DSM

Statement of Concern about the Reliability, Validity, and Safety of DSM-5

We, the undersigned, are concerned that the Diagnostic and Statistical Manual of Mental Disorders, fifth edition(DSM-5):

  • Includes many diagnostic categories with questionable reliability, which may lead to misleading assumptions about their scientific validity; 

Prior to the publication of DSM-III in 1980, psychiatric diagnoses were frequently criticized for their substandard reliability, as clinicians too commonly disagreed on diagnostic decisions even when presented with the same information. Thus, a major impetus for those who developed the diagnostic model featured in DSMs III and IV was the improvement of interrater reliability, or the likelihood that two or more professionals would agree on a particular diagnosis (Feighner et al., 1972; Spitzer, Forman, & Nee, 1979). This goal was achieved with marked success, leading to hopes that the next step of achieving validity (i.e., empirical support for the real-world legitimacy of DSM-defined mental disorders) was close within reach (see Robins & Guze, 1970). However, subsequent research –and in fact a large body of data collected over the last 30 years– did not yield consistent validity evidence for DSM-defined categorical diagnoses. Instead, the gradual accumulation of inconsistent data has led some researchers to call for a root-and-branch review of diagnostic classification in psychiatry.

The DSM-5 development process was –especially at first– one effort to conduct that review. In the words of DSM-5 Task Force members (Regier, Narrow, Kuhl, & Kupfer, 2009), “As we began the DSM-V developmental process in 1999, a major concern was to address a range of issues that had emerged over the previous 30 years,” including “the basic definition of a mental disorder” (para. 7). There was hope for a “paradigm shift” in psychiatric diagnosis (Kupfer, First, & Regier, 2002, p. xix), and even though that aspiration has since been pushed to the back-burner (Kendler et al., 2009), the new manual will be published with markedly liberal revisions to DSMs III and IV.

The DSM-5 field trials (conducted in one phase, due the cancellation of plans for a second phase revealed an unexpected change from the previous two editions of the manual: reliability estimates for many major disorder categories had dropped well below not only those for DSM-III/IV-designed disorders, but also below commonly accepted standards (see Frances, 2012c). Furthermore, primary care doctors (family physicians and general practitioners) were not included in the field trials (American Psychiatric Association, 2011), despite the fact that they provide the majority of mental health treatment (Wang et al., 2007) and prescribe the majority of psychiatric medications (Mark, Levit, & Buck, 2009).

A primary tenet of empirical research holds that reliability is a necessary precondition for validity, as scientists cannot make stable claims about a concept that fluctuates empirically or lacks consensus among observers. Thus, before achieving common reliability standards, it is premature and untenable to introduce the DSM-5 revisions into hospitals, clinics, and general practice. Clinical research, likewise, should seek to establish psychometric stability before proceeding on the assumption that DSM-5 diagnostic categories are valid empirical entities. Epidemiological investigations may suffer from inconclusive findings and lack of continuity with research conducted using previous diagnostic definitions.

  • Did not receive a much-needed and widely requested external scientific review;

We recognize and appreciate that numerous professionals have worked hard to produce DSM-5, and have done so in good faith. However, many experts in the field have also spoken out in good faith about flaws in the document, and most of these flaws have not been resolved by the DSM-5 Task Force.

On January 9, 2012, the Open Letter Committee of the Society for Humanistic Psychology (Division 32 of the American Psychological Association) called for an external scientific review of the DSM-5 proposals by an independent group of researchers who are not affiliated with DSM-5 or the American Psychiatric Association (the full text can be found here: http://dsm5-reform.com/the-open-letter-committee-calls-for-independent-review-of-dsm-5/). This request was made in light of widespread reservation about the scientific status and safety of DSM-5 among mental health professionals and patient advocacy groups. An open letter to the DSM-5 Task Force and the American Psychiatric Association detailing these concerns (http://www.ipetitions.com/petition/dsm5/) was endorsed by more than 14,000 individuals and over 50 professional organizations, including 16 divisions of the American Psychological Association.

  • May compromise patient safety through the implementation of lowered diagnostic thresholds and the introduction of new diagnostic categories that do not have sufficient empirical backing;

DSM-5 has been criticized for lowering numerous diagnostic thresholds, i.e., reducing the number and severity of diagnostic criteria that are considered sufficient for a diagnosis to be made. The anticipated result is broad increase in the number of persons who qualify for a diagnosis of mental disorder, especially among individuals whose symptoms would have been considered subclinical in DSMs III and IV.  In the third draft of the manual (formerly available for public viewing on dsm5.org), lowered diagnostic thresholds appeared in the proposed definitions for Generalized Anxiety Disorder, Somatic Symptom Disorder, Bulimia Nervosa, and Alcohol Use Disorder, among other diagnoses.

DSM-5 also introduces new disorders that did not appear in earlier editions of the manual. Among them: Premenstrual Dysphoric Disorder, Disruptive Mood Dysregulation Disorder, Somatic Symptom Disorder, and Mild Neurocognitive Disorder. These new diagnoses have generated significant controversy as a result of their questionable research backing and their potential for application to vulnerable populations, including children, the elderly, and persons with chronic medical illnesses. Some of the feared consequences of these new categories are as follows:

  • Somatic Symptom Disorder (a modification of the Somatoform Disorders in DSM-IV-TR) includes a new stipulation that will allow for the diagnosis of mental disorder in persons with chronic medical illness complaining of excessive pain. As a result, doctors may prematurely jump to the conclusion that “it’s all in the head” (Frances, 2012b, para. 3).
  • Disruptive Mood Dysregulation Disorder may be diagnosed in children and adolescents displaying significant mood swings (temper tantrums), which may be developmentally normal and resolve without treatment. Although the new category was invented with the aim of precluding the controversial practice of diagnosing Pediatric Bipolar Disorder, the latter diagnosis never existed in previous editions of the manual due to its questionable validity.
  • Mild Neurocognitive Disorder appears to describe normal cognitive decline that may be expected in elderly populations. Over-diagnosis of mental disorder and psychiatric treatment in the elderly –especially elderly populations in nursing homes– is already a nationwide problem in the US and other countries.
  • Premenstrual Dysphoric Disorder transforms severe PMS into a psychiatric disorder. In the past, similar proposals have been excluded from previous editions of the DSM due to substantial controversy and attention from women’s rights groups because of the risk of pathologizing women’s experience.

Altogether, the lowering of diagnostic thresholds and introduction of new disorders in DSM-5 has led to increasing concern about patient safety. Though it has been suggested that psychotropic medication may not be the first line of treatment for some of these diagnoses, the vast majority of psychiatric diagnoses are made in fast-paced treatment settings by general medical practitioners who do not have time to critically evaluate the research literature behind the DSM and often have few alternatives to prescribing medications.

Our duty in the medical and helping professions is, first and foremost, to do no harm. Thus, as mental health practitioners and researchers, we are greatly concerned about the introduction of empirically questionable diagnostic concepts into psychiatric and general medical practice.

  • Is the result of a process that gives the impression of putting institutional needs ahead of public welfare. 

Several aspects of the DSM-5 development process reflect an apparent prioritization of institutional needs above patient safety and general public welfare. For example, the DSM-5 field trials were designed for implementation in two stages; the first was intended to address reliability, the second quality control. The second stage of the field trials was ultimately cancelled due to delays in the development process. Despite the importance of assessing quality control before the manual is used in patient care, DSM-5 will proceed with its expected May 2013 publication. For more information about the conduct and findings of the DSM-5 field trials, see Frances (2012a).

Additional concerns about the DSM-5 development process include hiring of a pubic relations firm (GYMR) to influence public opinion about the manual through a PR website (http://dsmfacts.org/), the lack of external scientific evaluation of the proposals, and the lack of a formal forensic review.

Because of the above, we fear that DSM-5:

  • May result in the mislabeling of mental illness in people who would fare better without a psychiatric diagnosis;

We have no doubt that many of the issues considered by DSM-5 constitute clinical and societal problems. It is worrying that many people are so affected by economic crises that they contemplate taking their own lives. Excessive alcohol or recreational drug use are dilemmas for individuals and societies. We have aging populations, troubled and disruptive children. It is a marker of humanitarian progress that we attempt to help people in distress. But, to take one example of many, it is unhelpful to suggest that a child throwing temper tantrums or a woman experiencing ‘period pains’ is mentally ill. It is unhelpful to suggest that a consumer seeking help from medical doctors is, by virtue of multiple complaints or visiting multiple doctors, mentally ill. Clients and the general public are negatively affected by the continued and continuous medicalization of their natural and normal responses to their experiences; responses which undoubtedly have distressing consequences that demand helping responses, but which do not reflect illnesses so much as normal individual variation. The mental health professions are uniquely suited for helping to create a better global society. But the application of inappropriate psychiatric labels is not a solution.

  • May result in unnecessary and potentially harmful treatment, particularly with psychiatric medication;

It is highly likely that, if a person receives a diagnosis under DSM-5, recommended treatment will involve medication. Recently, mounting empirical evidence has suggested that psychiatric medication, though helpful when used properly, may lead to iatrogenic consequences when used inappropriately. For example, antipsychotic medications, which are increasingly used to treat non-psychotic symptoms such as depression and anxiety, may lead to metabolic syndrome, obesity, diabetes, Parkinson’s-like movement disorders, neurocognitive decline, psychotic symptoms, reduced brain volume, and significantly shortened lifespan (Ho, Andreasen, Ziebell, Pierson, & Magnotta, 2011; Olfman & Robbins, 2012; Robbins, Higgins, Fisher, & Over, 2011; Whitaker, 2010). .

  • May divert precious mental health resources away from those who most need them.

Mental health problems affect one person in every four, making them the leading cause of disability worldwide (World Health Organization, 2012), at an estimated cost of $2,500 billion in 2010 (Bloom et al, 2011). The provision of high-quality and appropriate mental health care is an urgent global issue. Although the aspiration to improve the well-being of all citizens may be laudable, the use of scientifically unstable diagnoses will only confuse a complex picture and lead to the inappropriate investment of scarce resources. Since mental health problems disproportionately affect poor and socially excluded people, questionable diagnostic systems risk further disadvantaging the most vulnerable.