611 W. Union Street
Benson, AZ 85602
(520) 586-0800

Health Choice Integrated Care crisis Line

NurseWise 24-Hour Crisis Line


611 W. Union Street
Benson, AZ 85602
(520) 586-0800

NurseWise 24-Hr Crisis Line


powered by centersite dot net

Getting Started
Here are some forms to get started. These can be printed and brought with you so that you can pre-fill out some known info ahead of time. More...

Health Policy & Advocacy
Basic InformationMore InformationLatest NewsQuestions and AnswersLinksBook Reviews
Related Topics

Health Insurance

by Thomas A. Widiger, Allen J. Francis, Harold Alan Pincus, Ruth Ross, Michael B. First, Wendy Wakefield Davis (editors)
American Psychiatric Association, 1996
Review by Jeffrey Poland, Ph.D. on Feb 25th 2002

DSM-IV Sourcebook

The DSM-IV Sourcebook, Volume 2 is the second of a projected four volume series aimed at presenting the empirical bases and rationales underlying the recommendations and decisions that informed the development of DSM-IV (The Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, developed and published by the American Psychiatric Association 1994.)  (Note that the fourth volume is out of print.)  Given the central role of the DSM in clinical practice and scientific research bearing on mental illness and given the considerable controversy that has traditionally accompanied psychiatric classification and has surrounded DSM based classification in particular, it was deemed important by the DSM developers to lay bare the empirical basis for the decisions leading to the production of the most recent DSM. In so doing, the scientific credibility of the classification system and of the practices based upon it would be affirmed. The DSM Sourcebook is a massive record of the results of literature reviews, data re-analyses, and field trials and of the deliberations based upon them.

In a prior review of the DSM-IV Sourcebook, Volume 1 (Metapsychology April 2001), I provided a general outline of the process by which the DSM-IV was developed and the various guidelines and standards that were supposed to inform that process. In addition, I outlined a number of lines of criticism of the DSM approach to classification. The main thrust of that review was to examine Volume 1 of the Sourcebook with an eye to determining how well the work recorded in that volume lived up to the standards of the developers and how well it responded to the concerns of the critics. In the present review, I will continue pursuing those objectives with respect to Volume 2.

As with Volume 1, this volume consists of a series of papers authored by members of the various work groups charged with identifying critical issues to be addressed in the development process, conducting literature reviews designed to collect relevant evidence, conducting data re-analyses and field trials to gather further evidence, framing development options for consideration, and eventually making recommendations for development of DSM-IV. The volume is divided into sections concerning five broad categories of mental disorder (Mood Disorders, Anxiety Disorders, Personality Disorders, Psychiatric System Interface Disorders, Sexual Disorders), and each of these sections is broken down into an introductory chapter (providing an “executive summary”) followed by a series of chapters summarizing the results of literature reviews relevant to issues related to specific disorders and presenting options and recommendations concerning those issues. A sixth section of the volume is devoted to Late Luteal Phase Dysphoric Disorder (LLPDD), and this section consists of a single long chapter with the same purposes as the other five sections. The existence of a separate section devoted to LLPDD is perhaps best explained in terms of the considerable controversy surrounding that category since its initial proposal in the mid 1980's.

The papers collected in this volume exhibit a number of shortcomings that lead one to conclude that the development process did not live up to its own standards and has not achieved some of its most important stated objectives. To begin, it does not seem likely that the DSM development process, as reflected in Volume 2 of the Sourcebook, has succeeded in effectively breaking with the long tradition (re-affirmed in both DSM-III and DSM-III-R) of basing decisions regarding the existence and nature of diagnostic categories on consensus among practitioners rather than on sound science, either in the form of carefully collected empirical data or well tested theories. There are at least three reasons for thinking this. First, throughout the volume there are repeated affirmations of the lack of relevant empirical research findings bearing on issues concerning the construct and predictive validity of the categories (i.e., their scientific meaningfulness.) Most of these issues could not be so much as seriously raised, let alone resolved during the DSM-IV development process. Second, as a consequence of the conservative approach to change (viz., to make changes only when there is a “solid basis” for doing so) and the lack of relevant empirical research, the inertia supplied by past editions of the DSM dominated this most recent effort. Since it is generally agreed that DSM-III and DSM-III-R were developed on the basis of a politically derived consensus (sometimes not even among “experts”; cf., p. 647) and without being informed by any serious research bearing on the validity of the categories, it can only be concluded that DSM-IV consists largely of categories which have been introduced and retained, not on the basis of science, but on the basis of political consensus unconstrained by empirical research. Third, the explicitly articulated guideline of making decisions that increased the consistency of DSM-IV with the ICD-10 (The International Classification of Diseases, 10th Edition), means that recommendations for change were made, not on the basis of empirical data addressing validity, but with an eye to enhancing agreement with a different system which, at least with respect to mental disorders, is also based on consensus largely uninformed by science.

A further shortfall with respect to the stated standards and objectives of the development process concerns the ideal of making significant changes only when there was a “solid basis” for doing so. Throughout this volume, the widespread lack of relevant and good quality research made the process quite vulnerable to loose, sometimes rambling and barely coherent, speculation. A frequent refrain throughout the volume involved vague appeals to “ease of use” and “clinical utility” without any clarification or justification. More serious problems perhaps are the imposition of dubious ideas onto the reasoning process (e.g., the idea that it is important to minimize differences between children, adolescents, and adults, p. 4-5) and pursuit of lines of reasoning which are barely relevant to an issue under discussion or barely coherent (e.g., discussions of whether pain is a mental disorder, p. 923-4). It is one thing to call for changes for which a solid basis exists, it is quite another to have and adhere to a serious standard of what counts as a solid basis.

Two specific sorts of change recommended for introduction into DSM-IV are especially noteworthy in both their significance and their lack of a solid basis. The first concerns the widespread introduction of a criterion of “clinical significance” (e.g., clinically significant distress or impairment in social, occupational or other important area of functioning) for most if not all categories. None of the work groups identified any relevant data to support such an introduction as improving the construct or predictive validity of the categories they considered. Apparently, the introduction of this criterion was mandated by the Task Force overseeing the work groups (e.g., p. 1082) perhaps because of some belief that raising the bar on distress and impairment (“clinical significance”) might have the effect of reducing false positives, creating more homogeneous categories, or reducing inflated prevalence rates. However, there is no reason to believe that a criterion of clinical significance will achieve either greater categorical homogeneity or fewer false positive diagnoses; and the proposal, as it bears on prevalence, seems quite arbitrary in the absence of an independent measure of the presence or absence of a disorder. Such a change clearly is significant insofar as it bears on who is diagnosed with a mental disorder for either clinical or research purposes; but it is far from solidly based.

The second sort of change, which is significant but lacking in a solid basis, concerns the indirect introduction into the official nomenclature for mental disorders of categories which were not deemed to have a sufficient basis for direct inclusion. This is typically accomplished by citing a category as an example of a residual category of mental disorder (e.g., Depressive Disorder Not Otherwise Specified.) Such categories exist so that someone who does not satisfy the criteria for any listed category can still be diagnosed as suffering from a mental disorder. At east two categories discussed in Volume 2 of the Sourcebook were determined not to have a sufficient basis to be included as an official diagnostic category, but were nonetheless included as examples of mental disorders which can fall under a residual label (i.e., they are real mental disorders that just don’t happen to satisfy one of the listed categories.) One example of this, “Telephone Scatalogia” (i.e., the making of obscene phone calls) demonstrates how something which certain clinicians believe “in their bones” is a sign of mental pathology can make it into the official classification system of mental disorders (under the heading, “Paraphilias Not Otherwise Specified”) despite there being no empirical or theoretical basis for believing this represents a legitimate mental disorder. A second example, “Premenstrual Dysphoric Disorder” (the new label for what was previously called “Late Luteal Phase Dysphoric Disorder”), is a clear case of the introduction into the diagnostic system through the backdoor of a category that was heatedly debated and found to lack a sufficient basis for direct inclusion. PMDD is now listed as an example of a mental disorder that qualifies one for a diagnosis of Depressive Disorder NOS.

Yet a third reason for questioning whether the DSM-IV development process, as reflected in Volume 2 of the Sourcebook, lived up to the stated ideals and objectives of the developers concerns its conformity to ideals of scientific objectivity frequently rehearsed throughout the volume. One such ideal concerns the openness of the development process: it is pointed out by the developers that the DSM development process involves many stages with ample opportunity for critical review and discussion of evidence, options, deliberations, and recommendations by the substantial number of consultants and the psychiatric community at large (e.g., by circulation of interim drafts, presentations at conferences, publications). In this way, the ideals of peer review and open critical discussion in the relevant scientific community are supposed to be realized. But, there is reason to be skeptical about just how open the process was. In particular, with respect to Late Luteal Phase Dysphoric Disorder (or, PMDD), there are grounds for concern that the process involved significant marginalization of critical perspectives and the stifling of dissent. Although the chapter concerning LLPDD mentions the controversy and provides a limited discussion of some of the concerns of critics, there is no very sophisticated discussion of the issues involved. Rather, the chapter is for the most part devoted to a long hashing over of a largely inconclusive research record and a clarification of just how little empirical support there is for direct inclusion of the category in the DSM. But, then, PMDD was included in the DSM nonetheless and despite the controversy surrounding the legitimacy of the category, a controversy which, if properly engaged, might have produced a solid basis for exclusion of the category entirely from the DSM. See Caplan 1995 for an alternative perspective on the process from that reflected in the Sourcebook. There it is suggested that, far from being open, the process ignored criticisms, stifled dissent, and marginalized dissenters. The importance of this episode is that it suggests that those controlling the DSM development process may not be very willing or able to grapple with issues that challenge implicit or explicit interests or presuppositions. Controversy may be permitted, but only within rigid assumptions to which many participants may well be blind and only subject to politically influenced procedural practices. At the least, this is something requiring closer scrutiny given the broad impact of the DSM system of classification.

Let us turn now to the question of how well the papers in Volume 2 of the Sourcebook, either individually or collectively, provide effective rejoinders to the critics of the DSM approach to classification of mental disorder. There are at least three areas of concern: foundational issues, validity of the categories, impact of the inertia of the system.

First, as with Volume 1, there was in Volume 2 a notable lack of serious attention to foundational issues and a serious mishandling of those that were discussed. There was frequent confusion about the nature of the entities being included in or excluded from the taxonomy (e.g., Are they empirically defined syndromes?, Are they disorders involving a core dysfunction?), and hence there was lack of clarity regarding the sorts of considerations relevant to making key decisions (e.g., p 5.) Further, discussions regarding the distinction between disorder and no disorder (normality and abnormality) were marred by a lack of understanding of and attention to conceptual issues raised by such questions. Hence, there was again an evident lack of clear understanding of what needs to be shown in order to effectively address important questions (e.g., p. 955-6.) And, as noted above, the broad incorporation of a clinical significance criterion was made without any serious attention to the issues raised by the problems of categorical heterogeneity, inflated prevalence rates, and false positives. Finally, an extended discussion regarding whether a categorical approach to classification should be replaced by a dimensional approach is found in the section concerning the personality disorders. The somewhat dismaying conclusion of the discussion was that, even though there is a broad consensus that the categorical approach needs to be replaced because it is riddled with problems and is not promoting either clinical or research purposes very effectively, the categorical approach will be retained in DSM-IV because it is the standard for research and clinical purposes and because changes would be too confusing and disruptive (cf., p. 647-650.) The deep problems with changing the DSM come into full view here: even when a solid basis for change exists, change proves impossible because of the powerful inertia of the interests and the tradition that supports the current form of the classification system. Never was there a more patent need for clarity about the foundations of a discipline than is evident here; and yet, the DSM development process, to date, has given no serious attention to such issues.

Second, the evident lack of empirical research bearing on the most critical issues concerning the development of a categorical system of classification, viz., the validity of the categories, meant that these issues could not so much as be meaningfully framed, let alone addressed by the Work Groups represented in Volume 2. As a consequence of both this lack of evidence and the conservative approach of the development process, the discussions represented in Volume 2 (as in Volume 1) displayed a very powerful bias to retention of categories (the vast majority of categories) which had been introduced in DSM-III and retained in DSM-III-R on the basis of no sound scientific evidence of validity. As a consequence, beginning in 1994 (the publication date of DSM-IV), we have had eight more years (and counting) of powerful influence on clinical and research practice of a system of classification containing categories with no demonstrated validity. Critics of the DSM approach to classification have contended for two decades that this approach (as initiated in DSM-III) is not a sound way to develop a system of classification that exhibits appropriate levels of validity, and the deliberations in Volume 2 of the Sourcebook provide nothing to encourage a more optimistic view.

Finally, since the deliberations and recommendations made in Volume 2 were rarely based on sound scientific considerations, they were inevitably made on the basis of considerations that tend to corrupt future scientific research. Thus, deliberations and recommendations based on consistency with ICD-10, clinical utility (“ease of use”), empirically unconstrained speculation, and the imposition of ideas and criteria with no empirical basis introduce arbitrary and unjustified factors that directly influence diagnostic practices in research settings and hence that influence the nature of the samples that are collected and studied. The failure of the DSM developers to appreciate this sort of impact has led to a continued endorsement of a flawed model of the relation between science and the clinic, a model that undermines research. The seemingly bankrupt research programs associated with the various categories discussed in Volume 2 are not likely to be revived given the way that diagnostic criteria are developed.

What is the significance of these shortcomings in the DSM development process as reflected in Volume 2 of the Sourcebook? The result of laying bare the nature of the scientific evidence and the deliberations supporting the decisions leading to the publication of DSM-IV is to bring into sharper relief just how deep the problems with the DSM system of classification go. The developers of DSM-IV appear to have been massively uninformed by sound scientific research and to have been deeply confused about a number of important foundational issues. Both of these shortcomings impact upon the coherence and credibility of the process, as well as upon its capacity to clarify and effectively cope with important problems as they arise. The result is that powerful inertial forces direct the development process and lead to retention of categories that have never had a scientific justification. These problems are compounded by concerns about the objectivity and scientific integrity of the DSM development process and about the potential for future research based upon DSM categories to be productive. Volume 2 of the Sourcebook has done nothing to allay such concerns, and has provided additional grounds for thinking it is time for some profound changes in how we conceptualize, investigate, and respond to mental illness, and in how a highly influential system of classification gets developed.


© 2002 Jeffrey Poland


Jeffrey Poland, Ph.D., University of Nebraska-Lincoln




Caplan, Paula (1995). They Say You’re Crazy: How the World’s Most Powerful Psychiatrists Decide Who’s Normal (New York: Addison-Wesley).