Dsm-v From The Perspective Of The Dsm-iv Experience

BACKGROUND

DSM-V from the Perspective of the DSM-IV Experience B. Timothy Walsh, MD*

ABSTRACT Objective: This article provides a brief overview of the development of the diagnostic criteria for eating disorders in DSM-IV. Method: The process by which DSM-IV was developed is reviewed, including perspectives on what constitutes diagnostic validity and clinical utility, and their importance in assessing proposed changes in diagnostic criteria. Results: The question of whether alterations in diagnostic criteria would clearly improve clinical utility was a major consideration in the DSM-IV process. Because of concerns that changes in diagnostic criteria would be disruptive and

Introduction As the process of developing DSM-V begins, it may be useful to look at the process which governed the development of DSM-IV. While history is not always an accurate guide to the future, it does seem likely that, in this instance, the best predictor of the future may well be the past. In that spirit, this article will briefly describe the process of the development of DSM-IV and some of the principles that emerged; these issues will be discussed from the perspective of the Eating Disorders Work Group for DSM-IV, which the author had the privilege of chairing.

The DSM-IV Process The full title represented by the DSM-IV acronym is the \Diagnostic and Statistical Manual of Mental Accepted 26 April 2007 *Correspondence to: B. Timothy Walsh, Department of Psychiatry, College of Physicians and Surgeons of Columbia University, 1051 Riverside Dr., Unit 98, New York, New York 10032. E-mail: [email protected] Department of Psychiatry, College of Physicians and Surgeons of Columbia University, and the New York State Psychiatric Institute, New York, New York Published online 15 June 2007 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/eat.20397 C 2007 Wiley Periodicals, Inc. V

might entail loss of established knowledge, the DSM-IV Task Force assumed a generally conservative stance vis-a`-vis change. Conclusion: The process of developing DSM-V is just beginning, and it is far from clear what alterations in diagnostic criteria for eating disorders will occur. However, the evolution of DSM-IV may provide a useful perspective on the deC 2007 by Wiley velopment of DSM-V. V Periodicals, Inc. Keywords: diagnosis; DSM-IV; DSM-V; anorexia nervosa; bulimia nervosa; binge eating disorder (Int J Eat Disord 2007; 40:S3–S7)

Disorders, Fourth Edition," and, while obvious, it is important to note that the DSM is a product and publication of the American Psychiatric Association. Consistent with the nature of this organization, comprised of some 35,000 physicians, the primary goal of the DSM is to enhance the care of individuals with psychiatric disorders. The first edition of DSM was published in 1952, and was followed by DSM-II (published in 1968), DSM-III (1980), DSM-IIIR (1987), and DSM-IV (1994). DSM-IV-TR (the TR indicating \text revision") was published in 2000; no revisions to the diagnostic criteria were made in DSM-IV-TR, but the accompanying text was updated based on extensive literature reviews. The publication of DSM-III in 1980 was a watershed event. Acknowledging that the etiology of most mental disorders was unknown, DSMIII attempted to be atheoretical in orientation, thereby becoming much more focused on descriptions of the clinical features of disorders. In an attempt to provide fuller descriptions and to increase reliability, DSM-III, unlike its predecessors, provided explicit criteria for each diagnosis. The number of diagnostic categories was significantly expanded, and, for the first time in the DSM system, Anorexia Nervosa and Bulimia were explicitly described. DSM-IIIR and DSM-IV remained consistent with the atheoretical and descriptive approach established by DSM-III. The timing of the publication of DSM-IV in 1994 was prompted, in part, by a wish to coordinate the

International Journal of Eating Disorders 40 S3–S7 2007—DOI 10.1002/eat

S3

WALSH

criteria and disorders of DSM-IV with the ongoing development of the 10th edition of the International Classification of Diseases (ICD-10) by the World Health Organization. Therefore, in 1988, the Board of Trustees of the American Psychiatric Association appointed a Task Force to develop DSM-IV. The Task Force membership included the chairs of thirteen Work Groups, each of which had responsibility for a category of disorders, such as Mood Disorders and Anxiety Disorders; among these was the Eating Disorders Work Group. Membership of each of the Work Groups was developed by the Work Group Chair in consultation with the leadership of the DSM-IV Task Force; the Eating Disorders Work Group was comprised of Paul Garfinkel, MD, Katherine Halmi, MD, James Mitchell, MD, G. Terence Wilson, PhD, and myself. Final membership was formally approved by the American Psychiatric Association. The Eating Disorders Work Group reviewed the DSM-IIIR criteria, solicited suggestions for revisions from colleagues in the field, and identified areas for further scrutiny. Examples included whether it would be useful to distinguish between restricting and binge-purge subtypes in Anorexia Nervosa, whether the frequency criterion for Bulimia Nervosa (at least twice a week for 3 months) was appropriate, and whether a new category, Binge Eating Disorder, should be introduced. In a series of methods conferences, it was decided that changes in the DSM criteria should be based as much as possible on empirical data, derived from extensive literature reviews, the reanalysis of existing data sets, and field trials. For the eating disorders, a few relevant data sets were identified and formally examined, but no field trials were carried out. Thus, changes recommended were based primarily on reviews of the literature.

Guidelines for Change Arguably, the salient question in the development of DSM-IV was \what guidelines should be followed in deciding to make a change?" In other words, what are the criteria for judging the diagnostic criteria? This issue prompted numerous discussions at Task Force meetings, and led to a number of publications describing the DSM-IV process (for example, Frances et al.,1 Widiger et al.2). S4

Cleave Nature at Its Joints?

Intuitively, an ideal method would be to identify features that distinguish among conditions as they exist naturally, and to utilize these features in constructing and revising diagnostic criteria. Presumably, if sufficient data were available, it would be clear from inspection where the dividing lines between disorders lie, and diagnostic criteria could be devised in a completely objective fashion. Unfortunately, it becomes quickly apparent that such an approach is fraught with problems. A fundamental problem with this approach is that it assumes such \joints" or discontinuities can be identified if the sufficient research is conducted. In fact, there are substantial data that the disturbances of emotion and behavior that the DSM system attempts to place in categories, in fact, occur along continua. A number of authors suggest that natural boundaries between mental disorders and normality simply may not exist (see Kendell and Jablensky3). An additional issue is that describing naturally occurring dividing lines that distinguish among disorders is, in fact, not the primary purpose of the DSM system. The critical goal of DSM is to improve the care of individuals with mental disorders. Thus, a diagnostic label should make it possible for clinicians to more easily and accurately obtain and apply information on course, complications, and effective treatment options. For example, it is possible to imagine a specific genetic or environmental factor that contributes importantly to the development of eating disorders, and helps to distinguish participants with eating disorders from others in the population. However, unless identification of this specific factor clearly improves the clinician’s ability to care for patients with eating disorders, it may not be useful to include it in the diagnostic criteria. These considerations reflect the difficult problem of establishing the validity of diagnostic criteria. Validity is a complex, if not somewhat nebulous, concept, and a variety of facets of validity have been described, including, for example, content, constructive, and predictive validity. But, as Kendell4 emphasizes, in the context of psychiatric diagnostic criteria, the crucial element of validity is predictive power. That is, having established a diagnosis, is the clinician better able to predict clinically important features, such as course of illness, complications, and treatment response? In other words, a crucial element of validity is clinical utility. Similarly, consideration of clinical utility may aid decisions about competing sets of diagnostic crite-

International Journal of Eating Disorders 40 S3–S7 2007—DOI 10.1002/eat

DSM-V FROM THE PERSPECTIVE OF THE DSM-IV

ria. As described by Kendler,5 the relative validity of different criteria sets for the diagnosis of schizophrenia may vary depending on the choice of the validator. There may be a greater degree of familial aggregation of the disorder among individuals receiving the diagnosis on the basis of one set of criteria, but individuals receiving the diagnosis based on another criteria set may have a more similar clinical outcome. If true, and if improved prediction of clinical outcome is judged to be a crucial component of clinical utility, the second criteria set would presumably be more appropriate for inclusion in DSM. As Kendler notes, the choice between the criteria sets may rest not on scientific validators, but on judgments about the fundamental goals of the diagnostic system. Two alternate criteria sets for a diagnostic category may both be scientifically valid; in the earlier example, one may more reliably predict familial association, while the other may more reliably predict course. A value judgment is required as to which better serves the purpose for which the diagnoses are being made; for DSM, the fundamental purpose is to enhance clinical utility. Clinical Utility?

If clinical utility is a crucial standard by which diagnostic criteria should be judged, it is important to attempt to identify some of its components.3,6 These include ease of use, reliability, and predictive value. The diagnostic criteria should be relatively easy for mental health professionals, who are likely to have numerous demands on their time, to use. As much as possible, the criteria should therefore be intuitively sensible and understandable, easy to remember, and brief. Diagnostic criteria that do not require the endorsement of a minimum number of a long list of possible symptoms are preferred. Similarly, the diagnostic criteria should not require specialized training or instruments to be used. In other words, the DSM criteria should serve the needs of busy clinicians caring for patients, not just of researchers. To be clinically useful, diagnostic criteria should be reliable. Presented with the same symptoms, different clinicians should arrive at the same diagnosis. Presumably, concrete fact-oriented criteria should be more reliable than criteria which require the assessment and interpretation of complex intrapsychic processes. Finally, and arguably most importantly, the assignment of a specific diagnosis should have specific clinical implications. The utility of a diagnostic

category is greatly strengthened if individuals receiving that diagnosis are likely to develop different complications, to have different short- and long-term outcomes, and to respond to different treatment interventions compared with individuals receiving another diagnosis. As an obvious example, consider the different implications of the diagnoses of schizophrenia and of manic-depressive illness. As suggested by First et al.,6 issues of clinical utility should ideally be addressed by empirical studies. Unfortunately, while data from such studies would be of enormous value in revising diagnostic criteria, few studies are available that carefully examine the comparative reliability and acceptability, or the relative advantages and disadvantages, of different diagnostic schemas or criteria sets. In DSM-IV, criteria for Binge Eating Disorder were included in an appendix containing proposals for new categories, and the publication of these criteria has certainly helped spur interest and valuable research. However, very few studies have examined an issue explicitly raised in the criteria, \Future research should address whether the preferred method of setting a frequency threshold is counting the number of days, on which binges occur, or counting the number of episodes of binge eating" (DSM-IV, p. 731). Such issues, which assume great significance when revisions to the diagnostic criteria are being considered, receive scant attention from investigators and funding agencies at other times.

Restraints on Change Even when empirical information is available suggesting that changes in and additions to the DSM would be worthwhile, a number of factors operate to inhibit alterations. Change will be Disruptive

Even changes that enhance the clinical utility of diagnostic categories will inevitably be disruptive to implement. Clinicians will be required to learn new terms, new criteria, and, possibly, new diagnostic conceptualizations. In most instances, they will have to acquire and study the newest version of the DSM. Insurers and third-party payers will have to review changes in diagnostic criteria and categories, and determine the impact on reimbursement. Separate of the financial costs involved, time will be required for clinicians and their patients to become familiar and comfortable with

International Journal of Eating Disorders 40 S3–S7 2007—DOI 10.1002/eat

S5

WALSH

the new system, and, along the way, there will be an inescapable period of some confusion and uncertainty. Knowledge will be Lost

Revisions in the diagnostic criteria for eating disorders will hopefully increase clinical utility, and thereby set the stage for important gains in the care of participants with eating disorders. On the other hand, if the criteria required for a diagnosis are altered, especially if they are altered significantly, information obtained from research utilizing the previous criteria may not apply. Obviously, the greater the change in the criteria, the greater the threat to previously established facts. If the criteria for Anorexia Nervosa and Bulimia Nervosa are substantially altered, how much uncertainty will arise concerning the utility, or lack thereof, of current psychological and pharmacological treatments? Might treatment guidelines be rendered invalid? While these extreme fears are unlikely to come to pass, some degree of uncertainty about the applicability of previous knowledge is an unfortunate but inevitable byproduct of change. DSM-V will Receive Extraordinary Scrutiny

The American Psychiatric Association and the DSM-IV Task Force carefully articulated and were committed to following a scientific and empirical, evidence-based approach to revising diagnostic categories, and undoubtedly the development of DSM-V will follow similar principles. Nonetheless, both specific decisions and the entire DSM enterprise have been subject to criticism from within and without the mental health field, including calls for Congressional hearings. Constructive criticism should improve both the process and the product, but it seems likely that the prospect of negative reactions will be an additional restraint on change. Available Information will Likely be Insufficient

The greatest restraint on the work of the DSMIV Eating Disorders Work Group was simply the lack of sufficient data to address crucial questions about the DSM-IIIR criteria. For example, concerns were expressed about the appropriateness of the \twice-a-week" frequency criterion required for Bulimia Nervosa, but the Work Group concluded that there were insufficient data available to justify a change. Similarly, while there was great interest and excitement about the S6

then-recently introduced Binge Eating Disorder, the Work Group believed that not enough was known about the clinical utility of the diagnosis to merit its inclusion in DSM-IV as a recognized disorder. A clear and conservative principle governing changes considered for DSM-IV was that, in the absence of convincing information, no changes should be made. Therefore, while the DSM-IV Eating Disorders Work Group had significant concerns about several aspects of the existing criteria, the lack of substantial data about alternative diagnostic formulations was a major limitation on change.

Implications for DSM-V The DSM-IV experience has several implications for the DSM-V process. The disruptive nature of change and the threat that important clinical information may be rendered obsolete by major alterations imply that the process of revising DSM-IV should be conservative. The Work Group and the Task Force should recommend changes only that, as best as can be judged, will clearly lead to significant improvements in the care of participants with eating disorders. The broad influence of the DSM system, not only on the health care system but also on societal views of behavior and mental illness, imply that the process of considering revisions to DSM-IV should be an open one. The Work Group and the Task Force will undoubtedly benefit from input from colleagues and critics, and, hopefully, consideration of different perspectives will lead to more useful diagnostic criteria. Openness will also serve to build consensus around proposed changes and blunt, to some degree, inevitable criticism. Finally, the process of revising DSM-IV should follow the field, not lead it. The principle that changes should be based on empirical evidence implies that it is not the task of the Work Group to devise new and untested diagnostic categories and criteria. Rather, the Work Group and Task Force should ideally serve as an educated but unbiased jury to consider ideas and proposals that the field has developed, and to evaluate these in light of the principles described earlier. The DSM-V criteria for eating disorders should therefore reflect the progress of the field in the years since DSM-IV, and

International Journal of Eating Disorders 40 S3–S7 2007—DOI 10.1002/eat

DSM-V FROM THE PERSPECTIVE OF THE DSM-IV

provide a foundation for future study of eating disorders and for improved care of individuals afflicted with them.

References 1. Frances AJ, Widiger TA, Pincus HA. The development of DSM-IV. Arch Gen Psychiatry 1989;46:373–375.

2. Widiger TA, Frances AJ, Pincus HA, Davis WW, First MB. Toward an empirical classification for the DSM-IV. J Abnorm Psychol 1991;100:280–288. 3. Kendell R, Jablensky A. Distinguishing between the validity and utility of psychiatric diagnoses. Am J Psychiatry 2003;160:4–12. 4. Kendell RE. Clinical validity. Psychol Med 1989;19:45–55. 5. Kendler KS. Toward a scientific psychiatric nosology. Strengths and limitations. Arch Gen Psychiatry 1990;47:969–973. 6. First MB, Pincus HA, Levine JB, Williams JB, Ustun B, Peele R. Clinical utility as a criterion for revising psychiatric diagnoses. Am J Psychiatry 2004;161:946–954.

International Journal of Eating Disorders 40 S3–S7 2007—DOI 10.1002/eat

S7