Sometime in 2013, we’ll presumably be forced to shell out a hundred bucks or so for the fifth edition of DSM. While we do not have the space to review every potential change here, we will cover those of most clinical importance. If you want to see all of the proposed changes, visit www.dsm5.org, which was our source for much of the information in this article.

A Nutshell History of the DSM
In 1943, the U.S. military published a list of mental disorders with the rather inscrutable title, “War Department Technical Bulletin 203.” The American Psychiatric Association adapted this document and published it in 1952 as the “Diagnostic and Statistical Manual: Mental Disorders,” or what we now know as DSM-I. History buffs can peruse a PDF of this document on the web at http://bit.ly/ZBCCG.

In 1968, DSM-I was superseded by DSM-II, which was based on psychoanalytic principles. Thus, for example, what we now call panic disorder was considered a form of anxiety neurosis, and was thought to be caused by the surfacing of forbidden material from the subconscious.

In 1980, Columbia University psychiatrist Robert Spitzer oversaw the publication of DSM-III, which banished the term “neurosis,” and with it any mention of psychoanalytic theory. In addition, DSM-III introduced the by-now familiar lists of criteria, along with minimal threshold numbers of symptoms required for patients to qualify for a disorder. A patient with panic disorder, for example, was no longer considered neurotic, but instead was defined as suffering at least four out of 13 possible symptoms—symptoms which, at least in theory, could reliably be recognized by any qualified psychiatrist. DSM-III also inaugurated the controversial multi-axial system of diagnoses.

As you can see from the table accompanying this article, each subsequent version of the DSM has expanded the number of diagnoses, but the essential structure of the manual has remained intact. That is, until now.

DSM-5 proposes two different types of changes: foundational and incremental. The foundational changes include radically transforming the multi-axial system; adding dimensional and severity scales to many diagnoses; and reformulating the way we think about personality disorders. The incremental changes are basically refinements in diagnostic criteria for a variety of disorders, with a few name changes here and there.

Foundational Changes
The multi-axial system would undergo radical surgery. We all know the multi-axial drill: axis I is for major mental disorders; axis II, personality disorders and mental retardation; axis III, medical illnesses; axis IV, psychosocial factors; and axis V, global assessment of functioning (GAF). In DSM-5, the first three axes would be truncated into a single axis, which would include all psychiatric and medical diagnoses. That would make our lives a bit easier—but don’t relax too much, because plans are afoot to turn axes IV and V into a paperwork marathon.

Evidently, the DSM-5 committee likes the way the International Classification of Disease (ICD) treats social functioning and disability issues and wants to adopt portions of it. But a quick glance at ICD’s disability questionnaire (which you can view online at http://bit.ly/9YEN20) will make you quake in your couch—it is a 15 page long checklist.

Admittedly, this version covers all physical and mental disabilities, but even filling out the psychiatrically relevant items for each patient would add hours to our days. Let’s hope common sense prevails in the end.

Questionnaires, scales, and tests, oh my! Many of you may already be using rating scales in your practice, such as the Patient Health Questionnaire or the Connors Rating Scale. Get ready for more. The DSM-5 is likely to recommend that you use so-called “cross cutting dimensional assessments” for most of your patients. The goal is to quantitatively measure commonly occurring symptoms that “cut across the boundaries of any single disorder” (hence, the name)—symptoms such as depressed mood, anxiety, anger, sleep problems, and substance use. In theory, these scales may be helpful, but would they truly add clinical utility to what we are already doing in our practices when we interview patients? Maybe or maybe not—but they would unquestionably increase our time spent on paperwork.

Personality disorders would get an extreme makeover. Prepare for a double-take when you see the personality disorders section of the DSM-5. A massive overhaul would include a new system to describe personality dysfunction, and the removal of a number of personality disorder types we’ve come to know so well. Bid farewell to schizoid, paranoid, histrionic, narcissistic, and dependent personality disorders, as well as that admittedly overused standby, personality disorder NOS. Surviving the cut are the following five: schizotypal, avoidant, borderline, antisocial/psychopathic, and obsessive compulsive personality disorders.

While the reduction in the personality disorder types represents a significant change from DSM-IV, it is the shift to a dimensional model of personality assessment that would likely leave many in the psychiatric community scratching their heads. In an effort to create an individualized, person-centered approach, the DSM-5 would include a four part assessment of personality, which would include:

1. Severity scale. Rated from zero (no impairment) to four (extreme impairment)


2. Type match. To what degree does a patient’s personality match one of the five remaining personality types, from one (no match) to five (a good match)


3. Trait domains and facets. Each personality type may have up to six “trait domains”—negative emotionality, introversion, antagonism, disinhibition, compulsivity, and schizotypy. Each trait domain is further broken down to more specific “trait facets.” You then rate each trait on a scale from zero (very little or not at all descriptive) to three (extremely descriptive)


4. Personality disorder. Finally, you determine: “Does the person meet criteria for a personality disorder?”

These changes would transform how we conceptualize and diagnose personality disorders, but they are so complex and multifaceted that many of us would be “lost in translation.”

Let’s use as an example a patient with the DSM-IV diagnosis of schizoid personality disorder. In DSM-5, our hypothetical patient might be described as having a personality disorder with a trait domain of introversion, a trait level of three on the trait facets of social withdrawal and intimacy avoidance, and a trait level of two on the trait facets of restricted affectivity and anhedonia, with a severity rating of four. No doubt, this is a far more accurate way to describe our patients, but it seems highly impractical for busy clinicians.

Incremental Changes

Mood disorders. Currently, DSM-IV treats depressive symptoms triggered by bereavement as a special case of major depression. We are supposed to wait for two months of symptoms before making the diagnosis (unless there is suicidal ideation or marked impairment of functioning). DSM-5 would remove this so-called bereavement exclusion. Those opposed to this change argue that it would pathologize normal grief, and lead to an overdiagnosis of major depression. Those in favor of the change argue that: 1) it is safer to err on the side of more false negatives (ie, treating someone without major depression), than more false positives (ie, not treating someone with major depression); and 2) there is little evidence to suggest that the loss of a loved one is different from any other stressor that may trigger depression.

In support of this view, a recent study reported that bereavement-related depression (aka, normal grief) and regular depression are much more similar than different. They did not differ in regard to the number of prior episodes, duration of first episode, risk for future episodes, pattern of comorbidity, and other factors (Kendler KS, Am J Psychiatry 2008;165(11):1449–1455). We believe this is a reasonable tweak to the depression criteria, and is likely to lead to more benefit than harm.

Psychotic disorders. Patients sometimes present with mild hallucinations or delusions, along with intact reality testing. Unsurprisingly, research has shown that some such patients eventually develop a full-blown psychotic disorder, such as schizophrenia. DSM-5 proposes that such patients be given a new diagnosis: “risk syndrome for first psychosis.” However, as we discussed in the December 2009 issue of TCPR, the largest studies have found that such patients have only a 35% conversion rate to frank psychosis, meaning that there is a 65% false positive rate (Canon TD et al, Arch Gen Psychiatry 2008;65(1):28–37).

The psychosis work group has recently changed the name to “attenuated psychotic symptoms syndrome,” in order to emphasize that such patients are not necessarily at risk to become schizophrenic. All in all, the science here is preliminary, and it’s likely that this entity will end up in DSM-5’s appendix as a “disorder for further study.”

Substance abuse, dependence, or disorder? The DSM-IV distinction between substance abuse and substance dependence has always been a bit fuzzy. “Alcohol abuse” is for patients who, for example, frequently get drunk to the extent that it causes life problems—but who do not predictably have withdrawal, tolerance, or loss of control over their use of alcohol. Think of the weekend binger. “Alcohol dependence,” on the other hand, requires the presence of tolerance and withdrawal, plus one of several examples of loss of control of use—the colloquial alcoholic. In the clinic, some patients fit pretty clearly in one camp or the other, but most are somewhere in between.

The DSM-5 would collapse abuse and dependence into one category called “substance use disorder.” The proposal has curb appeal, because it seems to make our jobs easier by eliminating needless diagnostic hair-splitting. On closer view, however, there are some problems.

One is that the proposed diagnosis consists of a long and sometimes redundant laundry list of 11 different symptoms or behaviors. In addition, from a practical standpoint, treating an alcohol abuser is different from treating someone who is physiologically dependent. For example, an alcohol abuser will rarely need detox and a subsequent prolonged rehab; an alcohol dependent often will.

If it sounds like we’re hemming and hawing a bit on this one, we are. There are potential upsides and downsides to this change.

Childhood bipolar disorder vs temper dysregulation with dysphoria. With widespread concern that bipolar disorder is being excessively diagnosed in children (see, for example, Moreno C et al, Arch Gen Psychiatry 2007;64(9):1032–1039), re-evaluating how we diagnose moody and explosive kids has become a high profile issue for DSM-5. The possible creation of “temper dysregulation syndrome with dysphoria” (TDSD) is an attempt to provide a more precise diagnostic home for kids who now are often classified as having bipolar disorder.

Proposed diagnostic criteria for TDSD are: out-of-proportion reactions to normal stressors, including violence or rage, occurring at least three times a week; a baseline mood of unhappiness for 12 months or more; and the absence of episodes of mania or hypomania. In plainer language, this diagnosis is for chronically unhappy kids with periodic attacks of rage.

Because some have pointed out that this label could be given to normal kids with temper tantrums, the DSM-5 work group is toying with the idea of tweaking the name to “disruptive mood dysregulation disorder”—you’ll notice that the word temper is absent. This possible name change is a good illustration of how DSM-5 is a mélange of science and public relations, perhaps inevitable given the ambiguous nature of psychiatric science itself. See the May 2010 issue of The Carlat Child Psychiatry Report for more details on DSM-5 changes relevant to children.

Dementia to be replaced by major neurocognitive disorder. In DSM-IV, there are four different types of dementia available: Alzheimer’s, vascular, due to a general medical condition, and not otherwise specified. Each has the same basic criteria—patients must demonstrate both memory loss and one of the following four cognitive problems: aphasia, apraxia, agnosia, or executive functioning disturbances.

DSM-5 would replace the loaded term “dementia” with the more neutral sounding term “major neurocognitive disorder.” Aside from making the diagnosis sound less scary, memory loss would no longer be a requirement—a nod tothe fact that some forms of cognitive decline present primarily with non-memory problems, such as the executive functioning impairments in fronto-temporal dementia.

What would happen to Alzheimer’s? It would still exist as a subtype of major neurocognitive disorder—and it would be by far the most common subtype. More controversial is the creation of “minor neurocognitive disorder” which would describe patients with mild cognitive impairment (MCI), a kind of predementia. MCI can be hard to diagnose. The usual quandary is deciding whether someone has benign forgetfulness of aging vs the more foreboding MCI. (For a review of screenings for cognitive impairment, see Alagiakrishnan K, Postgrad Med 2010;122(4):105-111.)

The danger is that many patients with the minor neurocognitive disorder diagnosis would likely end up taking a cholinesterase inhibitor (such as Aricept), even though there is no data showing that this prevents or slows down the eventual development of dementia. Nonetheless, minor neurocognitive disorder appears to be a real disorder, and the lack of an effective treatment is not a particularly good reason to omit it from DSM-5.

Some sensible tweaks to eating disorders. It has become clear over the years that some patients suffer severe anorexia without necessarily having amenorrhea or a fear of gaining weight. DSM- 5 would drop both of these as requirements for the diagnosis. In a similar vein, the requirement that bulimics binge and purge twice a week has proven to be arbitrary; DSM-5 would loosen the criteria to once a week, in line with research showing that this lesser frequency leads to significant dysfunction. Finally, binge eating disorder would become a bona fide disorder, which is a promotion from its current home in the appendix of DSM-IV.

Somatoform disorders. Somatoform disorders would become “somatic symptom disorders,” but the more meaningful change is the renaming of conversion disorder to “functional neurological symptoms,” which is certainly more descriptively accurate and a far cry from its 19th century moniker, hysteria. In addition, four of the seven current somatoform disorders—somatization disorder, undifferentiated somatoform disorder, hypochondriasis, and pain disorder—would be combined into the single diagnosis of somatic symptom disorder.

This makes sense, as all these afflictions share certain features, including prominent somatic symptoms, cognitive distortion, and chronicity. They are often comorbid—many of us have worked with a chronic somatizing patient who has met the criteria for hypochondriasis, somatization disorder, and pain disorder at one point or another in his or her treatment.

Understanding and acknowledging that impaired cognition is a shared element in all these diagnoses can be very useful in regard to treatment planning—not only for the psychiatrist but the collaborating primary care doctors, pain doctors, and other medical specialists as well.

New Additions
Whether it’s a new addition to one’s family, one’s workplace, or one’s therapy group, new additions are always interesting and, sometimes, exciting. Potential additions to DSM-5 include some diagnoses that are currently listed in DSM-IV but not as full-fledged disorders: seasonal affective disorder, premenstrual dysphoric disorder, and binge eating disorder are in this category.

Others are entirely new and seem to capture meaningful categories of real patients, such as “non-suicidal self injury,” “Internet addiction,” and “hoarding disorder.” Then, there are a few dicier proposals, such as “parental alienation disorder” and “male-to-eunuch gender identity disorder.” We’ll have to wait to see which conditions make the cut.

Conclusion: Our Humble Predictions
We predict that most of the changes described in this article will, in fact, take place, with the exception of two of the foundational changes—the cross-cutting dimensional assessments and the personality disorder overhaul. These changes are too radical and clinically onerous, and the supporting evidence base is shaky. Some symptom scales will probably survive into DSM-5 as “useful diagnostic adjuncts,” or some similar term. Overall, the new bible will represent a reasonable refinement of our nomenclature, and will surely provide new material for jokes at psychiatry’s expense!