“Clinical judgment” and the DSM-5 diagnosis of major depression

The introduction of “explicit diagnostic criteria” in psychiatry – initially only for research purposes and subsequently, with the DSM-III, also for use in ordinary clinical practice – had a main objective: to overcome the “vagueness and subjectivity inherent in the traditional diagnostic process” (1, p. 85), and in particular the variability in the inclusion and exclusion criteria used by clinicians when summarizing patient data into psychiatric diagnoses (“criterion variance”), which was regarded as the main source of the poor reliability of those diagnoses (2).

From the very beginning, however, there was some ambivalence in mainstream American psychiatry about the constraints that the use of fixed diagnostic criteria would pose to the exercise of clinical judgment. Spitzer et al (3), in an early paper reporting on the development of the DSM-III, acknowledged that “the use of specified criteria does not, of course, exclude clinical judgment”. They qualified this statement by adding that “the proper use of such criteria requires a considerable amount of clinical experience and knowledge of psychopathology”, thus giving the impression that clinical judgment was regarded as just instrumental to the proper use of the explicit diagnostic criteria. However, they also stated that “in any case, the criteria that may be listed in DSM-III would be ‘suggested' only, and any clinician would be free to use them or ignore them as he saw fit” (3, p. 1191).

Spitzer et al's prediction that operational criteria would appear in the DSM-III “under the heading ‘suggested criteria'” (3, p. 1190) did not come true. However, the DSM-III introduction emphasized that those criteria were provided as “guides for making each diagnosis”, in order not to leave the clinician “on his or her own in defining the content and boundaries of the diagnostic categories” (4, p. 8). As Spitzer commented later on (5, p. 403), the DSM-III diagnostic criteria were intended “as guides, not as rigid rules”.

This is further clarified in the DSM-IV introduction, where it is stated that explicit diagnostic criteria “are meant to serve as guidelines to be informed by clinical judgment and are not meant to be used in a cookbook fashion” (6, p. xxiii). The example is provided of a diagnosis which is made through the exercise of clinical judgment although the clinical presentation falls just short of meeting the full criteria. So, clinical judgment does not only inform the use of explicit criteria; it may also lead the psychiatrist to “force”, to a limited extent, those criteria if he finds this appropriate.

The text of the DSM-IV also mentions clinical judgment when it comes to the assessment of clinical significance, required for the diagnosis of several disorders: “assessing whether this criterion is met, especially in terms of role function, is an inherently difficult clinical judgment” (6, p. 7). The chair of the DSM-IV Task Force, A. Frances, emphasized that “this appeal to clinical judgment is a reminder to evaluate not only the presence of the symptoms in the criteria set, but also whether they are severe enough to constitute mental disorder”, though he acknowledged that an evaluation of clinical significance by using clinical judgment “contains the seeds of tautology” (7, p. 119).

However, as pointed out by Spitzer and Wakefield (8), there is no reference in the DSM-IV to the exercise of clinical judgment in the differential diagnosis between depression and the “normal” response to a significant loss. The text is very clear in stating that the diagnosis of major depression should be made whenever the severity, duration and distress/impairment criteria for that condition are met, even if the depressive state is the understandable response to a psychosocial stressor (6, p. 326). The only exception is bereavement: if the depressive state follows the loss of a loved one, the diagnosis of major depression should not be made even if the diagnostic criteria are fulfilled, unless some further elements are present (the symptoms persist for longer than 2 months, or are characterized by marked functional impairment, morbid preoccupation with worthlessness, suicidal ideation, psychotic symptoms, or psychomotor retardation). So, in any case – whether depression is related to bereavement or not – explicit criteria are provided, and no mention is made of the use of clinical judgment.

Actually, when J. Wakefield proposed to exclude “normal” responses to psychosocial stressors from the diagnosis of major depression, leaving to the clinician the decision on whether the depressive response was proportional or not to the preceding stressor (8,9), the rebuttal by K. Kendler, a protagonist of mainstream American psychiatry (and of the process of development of the DSM-5), was straightforward: this return to “what at basis will be the subjective criteria proposed by Jaspers in his old idea of ‘understandability'” would represent “more a step backward than forward for our field” (10, pp. 149–150).

This “step backward” has apparently been done in the DSM-5 (11). A note included in the DSM-5 criteria for major depressive disorder states that “responses to a significant loss (e.g., bereavement, financial ruin, losses from a natural disaster, a serious medical illness or disability) may include feelings of intense sadness, rumination about the loss, insomnia, poor appetite and weight loss, which may resemble a depressive episode”, and that the decision about whether a major depressive episode (or just a normal response to the loss) is present “inevitably requires the exercise of clinical judgment based on what the clinician knows about the individual in question and the individual's cultural norms for the expression of distress in the context of loss”.

This solution adopted by the DSM-5 Task Force should be seen within the context of the debate, taking place in both the scientific and the lay press (e.g., 12,13), about the elimination of the bereavement exclusion in the diagnosis of major depression. This development, announced on the DSM-5 website very early in the process (14), raised concerns about a possible trivialization of the concept of depression and consequently of mental disorder, since a depressive response to the death of a significant loved one is normative in several cultures (e.g., 15). It was also pointed out that, contrary to what reported on the DSM-5 website, the ICD-10 does exclude “bereavement reactions appropriate to the culture of the individual concerned” from the diagnosis of depression, and this exclusion is likely to be kept in the ICD-11 (16). The introduction of a note emphasizing the role of clinical judgment in the differential diagnosis between depression and a “normal” response to a significant loss has been thus regarded as a way to mitigate the consequences of the elimination of the bereavement exclusion and to facilitate the harmonization between the DSM-5 and the ICD-11.

As a matter of fact, this re-emphasis on clinical judgment is likely to be welcome by many clinicians worldwide, being perceived as a remarkable acknowledgement of the limitations of the operational approach, which arguably “does not reflect the complex thinking that underlies decisions in psychiatric practice” (17, p. 182). Indeed, in a large international WPA-World Health Organization (WHO) survey of practicing psychiatrists (18), more than two-thirds of respondents expressed the opinion that, for maximum utility in clinical settings, diagnostic manuals should contain flexible guidance allowing for clinical judgment rather than fixed diagnostic criteria.

So, the DSM-5 note does not come out of the blue, and can be seen as a further step in the articulated (and somewhat ambivalent) approach of mainstream American psychiatry to the issue of clinical judgment. However, the note leaves several questions open.

First, is it correct to assume that clinical judgment will have priority over operational criteria in determining whether the response to a significant loss is normal or pathological? In other terms, will it be possible not to make the diagnosis of major depression – in cases in which the severity, duration and distress/impairment criteria are completely fulfilled – because the depressive state appears, on the basis of what the clinician knows of the individual and his/her cultural background, a “normal” response to the loss? Or should we assume that the diagnosis of major depression will have to be made whenever the full criteria are met, and the exercise of clinical judgment be limited to doubtful or subthreshold cases? This is presently unclear, and this uncertainty is likely to introduce an “interpretation variance” in the application of the DSM-5 criteria for major depression which, added to the variance certainly produced by the exercise of clinical judgment, may substantially reduce the reliability of that diagnosis, already found to be “questionable” (kappa=0.20-0.35) in DSM-5 field trials (19) when using an early version of the criteria not including the note.

Second, do we assume that clinical judgment will be exercised by professionals with “specialized clinical training” (6, p. xxvii)? What about the non-specialist settings in which the diagnosis of depression is mostly made worldwide, where professionals may be unable to exercise an “expert” clinical judgment?

Third, the emphasis on the role of clinical judgment in the distinction between depression and “normal” responses to psychosocial stressors is likely to increase the burden of responsibility on clinicians in some contexts (e.g., community settings in areas heavily struck by the economic crisis) in which borderline cases are frequent and traditional differential diagnostic skills have become insufficient (see 20). The second note introduced in the DSM-5 definition of major depression – describing differential features between “normal” grief and depression – may be viewed as an attempt to support professionals in the exercise of clinical judgment. No similar guidance, however, is provided for the distinction between a depressive episode and “normal responses” to other psychosocial stressors, so that the clinician may be left again “on his or her own” (4, p. 8), exposed to several biases (see 21), when making a crucial and often delicate differential diagnosis.

Fourth, what will become of epidemiological research using lay interviewers, who by definition are unable to exercise clinical judgment when exploring whether a person has (or has had in the past) a period of “normal” sadness or a depressive episode? Can we afford using two different definitions of major depression, one for clinical purposes and the other for community epidemiological studies?

On the other hand, specifying those aspects of mental disorder which “are at present left to the uncertainties of clinical judgment” represents a challenge for psychiatry, since “reliance upon clinical skills implies that some aspects of psychiatric disorder are impossible at the moment to specify in an explicit manner” (22, p. 978). Indeed, the term “clinimetrics” (23) has been introduced to indicate “a domain concerned with the measurement of clinical issues that do not find room in customary clinical taxonomy” (17, p. 177). One could argue that the DSM-5 re-emphasis on clinical judgment may represent a stimulus to consider and develop this research line, which may be particularly relevant in the case of depression.