Anatomy of a Non-Epidemic - a Review by Dr. Torrey


How Robert Whitaker Got It Wrong

Review by E. Fuller Torrey, MD
Treatment Advocacy Center founder   

anatomy-of-an-epidemicIn 2010 Robert Whitaker published Anatomy of an Epidemic: Magic Bullets, Psychiatric Drugs, and the Astonishing Rise of Mental Illness in America (New York: Crown Publishers). The book has circulated widely, in large measure due to Marcia Angell’s surprisingly uncritical review of it in the New York Review of Books (Angell, 2011). In its 396 pages Whitaker got many things right, including criticism of the broad DSM diagnostic criteria for mental illnesses; the reckless prescribing of psychiatric drugs for children; and the prostitution of many psychiatric leaders for the pharmaceutical industry. Indeed, regarding the last, Whitaker may have understated the problem, based on recently released court documents detailing how the pharmaceutical industry secretly controlled the Texas Medication Algorithm Project.

When it came to schizophrenia and antipsychotic drugs, however, Whitaker got it mostly wrong. He made so many errors it is difficult to know where to begin, so I will begin where he did. In his Preface Whitaker says that his research on the book began when he encountered “two research findings that just didn’t make sense”: a 1994 schizophrenia outcome study (Hegarty et al. 1994) and the World Health Organization (WHO) studies “which had twice found that schizophrenia outcomes were much better in poor countries.”

diagnostic criteria and schizophrenia outcomes                                                                                                           

Whitaker summarized the 1994 outcome study by saying that “outcomes for schizophrenia patients in the United States had worsened during the past two decades and were now no better than they had been a century earlier,” and he later added that the worsened outcomes were due to the use of antipsychotic drugs. What the paper actually says, however, is quite different. It is an analysis of schizophrenia outcomes throughout the twentieth century, linking outcomes to diagnostic criteria. When a broad definition of schizophrenia was in vogue, outcomes were better but when a narrow definition was in vogue, outcomes were worse, as would be expected. The data showed a clear improvement in outcomes during the 1960s and 1970s following the introduction of antipsychotic drugs, then a worsening in the 1980s and 1990s. The authors attributed this to the introduction of a narrow definition of schizophrenia, requiring six months of symptoms to qualify for the diagnosis, in DSM-III introduced in 1980. They concluded “that diagnostic criteria have had a consistent and predictable impact on outcome before and during the era of modern biomedical therapeutics.” The authors did not claim that the outcome of schizophrenia had actually worsened.

The two multi-country WHO schizophrenia outcome studies, which claimed that individuals with schizophrenia in developing countries had much better outcomes than those in developed countries, were relied upon heavily by Whitaker, both in this book and in his previous writings (Whitaker, 2004). Whitaker claimed that the patients in developing countries did better because they were less likely to be treated with antipsychotic medication. The authors of the WHO studies reported “a marked predominance of favourable outcomes in the centres in developing countries” and a “considerably more favourable [course] in developing countries” (Sartorius et al., 1986). These claims were heavily criticized at the time they were first published (e.g. Stevens, 1986) because it was alleged that the WHO centers in the developing countries had included many individuals who did not have true schizophrenia. Rather they had included many patients with acute reactive psychosis which has a much better outcome than true schizophrenia.

There is much evidence to support this criticism. For example, the second WHO study reported that “49% of the patients in the developing countries had an acute onset, and this was the case in only 26% of the patients in the developed countries.” Consistent with this was the fact that 40% of the cases in the developing countries were initially diagnosed with an “acute schizophrenic episode” versus 23% of cases in developed countries. In addition, the inclusion criteria for diagnosing schizophrenia was very broad for these studies; for example, a patient could meet criteria for schizophrenia with a combination of “severe excitement” and “overwhelming fear” without having any delusions, hallucinations or thought disorder (Sartorius et al. 1986). Even more worrisome was the fact that a significant number of subjects in the developing countries were referred to the study by religious or traditional healers. All of us who have worked psychiatrically in developing countries have seen many acutely excited individuals with acute reactive psychosis, most of whom get well. This is the most probable explanation for the WHO findings.

The WHO claim of better outcomes for schizophrenia in developing countries has continued to be criticized over the years and has now been largely discredited. Cohen et al. (2008) examined 23 schizophrenia outcome studies in 11 low-and-middle income countries and concluded that there is “a need to reexamine the conclusions of the WHO studies.” Messias et al. (2007) suggested that an increased mortality among the sickest patients in developing countries may have created an illusion that outcomes among other patients were better. And most recently Teferra et al. (2011) reported five years outcome data on 321 schizophrenia patients in rural Ethiopia with results sharply at variance with the WHO results. Faced with such criticisms, the authors of the WHO studies have recently modified their claims, stating that “we do not argue that the prognosis of schizophrenia in developing countries is groupwise uniformly milder” and acknowledging that “the proportions of continuous unremitting illness…did not differ significantly across the two types [developed and developing] of settings: (Jablensky & Sartorius 2008).

Schizophrenia Outcomes and Medication

The WHO studies raise the more general issue of schizophrenia outcome studies, an issue that is critical for Whitaker’s thesis. It has been known for a century that approximately one-quarter of individuals who develop a schizophrenia-like psychosis will recover without treatment and not get sick again. For example, in 1938 Stalker reviewed 16 schizophrenia outcome studies and concluded that 22 percent of the patients had achieved complete or social remission, the latter meaning that the person had resumed employment. More recently Stephens (1978) analyzed 25 schizophrenia outcome studies in which follow-up had been for 10 years or longer. Although the percentages of good and poor outcomes varied considerably in the different studies depending on the diagnostic criteria used, consistent with the findings of Hegarty et al. (1994) noted above, on average one-quarter of the patients recovered completely, one-quarter had a continuous illness, and the other half had intermediate outcomes between these two extremes. Stephens thus “emphasized that there are relatively benign and malignant forms of illnesses generally diagnosed as schizophrenia.”

Since these schizophrenia outcome studies had been known for many years, it was a great surprise to many of us when the 2003 President’s New Freedom Commission on Mental Health “discovered” that many individuals with schizophrenia do recover. It was a discovery of recovery, a classic example of finding something that had not really been lost. It suggested that the potpourri of political appointees who made up the Commission did not include anybody with any knowledge of the past studies.

This is directly relevant for Whitaker’s book because he focuses much attention on studies in which some patients with schizophrenia recover without medication. One of Whitaker’s favorite studies to support his thesis is a 20-year follow-up study of 70 patients (61 with schizophrenia and 9 with schizoaffective disorder) followed by Martin Harrow, a psychologist at the University of Illinois, and his colleagues. Harrow et al. (2012) reported that over the 20-year period 15 (21 percent) of the patients recovered and never thereafter needed antipsychotic medication. In addition, at each of the six follow-up assessments over the 20 years, between 7 and 12 additional patients were not taking antipsychotic medication, although some were taking antidepressants or other psychiatric medication (Harrow & Jobe, 2007; Harrow et al. 2012). Harrow et al. note that those patients not taking antipsychotics had either been taken off the medications by their treating psychiatrist or were “self-selected” and had taken themselves off medication. In terms of function, the individuals who were not taking antipsychotics were functioning better than those taking antipsychotics, as would be expected since the former group included the 15 patients who initially recovered.

On non-medication and recovery

Since the Harrow et al. study reported a schizophrenia recovery rate entirely consistent with previous follow-up studies and the fact that patients who are able to stop antipsychotic medication at least temporarily were functioning better, it is a completely unremarkable study. Whitaker, however, calls it “the best study of the long-term outcomes of schizophrenia in the United States” and discusses it four separate times in his book. Using tortured logic, he asserts that the Harrow et al. study proves that long-term antipsychotic use causes brain damage and is responsible for many of the symptoms of schizophrenia, when in fact the study does nothing of the kind. Harrow et al. even explicitly state that their study provides no evidence on “whether very long-term use of antipsychotic medication produces undesirable effects for some SZ [individuals with schizophrenia].”

Another study held up by Whitaker to demonstrate that individuals with schizophrenia who are not taking medication do better is Courtney Harding et al.’s Vermont outcome study. As Whitaker described it, among 168 patients with schizophrenia followed up 20 years after hospital discharge, “34 percent were recovered” which he claims is “a startling good long-term outcome.” Whitaker attributes this outcome to the fact that “they all had long since stopped taking their medications.” The original study stated that half of the patients were taking antipsychotics continuously or intermittently (Harding et al, 1987). Since the average age of these patients was 61, it is not surprising that many others had been able to discontinue medication and were doing well since it is known that the need for antipsychotic medication decreases with age for many patients. The other factor not mentioned by Whitaker is that Dr. George Brooks, the second author on the Harding et al. study and former director of the hospital, was known to be an extremely caring and conscientious psychiatrist who personally followed up his patients in the community. He was essentially a one-man ACT team. The important lesson from the Vermont study is thus not its outcome data, but rather the importance not only of continuity of care but also of continuity of caretaker, as described elsewhere (Torrey 1986).

Yet another study invoked by Whitaker to illustrate the importance of treating schizophrenia without antipsychotics is a treatment program in northern Finland covering a population of 70,000 people. The program was begun in 1969 by a Finnish psychoanalyst. Using “group family therapy,” individual psychotherapy, a belief that psychosis “arises from severely frayed social relationships” and a philosophy of “no immediate use of neuroleptics,” Whitaker claims that this group has achieved remarkable outcomes with 79 percent of first-episode patients being “asymptomatic” at the end of five years (p.340). He even claims that “schizophrenia is now disappearing from the region”(p.343). Most revealing and remarkable, however, is the fact that more than 40 years after this treatment program began, there are almost no publications describing its results and nobody in Finland or elsewhere has tried to replicate it. Robert Whitaker appears to be the person most impressed by it.

Whitaker clearly believes that schizophrenia should be treated without medication if at all possible. However he fails to focus any attention on the fact that on any given day in the United States half of all individuals with schizophrenia, or about one million people, are not being treated. This is a huge natural experiment to test his thesis. Many of these individuals are found in public shelters, sleeping under bridges, in jails, and in prisons. If Whitaker had spent more time in these settings observing the outcome of this natural experiment, instead of delivering lectures on his vision of the impending antipsychotic apocalypse, he would have written a very different book.

What SSI/SSDI Rates Say

Whitaker’s main thesis is that antipsychotic drugs are responsible for causing many, if not most, of the symptoms of schizophrenia. It is a “drug-induced epidemic of disabling mental illness” (p.361). His exhibit A to prove his thesis is the number of individuals receiving SSI and SSDI disability for mental illness. In 1987 there were 1.25 million Americans receiving such disability, or 1 in every 184 Americans. By 2007 this number had increased to 3.97 million people, or 1 in every 76 Americans, more than doubling over the 20 years. During these years antipsychotic and antidepressant use increased markedly in the United States, which Whitaker claims is the cause of the increase in the disability numbers.

The number of individuals on disability for mental illness under SSI and SSDI has indeed increased alarmingly, but not for the reason proposed by Whitaker. One reason is that virtually all public psychiatric beds have been closed so patients previously living in state mental hospitals are now living on SSI and SSDI in the community. In 1987 there were still 107,531 state hospital beds available for the nation’s 242.3 million people or 44.4 beds per 100,000 population. By 2007 the number of beds had been reduced to 37,376 for 301.6 million people, or 12.4 beds per 100,000 population. Almost all seriously mentally ill individuals who had become sick during that period had been put on mental disability on SSI and SSDI.

The other reason for the dramatic increase in SSI and SSDI disability numbers is that SSI and SSDI have become alternatives to welfare for poor and unemployed individuals who have any kind of psychiatric problem. During the Reagan administration there was an attempt to purge the SSI and SSDI roles of individuals who didn’t really qualify. This attempt resulted in court suits, bad publicity, and it thus politically backfired on President Reagan. Since that time the federal government has made no serious attempt to restrict SSI and SSDI with the result that most people who apply are ultimately approved if they persist through the appeals process. Indeed, a small industry of social workers and lawyers has grown up to help people get SSI and SSDI disability. They advertise on billboards in low income areas and on the Internet. In fact, coincidentally on the day I was writing this review, I received an Internet solicitation from a firm in Omaha, Nebraska offering to help me “if you were denied your Social Security or Disability benefits, or you would like to know if you qualify for them.” SSI and SSDI disability benefits are attractive; SSI currently pays $698 per month if you are single and $1,048 if you are married and SSDI pays more; the benefits continue indefinitely; and SSI eligibility also qualifies the person for Medicaid. It is thus no mystery why these disability programs have expanded so quickly.

The Dopamine Receptor Story

In addition to believing that antipsychotic drugs cause many, if not most, symptoms of schizophrenia, Whitaker thinks he knows how this happens. Almost forty years ago it was shown that when rats are given antipsychotic drugs, a subset of their dopamine-2 receptors markedly increase in number. Although it was not then, and still is not clear whether or not this also occurs in humans, it has led to a theory of supersensitivity psychosis. It is proposed that antipsychotic drugs cause an increase in the receptors, and when the drugs are stopped this causes symptoms that are identical to those of schizophrenia. This causes the person to be treated with more antipsychotics, causing a further increase in receptors, causing more symptoms when the drugs are withdrawn etc. Whitaker is correct that this could potentially be a serious problem, but at this point in time the reality of the problem in humans is unknown. That does not stop Whitaker from simply assuming its validity, as when he describes a patient who is “screaming and tearing at her hair” as being “deep into a withdrawal psychosis” (p.250). Whitaker does not say how he knows that this is a “withdrawal psychosis” rather than the symptoms of schizophrenia, a distinction that no professional has yet been able to make.

In fact, the dopamine receptor story is much more complicated than Whitaker described. The increase in rat dopamine receptors which can be caused by administering antipsychotic drugs can also be caused by administering caffeine, steroids, cocaine, amphetamines, by birth injuries, or even by social isolation (Seeman 2011). If rats, who have been administered amphetamines causing a marked increase in dopamine 2 receptors are then given antipsychotic drugs, this causes a decrease in the number of such receptors (Seeman 2008). It is not even known for certain how dopamine is causally related to the symptoms of schizophrenia. The recent finding that the parasite Toxoplasma gondii, which has been linked to schizophrenia in some studies (Torrey et al. 2008), can also produce dopamine (Gaskell et al. 2009), has further complicated dopamine’s role.

What is perhaps most surprising in Whitaker’s book, given his past career as a respected journalist, is his willingness to uncritically accept anything he has been told, as long as it fits his thesis and his wish to blame antipsychotics for everything except global warming. An example is his claim that in the 1970s “Loren Mosher was ousted from the NIMH for having run his Soteria experiment” (p. 304), an experiment in which some patients with schizophrenia were treated without drugs. Mosher at the time was the director of NIMH’s Schizophrenia Research Branch. The loss of his position had nothing to do with Soteria but rather was due to other factors. Most important was the fact that the schizophrenia research field was moving strongly in a biological direction at the time, with the approval of a new NIMH director. Mosher did not view schizophrenia as a brain disease, a view that put him increasingly at odds with both NIMH and the vast majority of researchers. NIMH continued to support the evaluation of the Soteria experiment after Mosher’s departure.

An example of Whitaker’s attribution of all adverse events to antipsychotics is in his description of a woman who was treated with several antipsychotics. She later got ovarian cancer and, according to Whitaker, “it’s possible that illness was related to the psychiatric medications” (p.214). On the contrary, multiple studies have reported that individuals with schizophrenia have a reduced risk of cancer, with some studies have even suggested that antipsychotic drugs have a cancer protective value (Wagner & Mantel 1978; Mortensen 1994; Caraillo & Benitez 1999). Similarly, when short on data Whitaker tends to invoke hyperbole to make his points: the use of antidepressants “routinely manufactures bipolar patients” (p. 181) and antipsychotics are causing “an astonishing medical disaster” (p. 193).

Schizophrenia Before Medications

So what does Whitaker believe about people diagnosed with schizophrenia? He never makes this clear but hints that he doubts that schizophrenia is a real disease. In his previous book, Mad in America, he described schizophrenia as a term “loosely applied to people with widely disparate emotional problems” and accused the medical profession of “minting ‘schizophrenics’ from a troubled cast of people.” In that book he recommended that such individuals be treated with “love and food and understanding, not drugs.” In the present book he recommends as treatment “rest, psychological therapies etc.” (p.19). Other than reciting studies showing that antipsychotics change some brain structures, as would be expected from an effective medication, Whitaker completely ignores the more than 100 studies demonstrating brain changes in individuals before they have ever been treated with antipsychotics.

But if schizophrenia is largely a product of antipsychotic drug use, who were those millions of people in state psychiatric hospitals beginning in the early 1800s until antipsychotics were introduced in the 1960s? This question seems to perplex Whitaker and he has no good answer. He weakly offers an explanation put forth by Mary Boyle, an English psychologist and author of Schizophrenia: A Scientific Delusion? Citing a 1990 paper by Boyle, Whitaker claims that “many of Kraepelin’s dementia praecox [schizophrenia] patients were undoubtedly suffering from a viral disease, encephalitis lethargica, which in the late 1800s had yet to be identified” (pp. 90-91). In fact, encephalitis lethargica is a well-defined syndrome which followed the 1917 influenza pandemic (Ravenholt & Foege, 1982). There is no evidence whatsoever that it existed prior to that time or continued beyond about 1930.

But something was going on in the nineteenth century to cause the number of insane persons, as they were then called, to increase markedly. As I detailed in my book The Invisible Plague: The Rise of Mental Illness from 1750 to the Present, the number of insane persons per 1000 population increased from approximately 0.5 per 1,000 in 1850 to 3.2 per 1,000 in 1950, a six fold increase. The largest percentage of these insane individuals had the disease we now call schizophrenia. The greatest irony of Whitaker’s book, in my opinion, is that there really was an epidemic of schizophrenia, but it occurred a century prior to the introduction of antipsychotic drugs, claimed by Whitaker to have caused it. The reason why his book is weakest in discussing this period is because literally millions of individuals with schizophrenia were not treated with drugs, as he now advocates. The outcome of this lack of treatment was not pretty, as he is aware.

Anatomy of an Epidemic is not without merit, however. In addition to detailing the many wrongs of American psychiatry, it reminds us what good psychiatric practice should be regarding the use of antipsychotic drugs. Use them in as low a dose as possible for no longer than necessary. Patients with a first episode of psychosis should be taken off the drugs several months after they go into remission to ascertain whether they are among the subgroup of patients who will not need maintenance medication. As patients age their medication can often be reduced and sometimes discontinued. And we need better research to be able to identify which patients need which drugs, who will develop which side effects, and who no longer needs medication. As psychiatrists we shouldn’t need to have a journalist remind us of these things; we should already be doing them.


  • Angell, M. The epidemic of mental illness and The illusions of psychiatry, a two-part review in the New York Review of Books, 2011.
  • Boyle M. Is schizophrenia what it was? A re-analysis of Kraepelin's and Bleuler's population. J Hist Behav Sci. 1990 Oct;26(4):323-33.
  • Carrillo JA, Benítez J. Are antipsychotic drugs potentially chemopreventive agents for cancer? Eur J Clin Pharmacol. 1999 Aug;55(6):487-8.
  • Cohen A, Patel V, Thara R, Gureje O. Questioning an axiom: better prognosis for schizophrenia in the developing world? Schizophr Bull. 2008 Mar;34(2):229-44. Epub 2007 Sep 28.
  • Gaskell EA, Smith JE, Pinney JW, Westhead DR, McConkey GA. A unique dual activity amino acid hydroxylase in Toxoplasma gondii. PLoS One. 2009;4(3):e4801. Epub 2009 Mar 11. doi:10.1371/journal.pone.0004801
  • Harding CM, Brooks GW, Ashikaga T, Strauss JS, Breier A. The Vermont longitudinal study of persons with severe mental illness, II: Long-term outcome of subjects who retrospectively met DSM-III criteria for schizophrenia. Am J Psychiatry. 1987 Jun;144(6):727-35.
  • Harrow M, Jobe TH. Factors involved in outcome and recovery in schizophrenia patients not on antipsychotic medications: a 15-year multifollow-up study. J Nerv Ment Dis. 2007 May;195(5):406-14.
  • Harrow M, Jobe TH, Faull RN. Do all schizophrenia patients need antipsychotic treatment continuously throughout their lifetime? A 20-year longitudinal study. Psychol Med. 2012 Feb 17:1-11.
  • Hegarty JD, Baldessarini RJ, Tohen M, Waternaux C, Oepen G. One hundred years of schizophrenia: a meta-analysis of the outcome literature. Am J Psychiatry. 1994 Oct;151(10):1409-16.
  • Jablensky A, Sartorius N. What did the WHO studies really find? Schizophr Bull. 2008 Mar;34(2):253-5. Epub 2008 Jan 18.
  • Messias EL, Chen CY, Eaton WW. Epidemiology of schizophrenia: review of findings and myths. Psychiatr Clin North Am. 2007 Sep;30(3):323-38. Review.
  • Mortensen PB. The occurrence of cancer in first admitted schizophrenic patients. Schizophr Res. 1994 Jun;12(3):185-94.
  • Ravenholt RT, Foege WH. 1918 influenza, encephalitis lethargica, parkinsonism. Lancet. 1982 Oct 16;2(8303):860-4.
  • Sartorius N, Jablensky A, Korten A, Ernberg G, Anker M, Cooper JE, Day R. Early manifestations and first-contact incidence of schizophrenia in different cultures. A preliminary report on the initial evaluation phase of the WHO Collaborative Study on determinants of outcome of severe mental disorders. Psychol Med. 1986 Nov;16(4):909-28.
  • Seeman P. All roads to schizophrenia lead to dopamine supersensitivity and elevated dopamine D2High receptors. CNS Neurosci Ther. 2011 Apr;17(2):118-32. doi: 10.1111/j.1755-5949.2010.00162.x.
  • Seeman P. Schizophrenia model of elevated D2High receptors: haloperidol reverses the amphetamine-induced elevation in dopamine D2High. Schizophr Res. 2009 Apr;109(1-3):191-2. Epub 2009 Jan 25.
  • Stalker H. The prognosis in schizophrenia. Based on a follow-up study of 129 cases treated by ordinary methods. J Ment Sci. 1939;85:1224-40.
  • Stephens JH. Long-term prognosis and followup in schizophrenia. Schizophr Bull. 1978;4(1):25-47.
  • Stevens J. Brief psychoses: do they contribute to the good prognosis and equal prevalence of schizophrenia in developing countries? Br J Psychiatry. 1987 Sep;151:393-6.
  • Teferra S, Shibre T, Fekadu A, Medhin G, Wakwoya A, Alem A, Jacobsson L. Five-year clinical course and outcome of schizophrenia in Ethiopia. Schizophr Res. 2012 Apr;136(1-3):137-42. Epub 2011 Nov 21.
  • Torrey EF. Continuous treatment teams in the care of the chronic mentally ill. Hosp Community Psychiatry. 1986 Dec;37(12):1243-7.
  • Torrey EF, Bartko JJ, Lun ZR, Yolken RH. Antibodies to Toxoplasma gondii in patients with schizophrenia: a meta-analysis. Schizophr Bull. 2007 May;33(3):729-36. Epub 2006 Nov 3.
  • Wagner S, Mantel N. Breast cancer at a psychiatric hospital before and after the introduction of neuroleptic agents. Cancer Res. 1978 Sep;38(9):2703-8.
  • Whitaker R. The case against antipsychotic drugs: a 50-year record of doing more harm than good. Med Hypotheses. 2004;62(1):5-13.

Visit Your State