Psychiatric Annals

CME 

Lessons from Placebo Arm Wrestle in Clinical Research

Devdutt Nayak, MD; Noshin Chowdhury, MD

Abstract

In the last half-century, randomized, placebo-controlled, double-blind design has become the norm for clinical trials and has created a windfall for the drug manufacturers. It has also created extra headaches for the statisticians measuring study outcomes. When compared with the placebo pill, there has been a steady decline in the effectiveness of new agents during last 25 years, whereas placebo responsiveness has increased in studies lasting less than 8 weeks. Placebo effect rates have risen approximately 7% per decade during the last 30 years, and more recent trials have shown an approximate 1.6-fold greater risk for placebo effects. There is a high, unpredictable, and heterogeneous responsiveness in many psychiatric conditions. True placebo effects can only be separated from non-specific treatment effects if a no-treatment control group is included; however, such trials are rarely performed in psychiatric conditions for ethical reasons.

In clinical practice, the positive-care effect can increase the magnitude of placebo component of the pharmacodynamic response of the drug. It is difficult to apply a placebo condition to psychotherapy as the therapist knows what “brand” of therapy is given, and some of the essential ingredients (ie, hope, belief, expectations, healing ritual, and emotionally charged interaction with an expert) are also common to placebo condition. It has been shown that the therapeutic relationship itself can produce benefits in psychotherapy, independent of the specific techniques employed. Experimental conditions affect both the subjects and the researchers in different and unpredictable ways. Blinded, randomized, controlled trials may be impartial and truth-seeking, but they are still vulnerable to measurement and reporting bias, and some scholars are in favor of replacing them with well-designed observational studies because their results do not overestimate the magnitude of therapeutic effect as once thought.

Abstract

In the last half-century, randomized, placebo-controlled, double-blind design has become the norm for clinical trials and has created a windfall for the drug manufacturers. It has also created extra headaches for the statisticians measuring study outcomes. When compared with the placebo pill, there has been a steady decline in the effectiveness of new agents during last 25 years, whereas placebo responsiveness has increased in studies lasting less than 8 weeks. Placebo effect rates have risen approximately 7% per decade during the last 30 years, and more recent trials have shown an approximate 1.6-fold greater risk for placebo effects. There is a high, unpredictable, and heterogeneous responsiveness in many psychiatric conditions. True placebo effects can only be separated from non-specific treatment effects if a no-treatment control group is included; however, such trials are rarely performed in psychiatric conditions for ethical reasons.

In clinical practice, the positive-care effect can increase the magnitude of placebo component of the pharmacodynamic response of the drug. It is difficult to apply a placebo condition to psychotherapy as the therapist knows what “brand” of therapy is given, and some of the essential ingredients (ie, hope, belief, expectations, healing ritual, and emotionally charged interaction with an expert) are also common to placebo condition. It has been shown that the therapeutic relationship itself can produce benefits in psychotherapy, independent of the specific techniques employed. Experimental conditions affect both the subjects and the researchers in different and unpredictable ways. Blinded, randomized, controlled trials may be impartial and truth-seeking, but they are still vulnerable to measurement and reporting bias, and some scholars are in favor of replacing them with well-designed observational studies because their results do not overestimate the magnitude of therapeutic effect as once thought.

“The most effective medication known to science, subjected to more clinical trials than any other medicament, yet nearly always does better than anticipated. The range of susceptible conditions appears to be limitless.”

The quirky definition cited above wrestles with the complexities of investigating placebo effects in clinical research. Shorter2 recently reviewed the history of placebo trials. In 1747, James Lind discovered the cure for scurvy by random assignment of 12 patients to five different remedies. The first placebo-controlled, single-blinded trial in medicine was published by Harry Gold in 1937, which compared xanthenes and placebo in patients with cardiac pain. In 1938, Dub and Lurie introduced the double-blinded trial in psychiatry on the effectiveness of benzedrine for depression. Since the 1950s, the randomized, placebo-controlled, double-blind design has become the “gold standard” for pharmacological trials. Randomized, controlled trials (RCT) were initially designed to weed out the treatments that did not work and not to provide proof of whether treatment worked.3 There is a high but heterogeneous placebo responsiveness in many psychiatric conditions.4

Separating Placebo Effects from Non-Specific Treatment Effects

Ernst and Resch5 helped to distinguish the true placebo effect from perceived placebo effect. The true placebo effect is the whole effect in the placebo group minus nonspecific factors. From a methodological point of view, a true placebo effect can only be measured if a no-treatment control group is included; however, such trials are rare. Ernst and Resch came across only 12 reports of three-arm trials; six of them reported on the ratings of pain measurement. Of these, a substantial true placebo response was found in four studies, a minimal true placebo response was found in one, and one study showed no true placebo responses.5 Kienle and Keine6 presented 21 factors that can create false impressions of a placebo effect. These included natural course of the disease; concomitant interventions; patient and researcher personalities, expectations, and biases; false positives; regression to the mean; environmental conditions; and the Hawthorne effect. Kienle and Keine argued that significance of placebo effect has been grossly overestimated and consists of all factors in the spectrum of clinical treatment.6 Hrobjartsson and Gotzsche7 have completed many studies on the importance and magnitude of placebo effects. They have reviewed RCTs that included both a placebo and a no-treatment control group. In a 2010 update, they reviewed 202 trials covering 60 health care problems in 16,566 patients and concluded that placebo interventions did not have important clinical effects in general. However, the larger group allowed them to do subgroup analyses, which revealed significant placebo effects for physical placebo interventions (eg, sham acupuncture), patient-reported subjective outcomes, trials that did not inform patients about the possible placebo intervention, small trials, and trials specifically designed to study placebo effects.7

Placebo Response Rates with Psychiatric Conditions

Walsh et al.8 reported that placebo effect rates have risen 7% per decade during the past 30 years. Among the recently published trials, the mean percentage rates of placebo improvement in major depression was 30% (range, 13% to 52%) and for schizophrenia, the mean percentage rate was 25% (range, 0% to 41%).9 Moreover, recent trials in schizophrenia had an approximate 1.6 times greater risk for placebo effects.10 Efficacy data from depression and schizophrenia trials submitted to the U.S. Food and Drug Administration (FDA) showed a high and incremental placebo response, falling treatment effects, high study failure rates during last 25 years, and an even higher rate in trials performed in the U.S.11–13 Furthermore, response to active drugs has decreased during the last 20 years.9

A recent meta-analysis of placebo response in 40 years of antipsychotic drugs for schizophrenia trials found new variables not found in the previous work.14 These are: younger age of patients, shorter illness duration, more severe baseline symptoms, and greater number of trial sites (particularly if the sites were non-academic or non-VA affiliated). All of the aforementioned variables resulted in increasing sample sizes, increased number of failed studies, much higher developmental costs, delays in the development of new psychotropic drugs, and even abandonment of new drug research due to prohibitive costs.10

Non-Specific Response Factors Unequal in Drug and Placebo Arms

Placebo response rates are dissimilar in the drug arm and in the placebo arm. The commonsense assumption of equality in both the drug and the placebo arm may not be true.15 In all RCTs, performed to date, drug efficacy has been derived by subtracting the response rates in the placebo arm from that in the drug arm. Meta-analysis of conditions with high placebo response rates (eg, psychiatric and functional disorders, pain) has identified some common placebogenic factors, such as improvement of symptoms in placebo run-in phase, or lower symptom severity at study entry.16,17 The same factors did not contribute to the drug’s efficacy. Hence, the placebo response in the drug arm must be driven by other mechanisms. Muthen and Brown18 reanalyzed data from depression and schizophrenia trials using a mathematical model and continuous-symptom monitoring data (rather than fixed endpoints) to segregate “always responders” from “never responders” in both drug and placebo conditions. They found discrepant response rates in both drug and placebo conditions, which could not be explained by the simple additive model.18 Thus, it seems one is not comparing “apples to apples” when analyzing placebo responsiveness in the two trial arms.

Response Rates in the Placebo Arm

As shown in the meta-analyses of depression and schizophrenia trials, the incremental rise in possibility of receiving an active drug during an RCT produces incremental placebo responses, up to a point. The classical placebo design assigns 50% of patients to either placebo or a drug arm (so-called 1:1 ratio), whereas in an RCT, when two or more active treatment arms exist compared to placebo (called 2:1 or more than 2:1 ratio), more patients are willing to participate in the trial.19 But the 1:1 ratio still yielded maximal recruitment efficiency and maximal ability to differentiate between a drug and placebo arm, as shown in trials in depression, schizophrenia, migraine headache, and irritable bowel syndrome (IBS).15 Among approximately 100 IBS trials that used the 2:1 or more than 2:1 ratio, a significantly higher placebo response rate was observed compared to the 1:1 ratio.20

In active comparators trials, the drug response goes up, but so does the true placebo response. Rutherford et al.21 estimated the placebo response in comparator trials when a novel drug is tested against drug already marketed. They found response rates to be an average of 15% higher in the active comparator group versus the 34% response in the placebo-controlled trial group. Rutherford and colleagues cautioned about implications of this for using placebo-controlled medication conditions for combined psychotherapy and medication trials.21

‘Active’ Placebo Pill Increases Placebo Response Rates

When active placebos (such as atropine or nicotine) are included, the differences between antidepressants and active placebos were small. This suggests that side effects (and the resultant unblinding with expectancy of improvement) may inflate the efficacy of antidepressants in trials using inert placebos.

Another emerging pattern from placebo control suggests that drug-placebo differences may be decreasing due to smaller percentages of patients assigned to placebos. This potentially occurs because patients involved in research studies want to have greater assurance of receiving an active treatment.22,23 Depression studies have shown that trials with more severely depressed patient showed larger drug-placebo differences.24 Interestingly, in schizophrenia trials, a larger placebo response was seen in patients with more severe baseline symptoms and/or shorter duration of trial. Once again, a significant heterogeneity was demonstrated.15

Physical Placebo Interventions Produce Longer-Lasting Improvements

Abrupt, early, and short-lived improvement appears to be an imprimatur of placebo response, as shown in two studies.25,26 In a study by Stewart et al.,26 even a quick response to the active drug (“a placebo pattern of response”) resulted in relapse, regardless of whether fluoxetine or placebo condition was continued in a subsequent 50-week maintenance trial; however, patients who responded gradually and incrementally to the drug maintained improvement on maintenance medication. Only those who switched to a placebo relapsed.

In a meta-analysis of antidepressant maintenance trials, 79% of placebo responders remained well on placebo compared to 93% of drug responders who remained well on drug continuation.27 In a meta-analysis of antipsychotic maintenance trials,28 36% remained well on placebo and 73% did well on the active drugs; so, both drug and placebo responses were less robust than in antidepressant trials. As shown in Hrobarjtsson and Gotzsche’s study,7 physical interventions (ie, sham surgery, fake acupuncture) often brought out robust, long-lasting placebo effects dependent on the patient’s perception of receiving active treatment, belief in the procedure, and expectations of benefit. The benefits (ie, symptom relief, increase in quality of life) often continue to persist in follow-up at 6 months, 12 months, or even more,29,30 quite possibly due to the ritualistic, invasive, and often stressful nature of the intervention and the steadfastness of patients’ belief.

Therapeutic Relationship Increases a Drug Response’s Placebo Effect

Clinicians are well aware that results from an RCT will advise us about statistical significance but rarely about clinical relevance. The high rate and ubiquitousness of placebo responses in psychiatric drug trials can translate into zero difference between the active drug and placebo.31 In a harried clinical practice, perhaps less time and attention, less frequent monitoring, and fewer measurements and concomitant therapies are given to the patient compared to trial subjects. Still, a clinical trial is less personal and of shorter duration than a relationship with a professional practice. As shown in the study of patients with IBS, time-augmented, warm, interpersonal relationships produced incremental therapeutic responses in sham acupuncture treatment.32 Certain strategies available to a psychiatrist (ie, watchful waiting, dosing adjustment, switching of drug) may alter the placebo component of an active drug. On the other hand, the patient is always assured of active treatment, and there is no element of uncertainty dampening the beneficial effect in clinical practice. Benedetti et al.33 consistently showed greater effect when patients knew they were receiving the drug openly compared with the same drug/ dosage given in hidden form. In psychiatric practice, the often long-term, trusting relationship many patients develop with their doctors maximizes the placebo component of a drug’s pharmacodynamic response.

Therapeutic Relationships Can Produce Benefits Irrespective

The notion of fleshing out placebo effects in psychotherapy is logistically complex. Some of the essential ingredients of psychotherapy (ie, hope, belief, expectations, therapeutic ritual, meaningful interaction with an expert) are also shared by placebos. Double-blind studies in psychotherapy are difficult to perform, as the therapists, by definition, know what treatment is being given or to design a control that is indistinguishable from active treatment. The closest placebo-like conditions include: waiting list; treatment as usual; no treatment; or miscellaneous controls, such as yoga, relaxation, and recreation. A review of 94 psychodynamic psychotherapy RCTs mostly showed superiority to inactive comparators (ie, the aforementioned placebo-like conditions)34 but did not differentiate from other empirically validated treatments such as supportive therapy, family therapy, group therapy, or cognitive behavioral therapy. Wampold35 called this an equivalence paradox and argued that all bona fide psychotherapies have equally efficacious global outcomes, independent of their partisan practitioners’ claim of technique specificity. A recent 16-week study of dynamic psychotherapy that included 156 patients randomized to psychotherapy alone (28% response rate) versus pharmacotherapy plus clinical management (31% response rate) versus placebo pills (24% response rate) failed to show any difference between the three groups.36 Kirsch and Saperstein37 analyzed data from no-treatment groups in psychotherapy studies and compared them with placebo groups in antidepressant studies. The drug group showed 33% more improvement over the patients in the placebo group; however, the placebo group showed 200% more improvement over the no-treatment group. Kirsch and Saperstein calculated that up to 50% of the drug effect may be due to placebo effect.

Clinical research is vulnerable to measurement and reporting bias, which is pervasive throughout medicine. Without having a no-treatment arm, placebo-controlled trials are mute as to comparative efficacy.2 The gold standard has lost its sheen, but it also has dulled the luster of the “miracle” of new psychoactive agents. In psychopharmacology, there has been a steady decline in the effectiveness of new agents during last 25 years. Common use of placebos in clinical trials may be hampering our ability to study the value of a pragmatic intervention for improving the lives of our patients.38

References

  1. Oh VM. The placebo effect: can we use it better?BMJ. 1994;309(6947):69–70. doi:10.1136/bmj.309.6947.69 [CrossRef]
  2. Shorter E. A brief history of placebos and clinical trials in psychiatry. Can J Psychiatry. 2011;56(4):193–197.
  3. Healy D. Trussed in evidence? Ambiguities at the interface between clinical evidence and clinical practice. Transcult Psychiatry. 2009;46(1):16–37. doi:10.1177/1363461509102285 [CrossRef]
  4. Khan A, Kolts RL, Rapaport MH, et al. Magnitude of placebo response and drug-placebo differences across psychiatric disorders. Psychol Med. 2005;35(5):743–749. doi:10.1017/S0033291704003873 [CrossRef]
  5. Ernst E, Resch KL. Concept of true and perceived placebo effects. BMJ. 1995;311(7004):551–553. doi:10.1136/bmj.311.7004.551 [CrossRef]
  6. Kienle GS, Kiene H. The powerful placebo effect: fact or fiction?J Clin Epidemiol. 1997;50(12):1311–1318. doi:10.1016/S0895-4356(97)00203-5 [CrossRef]
  7. Hróbjartsson A, Gotzsche PC. Placebo interventions for all clinical conditions. Cochrane Database Syst Rev. 2010;(1):CD003974.
  8. Walsh BT, Seidman SN, Sysko R, Gould M. Placebo response in studies of major depression: variable, substantial, and growing. JAMA. 2002;287(14):1840–1847. doi:10.1001/jama.287.14.1840 [CrossRef]
  9. Kinon BJ, Potts AJ, Watson SB. Placebo response in clinical trials with schizophrenia patients. Curr Opin Psychiatry. 24(2):107–113.
  10. Alphs L, Benedetti F, Fleischhacker WW, Kane JM. Placebo-related effects in clinical trials in schizophrenia: what is driving this phenomenon and what can be done to minimize it?Int J Neuropsychopharmacol. 2012;15(7):1003–1014. doi:10.1017/S1461145711001738 [CrossRef]
  11. Khin NA, Chen YF, Yang Y, Yang P, Laughren TP. Exploratory analyses of efficacy data from major depressive disorder trials submitted to the U.S. Food and Drug Administration in support of new drug applications. J Clin Psychiatry. 2011;72(4):464–472. doi:10.4088/JCP.10m06191 [CrossRef]
  12. Khin NA, Chen YF, Yang Y, Yang P, Laughren TP. Exploratory analyses of efficacy data from schizophrenia trials in support of new drug applications submitted to the US Food and Drug Administration. J Clin Psychiatry. 2012;73(6):856–864. doi:10.4088/JCP.11r07539 [CrossRef]
  13. Kemp AS, Schooler NR, Kalali AH, et al. What is causing the reduced drug-placebo difference in recent schizophrenia clinical trials and what can be done about it?Schizophr Bull. 36(3):504–509.
  14. Agid O, Siu CO, Potkin SG, et al. Meta-regression analysis of placebo response in antipsychotic trials, 1970–2010. Am J Psychiatry. 2013;170(11):1335–1344. doi:10.1176/appi.ajp.2013.12030315 [CrossRef]
  15. Enck P, Klosterhalfen S, Weimer K, Horing B, Zipfel S.. The placebo response in clinical trials: more questions than answers. Philos Trans R Soc Lond B Biol Sci. 2011;366(1572):1889–1895. doi:10.1098/rstb.2010.0384 [CrossRef]
  16. Evans KR, Sills T, Wunderlich GR, McDonald HP. Worsening of depressive symptoms prior to randomization in clinical trials: a possible screen for placebo responders?J Psychiatr Res. 2004;38(4):437–444. doi:10.1016/j.jpsychires.2003.11.006 [CrossRef]
  17. Bridge JA, Birmaher B, Iyengar S, Barbe RP, Brent DA. Placebo response in randomized controlled trials of antidepressants for pediatric major depressive disorder. Am J Psychiatry. 2009;166(1):42–49. doi:10.1176/appi.ajp.2008.08020247 [CrossRef]
  18. Muthén B, Brown HC. Estimating drug effects in the presence of placebo response: causal inference using growth mixture modeling. Stat Med. 2009;28(27):3363–3385. doi:10.1002/sim.3721 [CrossRef]
  19. Halpern SD, Karlawish JH, Casarett D, et al. Hypertensive patients’ willingness to participate in placebo-controlled trials: implications for recruitment efficiency. Am Heart J. 2003;146(6):985–992. doi:10.1016/S0002-8703(03)00507-6 [CrossRef]
  20. Enck P, Klosterhalfen S, Weimer K, Horing B, Zipfel S. The placebo response in clinical trials: more questions than answers. Philos Trans R Soc Lond B Biol Sci. 2011;366(1572):1889–1895. doi:10.1098/rstb.2010.0384 [CrossRef]
  21. Rutherford BR, Sneed JR, Roose SP. Does study design influence outcome?. The effects of placebo control and treatment duration in antidepressant trials. Psychother Psychosom. 2009;78(3):172–181. doi:10.1159/000209348 [CrossRef]
  22. Papakostas GI, Fava M. Does the probability of receiving placebo influence clinical trial outcome? A meta-regression of double-blind, randomized clinical trials in MDD. Eur Neuropsychopharmacol. 2009;19(1):34–40. doi:10.1016/j.euroneuro.2008.08.009 [CrossRef]
  23. Mallinckrodt CH, Zhang L, Prucka WR, Millen BA. Signal detection and placebo response in schizophrenia: parallels with depression. Psychopharmacol Bull. 2010;43(1):53–72.
  24. Khan A, Brodhead AE, Kolts RL, Brown WA. Severity of depressive symptoms and response to antidepressants and placebo in antidepressant trials. J Psychiatr Res. 2005;39(2):145–150. doi:10.1016/j.jpsychires.2004.06.005 [CrossRef]
  25. Quitkin FM, Rabkin JG, Stewart JW, et al. Heterogeneity of clinical response during placebo treatment. Am J Psychiatry. 1991;148(2):193–196.
  26. Stewart JW, Quitkin FM, McGrath PJ, et al. Use of pattern analysis to predict differential relapse of remitted patients with major depression during 1 year of treatment with fluoxetine or placebo. Arch Gen Psychiatry. 1998;55(4):334–343. doi:10.1001/archpsyc.55.4.334 [CrossRef]
  27. Khan A, Redding N, Brown WA. The persistence of the placebo response in antidepressant clinical trials. J Psychiatr Res. 2008;42(10):791–796. doi:10.1016/j.jpsychires.2007.10.004 [CrossRef]
  28. Leucht S, Tardy M, Komossa K, et al. Antipsychotic drugs versus placebo for relapse prevention in schizophrenia: a systematic review and meta-analysis. Lancet. 2012;379(9831):2063–2071. doi:10.1016/S0140-6736(12)60239-6 [CrossRef]
  29. Linde K, Witt CM, Streng A, et al. The impact of patient expectations on outcomes in four randomized controlled trials of acupuncture in patients with chronic pain. Pain. 2007;128(3):264–271. doi:10.1016/j.pain.2006.12.006 [CrossRef]
  30. Buchbinder R, Osborne RH, Ebeling PR, et al. A randomized trial of vertebroplasty for painful osteoporotic vertebral fractures. N Engl J Med. 2009;361(6):557–568. doi:10.1056/NEJMoa0900429 [CrossRef]
  31. Keefe RS, Kraemer HC, Epstein RS, et al. Defining a clinically meaningful effect for the design and interpretation of randomized controlled trials. Innov Clin Neurosci. 2013;10(5–6 Suppl A):4S–19S.
  32. Kaptchuk TJ, Friedlander E, Kelley JM, et al. Placebos without deception: a randomized controlled trial in irritable bowel syndrome. PLoS One. 2010;5(12):e15591. doi:10.1371/journal.pone.0015591 [CrossRef]
  33. Colloca L, Lopiano L, Lanotte M, Benedetti F. Overt versus covert treatment for pain, anxiety, and Parkinson’s disease. Lancet Neurol. 2004;3(11):679–684. doi:10.1016/S1474-4422(04)00908-1 [CrossRef]
  34. Gerber AJ, Kocsis JH, Milrod BL, et al. A quality-based review of randomized controlled trials of psychodynamic psychotherapy. Am J Psychiatry. 168(1):19–28.
  35. Wampold BE, Mondin GW, Moody M, et al. A meta-analysis of outcome studies comparing bona fide psychotherapies. Psychological Bull. 1997;122(3):203–215. doi:10.1037/0033-2909.122.3.203 [CrossRef]
  36. Barber JP, Barrett MS, Gallop R, Rynn MA, Rickels K. Short-term dynamic psychotherapy versus pharmacotherapy for major depressive disorder: a randomized, placebo-controlled trial. J Clin Psychiatry. 2012;73(1):66–73. doi:10.4088/JCP.11m06831 [CrossRef]
  37. Kirsch I, Sapirstein G. Listening to Prozac but hearing placebo: a meta-analysis of antideprssant medication. Prev Treat. 1998;1(2):No pagination specified; article 2a.
  38. Avins AL, Cherkin DC, Sherman KJ, Goldberg H, Pressman A. Should we reconsider the routine use of placebo controls in clinical research?Trials. 2012;13:44. doi:10.1186/1745-6215-13-44 [CrossRef]
Authors

Devdutt Nayak, MD, is Director, Consultation Psychiatry, Richmond University Medical Center. Noshin Chowdhury, MD, is Staff Psychiatrist, South Beach Psychiatric Center.

Address correspondence to: Devdutt Nayak, MD, 842 President Street, Brooklyn, NY 11215; email: sykofarm@gmail.com.

Disclosure: The authors have no relevant financial relationships to disclose.

10.3928/00485713-20140205-06

Sign up to receive

Journal E-contents