From Placebo to Panacea: Putting Psychiatric Drugs to the Test 1st Edition

Contemporary Psychology: A Journal of ReviewsSeptember 1998, Vol. 43, No. 9, 628-630Not a Panacea for Psychiatric DrugsA Review ofFrom Placebo to Panacea: Putting Psychiatric Drugs to the Testby Seymour Fisher and Roger P. Greenberg (Eds.)New York: Wiley, 1997. 404 pp. ISBN 0-471-14848-2. $49.95doi: 10.1037/001769Reviewed byDonald F. Klein, Frederic QuitkinThis edited book purports to be a reasoned criticism of recent scientific developments in psychiatry and psychopharmacology. It claims to find flaws with the conceptual bases of systematic diagnosis and placebo-controlled clinical trials, to demonstrate their self-serving, distorted therapeutic conclusions, and to reveal suppressed facts about alternative, superior, and safer psychosocial treatments. The topics addressed include the placebo, psychiatric comorbidity, the deficiencies of the Diagnostic and Statistical Manual of Mental Disorders (4th ed., American Psychiatric Association, 1994) approach, and ineffectual medication use in schizophrenia, mood, anxiety, and personality disorders and children.This might be a real contribution if carried out in the spirit of objective inquiry. Unfortunately, rather than a systematic analysis of pros and cons, a counterfeit of reasoned discussion is presented. This is made clear by reviewing expository tactics. Supposedly supportive data are briefly cited in conclusory, laudatory terms. Contradictory material receives tangential, lengthy, confusing, and critical barrages or is simply ignored.Treating schizophreniaFor instance, Cohen, critiques antipsychotic drugs (termed, in outdated terminology, neuroleptic) by citing the editors Ross and Pam from the 1995 book, Pseudoscience in Biological Psychiatry; Blaming the Body, who state:The conceptual system of biological psychiatry is organized [such that the] tautologies, positive feedback loops, closure to alternative hypotheses, pervasive overgeneralization, use of dissociation to eliminate cognitive dissonance, and other structural and functional properties of the system maintain it in a dysfunctional homeostasis. (p 178)But, unfortunately, the grounds for these sweeping conclusions are never presented. Pejorative terms such as "chemical straight jacket," "zombie effect," and "chemical lobotomy," are used in arguing "that most contemporary writers display a blind spot about NLP [neuroleptic] toxicity" (p 183). However, Cohen also describes the sudden popularity of risperidone which occurred after initial reports of equivalent efficacy but low extrapyramidal side effects. Cohen does not perceive the contradiction between a supposed psychiatric blindness to antipsychotic side effects and the rapid popularity of the less toxic new antipsychotics.In discussing antipsychotic effectiveness, Cohen ignores the overwhelming controlled demonstrations of acute benefit because these are apparently nullified by a 1961 French study (see p. 190). Citing Davis, Kane, Marder, Brauder, Gierl, Schouler, et al., "Patients on placebo relapse at a rate of 55%, whereas only 21% of schizophrenic patients relapse when they are on maintenance therapy." Cohen declares a 34 percent net effectiveness rate that is "by no means insignificant--given the disturbing impact of psychosis on the individual and his or her social network," but then claims that this "does not give an accurate picture of the NLP's role in helping the schizophrenic function better in society" (p 192). This last statement is true, irrelevant, and a common tactic in this book. Important facts are dismissed because other issues have not been addressed at that point, even though addressed later in the same article. Cohen also does not cite Davis et al.'s conclusion that virtually all those with schizophrenia will relapse within three years unless maintained on medication.That antipsychotics benefit acute negative symptoms (e.g., withdrawal, perplexity), has been known since the 1960s when the National Institute of Mental Health studies showed chlorpromazine particularly beneficial for such symptoms. This is ignored in favor of the "chemical straight-jacket" characterization.Controlled studies are relentlessly criticized for minor design problems, but when it comes to touting the advantages of nondrug approaches, anecdotes and conclusory language suffices. Cohen states Karon has performed "the most careful and detailed review of studies comparing psychotherapeutic and pharmacotherapeutic treatment of schizophrenia" (p 120). He cites Karon's remarkable conclusion that the lack of positive results found for the psychotherapy of schizophrenic patients can be explained by any of six different flaws, but that nevertheless "most studies reviewed showed psychotherapy to be at least as effective as NLP's" (p. 200). However, study content goes undiscussed. The negative controlled inpatient and follow-up work of May or the controlled outpatient studies of Gunderson, Frank, Katz, Vannicelli, Frosch, and Knapp are not mentioned.Danton and Antonuccio dismiss studies indicating that medication enhances behavioral therapies as "methodologically flawed" (p. 239) without further discussion, simply referring to Michelson and Marchione. Although that article states that these studies "design, assessment, treatment and statistical limitations rendered them more of historical interest," (p. 102) none of these flaws were detailed. A contagious critical vacuity somehow yields conclusory precision.The active placebo track recordSimilar dubious tactics are used regarding a supposed fatal flaw in placebo controlled trials. "All the published results in this area are probably seriously flawed because of failures of the double-blind" (p. 370). Thomson is cited by Fisher and Greenberg five times. That is,Thomson appraised all the double-blind, placebo-controlled studies completed between 1958 and 1972 that he could find. He noted that 68 used an inert placebo as a control while 7 employed an active one (atropine), which produced an array of body sensations. Comparing study outcomes, he discovered that tricyclic antidepressants had a superior therapeutic effect in 59% of the studies in which an inert placebo comparison occurred, but only one study (14% of the designs) showed the antidepressant to be superior when an active placebo was employed.... A few additional studies have shown results consistent with those reported by Thomson. (p. 135)No attempt is made to critically assess this report. Friedman, Granick, Cohen, and Cowitz studied psychotic depressed patients for 3 weeks using a maximum daily antidepressant dose of 200 mg. Four weeks is the bare minimum length for an antidepressant trial (Quitkin, Rabkin, Ross, & McGrath, 1984). Furthermore, 200 mg of imipramine, as used here, is considered ineffective for psychotic depression. This flawed study is incompetent to support their assertion.Hollister, Overall, Johnson, Pennington, Katz, and Shelton studied a heterogeneous mood disorder group. Thomson notes a drug effect was evident when the analysis was limited to "severely depressed subjects" but not the total sample. Greenberg and Fisher fail to note this contradictory caveat.Hussain reports on the following 3 groups: amitriptyline plus perphenazine, perhpenazine, or a placebo containing atropine. An 11/15 of the amitriptyline-per-phenazine-treated patients versus 7/19 placebo-treated patients were rated as at least much improved (X2 = 4.48, p < .03). This study shows medication superior to active placebo.Uhlenhuth and Park compared imipramine and atropine placebo in a crossover study. At 4 weeks, imipramine was statistically superior on the morale loss score. Inexplicably, Thomson concludes that these results suggest no imipramine advantage and Greenberg and Fisher agree. This study is compromised by low dose (imipramine 100 mg), short duration, and inclusion of psychoneurotics (poor responders to tricyclic antidepressant, Bielski & Friedel, 1976).Wilson, Vernon, and Sandiler did a 5-week, 4-group study. The relevant contrast was with anesthesia plus imipramine versus anesthesia plus placebo. In this low power study (N = 6 per group) demonstrating the differential utility of any treatment is virtually impossible, yet Greenberg and Fisher assert this affirms the null hypothesis.Daneman showed a robust imipramine superiority to atropine. Greenberg and Fisher (1989) cited four other instances in which active placebos "or reasonable equivalent" supposedly support their views. Fahy, Imlah, and Harrington found imipramine, electroconvulsive shock therapy (ECT), and sleep treatment equally effective. That ECT was not superior to sleep treatment suggests these patients did not require medication because the marked efficacy of electroconvulsive therapy is clear. McLean and Hakstian compared amitriptyline, short-term psychotherapy, and behavior therapy and used progressive relaxation as a control ("an active placebo"). The study was flawed, because dropouts were replaced, thus destroying randomization and the measures consisted of mailed self-reports, whose date of completion is not specified. Weintraub and Aronson compared imipramine and active placebo. On two of their three measures, imipramine was significantly superior to active placebo, thus refuting Greenberg and Fisher.In Friedman's study of outpatients receiving marital therapy, patients received 100 mg per day of amitriptyline that was compared with atropine placebo in "neurotic or reactive depression." The inadequate dose, context, and diagnostic group do not provide a competent base for generalizing about lack of drug efficacy. The theory behind active placebo is that it will enhance efficacy but, in fact, these treatment groups did not do well. The null results were due to the bad antidepressant outcomes.Fisher and Greenberg's summary of the active placebo versus antidepressant data glosses over substantial contradictory findings and converts minuscule, incompetent evidence into a firm, fallacious affirmation of the null hypothesis.In studies of putative new agents, at least half are no more effective than inactive placebo. This would be surprising if side effects elicit overwhelming bias or actual benefits as Fisher and Greenberg claim.Fisher and Greenberg affirm that anti-depressants must really be a placebo because there is no dose effect, citing Brugha, Bebbington, MacCarthy, Sturt, and Wykes's naturalistic study in which clinicians could treat by any means.Looking only at those receiving antidepressants, there was no significant differences in outcome between those placed on higher or lower anti-depressant doses. The lack of statistical significance for dosage difference remained even when other variables such as depression severity, duration, compliance, patient age, and gender were taken into account. (p. 130)But in a naturalistic study, that is exactly what one expects because more refractory patients eventually receive higher doses. One must distinguish between inferences allowable from experimental, as opposed to naturalistic studies or promote obfuscation.A balanced chapter on stimulant phar-macotherapy of attention deficit hyperactive disorder by Whalen and Henker (that somehow was included) is refuted by Fisher and Greenberg who cite McGuiness by saying that "the data consistently fail to support any benefits from stimulant medication" (p. 367). Consistently, the grounds for this conclusion are not presented.Book reviews do not usually include detailed literature analyses, but this was required to illustrate the authors' distortions. A complete refutation of this thick book requires an extended review, however, the cited material is representative. Much of the book consists of self-citations from the "peer reviewed" literature. This is ample evidence that the intellectual rigor of the peer review process is an assumption rather than a fact. Quitkin, Davis, Gerald, Davis, and Klein (in press) have presented a more extensive analysis.To paraphrase the American author Dorothy Parker, this book is not to be tossed aside lightly but hurled forcefully away.References* American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author.* Bielski, R. J., & Friedel, R. O. (1976). Predictions of tricyclic antidepressant response. Archives of General Psychiatry, 33, 1479-1489.* Greenberg, R. P., & Fisher, S. (1989). Examining antidepressant effectiveness: Findings, ambiguities, and some vexing puzzles. In S. Fisher & R. P. Greenberg (Eds.), The limits of biological treatments for psychological distress: Comparisons with psychotherapy and placebo (pp. 1-37). Hillsdale, NJ: Erlbaum.* Quitkin, F., Davis, J., Gerald, J., Davis, J. M., & Klein, D. F. (in press). Clinical trial validity of antidepressants. Archives of General Psychiatry.* Quitkin, F. M., Rabkin, J. G., Ross, D., & McGrath, P. J. (1984). Duration of antidepressant drug treatment: What is an adequate trial? Archives of General Psychiatry, 41, 238-245.Seymour Fisher (deceased), was professor of psychology at the State University of New York (SUNY) Health Science Center at Syracuse. Fisher was author of Development and Structure of Body Image and Sexual Images of the Self and coauthor, with R. Fisher, of The Psychology of Adaptation to Absurdity.Roger P. Greenberg is professor and head of the Division of Clinical Psychology in the Department of Psychiatry and Behavioral Science at SUNY Health Sciences Center at Syracuse. Greenberg is the author of more than 125 published articles and books, including The Scientific Credibility of Freud's Theories and Therapy (with S. Fisher), which was selected by both the National Library Association and Psychology Today as one of the best books in the behavioral sciences.Donald F. Klein, professor of psychiatry at Columbia University, College of Physicians and Surgeons (New York City), and director of research and the Department of Therapeutics at New York State Psychiatric Institute, is recipient of several awards, including the William R. McAlpin Award for Research Achievement (1988), The Society of Biological Psychiatry Gold Medal Award (1990), and the Heinz E. Lehman Research Award from New York State (1991).Frederic Quitkin is clinical professor of psychiatry at Columbia University and coauthor, with D. F. Klein, R. Klein, and A. Rifkin, of Diagnosis and Drug Treatment of Psychiatric Disorders.