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

Background:

In Iceland, treatment with ADHD medication can only be initiated by psychiatrists or pediatricians with experience in diagnosing neurodevelopmental disorders. The diagnostic evaluation is most often carried out by a psychologist or psychiatrist, and must be confirmed by a psychiatrist. 

Some previous studies have suggested a small but significant increased risk of psychosis or mania associated with ADHD medication, while others have not. 

Iceland has a single-payer national healthcare insurance system that tracks virtually its entire population. An Icelandic research team accessed two administrative databases with nationwide coverage – the Icelandic Prescription Medicines Register and the Icelandic Hospital Discharge Register – to explore this relationship among all adults from 2010 through 2022. 

They included three categories of ADHD medications prescribed in Iceland: amphetamines, including dexamphetamine and lisdexamphetamine; methylphenidate; and atomoxetine. In Iceland, methylphenidate or atomoxetine must be prescribed and tried first before switching to lisdexamphetamine or dexamphetamine. 

Method:

Diagnoses of mania or psychosis recorded in electronic health records were used to identify individuals who were admitted to a psychiatric ward within one year of starting treatment with a specific class of ADHD medication. First-onset psychosis or mania was defined as the emergence of these conditions in individuals with no prior history, diagnosis, or hospitalization for psychosis or mania. 

A total of 16,125 adults began using an ADHD medication for the first time during the 13-year study period. 

Methylphenidate was the most used ADHD medication among those admitted for psychosis or mania (25 out of 61; 41%), reflecting its status as the most frequently prescribed stimulant during the study period. It was followed by amphetamines (21 out of 61; 34.4%) and atomoxetine (15 out of 61; 24.6%). 

Half of those hospitalized had previously received a diagnosis of substance use disorder. One in nine (11%) of those hospitalized acknowledged misuse of the type of ADHD medication they had been prescribed. 

Within a year of discharge, 42 out of the 61 patients (68.9%) had been prescribed an ADHD medication again. Among those, one in four (11 out of 42; 26%) were readmitted for psychosis or mania within the following year.  

The team noted, “It is concerning that most patients (68.9%) in our study resumed ADHD drug treatment within a year of hospital discharge … However, some studies have reported that the use of psychostimulants or atomoxetine to treat ADHD in individuals with psychotic disorders did not increase the risk of hospitalisation for psychosis if used concurrently with antipsychotic medication or that such use might even reduce this risk.”  

Findings: 

By comparison with the general population, adults initiating ADHD medications had eight times the relative risk of being admitted for psychosis or mania within the first year.  

The absolute risk was low: 0.38% overall for those initiating ADHD medication.  Adjusting for the general population risk of hospitalization for first-onset psychosis or mania, more than 300 patients would need to be initiated to ADHD medication to generate one hospital admission for psychosis or mania. 

The team conceded, “Confounders of real-life clinical settings, such as non-disclosed ADHD drug abuse or misuse or some degree of substance abuse, may have influenced our findings.” 

A further, unmentioned, limitation is that the team did not perform any of the usual adjustments for confounding variables, critically including co-occurring (comorbid) psychiatric disorders known to be common with ADHD, and likely to have a major effect on the relative risk of hospitalization. 

Given the very small increase in risk along with the methodological flaws, the team’s suggestion of a “potential causal role of ADHD drugs in the development of first-onset psychosis or mania” is unsubstantiated and speculative.  This is especially so given other studies suggesting no increased risk for psychosis due to these medications.  

In any event, causation cannot be established through observational studies.

‍

The Spanish National Health Survey tracks health care outcomes through representative samples of the Spanish population. 

A Spanish research team used survey data to explore the relationship between ADHD symptoms and dental and gum health in a representative sample of 3,402 Spanish children aged 6 to 14.

While previous studies have found associations between ADHD and poor dental health, they have not fully accounted for such important determinants of poor oral health as socioeconomic status, dental hygiene, or diet. 

The team therefore adjusted for sociodemographic factors, lifestyle variables, and oral hygiene behaviors. More specifically, they adjusted for sex, age, social class, parental education, exposure to tobacco smoke, consumption of sweets, consumption of sugary drinks, use of asthma or allergy medication, adequate oral hygiene behavior of children, adherence to regular dental visits, parental adequate oral hygiene behavior, and parental adherence to regular dental visits.

With those adjustments, children with ADHD symptoms had over twice the incidence of dental caries (cavities) as their counterparts without ADHD symptoms.

Tooth extractions and dental restorations also occurred with over 40% greater frequency in children with ADHD symptoms.

Gum bleeding, a sign of gum disease, was more than 60% more common among children with ADHD symptoms than among their non-ADHD peers.

Importantly, excluding children with daily sugar consumption, which left 1,693 children in the sample, made no difference in the outcome for cavities.

Excluding children with poor oral hygiene habits, which left 1,657 children in the sample, those with ADHD had 2.5-fold more caries than their non-ADHD counterparts.

Excluding children of low social class, which left 1,827 children in the sample, those with ADHD had 2.6-fold more caries than their non-ADHD counterparts.

Turning to a different method to address potential confounding factors, the team used nearest-neighbor propensity score matching to create virtual controls. This compared 461 children with ADHD to 461 carefully matched children without ADHD.

This time, children with ADHD symptoms had just under twice the incidence of cavities as their counterparts without ADHD symptoms, but 60% more tooth extractions and about 75% more dental restorations. The difference in gum bleeding became nonsignificant.

Noting that “The increased risk of caries was maintained when the analyses were restricted to middle/high social class families and children with low sugar intake, good oral hygiene behaviors and regular dental visits,” the team concluded, “Children with ADHD symptoms in Spain had worse oral health indicators than those without ADHD symptoms. Our results suggest that the association of ADHD symptoms with caries was independent of socioeconomic level, cariogenic diet, frequency of toothbrushing, and dental visits.”

‍

This new meta-analysis confirms what other meta-analysis have already shown, i.e, that there exists in the population an association between maternal smoking during pregnancy and ADHD in their offspring.  But reader beware, association does not mean causation.

The team identified 55 studies with quantitative data suitable for meta-analysis, including 11 case-control, 13 cross-sectional, and 31 retrospective/prospective cohort studies. 

Altogether they combined more than four million persons in countries spanning six continents, including the United States, Finland, Sweden, Brazil, the Netherlands, Japan, the UK, Spain, China, Australia, New Zealand, Norway, Canada, France, Sweden, South Korea, Turkey, Romania, Bulgaria, Lithuania, Germany, Denmark, Egypt, and India.

Meta-analysis of all 55 studies found that offspring of mothers who smoked tobacco during pregnancy were about 70% more likely to develop ADHD than offspring of mothers who did not smoke during pregnancy.

Because variation in outcomes across studies was very high, the team performed subgroup analyses to explore potential sources of this heterogeneity. 

Comparing study designs, cohort studies reported roughly 50% greater odds of ADHD among children of mothers who smoked during pregnancy, whereas case-control studies reported roughly 70% greater odds and cross-sectional studies 2.3-fold greater odds.

Studies using the most reliable method of determining ADHD – clinical interview/professional diagnosis – reported 90% greater odds, contrasting with 66% through medical records/databases and 58% through self-report by child/parent or through teacher report.

Good quality studies reported roughly 75% greater odds. 

Studies with sample sizes above two thousand similarly found 70% greater odds.

There was no sign of publication bias using the more commonly used Egger’s test, but a marginal indication of publication bias using Begg’s test. Performing a standard correction reduced the effect size, indicating that the offspring of mothers who smoked tobacco during pregnancy were over 50% more likely to develop ADHD than the offspring of mothers who did not smoke during pregnancy.

The team concluded, “This systematic review and meta-analysis of 55 studies, encompassing over four million participants, provides compelling evidence that maternal tobacco smoking during pregnancy significantly increases the odds of ADHD in children … These findings underscore the critical need for public health interventions aimed at reducing tobacco smoking during pregnancy.”

However, we disagree with this conclusion; The authors ignore substantial evidence showing that maternal smoking during pregnancy is confounded by maternal ADHD. These mothers transmit ADHD via genetics, not via their smoking. This study should be seen not as "...[further evidence that smoking during pregnancy causes ADHD.] ", but as a lesson in how easy it can be to see correlation as causation.

‍

‍

------

Struggling with side effects or not seeing improvement in your day-to-day life? Dive into a step-by-step journey that starts with the basics of screening and diagnosis, detailing the clinical criteria healthcare professionals use so you can be certain you receive an accurate evaluation. This isn’t just another ADHD guide—it’s your toolkit for getting the care you deserve. This is the kind of care that doesn’t just patch up symptoms but helps you unlock your potential and build the life you want. Whether you’ve just been diagnosed or you’ve been living with ADHD for years, this booklet is here to empower you to take control of your healthcare journey.

Proceeds from the sale of this book are used to support www.ADHDevidence.org.

Get the guide now– Navigating ADHD Care: A Practical Guide for Adults

‍