Saving Normal : An Insider's Revolt Against Out-of-control Psychiatric Diagnosis, Dsm-5, Big Pharma, and the Medicalization of Ordinary Life (9780062229274)
Page 4
Elusive reality does not discourage Umpire Two. We don’t have to fully perceive or understand the underlying nature of our world to negotiate it well. Our senses and reasoning powers evolved as they did because they work just fine in the everyday, nonphilosophical business of survival. Mental constructs of reality are imperfect, but indispensable, ways to organize the otherwise bewildering phenomena of the world.
Umpire Two “calls them as he sees them.” Mental disorders are not real diseases as Umpire One might wish; but neither are they the dangerous myths feared by Umpire Three. Instead he follows a down-to-earth brand of utilitarian pragmatism. His umpire’s eye is fixed on what works best—not distracted by biological reductionism or rationalist doubt. He accepts that we are constantly constructing perceptions and finding temporary meanings that are useful, but never completely accurate. Our classification of mental disorders is no more than a collection of fallible and limited constructs that seeks but never finds the truth—but this remains our best current way of communicating about, treating, and researching mental disorders.
Schizophrenia is a useful construct—not myth, not disease. It is a description of a particular set of psychiatric problems, not an explanation of their cause. Someday we will have a much more accurate understanding and more precise ways of describing these same problems. But for now, schizophrenia is very valuable in our day-to-day work. And so are the other DSM disorders. It is good to know and use the DSM definitions, but not to reify or worship them.31,32
Defining Disorder Around the World
What about the potentially distorting lens of culture? Do mental disorders present the same way everywhere or does each culture need its own diagnostic system? The answer seems to be that one size usually fits almost all. Although “normal” behavior is variable across cultures, the specific mental disorders are pretty uniform. Dementia, psychosis, mania, depression, panic attacks, anxiety, obsessive-compulsive disorder, and the personality disorders have been described in all past ages and in all places and are found today in epidemiological studies wherever in the world these are conducted. When rates of disorder differ (e.g., blacks being diagnosed more often with schizophrenia in the United States), it is because of bias or cultural blind spots in the raters, not real differences in the patients they are rating.33
There are two diagnostic systems in current, overlapping use around the world—DSM-5 (soon to be translated into about twenty-two languages) and ICD-10, developed by the World Health Organization (translated into forty-two languages).34 DSM-5 and ICD-10 are really very similar; which is not surprising, since they are closely related sibs. Both are no more than minor modifications of the same parent (DSM-III) and were prepared at the same time and with some efforts to achieve harmony. As with sibs, there is a rivalry between systems. The DSMs have so far been more influential, but it will be several years before we can judge whether DSM-5 or ICD-11 (planned for publication around 2016) will win the next round of the competition. For now, the relative merits of DSM and ICD are pretty obvious—DSM is used much more often in research; they are about equal for clinical work in the developed countries; and ICD works better when a simpler system is needed in the developing world.35
The more fascinating question is why both diagnostic systems have gained such universal applicability across all the races and cultures of the world. Clearly we humans are more alike than we are different, resembling one another closely in the things that count toward defining normal and mental disorder.
There are no genetically caused racial differences in mental disorder. How come there is such uniformity? Compared to other species, humans have a remarkably homogeneous gene pool. Genetic and geologic evidence converge on the theory that there was a catastrophic die-off of humans about 70,000 years ago caused by a giant volcanic super eruption in what is now Indonesia.36 Our species was almost wiped out by the protracted climate change, and most of us are the closely related descendants of the few thousand breeding pairs who survived. Racial differences, for all the trouble they cause, are literally skin deep, of recent vintage, and result in relatively few differences in how medical and mental problems express themselves.
Culture plays a much bigger role but influences only the surface presentations. Brief psychotic disorders and physical symptom presentations are much more common in poorer parts of the world and anorexia nervosa and attention deficit in the richer. In diagnosing and treating, it is crucial to be sensitive to cultural differences, but they are not so great as to require different diagnostic systems for different parts of the world. Across the board, humans are alike enough genetically and culturally that one diagnostic system (either DSM or ICD) is elastic enough to fit all the possibilities.
Defining the Individual Mental Disorders
The bad news that we can’t develop a useful definition for the general concept “mental disorder” is balanced by the very good news that we can quite easily define each one of the specific mental disorders. The method, introduced by DSM-III in 1980, is simple and effective. The description of each DSM disorder is accompanied by a criteria set that lists in fairly precise terms which symptoms define it, how many must be present, and their required duration. For example, a major depressive episode is defined as five or more of the following symptoms, presenting together for more than two weeks and causing clinically significant distress or impairment: depressed mood; loss of interest; reduced appetite; changed sleep; fatigue; agitation; guilt; trouble thinking; and suicidal feelings. Clinicians everywhere have been using this as a consensus definition for more than thirty years. Clinical depression is not diagnosed if there are only four instead of five of these very same symptoms, or if they are present for only one week, not two, or if the impairment they cause is not all that big a deal. There are about two hundred criteria sets in the DSM—one for each disorder. These establish the boundaries that separate the mental disorders from one another and from normality. Each criteria set has the symptoms that define that particular disorder (panic, generalized anxiety, obsessive-compulsive, attention deficit, autism, etc.) and the required threshold. When clinicians follow the criteria, they achieve reasonable agreement. Without them, there is poor agreement. Each clinician becomes a law unto himself, and the result is a confusing Babel of clashing, idiosyncratic voices.
But there is a catch. The boundaries demarcating the different disorders are ever so much fuzzier in real life than they appear to be on paper. There is really nothing magical or preordained about any of the DSM thresholds—shades of gray exist between their seemingly black and white cutoff points. Requiring five symptoms and two weeks for major depressive disorder derives from a fairly arbitrary choice, not a scientific necessity. Just as easily, the set points could have been set higher—say at six symptoms and four weeks. With a more demanding threshold, we would lose in “sensitivity” (thereby missing some sick people who are in need of diagnosis) but would gain in “specificity” (mislabeling fewer normal people). Sensitivity and specificity are reciprocally intertwined—you can’t help one without hurting the other. There is an inevitable trade-off between them that requires a proper balancing of the risks and benefits of overdiagnosis versus underdiagnosis. The final decision where to set the bar is always a judgment call; the research never renders a clear and compelling answer forcing the choice of one particular threshold in preference to other possibilities.
Once a criteria set is established, there should be good reasons for changing it; otherwise the system will be not only arbitrary, but also inconsistent and confusing. But this leads to a problem. Many of our current categories and thresholds were created thirty-five years ago when achieving sensitivity was the more important goal—too many people who needed a diagnosis were being missed. Circumstances have now changed dramatically, and poor specificity is now the biggest issue. Before DSM-III, there were too few diagnoses—now, because of diagnosis inflation, there are far too many. Raising severity and duration thresholds would help “save normal” and cure excessive diagnos
is—but would create instability and reduce sensitivity. You can’t have it both ways.
The method of defining disorders by criteria sets has another inherent set of difficult-to-balance trade-offs—between “reliability” and “validity.” Reliability means agreement and consistency—will different clinicians seeing the same patient arrive at the same diagnosis. Validity means truth—will the diagnosis tell you what you want to know. Ideally, of course, the definition of a disorder would do both—be both reliable and valid. But to meet the goal of reliability, the defining symptoms must be extremely simple, obvious, and generalize easily across all the people with that particular disorder. If the criteria set includes items that are inferential or complicated, different clinicians will disagree on whether or not they are present. Worshiping at the temple of reliability, the DSM criteria sets are as simple as they can be—a catalog only of what is most surface and common in mental disorders. This was a necessary choice, but it necessarily compromises validity—constraining ourselves to the simple blinds us to subtlety, nuance, and individual variability. A great deal is lost in the translation between the rich diversity of different individual experiences of depression and the bland five-of-nine criteria set chosen to define it. In describing the characteristics shared by those who meet the criteria for a given mental disorder, the DSM definitions must obscure the ways they are individual and different. DSM definitions do not include personal and contextual factors, such as whether the depressive symptoms are an understandable response to a loss, a terrible life situation, psychological conflict, or personality factors.
DSM has to stay simple, but psychiatry doesn’t. DSM diagnosis should be seen as just one small part of an overall evaluation that would also comprehensively account for the more complicated and individual aspects of each patient. Unfortunately, the DSM approach has been far too influential—dominating the field in a way we never intended. Nuanced psychiatry has become checklist psychiatry, homogenizing individual differences and custom-tailored treatments. Psychiatry, once too idiosyncratic and chaotic, has become too standardized and simpleminded. Training programs focus excessive attention on teaching diagnosis and not enough on understanding everything else about the patient.37 People forget the wisdom of Hippocrates: “It is more important to know what sort of person has a disease than to know what sort of disease a person has.” Of course, best practice is to pay close attention to both. DSM diagnosis has a necessary place in every evaluation, but it never tells the whole story.
Then there is the problem of knowing which criteria to choose and pretesting for their safety. Before we go prime time with a criteria set, the safe play is to audition it in a field trial. A test drive reduces uncertainty about how it will eventually perform, reducing the risk of unpleasant surprises and the dangers of unwanted fads. The idea is to have clinicians try out the new definition under conditions that approximate real-life circumstances. If the proposal performs well, it becomes official; if poorly, it is revised or scrapped. But again there is a catch. Really several different catches.
First, even the best field trial is performed in the present and can’t fully anticipate the future. The carefully done DSM-IV attention deficit predicted that our proposed changes would cause only a 15 percent increase in rates. This was probably a fairly accurate estimate given the reality when the data were gathered in the early 1990s. We couldn’t foresee the abrupt switch in this reality that occurred in 1997, when drug companies brought new and expensive medicine for ADD to market and were simultaneously set free to advertise them directly to parents and teachers. Soon the selling of ADHD as a diagnosis was ubiquitous in magazines, on your TV screen, and in pediatricians’ offices—an unexpected epidemic was born, and the rates of ADHD tripled.
Next, there is the problem of generalizability. The best setting for doing field trials would be the offices of the clinical psychiatrists and of the primary care physicians who actually write most of the prescriptions for psychiatric drugs. But because it is easier, field-testing is instead done in samples of convenience drawn from university research clinics that are very unlike the sites of eventual greatest misuse. The results generated in these cloistered settings will always be much better than what will be obtained in the hustle and bustle of the real world. Then there is the distorting effect of observing the thing observed. It is inherently impossible to learn everything about an electron because the act of observing an electron changes its momentum. Similarly, the act of observing everyday diagnostic practice distorts it so that it is no longer everyday. The selection and training of the clinicians, and their focused attention, make them better diagnosticians within a study than they will be outside of one.
Patient selection in a field trial also artificially raises reliability. In real life, making a diagnosis is like finding a needle in a haystack containing hundreds of possible choices. The field test presents a much easier challenge to the clinician. He knows he is selecting from among only a handful of different choices. Bottom line: Field trials are absolutely necessary but extremely fallible. New suggestions will perform much better in the trial than in real life. Possible future misuses may be entirely undetectable and unpredictable. At their best, field trials will help you avoid some, but certainly not all, the possible future trouble spots.
One final interesting question: Should we use names or numbers in defining mental disorders? The DSM system uses only names. Psychologists have developed thousands of rating scales that instead use numbers. Which is better? Like most things, there is no one right answer—it depends on your purpose. Numbers are much more accurate than names—that’s why we use them to measure height, weight, IQ, or wavelength in physics. When describing someone’s place on a graded continuum, it is ever so much more precise to give a number than a name. “He is six feet tall” saves information lost if the description is reduced simply to “He is tall.” Computers love numbers. And so do the researchers who use them.
But most people think names, not numbers. Evolutionary pressures shaped our minds to give simple names, not to make fine-grained mathematical distinctions. We are adapted to a world that required quickly choosing between a yes or no—trying to quantify predators too closely might get you eaten. It’s no surprise that statistics as a branch of mathematics arose so late in the game—just a few hundred years ago.
In everyday life, we still usually prefer names to numbers—even though names are admittedly inexact and we do have eager computers ready to crunch whatever numbers we throw at them. We label a color “red” (rather than calling out its exact wavelength) because this is quicker, easier, clearer, and usually serves the purpose. A vivid name, not a confusing bunch of numbers, remains much more convenient for most tasks and provides a clearer and more readily understood image. Clinicians are busy people who have thought in names every step of their natural lives and in all their training. They will not switch to numbers easily, and patients wouldn’t understand them if they did. Computer-assisted dimensional diagnosis is certainly the wave of the future, but it is premature and impossible to implement in the present. For now, we will stick to naming the mental disorders, not numbering them.
At the extremes, the distinction between the completely well and the clearly sick is perfectly plain and not the least bit amenable to fudging. In contrast, the much fuzzier distinction between the mildly ill and the probably well is easily and frequently manipulated. Most normal people have at least occasional mild and transitory symptoms (e.g., sadness, anxiety, sleeplessness, sexual dysfunction, substance use) that can easily be misconstrued as mental disorder. The business model of the pharmaceutical industry depends on extending the realm of illness—using creative marketing to expand the pool of customers by convincing the probably well that they are at least mildly sick. Disease mongering is the fine art of selling psychiatric ills as the most efficient way of peddling very profitable psychiatric pills. Manipulating the market is particularly easy in the United States because we are the only country in the entire world that
allows drug companies the freedom to advertise directly to consumers.
Disease mongering cannot occur in a vacuum—it requires that the drug companies engage the active collaboration of the doctors who write the prescriptions, the patients who ask for them, the researchers who invent the new mental disorders, the consumer groups that advocate for more treatment, and the media and Internet that spread the word. A persistent, pervasive, and well-financed “disease awareness” campaign can create disease where none existed before. And psychiatry is especially vulnerable to manipulation of the normal/disease boundary because it lacks biological tests and relies heavily on subjective judgments that can be easily influenced by clever marketing.
The primary loyalty of any corporation is to its shareholders and to its own-survival, not to the public weal. General Motors sells cars, Anheuser-Busch sells beer, Apple sells computers, the drug cartels sell cocaine, and drug companies sell pills all for the same reason—to generate as big a profit as possible. The profitability of any corporation depends on increasing the size of its market and its margins on each sale. Drug companies are exemplary profit-making machines because of their highly developed skill in pushing product and their ability to maintain monopoly pricing. Pumping up diagnostic inflation is absolutely key to drug company success. Full saturation requires having the widest demographic reach, from the youngest child to the oldest adult. Casting the broadest net is always great for shareholders, but is very often bad news for the mislabeled normals who are subjected to the unnecessary medication and stigma that comes with fake diagnosis.
Is Normality Resilient or Fragile?
The paradoxical answer is both. Resiliency is built into every aspect of our biological, psychological, and social being. We are hardwired to work remarkably well, but are far too complicated always to work perfectly and we can lose purchase on normality by mislabeling as mental disorder each and every one of our glitches.