AI Technologies May Help to Improve Early Identification of Youth with ADHD At Risk of Substance Use

Substance use disorder (SUD) – also known as drug addiction – is one of the most problematic co-occurring conditions with ADHD. Children with ADHD are twice as likely to develop SUD later in life, compared to those without ADHD [1, 2].  Identifying children and adolescents who are at risk, even before they ever use substances, can be an important step in preventing SUD. But how do you know who is at risk? Artificial Intelligence (AI) technologies are promising tools to identify risk factors and predict the likelihood of an individual developing SUD later in life.

Image by rebcenter-moscow

SUD is a mental illness that involves impulsive use of substances such as alcohol, marijuana, nicotine and opioids. Substance use often starts with experimentation in adolescence, such as cigarette smoking and drinking. Studies have found that adolescents with ADHD often start these experiments at younger ages, more quickly become heavier users and develop more severe functional impairments [3].  Indeed, younger ages of first use of a substance is associated with a higher risk of abusing the substance later in life. 

While SUD causes significant harm, and personal and societal burdens, it is actually a preventable and treatable condition. Early identification of at-risk youth is especially important because it would allow for more targeted early interventions.  Studies have shown that school- and community-based prevention programs can be highly effective at reducing substance use. Treating and managing ADHD, both with medication and behavioral therapy, has also been found beneficial in terms of reducing substance use [4, 5].  Being able to screen and identify the high-risk adolescents before their first attempt and being able to further “discourage” and “prevent” such onsets of substance use would probably be the most cost-effective prevention program for SUD.

But how do you know who is high-risk and who is not, even before they start using drugs? One method is to use Artificial Intelligence (AI) technologies in combination with very large databases.

Machine Learning Prediction Models for SUD

Machine learning is a special subset of AI technologies, that can learn hidden patterns and attributes in a completely data-driven fashion. Machine learning algorithms learn from the iterative exposures to a large amount of data, and subsequently make predictions based on what the model has learned from these training examples.

In this recently published EU-funded study in the Journal of Child Psychiatry and Psychology [6],  my colleagues and I applied machine learning models to data collected from the Swedish national registers. These registries contain family and health data of millions of people, including information on clinical diagnoses such as ADHD and SUD. For our research, we focused on children with ADHD and trained various models to predict those who would eventually have a diagnosis of SUD, and those who would not. More than 19,000 children with ADHD were used in the study. The collected information that we used to train the models included psychiatric and somatic diagnoses, family history of these disorders, socioeconomic status, and birth complications.

The machine learning algorithm produced two useful models to predict SUD in children or adolescents with ADHD. The first one makes a prediction at age 17 for future SUD diagnosis between age 18-20. This is an important period when young adults, often leaving home for the first time, are more subjective to peer influence and start their first use of substance. The second model, a longitudinal model, makes a yearly prediction at every age from age 2 to 17 for SUD diagnosis in up to 10 years in the future.  We found that both models were able to make significant predictions. This means that when we tested the models on part of the dataset (that was not used to train the model), the model was more often correct to predict SUD outcome than could be expected from chance. You can interpret this as that the model had learned certain parameters in the data that predict whether a person with ADHD will develop SUD or not.

Early Detection of SUD Risk

One important discovery was that the longitudinal model was able to make significant prediction at as early as age 2 for up to 10 years into the future. The earliest age of a valid SUD diagnosis in our dataset was at age 12. Using a method called “supervised learning”, we “taught” the model to identify those children at age 2 who would be diagnosed with SUD at age 12.  Such supervised training was carried out for each age for any given outlook of 1, 2, 5 or 10 years.  Being able successfully predict at-risk children years before their first attempt of substance use is a really promising result.  Furthermore, the yearly prediction at every age provides a monitoring system tracking the risk for SUD, either increasing or decreasing risk over years.

What have we learned from our prediction models?

It is important to stress that the predictions of these models were not perfect. The algorithms are not always correct at identifying who will develop SUD and who does not. However, they do give insight into some important predictors that are more informative than others. For example, the early detection and diagnosis of ADHD and socioeconomic status.  At the population level, the early diagnosis of ADHD was associated with a lower risk of developing SUD later in life, while adverse socioeconomic status was associated with a higher risk.  However, we could not identify any predictors at the individual level that contributed their increased or decreased risks.  The ability to identify such specific risk for each individual would be extremely useful when targeted interventions can be most effectively applied.

Despite that the prediction accuracies were modest and not ready for real-life deployment yet, these findings clearly have many, broad implications for policy-makers, parents, teachers and clinicians. Given that more and more data will become available to train these AI models, more accurate and generalizable predictions can be made at early ages regarding the disease risk, it is imperative to develop more effective and personalized measures for prevention and risk-reduction. We will therefore continue our work to expand the prediction accuracy of our models.

Overall, we provided a proof-of-concept that machine learning AI technologies – by leveraging the large volume of data, such as those of national registers and other electronic health records – can be used to predict disease risk, such as SUD, long before the disease onset.

Dr. Yanli Zhang-James is an associate professor at SUNY Upstate Medical University Department of Psychiatry and Behavioral Sciences. She is involved in the CoCA Project.

1.         Ercan, E.S., et al., Childhood attention deficit/hyperactivity disorder and alcohol dependence: a 1-year follow-up. Alcohol Alcohol, 2003. 38(4): p. 352-6.

2.         Wilens, T.E., et al., Does ADHD predict substance-use disorders? A 10-year follow-up study of young adults with ADHD. J Am Acad Child Adolesc Psychiatry, 2011. 50(6): p. 543-53.

3.         Kousha, M., Z. Shahrivar, and J. Alaghband-Rad, Substance use disorder and ADHD: is ADHD a particularly “specific” risk factor? J Atten Disord, 2012. 16(4): p. 325-32.

4.         Molina, B.S., et al., Delinquent behavior and emerging substance use in the MTA at 36 months: prevalence, course, and treatment effects. J Am Acad Child Adolesc Psychiatry, 2007. 46(8): p. 1028-40.

5.         Schoenfelder, E.N., S.V. Faraone, and S.H. Kollins, Stimulant treatment of ADHD and cigarette smoking: a meta-analysis. Pediatrics, 2014. 133(6): p. 1070-80.

6.         Zhang-James, Y., et al., Machine-Learning prediction of comorbid substance use disorders in ADHD youth using Swedish registry data. J Child Psychol Psychiatry, 2020.

How psychiatric genetics can help to guide diagnostic practice and therapy

Recently, professor Stephen Faraone from SUNY Upstate University in the USA gave a webinar about genetic research in psychiatry (especially ADHD) and how this can help to better understand diagnosis and provide better treatment. In this blog I will share with you some highlights from this webinar.

  1. ADHD is a continuous trait in the population

ADHD is not something that you either have or don’t have. Rather, symptoms or characteristics of ADHD are present in the entire population, in varying severity. The system for psychiatric diagnoses is however based on categorical definitions that determine when a certain combination of symptoms and severity can be classified as a particular disorder. Although these categories can be of great help to provide public health data or determine insurance coverage, they often don’t really match individual cases. Hence there arise problems with heterogeneity, subtypes, subthreshold cases and comorbidity.

Genetic research has shown that psychiatric conditions such as ADHD are not caused by a few single genes, but rather by thousands or tens of thousands genetic variants that each contribute slightly to the ADHD risk. These so-called polygenic risk scores form a normal distribution across the entire population, with the majority of people having low polygenic risk scores (so a low to average risk of ADHD), while a small portion of individuals have a very low or very high risk. This adds to our understanding that ADHD is a continuous trait in the population.

Image from the webinar by prof. Stephen Faraone. The higher the number on the x-axis, the higher the genetic risk of having ADHD. Negative numbers mean reduced genetic risk of ADHD.

2. Comorbidity in psychiatry is the norm, rather than the exception

In the webinar, Stephen Faraone explains that in 90’s it was thought impossible that an individual can have both ADHD and depression. Now, we know better than that. There are substantial genetic correlations between different psychiatric disorders, meaning that the genes that increase the risk of for instance ADHD, also increase the risk of schizophrenia, depression, bipolar disorder, autism and tic disorder. This is further evidence that psychiatric conditions are not separate, categorial entities but rather arise from similar biological mechanisms.

3. Personalised medicine and pharmacogenetics are not yet sufficiently established to adopt widely and replace current medication on a broad scale

The second part of the webinar was about pharmacogenetic testing. This means that an individual’s genetic profile is used to determine whether a drug will be effective, and in what dose. Although this sounds promising, there is still a lot of discussion about the validity of such tests. This is due to varying results, differing protocols and large heterogeneity between studies. In some cases, pharmacogenetic testing can help to find the right treatment for an individual, for instance when this person is not responding well to regular treatment, but it is definitely not a fool-proof method yet. Better randomized controlled clinical trials are needed to improve reliability of these tests.

You can watch the full webinar here:

These are the world’s most high ranking experts on ADHD

Who are the most knowledgeable people about ADHD in the world? According to the website, these are professors Stephen Faraone (SUNY upstate University), Samuel Cortese (University of Southampton) and Jan Buitelaar (Radboud University Nijmegen).

What’s more, several scientists who are involved in our research consortia that investigate ADHD (i.e. Aggressotype, CoCA, IMpACT, Eat2beNICE) are top-ranked in this list of more than 30.000 possible experts in the field. These include Stephen Faraone, Jan Buitelaar, Philip Asherson, Barbara Franke, Joseph Antoni Ramos-Quiroga, Henrik Larsson, Catharina Hartman and Pieter Hoekstra. What this means is that the ADHD research that we do, and that is often reported on in this blog, is lead by the world’s top ADHD experts.

‘Our’ top-ranked ADHD experts. From left-to-right: Stephen Faraone, Jan Buitelaar, Philip Asheron, Barbara Franke, Joseph Antoni Ramos-Quiroga, Henrik Larsson, Catharina Hartman, Pieter Hoekstra.

How is an expert defined?

The website expertscape was started by John Sotos when he was looking for an expert on Parkinson’s disease to treat his uncle. This turned out to be more difficult than he thought. As John Sotos was a doctor himself, he luckily had a large network of doctors that he could contact about this. But this made him realise that people who don’t have such a network, would not be able to find out who the most knowledgeable persons are on a particular topic. He therefore created this website

The way the website works is quite simple: it searches for academic, peer-reviewed publications by a certain person on a certain topic. The more someone has published on a topic, the higher this person is ranked. Thus,  “[a]n expert is not just someone who knows a lot about a particular topic. We additionally require that the expert write about the topic, and be involved at the leading edge of investigation of the topic.”

This means that the site is actually not a very good tool to find a good doctor. As the website acknowledges “a great doctor has many important qualities beyond expert knowledge of your very specific medical condition.” However, it does mean that the website is pretty good at providing a simple overview of who has a lot of scientific knowledge about a specific topic.

So are they really experts?

In the past years I have met with most people in the top of this list, and I dare say that they are very knowledgeable indeed. Each of them has been working in the ADHD field for a considerable amount of time and has added important new insights into ADHD through research and publications. What I find most striking from this list however, is that most of these experts work together in consortia and international networks. And that is how the field really moves forward: by combining the knowledge of all these experts.

Several of these experts have also written for this blog:




This blog was written by Jeanette Mostert. Jeanette studied brain connectivity in adult ADHD during her PhD. She is now dissemination manager of the international consortia CoCA and Eat2beNICE. 


Pay Attention to ADHD – Podcast with prof. Stephen Faraone

Professor Stephen Faraone – professor in Psychiatry at SUNY Upstate University and expert on ADHD – was interviewed by dr. Therese Markow for the podcast series ‘Critically Speaking’. In this podcast they discuss myths about ADHD and the scientific evidence that debunks these myths. Stephen Faraone explains why it is so important to diagnose and treat ADHD early. He also explains why ADHD is often undiagnosed in girls, and why sometimes adults are diagnosed with ADHD who have not sought treatment earlier in their life.

Critically Speaking is a podcasts series hosted by dr. Therese Markow who interviews experts to discuss in plain language complex issues that concern our health, society and planet.

You can listen to the podcast here:


ADHD Is A Risk Factor For Type Two Diabetes And High Blood Pressure, As Well As Other Psychiatric Disorders

All Swedish residents have their health records tracked through unique personal identity numbers. That makes it possible to identify psychiatric and medical disorders with great accuracy across an entire population, in this case encompassing more than five and a half million adults aged 18 to 64. A subgroup of more than 1.6 million persons between the ages of 50 and 64 enabled a separate examination of disorders in older adults.

Slightly over one percent of the entire population (about 61,000) were diagnosed with ADHD at some point as an adult. Individuals with ADHD were nine times as likely to suffer from depression as were adults not diagnosed with ADHD. They were also more than nine times as likely to suffer from anxiety or a substance use disorder, and twenty times as likely to be diagnosed with bipolar disorder.  These findings are very consistent with reports from clinical samples in the USA and Europe.

Adults with ADHD also had elevated levels of metabolic disorders, being almost twice as likely to have high blood pressure, and more than twice as likely to have type 2 diabetes. Persons with ADHD but without psychiatric comorbidities were also almost twice as likely to have high blood pressure, and more than twice as likely to have type 2 diabetes.

Similar patterns were found in men and women with ADHD, although comorbid depression, bipolar disorder, and anxiety were moderately more prevalent in females than in males, whereas substance use disorder, type 2 diabetes, and hypertension were more prevalent in males than in females.

ADHD was less than a third as prevalent in the over-50 population as in the general adult population. Nevertheless, individuals in this older group with ADHD were twelve times as likely to suffer from depression, anxiety, or substance use disorders, and more than 23 times as likely to be diagnosed with bipolar disorder as their non-ADHD peers. They were also 63% more likely to have high blood pressure, and 72% more likely to have type 2 diabetes.

The authors noted, “Although the mechanisms underlying these associations are not well understood, we know from both epidemiologic and molecular genetic studies that a shared genetic predisposition might account for the co­existence of two or more psychiatric conditions. In addition, individuals with ADHD may experience increased difficulties as the demands of life increase, which may contribute to the development of depression and anxiety.” As for associations with hypertension and type 2 diabetes, these “might reflect health ­risk behaviors among adult patients with comorbid ADHD in addition to a shared biological substrate. As others have noted, inattention, disinhibition, and disorganization associated with ADHD could make it difficult for patients to adhere to treatment regimens for metabolic disorders.” They concluded that “Clinicians should remain vigilant for a wide range of psychiatric and metabolic problems in ADHD affected adults of all ages and both sexes.”

Stephen Faraone is distinguished Professor of Psychiatry and of Neuroscience and Physiology at SUNY Upstate Medical University and is working on the H2020-funded project CoCA. 


Qi Chen, Catharina A. Hartman, Jan Haavik, Jaanus Harro, Kari Klungsøyr, Tor­Arne Hegvik, Rob Wanders, Cæcilie Ottosen, Søren Dalsgaard, Stephen V. Faraone, Henrik Larsson, “Common psychiatric and metabolic comorbidity of adult attention-deficit/hyperactivity disorder: A population-based cross-sectional study,” PLoS ONE (2018), 13(9): e0204516.