Diagnostic Nomenclatures in the Mental Health Professions as Public Policy

It Is Practical for Curated Classifications to Be Developed by Smaller Committees of Experts

During my time in academia I have been very involved in shared governance, particularly with respect to formulating academic policy. I have chaired committees that have written or rewritten policies for tenure and promotion, post–tenure review, academic program review, sabbatical leave, and employing non–tenure track faculty. These are important and potentially controversial policies. My experience has taught me that even when a committee is composed of 10 to 14 members, the bulk of the work on these kinds of policies is often done by three to five members. That initial work is nearly always modified by wider input, so one cannot get too invested in initial drafts, but often success begins and ends with the smaller group.

With this in mind, my concern is that to actually treat some of the Standards and Guidelines’ statements about committee membership as best practices would paralyze the development process. Among those listed who should have decision-making authority are professionals whose intellectual commitments counterbalance other professionals’ intellectual commitments, and all relevant mental health professions including social workers, counselors, peer specialists, and family support specialists. In addition, it is stated that there should be representatives of service users, affected individuals who have elected not to receive services, and family members. With respect to service users, the Standards and Guidelines imply that enough of them should be selected to adequately represent that population, including the diverse cultural and demographic subgroups it contains.

The different stakeholders listed should have input and their perspectives considered, but it would be impractical to require that, in order to be representative, a development committee should be stratified like a norm group in psychological testing. This may be an established best practice for norm groups, but not for development committees. For a task as complicated as developing a classification system, it is important that that structure include many working committees, but those committees would work best with 7 to 14 members, and realistically not all members will contribute equally on every committee.

The most complicated part of the process, politically, is what happens when the initial proposal is assessed by the different stakeholder groups outside the committee, but to make those external perspectives an inherent part of the committee itself would be self-defeating. One example of how incompatible commitments cannot always be overcome, even among people relatively high on agreeableness, might be the committee that was tasked with deciding whether to include “premenstrual dysphoric disorder” in DSM-IV (Zachar & Kendler, 2014). Given my earlier comments about higher and broader standards, a big tent model is necessary, but there are more and less practical places to erect the tent.

Unless one is committed to a strong form of relativism, the insinuation that every perspective deserves to be weighted equally strikes one as missing the mark. Which perspectives should be weighted higher partly depend on the goals for the classification. For example, the Standards and Guidelines recognize that a classification that is constructed for the purpose of predicting clinical course may look different than one that is constructed to enhance communication between clinician and client. Classifications like those in the DSM, however, are coddled together to address a variety of scientific and professional goals. That coddling together occurs over time because it is not realistic to emphasize all possible goals at once.

When working on a classification system that will have so many uses, which perspectives do we weight higher and lower? Two basic principles come to mind. One is that committees should be composed of people who value classification, not those who are classification skeptics. The second is that committee members should be thoughtful people who have some awareness of how classifications work, including knowledge of the flaws and limits of classification. This kind of expertise is loosely analogous to institutional memory in an organization. Institutional memory includes understanding why certain policies were implemented in the past and knowing about policies that were implemented, but did not work. To this I would add as aspirational the Standards and Guidelines’ recognition that evaluators overall should possess the expertise and knowledge for assessing the scientific status, ethicality, and potential public health implications of the proposed revisions.

In addition to making suggestions about who should be included in the development process, the Standards and Guidelines also make suggestions about who should be disqualified. In doing so, the document borrows some language from The Institute of Medicine’s statement on conflict of interest regarding financial ties between physicians and industry (Institute of Medicine, 2009).

The Standards and Guidelines, however, expand conflict of interest to include any institutional, intellectual, ideological, or other conflict of interest. My worry is that that this is so open-ended that it could also paralyze the process. Any secondary interest that could bias a professional judgement about the goals of classification system could refer to any putative bias at all.

One problem with using this open-ended standard in the selection of committee members is that the assessment of whether the putative bias will affect someone’s judgement is often a guess, and potentially applied selectively. For instance, in DSM-5 development, judgments were made about potential intellectual conflicts of interest with respect to the mission of the Personality and Personality Disorders Work Group. In retrospect, some judgments about who would and would not be able to put aside prior intellectual commitments were likely mistaken as some members resigned from the group because they would not accept a compromise model that was a hybrid of dimensions and categories (Zachar, Krueger, & Kendler, 2016).

Even the Standards and Guidelines’ aspiration to completely eliminate any financial conflicts of interest with industry, while attractive, has practical drawbacks. It is certainly important to not tolerate self-serving corruption in which industry profits are given equal or greater weight than patient care or scientific progress. It would also be naïve to not acknowledge that a past positive relationship with an industry or its personnel can sensitize someone to corporate interests. Acquiring industry support for your research, however, does not mean you inevitably adopt the profit motive in favor of patient care or scientific progress. Context matters: Whether a financial relationship is disqualifying depends on how much support, for how long, for what project.

It Is Not Practical to Expect a Classification Manual to Be a Handbook of Psychopathology

A common criticism of the DSM is that the architects keep adding disorders with each new edition, thus expanding what falls under the purview of the mental health professions. For example, the DSM-II was 134 pages long, the DSM-III was 494 pages long, the DSM-II-R was 567 pages long, and so on up to the 947-page DSM-5. Such a comparison can be deceptive. A more relevant comparison would be to compare the number of disorders across the manuals. For example, Regier, Narrow, and Kupfer’s (2014) count of the number of disorders in each DSM was as follows: 80 in DSM-I, 117 in DSM-II, 145 in DSM-III, 172 in DSM-IV, and 157 in DSM-5.

What accounted for so many additional pages in the transition to DSM-III? It is not just the addition of disorders. For example, a pocket version of the DSM-III that contained only the diagnostic criteria was not much larger than the DSM-II (American Psychiatric Association, 1980). What accounted for the additional pages is that the DSM-III was written more like a textbook. It expounded on the diagnostic criteria with lengthy narratives and added several sections including age of onset, gender differences, time course, and differential diagnosis.

One unintended consequence of being more like a textbook is that the classification became subject to criticisms for those aspects of psychopathology that it did not address, that is, it is criticized for failing to be an ideal textbook of psychopathology. The first-person experiences, personal narratives, and phenomenological analyses called for in the Standards and Guidelines are important features of an informed understanding of psychopathology. They may also be relevant to the development of classifications, but that does not mean they should be incorporated as standards into the classification. It is not realistic for a classification system to do the work of an ideal textbook of psychopathology, and more.

An informed understanding of psychopathology is a prerequisite to the competent use of the classification system. It is the understanding of psychopathology, not an inert classification system, that does the important work. This message is part of what is being communicated in the rarely read introductions to the diagnostic manuals. Indeed, training programs can focus too much on learning the DSM or the ICD, and not enough on studying psychopathology. If the aspirations articulated were directed at the knowledge of psychopathology rather than its classification, the Standards and Guidelines’ call for inclusiveness would gain force. Still, no book or single person can acquire expertise in all aspects of psychopathology. That expertise necessarily resides in a community.

Conclusions

During the development of the DSM-5, some of the proposed changes were not approved by the American Psychiatric Association Board of Trustees. One of these was a proposal to include an assessment of symptom measures that cut across the diagnostic categories. These included depressive symptoms, symptoms of anxiety, and somatic symptoms. Although a desirable addition, the reason the cross-cutting symptom measures were not included in the main section of the manual was that the American Psychological Association. Assembly (a group of practitioners) was worried that their inclusion would establish a standard of practice, and a failure to have these screening results documented in a medical record could expose psychiatrists to malpractice liability. Whether the cross-cutting measures represented a standard, a suggested best practice, or an aspiration was (and remains) a matter of dispute.

As aspirational, the Standards and Guidelines deserve careful study by any person or group of persons working on developing a classification or a classification system. They usefully orient developers to the human consequences of their proposals. But if marketed as standards of care or best practice standards, I fear that they will be considered impractical, and ignored. I recognize that where I have drawn the lines between guidelines and standards may also be a matter of dispute. That is inherent to public policy.