Attention deficit hyperactive disorder diagnosis continues to fail the reliability and validity tests

Professor Florence Levy’s November 2014 opinion piece, DSM-5, ICD-11, RDoC, and ADHD Diagnosis, represents a shift in tone, but not core belief, from 2002 when she was 1 of 84 self-declared ‘leading scientists’ who signed the International Consensus Statement on ADHD. In the International Consensus Statement, the evidence supporting the validity of attention deficit hyperactive disorder (ADHD) was deemed to be equivalent to that supporting ‘the laws of gravity’ and the ‘periodic table’, and critics were dismissed as ‘flat earthers’.

In response ADHD critic, British psychiatrist Sami Timimi wrote,

Not only is it completely counter to the spirit and practice of science to cease questioning the validity of ADHD as proposed by the consensus statement, there is an ethical and moral responsibility to do so … The evidence shows that the debate is far from over. (Timimi et al., 2004)

Levy’s recent piece acknowledges the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV) Taskforce leader, Professor Alan Frances’, criticisms of the expanded Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM-5) definition of adult ADHD. However, she fails to address Frances’ concern that the ‘explosion’ in ADHD child prescribing that occurred prior to the publication of DSM-5 is a dangerous ‘fad’ that ‘requires taming’ (Frances, 2012). The closest Levy came is her recognition that the ‘DSM and ICD’ have only been moderately successful ‘in establishing diagnostic reliability’.

This is a massive understatement. ADHD diagnoses are notoriously unreliable. As detailed below, a child’s chances of being prescribed ADHD drugs are significantly influenced by birth date relative to classroom peers. Gender, postcode, clinician, teacher and parental attitudes, none of which have anything to do with a child’s neurology, are also significant factors.

Levy promotes the yet-to-be-developed National Institute of Mental Health’s Research Domain Criteria (RDoC), which are to be based on ‘underlying neurobiological components, conceptualized as disorders in brain circuitry’, as ‘promising greater reliability of diagnosis’. If, and it is a very big if, the RDoC develops hypothesised differences in ‘brain circuitry’ into a diagnostic tool, then ADHD may become a reliable diagnosis. However, this would not guarantee that it was a valid diagnosis.

Nothing better demonstrates the invalidity of an ADHD diagnosis than an 11-year study of 938,000 Canadian children that confirmed the results of two earlier large-scale American studies. The Canadian research found that children born in the last months of a school year are far more likely to be ‘medicated for ADHD’ (boys 41% and girls 77%) than their oldest classmates (Morrow et al., 2012).

The ‘Holy Grail’ for many proponents of ADHD is establishing its ‘genetic basis’, the logic being that this would validate it as a psychiatric disorder. In 2010, there was widespread international coverage of British psychiatric researcher Anita Thapar claim that ‘now we can say with confidence that ADHD is a genetic disease’ (ABC Online News, 2010). The study that Thapar claimed established that ADHD as a ‘genetic disease’ involved the comparison of the genetic codes of 366 children ‘with ADHD’ with those of 1047 ‘non-ADHD’ control children. Researchers found 13.9% (51) of the children with ADHD had short lengths of their genetic codes that were either duplicated or missing. This compared with 7.4% (78) of the ‘control children’ (Williams et al., 2010).

The average recorded IQ of the 366 children ‘with ADHD’ was 86, 14 points below the general population average of 100. While the IQ of the 1047 ‘non-ADHD children’ was not specified, presumably they were as intelligent as the general population. Furthermore, when 33 intellectually impaired ‘ADHD children’ (IQ lower than 70) were excluded from the ADHD cohort, only 11% of the remaining 333 had the hypothesised ADHD genetic abnormality (Williams et al., 2010). Even with the intellectually impaired children removed, the average IQ (89) of the 333 remaining in the ADHD group was significantly lower than the general population mean of 100. This evidence is more suggestive of a relationship between the identified genetic abnormality and intellectual disadvantage rather than ADHD.

Ultimately, as defined in both the DSM and ICD, ADHD is a collection of behaviours, with children diagnosed as being less attentive and/or more impulsive or hyperactive than their peers. Finding a genetic basis for ADHD would, therefore, mean finding a genetic basis for inattentive and/or impulsive or hyperactive behaviour. It is entirely reasonable to think behaviour is a combination of nature and nurture. However, conceding that ADHD may be in part a ‘genetic difference’ is vastly different from accepting it is a ‘genetic disease’.

Even ADHD proponents acknowledge that many people without ADHD exhibit the 18 behaviours at the core of a DSM or ICD diagnosis. What is supposed to distinguish ADHD sufferers from the rest of the population is their level of behavioural impairment or dysfunction. All 18 behavioural diagnostic criteria include the word ‘often’. How often a child ‘fidgets or squirms in their seat’ or ‘interrupts’ or ‘avoids homework’ or ‘fails to remain seated when remaining seated is expected’ or ‘is distracted by external stimuli’, etc., so that they exhibit ‘some impairment’ is not defined in DSM-IV. The DSM-IV diagnostic criteria do not specify age-appropriate levels of attention or impulsivity control.

Furthermore, DSM-IV states,

Signs of the disorder may be minimal or absent when the person is receiving frequent rewards for appropriate behaviour, is under close supervision, is in a novel setting, is engaged in especially interesting activities, or is in a one-to-one situation (e.g., the clinician’s office.). (American Psychiatric Association, 2000)

Therefore, according to the DSM-IV, children with ADHD may behave appropriately and not display ADHD symptoms when they are rewarded, when people pay attention to them (close supervision) and when they are having interesting experiences. Conversely, they will be easily distracted and display ADHD symptoms when their good behaviour goes unrewarded, no one pays any attention to them or they are bored. Surely, the nonsense of this proposition is self-evident?

As Levy identifies, the DSM-5 has broadened ADHD diagnostic criteria by raising the age from 7 to 12, before which ‘several inattentive or hyperactive/impulsive symptoms’ should be present and by reducing from 6 to 5 the number symptoms required for a diagnosis in those 17 or older. This is a well-established long-term pattern with successive versions of the DSM expanding both the criteria and the markets for stimulants. Nonetheless, Levy concludes her recent piece stating ‘while the RDoC approach promises the greater reliability of objective measurement, it will still be important to maintain careful phenomenological description … Ideally, DSM-5, ICD-11 and RDoC should enrich each other’.

In summary, Levy’s prediction of improved reliability comes 12 years after she and other signatories of the International Consensus Statement equated the science supporting the ADHD with the laws of gravity. In relation to the house of cards supporting the ever-increasing ADHD prescribing rates, is it not the time for the laws of gravity to prevail?