Diagnosis in child mental health

Exploring the benefits, risks and alternatives
CFCA Paper No. 48 – July 2018

3. Overdiagnosis of child mental health conditions

What should we make of the oft-repeated claims that too many children are diagnosed with mental health conditions, that the field of paediatric mental health is plagued by "false epidemics" and "psychiatric fads" (e.g. Frances & Batstra, 2013; Greenberg, 2013)?

Three mental health conditions in particular have inspired such claims.

The first is ADHD, which is widely believed to be overdiagnosed. Indeed, a study exploring attitudes of the general public in Queensland found that "the overwhelming majority of participants [78.3%] agreed that that too many children are diagnosed with ADHD when they don't really have it" (Partridge, Lucke, & Hall, 2014, p. 4).

Such attitudes are fuelled by media reports of research findings that can be interpreted as providing evidence of ADHD overdiagnosis, such as the recent and widely reported finding that the youngest children in Western Australian primary school classrooms were roughly twice as likely as their oldest classmates to be prescribed psychotropic medication for ADHD (Whitely et al., 2017).

The second is Autism Spectrum Disorder (ASD), the diagnostic criteria for which have undergone significant changes in the DSM-5, partly as an attempt to reduce the ballooning number of children being diagnosed (Basu & Parry, 2013).

The third is the highly contentious Paediatric Bipolar Disorder (PBD), the prevalence rates of which increased by over 4,000% between 1994-95 and 2002-03 (Moreno et al., 2007). Most of this massive rise in diagnosis was contained within the US however, and may, at least in part, be attributed to US policies allowing direct-to-consumer advertising of psychotropic medications (Parry & Levin, 2012). A survey of Australian and New Zealand child psychiatrists found that 90% thought that PBD was overdiagnosed in the US (Parry, Furber, & Allison, 2009).

While public attitudes and dramatic rises in prevalence rates suggest that overdiagnosis may be occurring, they do not in themselves provide conclusive evidence. Perhaps rates have risen because child mental health difficulties were previously under-recognised. Maybe negative community attitudes simply reflect a lack of awareness that children can develop mental health difficulties.

The different meanings of overdiagnosis

In order to properly investigate claims of overdiagnosis, it is necessary to distinguish between two different uses of the term "overdiagnosis".

1. As a pathologising of normal human experience

First, the term "overdiagnosis" serves to make the point, discussed earlier in this paper, that the diagnostic criteria or thresholds in the DSM lead to normal human experiences being labelled as pathological.

As already mentioned, it is difficult to provide conclusive evidence in support or refutation of this argument because it is not possible to know the "true" prevalence of DSM disorders. Unlike many physical conditions, which can be identified using objective measures, such as X-rays for broken bones, DSM disorders lack external criteria for examining their validity (Frances & Widiger, 2012).

Arguments for or against particular diagnoses cannot be settled using scientific evidence alone, but must also draw on professional judgement, public opinion, philosophical investigation, risk-benefit analyses, and various practical lines of reasoning (Frances & Widiger, 2012; Maj, 2015).

A historical example is the removal of homosexuality as a disorder from the seventh edition of DSM-II in 1974, which arguably had much more to do with pressure from gay rights activists - who drew on philosophical concepts such as "human rights", "equality" and "justice" - than it did with appeals to empirical evidence (R. Bayer, 1987).

More recently, a number of prominent mental health professionals have drawn on different forms of evidence and argument to assert that current diagnostic systems lead to far too many people in developed countries being diagnosed with mental health conditions (e.g. Brinkmann, 2016; Frances, 2013; Greenberg, 2013; Paris, 2015).

2. As misadministration of diagnostic criteria

Second, the term can be used to indicate that too many people are given psychiatric labels because diagnosticians do not properly adhere to diagnostic criteria. In other words, this use of the term is intended not as a criticism of any particular diagnostic system, but rather points to problems with their administration in clinical settings.

This form of overdiagnosis is easier to prove or disprove, because once the diagnostic criteria in the DSM or ICD are taken as a given, they can provide a relatively reliable baseline against which to measure the number of diagnosed cases in a given cohort.

In a recent meta-analysis of research investigating overdiagnosis of mental health conditions in children and young people, Merten and colleagues (2017) identified 17 studies that focused on particular conditions, including ADHD, PBD, ASD and anxiety disorders. Although the majority of these studies found evidence of overdiagnosis, all but one failed to meet the strict criteria of the meta-analysis. The single study that met the inclusion criteria, by Bruchmüller and colleagues (2012) in Germany, involved sending 473 registered child and adolescent psychotherapists (who included psychologists, psychiatrists and social workers) case vignettes describing a child fulfilling or not fulfilling the diagnostic criteria for ADHD. The authors found evidence of overdiagnosis: while 16.7% of participants diagnosed ADHD although diagnostic criteria were not fulfilled, only 7% gave no diagnosis for vignettes that fulfilled the criteria for ADHD. In sum, while the available research suggests that in developed countries certain child mental health conditions are overdiagnosed (in the second sense of the term), more research is needed before definitive statements on this matter can be made.

The causes of overdiagnosis

There is no single cause of overdiagnosis. If and when it does occur, it is most likely due to numerous converging pressures (Coon, Quinonez, Moyer, & Schroeder, 2014). The most commonly cited of these pressures are discussed below.

Influence of the pharmaceutical industry

Drug companies, who have a clear interest in increasing the number of children who use psychotropic medications, have in recent years marketed to younger and younger customers (Frances & Widiger, 2012). The scope of this marketing exercise is difficult to underestimate, as Batstra and Frances (2012, p. 7) describe:

Drug companies are ubiquitous players in psychiatry, influencing thought leaders and consumer advocacy groups, monopolizing professional "education", lavishing trips and meals, clogging physicians waiting rooms with attractive salespeople, stocking doctors with free samples, and (in the USA) conducting direct-to-consumer promotion campaigns in the print media, on TV, and on the Internet.

Numerous authors have argued that pharmaceutical industry influence has been a key driver in the dramatic rises in the prevalence rates of childhood conditions such as ADHD, PBD and ASD (e.g. Batstra & Frances, 2012; Greenberg, 2013). While these authors have clearly demonstrated the pharmaceutical industry's influence on psychiatric practice in the US, laws prohibiting direct-to-consumer marketing of medications in Australia suggest that this influence may be tempered somewhat in the Australian context. Further, it is unclear to what extent drug companies influence the diagnostic practices of non-psychiatric professionals, such as paediatricians.

Diagnostic inflation

"Diagnostic inflation" is another way of referring to the criticism that has recurred throughout this paper: that by introducing new disorders, and reducing the thresholds for existing disorders, diagnostic systems such as the DSM pathologise human experiences that were previously considered normal. Even small changes in the diagnostic criteria for mental health conditions can lead to huge increases in the number of children being diagnosed (Frances & Widiger, 2012).

Diagnostic upcoding

"[Diagnostic upcoding] occurs wherever medical practitioners are under pressure to give a diagnostic label in order to provide treatment and be reimbursed" (Parry & Levin, 2012, p. 59). Diagnostic upcoding is less of an issue in Australia as the health system does not normally require diagnosis for reimbursement for therapy, unlike in the US. However, it is still relevant in the case of ASD, because a positive diagnosis confers education and family financial welfare (Parry & Levin, 2012). Indeed, a study involving specialist medical clinicians in Queensland found that, in the face of uncertainty, many would provide an ASD diagnosis even though the criteria had not been met (Skellern et al., 2005).

Service provider factors

Some authors have explored the qualities of service providers (e.g. general practitioners, psychiatrists and psychologists) that may lead them to provide unwarranted diagnoses. For example, some have suggested that intolerance of uncertainty can be a powerful motivator for diagnostic testing (Coon et al., 2014). As Coon and colleagues (2014, p. 1017) observed, "the culture of medical education is an early impetus for training providers to find comfort in commission and fear in uncertainty". Research also suggests that clinical judgements tend to be influenced by various heuristics and biases (Bruchmüller et al., 2012). Take ADHD, which is more likely to affect boys than girls, as an example. A boy with certain ADHD symptoms is more likely to be seen as a prototypical ADHD case than a girl with the same symptoms, and thus is more likely to be diagnosed with the condition, even if he meets exclusion criteria (Bruchmüller et al., 2012).

Critical reflections

  • How would you respond to parents who were concerned that their child had been misdiagnosed?
  • Can you think of a time when you adjusted your practice to work with a child who you believe had been misdiagnosed?
  • Do you have first-hand experience of the factors that may lead to overdiagnosis, such as diagnostic upcoding?