medication Treatment of ADHD Patients: A Lesson in Science Media Literacy

A recent Wall Street Journal article raised alarms by concluding that many children who start medication for ADHD will later end up on several psychiatric drugs. It’s an emotional topic that will make many parents, teachers, and even doctors worry: “Are we putting kids on a conveyor belt of medications?”

The article seeks to shine a light on the use of more than one psychiatric medication for children with ADHD.   My biggest worry about the article is that it presents itself as a scientific study because they analyzed a database.  It is not a scientific study.  It is a journalistic investigation that does not meet the standards of a scientific report..

The WJS brings attention to several issues that parents and prescribers should think about. It documents that some kids with ADHD are on more than one psychiatric medication, and some are receiving drugs like antipsychotics, which have serious side effects.  Is that appropriate? Access to good therapy, careful evaluation, and follow-up care can be lacking, especially for low-income families.  Can that be improved?  On that level, the article is doing something valuable: it’s shining a spotlight on potential problems.

It is, of course, fine for a journalist to raise questions, but it is not OK for them to pretend that they’ve done a scientific investigation that proves anything. Journalism pretending to be science is both bad science and bad journalism.

Journalism vs. Science: Why Peer Review Matters

Journalists can get big datasets, hire data journalists, and present numbers that look scientific.  But consider the differences between Journalism and Science. These types of articles are usually checked by editors and fact-checkers. Their main goals are:

 Is this fact basically correct?

 Are we being fair?

 Are we avoiding legal problems?

But editors are not qualified to evaluate scientific data analysis methods.  Scientific reports are evaluated by experts who are not part of the project.  They ask tough questions like: 

Exactly how did you define ADHD? 

How did you handle missing data? 

Did you address confounding? 

Did you confuse correlation with causation?

If the authors of the study cannot address these and other technical issues, the paper is rejected.

The WSJ article has the veneer of science but lacks its methodology.  

Correlation vs. Causation: A Classic Trap

The article’s storyline goes something like this:  A kid starts ADHD medication.  She has additional problems or side effects caused by the ADHD medications.   Because of that, the prescriber adds more drugs.  That leads to the patient being put on several drugs.  Although it is true that some ADHD youth are on multiple drugs, the WSJ is wrong to conclude that the medications for ADHD cause this to occur.  That simply confuses correlation with causation, which only the most naïve scientist would do.

In science, this problem is called confounding. It means other factors (like how severe or complex a child’s condition is) explain the results, not just the thing we’re focused on (medication for ADHD). 

The WSJ analyzed a database of prescriptions.  They did not survey the prescribers who made the prescriptions of the patients who received them.  So they cannot conclude that ADHD medication caused the later prescriptions, or that the later medications were unnecessary or inappropriate. 

Other explanations are very likely.   It has been well documented that youth with ADHD are at high risk for developing other disorders such as anxiety, depression,  and substance use.  The kids in the WSJ database might have developed these disorders and needed several medications.  A peer-reviewed article in a scientific journal would be expected to adjust for other diagnoses. If that is not possible, as it is in the case of the WSJ’s database, a journal would not allow the author to make strong conclusions about cause-and-effect.

Powerful Stories Don’t Always Mean Typical Stories

The article includes emotional accounts of children who seemed harmed by being put on multiple psychiatric drugs.  Strong, emotional stories can make rare events feel common.  They also frighten parents and patients, which might lead some to decline appropriate care. 

These stories matter. They remind us that each data point is a real person.  But these stories are the weakest form of data.  They can raise important questions and lead scientists to design definitive studies, but we cannot use them to draw conclusions about the experiences of other patients.  These stories serve as a warning about the importance of finding a qualified provider,  not as against the use of multiple medications.  That decision should be made by the parent or adult patient based on an informed discussion with the prescriber.

Many children and adults with ADHD benefit from multiple medications. The WSJ does not tell those stories, which creates an unbalanced and misleading presentation.  

Newspapers frequently publish stories that send the message:  “Beware!  Doctors are practicing medicine in a way that will harm you and your family.”   They then use case studies to prove their point.  The title of the article is, itself, emotional clickbait designed to get more readers and advertising revenue.  Don’t be confused by such journalistic trickery.

What Should We Conclude?

Here’s a balanced way to read the article.  It is true that some patients are prescribed more than one medication for mental health problems.  But the article does not tell us whether this prescribing practice is or is not warranted for most patients.  I agree that the use of antipsychotic medications needs careful justification and close monitoring.  I also agree that patients on multiple medications should be monitored closely to see if some of the medications can be eliminated.  Many prescribers do exactly that, but the WSJ did not tell their stories.  

It is not appropriate to conclude that ADHD medications typically cause combined pharmacotherapy or to suggest that combined pharmacotherapy is usually bad. The data presented by the WSJ does not adequately address these concerns.  It does not prove that medications for ADHD cause dangerous medication cascades.

We have to remember that even when a journalist analyzes data, that is not the same as a peer-reviewed scientific study. Journalism pretending to be science is both bad science and bad journalism.

Oppositional Defiant Disorder, Autism, and ADHD: New Research Examines the Connection

Oppositional Defiant Disorder (ODD)—a pattern of chronic irritability, anger, arguing, or defiance—is one of the most challenging behavioral conditions families and clinicians face. 

A new study involving 2,400 children ages 3–17 offers one of the clearest pictures yet. Using parent-reported data from the Pediatric Behavior Scale, researchers compared how often ODD appears in Autism spectrum disorder (ASD), ADHD-Combined presentation (ADHD-C), ADHD-Inattentive presentation (ADHD-I), and those with both ASD and ADHD.

Results

ADHD-Combined + ODD: The Highest-Risk Group

Children with ADHD-Combined presentation show both hyperactivity/impulsivity and inattention.  They had the highest ODD rates of any single diagnosis: 53% of kids with ADHD-Combined met criteria for ODD.

But when autism was added to ADHD-Combined, the prevalence jumped to 62%. This group also had the highest overall ODD scores, suggesting more severe or more impairing symptoms. 

This synergy matters: while autism alone increases ODD risk, the presence of ADHD-Combined is what pushes prevalence into the majority range. Other groups showed lower, but still significant, rates of ODD:

  • Autism + ADHD-Inattentive: 28%
  • Autism Only: 24%
  • ADHD-Inattentive Only: 14%

These findings echo what clinicians often see: children with inattentive ADHD, while struggling significantly with attention and learning, tend to show fewer behavioral conflict patterns than those with hyperactive/impulsive symptoms.

It is important to note that ODD is considered to have two main components. Across all diagnostic groups, ODD consistently broke down into these two components: either Irritable/Angry (emotion-based) or Oppositional/Defiant (behavior-based). But the balance between these components differed depending on diagnosis. Notably, Autism + ADHD-Combined showed higher levels of the irritable/angry component than ADHD-Combined alone. The oppositional/defiant component did not differ much between groups. This suggests that autism elevates the emotional side of ODD more than the behavioral side, which is important for clinicians to note before tailoring interventions.

Understanding ADHD , ASD, & Comorbidity:

The study notes that autism, ADHD, and ODD often cluster together, with 55–90% comorbidity in some combinations.

As the authors explain, The high co-occurrence of ADHD-Combined in autism (80% in our study) largely explains the high prevalence of ODD in autism.” 

Clinical Implications: Why This Study Matters

The researchers point to a straightforward recommendation: clinicians shouldn’t evaluate these conditions in isolation. A child referred for autism concerns might also be struggling with ADHD. A child referred for ADHD might have undiagnosed ODD. And ignoring one disorder can undermine treatment for the others.

Evidence-based interventions (behavioral therapy, parent training, school supports, and/or medication) can reduce symptoms across all three diagnoses while improving long-term outcomes, including overall quality of life.

November 21, 2025

What Sleep Patterns Reveal About Mental Health: A Look at New Research

Background:

Sleep is more than simple rest. When discussing sleep, we tend to focus on the quantity rather than the quality,  how many hours of sleep we get versus the quality or depth of sleep. Duration is an important part of the picture, but understanding the stages of sleep and how certain mental health disorders affect those stages is a crucial part of the discussion. 

Sleep is an active mental process where the brain goes through distinct phases of complex electrical rhythms. These phases can be broken down into non-rapid eye movement (NREM) and rapid eye movement (REM). The non-rapid eye movement phase consists of three stages of the four stages of sleep, referred to as N1, N2(light sleep), and N3(deep sleep). N4 is the REM phase, during which time vivid dreaming typically occurs. 

Two of the most important measurable brain rhythms occur during non-rapid eye movement (NREM) sleep. These electrical rhythms are referred to as slow waves and sleep spindles. Slow waves reflect deep, restorative sleep, while spindles are brief bursts of brain activity that support memory and learning.

The Study: 

A new research review has compiled data on how these sleep oscillations differ across psychiatric conditions. The findings suggest that subtle changes in nightly brain rhythms may hold important clues about a range of disorders, from ADHD to schizophrenia.

The Results:

ADHD: Higher Spindle Activity, Mixed Slow-Wave Findings

People with ADHD showed increased slow-spindle activity, meaning those brief bursts of NREM activity were more frequent or stronger than in people without ADHD. Why this happens isn’t fully understood, but it may reflect differences in how the ADHD brain organizes information during sleep. Evidence for slow-wave abnormalities was mixed, suggesting that deep sleep disruption is not a consistent hallmark of ADHD.

Autism: Inconsistent Patterns, but Some Signs of Lower Sleep Amplitude

Among individuals with autism spectrum disorder (ASD), results were less consistent. However, some studies pointed to lower “spindle chirp” (the subtle shift in spindle frequency over time) and reduced slow-wave amplitude. Lower amplitude suggests that the brain’s deep-sleep signals may be weaker or less synchronized. Researchers are still working to understand how these patterns relate to sensory processing, learning differences, or daytime behavior.

Depression: Lower Slow-Wave and Spindle Measures—Especially With Medication

People with depression tended to show reduced slow-wave activity and fewer or weaker sleep spindles, but this pattern appeared most strongly in patients taking antidepressant medications. Since antidepressants can influence sleep architecture, researchers are careful not to overinterpret the changes.  Nevertheless, these changes raise interesting questions about how both depression and its treatments shape the sleeping brain.

PTSD: Higher Spindle Frequency Tied to Symptoms

In post-traumatic stress disorder (PTSD), the trend moved in the opposite direction. Patients showed higher spindle frequency and activity, and these changes were linked to symptom severity which suggests that the brain may be “overactive” during sleep in ways that relate to hyperarousal or intrusive memories. This strengthens the idea that sleep physiology plays a role in how traumatic memories are processed.

Psychotic Disorders: The Most Consistent Sleep Signature

The clearest and most reliable findings emerged in psychotic disorders, including schizophrenia. Across multiple studies, individuals showed: Lower spindle density (fewer spindles overall), reduced spindle amplitude and duration, correlations with symptom severity, and cognitive deficits.

Lower slow-wave activity also appeared, especially in the early phases of illness. These results echo earlier research suggesting that sleep spindles, which are generated by thalamocortical circuits, might offer a window into the neural disruptions that underlie psychosis.

The Take-Away:

The review concludes with a key message: While sleep disturbances are clearly present across psychiatric conditions, the field needs larger, better-standardized, and more longitudinal studies. With more consistent methods and longer follow-ups, researchers may be able to determine whether these oscillations can serve as reliable biomarkers or future treatment targets.

For now, the take-home message is that the effects of these mental health disorders on sleep are real and measurable.