Saving Normal : An Insider's Revolt Against Out-of-control Psychiatric Diagnosis, Dsm-5, Big Pharma, and the Medicalization of Ordinary Life (9780062229274)

by Frances, Allen

  DSM-IIIR—Too Much, Too Soon

  DSM-IIIR was the revision of DSM-III published just seven years later in 1987. It was meant to be a small makeover, no more than a minor midterm correction of the errors and omissions noted after the publication of DSM-III. Bob Spitzer was in charge again, and this time his enormous energy and enthusiasm got out of hand. The project went on twice as long as originally scheduled and made many major changes—all in the direction of making it easier to get a psychiatric diagnosis. I was part of the inner circle advising Bob on DSM-IIIR,23 but his eagerness for change and enthusiasm for expansion were impossible to contain.

  DSM-IIIR was a mistake and a distraction. One goal of DSM-III was to create a kind of Linnaean classification (or “periodic table”) of objectively defined mental disorders that would then stimulate clinical and basic science research. The system was meant to be iterative and self-correcting—taking what were essentially the made-up criteria sets in DSM-III as the starting point, but then confirming or altering them based on the research they promoted. This circular process could never proceed if the diagnostic system offered a moving research target, constantly shifting capriciously, based on arbitrary opinion.

  The lesson for me was that diagnosis needed to rest in order to let research catch up. It made no sense to keep rearranging the furniture of descriptive psychiatry, creating new diagnoses or altering the thresholds of existing ones, based only on the whims of the experts who happened to be in the room. The method that was necessary for DSM-III had been inappropriate for DSM-IIIR. Changes in diagnoses should be few and far between until we gained a much deeper understanding of what causes the mental disorders and how best to define and treat them. DSM-III had of necessity come mostly out of the heads of the experts as translated into “manualese” by Bob Spitzer. The future required more caution, tighter scientific standards, less personal opinion, and no change just for change’s sake.

  The full reality of diagnostic inflation was years away, but the threat was in the air. Prozac and DSM-IIIR were both introduced in 1987. Prozac’s sales took off at least in part because the DSM definition of major depressive disorder was so loose. The message was clear—psychotropic drugs offered vast market potential, and sales could be greatly influenced by DSM decisions. It was important that the diagnostic system not become an unwitting tool of drug company marketing.

  The DSM-IV Story

  When asked why I was picked to head the DSM-IV Task Force, I usually joke it was a punishment inflicted for sins in a prior life. The serious point is that I don’t really know why I was picked and the process was remarkably casual. The medical director of the American Psychiatric Association asked if I would accept the appointment. I thought about it for a day and said yes. Simple as that. No muss, no fuss, no vetting, no interviews, no competition, no testing of my qualifications, goals, or methods.

  By illustrative contrast, the selection marathon that preceded my appointment as chair of psychiatry at Duke University took four months and included: three search committee meetings; numerous conversations with the dean; meetings with the head of the hospital, the president of the university, the financial officer, the chairs of all the other departments, and with about fifty members of the department. Even before all this, a representative of the search committee had called at least thirty people to check me out. Before hiring me, Duke knew more about me than I knew about myself. The careful vetting process guaranteed that we would have a jointly worked-out agenda and plan of action.

  In contrast, the APA took me mostly on faith, with little idea how I would proceed. They probably chose me because I had worked on the two previous DSMs and was chair of the APA Committee on Diagnosis; to that degree, I was a known quantity. And there weren’t that many people who had the background to do the job or would be willing to. But a formal search would have ensured I was the best choice and would have made my plans and the charge much clearer. The casual selection process left me with far too much discretion.

  There was also almost no direction from APA on picking the colleagues who would work with me on DSM-IV and how to proceed in preparing it. The natural tendency is to recruit your friends. It will be more fun working with people you like and much easier to lean on them to put in the many unpaid hours needed to get the job done. I decided instead to include all the opposing views on what I anticipated would be the most controversial questions. But no one made me do this.

  We set goals for DSM-IV that were extremely modest—to introduce rigor, objectivity, and transparency in how decisions were made, and not to innovate or add personal touches to the system.24, 25 I knew that if we set a high scientific burden of proof, few changes would be made because there would not be convincing evidence to support them. Data rarely jump off the page, grab you by the throat, and insist you make a change. The strange thing is that APA had so little idea what my goals would be and so little input in shaping the direction of DSM-IV. This was a mistake. No single person should be left free to determine the future of a diagnostic system that has such wide influence. The numerous problems that afflicted DSM-5 illustrate the risks inherent to unchecked, potentially idiosyncratic leadership.

  Which brings up the question of leadership style. I was following in the footsteps of a charismatic leader who took a passionate interest in each and every question and loved the controversies that surrounded them. Bob’s approach was appropriate to the task of creating out of whole cloth a radically new diagnostic system, but creativity and innovation were not the skills best suited to the minor course corrections required for DSM-IIIR or DSM-IV. The seven years each between DSM-III, DSM-IIIR, and DSM-IV (1980, 1987, and 1994) hadn’t generated compelling research findings to warrant a significant revision of the diagnostic system. I therefore saw my role as one of refiner/conservator rather than innovator. Requiring a high level of scientific proof before any change took most of the personality and leadership issues out of the equation. I set up an impersonal system that would work automatically to avoid and solve controversies. Decision making was to follow rules rather than clashing personal convictions. The role of work group member was defined as “consensus scholar” quietly reviewing tables of available data, not advocating for a cause or aspiring to be an innovative diagnostic pathfinder.

  I don’t like controversies and find them unproductive—almost all heat, very rarely any light. So we took the realistic position that any prolonged disagreements on the interpretation of data meant the scientific literature was too sparse or ambiguous to support change. A change would be made only when it was compellingly necessary and there was an overwhelming consensus on the science. This didn’t happen very often. Anything close to a hung jury was decided in favor of the status quo. There were no votes, and I don’t remember any serious disagreements.

  Having hundreds of strong-minded experts working on a project could be a recipe for anarchy. The antidote was to set up standard operating procedures and to make sure that everyone followed them closely. All of the timelines for doing DSM-IV were established well before we began work on it. A series of methods conferences ensured that every procedure would be performed in a consistent way. The threshold for making changes was clear, uniform, and rigorous. We established a three-stage obstacle course intended to weed out new suggestions. Stage 1 was a searching literature review that would painstakingly gather the available scientific data, with special consideration given to the possible risks and unintended consequences of any change. Stage 2 consisted of data reanalyses funded by the MacArthur Foundation. This allowed us to access already collected but not yet analyzed data sets that lived in the computers of investigators around the world. We could ask questions pertaining to DSM-IV decisions that had not yet been answered in the published literature. Stage 3 consisted of NIMH-funded, peer-reviewed field trials covering twelve disorders where changes were contemplated. The goal was to test-drive the alternative criteria sets to see how they compared under conditions that approximated (but did not duplicate) real life.
/>
  The role of the central leadership was quality control. Our job was not to decide issues, but rather to ensure that the rules were followed. It was also our job to guarantee that all deadlines would be met. The result of all this methodological rigor was exactly what we expected. The pet proposals of the experts were consistently shot down because the science wasn’t there to support them. DSM-IV was faithful to DSM-IIIR.

  All our work was a completely open book with extremely porous boundaries. Anyone showing interest in DSM-IV could become an advisor to the process. Work group members were encouraged to tell all and get full input from colleagues. At midpoint an Options Book was published that gave everyone a chance to vet and shoot down all proposed changes. We also published an extensive four-volume DSM-IV Sourcebook as an archival resource, including all of the literature reviews, data reanalyses, field trials, and rationales for decisions.26 As we shall see, DSM-5 went far wrong in large part because it was secretive and closed to outside correction.

  We saw DSM-IV as a guidebook, not a bible—a collection of temporarily useful diagnostic constructs, not a catalog of “real” diseases. We tried to make this abundantly clear in the introduction to DSM-IV and at greater length in the DSM-IV Guidebook. Unfortunately, I am not sure anyone ever reads the introduction, and I know that few people have read the Guidebook.27, 28 People shouldn’t worship the DSM categories, but it does make you a better clinician to know them.

  DSM-IV did not save normal, or even protect it very well. Three years after its publication, drug company lobbyists won a huge victory over sensible regulation. The United States became the only country in the world that allows direct-to-consumer advertising of pharmaceuticals. Pretty soon the airwaves and print were filled with glowingly misleading representations that everyday problems were in fact unrecognized psychiatric disorder. DSM-IV turned out to be a very weak dike unable to block the flood of false demand instigated by the aggressive and devilishly clever drug company push. Although we had consistently rejected suggestions that would have benefited drug companies, we failed to predict that even our conservative manual could provide such easy fodder for advertising gold. Within a few years, it was clear the drug companies had won and we had lost.

  There were steps we could (and probably should) have taken to help curb diagnostic inflation. Most important, we could have tightened the DSM-IV diagnostic thresholds (required more symptoms, longer durations, and greater impairment) to make it more difficult for companies to sell diagnoses. But we had hoisted ourselves on our own perhaps too evenhanded, conservative petard—our strict evidentiary rules required the presence of extensive and compelling scientific data before we could make any change in either direction. Following our rules made it just as difficult to deflate the diagnostic system as to inflate it. The rules were a necessary constraint to contain arbitrary decision making and the experts’ natural tendencies to expand their own domains. Being evidence-based rather than opinion-driven helped us avoid contributing to new inflation, but it also prevented us from reducing the inflation that was already in place. Knowing what I know now, this was probably a mistake. It would have been better to create a double standard, with less evidence required to effect deflation than to promote inflation. Admittedly, deflation would have been difficult and arbitrary, but it now seems preferable to the excessive diagnosis and treatment that have since been protected by our adherence to evidence.

  And there was much more we could have done to sound warnings. We should have been far more active in educating the field and prospective patients about the risks of overdiagnosis. There should have been prominent cautions in DSM-IV warning about overdiagnosis and providing tips on how to avoid it. We should have organized professional and public conferences and educational campaigns to counteract drug company propaganda. None of this occurred to anyone at the time. No one dreamed that drug company advertising would explode three years after the publication of DSM-IV or that there would be the huge epidemics of ADHD, autism, and bipolar disorder—and therefore no one felt any urgency to prevent them. We thought we had done a reasonably good job of writing the manual and didn’t see it as our responsibility to make sure it was used responsibly. We missed the boat. Even if we had been smarter and tougher we probably couldn’t have stemmed the tide of overdiagnosis. Big Pharma was simply too big, too rich, and too politically powerful. But I do very much regret that we didn’t try harder.

  So what is my final scorecard on DSM-IV? Decidedly mixed. On the positive side: We made very few changes; developed and implemented a meticulous method of scientific review; improved the precision of the manual’s writing and coding; and made only one obvious mistake. On the negative side: Our changes contributed directly to the false epidemics of autistic, attention deficit, and adult bipolar disorder; we did nothing to prevent the overdiagnosis of several other disorders that have been puffed up by the drug companies; and our one outright mistake was a disaster, a sloppily worded paraphilia section that has allowed the widespread unconstitutional abuse of involuntary psychiatric hospitalization. We could have done a lot worse, but we should have done better in predicting unintended consequences and preventing continued diagnostic inflation. Trying hard to be rigorous and “do no harm” protected us from most errors of commission but forced us into serious errors of omission. We didn’t do much harm, but we weren’t of much help. When it was completed, I was pretty happy with DSM-IV. Now I wish we had done more to save normal and reduce the ease with which the drug companies were able to sell sickness.

  A legitimate question has been raised about the motivations of the people working on DSM-IV—did we go soft on diagnostic inflation because of a financial conflict of interest? The concern arises from a recent study showing that 56 percent of our experts had some financial connection to drug companies. The assertion has been made that the companies were pulling strings behind the scenes, directly or subtly, to bend decisions toward more diagnosis and more treatment. The question is certainly legitimate because we had no formal conflict of interest policy or vetting system. This omission was a silly mistake on our part—the necessity simply never occurred to any of us when we began work in the fairly innocent pre-Prozac days of 1987. There is no excuse for our failure to put in place the formal tools of protection against conflict of interest. I apologize for this but don’t agree that financial conflict of interest compromised any of our decisions. The proof is in the pudding. The results prove that our rigorous method of review successfully protected against any potential conflict of interest just as surely as the most thorough vetting ever could. Dozens of proposals that would have favored drug companies were shot down. Only two of our decisions wound up helping drug companies—slightly loosening the requirements for attention deficit disorder and introducing bipolar II. Both of these filled an important clinical niche, both were supported by substantial evidence, and neither had much obvious commercial value when the decisions were made. Unfortunately, both decisions were later exploited by the drug companies when they gained the right to advertise to consumers and had developed new and expensive products to sell—but this happened in ways we could not have predicted or prevented. The drug industry played no role whatever in how DSM-IV was written, but played the deciding role in how it was misused. I agree that Caesar’s wife should be above suspicion—but also feel certain that in this case the suspicion is misplaced.29,30

  The DSMs have a mixed record. They have served an extremely valuable function in improving the reliability of psychiatric diagnosis and in encouraging a revolution in psychiatric research. But they have also had the very harmful unintended consequence of triggering and helping to maintain a runaway diagnostic inflation that threatens normal and results in massive overtreatment with psychiatric medication.

  CHAPTER 3

  Diagnostic Inflation

  Alice: “But I don’t want to go among mad people.”

  Cheshire Cat: “Oh, you can’t help it, we’re all mad here.”

  LEWIS CARROLL, Alice’s Adve
ntures in Wonderland

  Medical research has made such enormous advances that there are hardly any healthy people left.

  ALDOUS HUXLEY

  DIAGNOSTIC INFLATION HAS many, many causes and will require many cures. Some of the problems are inherent to psychiatry and need curing from within the profession. But a number of powerful outside forces flexed their muscles, grabbed hold of DSM-IV, and used clever methods to encourage its misuse. They succeeded in changing diagnostic habits in ways we never imagined possible and lacked the tools to control. The past thirty years have witnessed a frightening vicious cycle. Diagnostic inflation has led to an explosive growth in the use of psychotropic drugs; this then produced huge profits that have given the pharmaceutical industry the means and the motive to blow up the diagnostic bubble into an ever-expanding balloon. The coinage of psychiatric diagnosis has been cheapened, making “normal” a scarce commodity. Just as happens with monetary inflation, the bad money drives the good out of circulation and resources allocation is distorted. The wasted effort devoted to those who don’t have real disorders deprives those who do from receiving badly needed psychiatric diagnosis and care.

  Causes of Diagnostic Inflation

  Following the Pied Piper of Diagnostic Inflation in the Rest of Medicine

  Psychiatry didn’t invent diagnostic inflation—it has just blindly followed the crowd in turning seeming wellness into dreaded sickness.The rest of medicine got there first by promoting the idea that all of us need to have regularly repeated batteries of screening tests to find out what might be going wrong with our bodies, even before we are experiencing any symptoms of illness. What an appealing goal—screen early to prevent disease before it can gain traction and cause harm. Not only would this reduce human suffering, but it would also save money. Detect cancer before it can spread; treat the slightest increase in blood pressure before it causes heart disease; control blood sugar before it becomes full-fledged diabetes; stop bone loss before osteoporosis causes fractures. At its most telling extreme, people subjected themselves to yearly full-body CAT scans to clearly visualize every internal nook and cranny—until it became clear that the test was worse than the disease. The risks of getting cancer from the X-ray exposure far outweighed any potential benefit in picking up early cancers. Preventive intervention would be wonderful, if only we had an accurate way to identify who needs it. But most early screening picks up lots of people who are better left alone.1