Testing a drug against a placebo in a randomized double-blinded trial is the gold standard in the development of new treatment options in medicine. If the results show the drug to be significantly better than the placebo, the drug can be added to the medical armamentarium; if not, this does not demonstrate the “power of placebo,” as the placebo effects is a compound of different mechanisms, expectations, and previous experience, but also methodological biases, measurement inaccuracies, and regression to the mean effects (1).
Many medical therapies, however, do not provide pills but other interventions: surgery, technical procedures such as acupuncture, electrical or magnetic nerve or brain stimulation, manual therapies, e.g. physical therapy and massage, or nutritional therapy or psychotherapy. They require different control strategies than a placebo (pill). These are often called “sham” procedures.
Designing such control conditions is not easy, as it has to fulfill certain criteria: it should blind the patient so that he/she does not know whether he/she receives the true or the sham treatment; it should also blind the treating doctor (double-blinded); and should separate treatment and treatment evaluation and blind the data evaluator (triple-blinded). And it should allow randomization, random allocation to one or the other treatment irrespective of the wish of doctor or patient, and a few other signatures of evidence-based medicine (EBM). Last but not least, it should allow testing for the efficacy of blinding and, thereby, the quality of the trial. And this is often violated, noticed or unnoticed, especially in instrumental (technical) therapies (2).
Designing, for instance, a valid sham acupuncture needle (3) has taken efforts over many years, and different technical solutions are available claiming superiority over each other (4); and not all can be seen as “inert” (neutral) (5). Most require patients to be naïve, as once they have experienced “true” treatment they may easily recognize “snake oil.” Different from drug treatment, these more technical interventions cannot be used in what is called “cross-over” trials, where patients receive both treatments (true and placebo) but in a random order for a period of time. However, cross-over trials have their advantages and pitfalls — as we will discuss another time.
It is well established that the more invasive a medicine is, the higher patients’ expectations that it may work, and therefore, the higher may be the placebo response (or, as was said in the old days: effective medicine has to taste bitter): skin ointments work less than pills taken, pills are less effective than injections, the bigger a pill is the more it may be seen as helping, and two pills are better than one. Technical interventions, as discussed above, are specifically prone to these enhancing placebo effects, as has been shown in meta-analyses (6).
The ultimate invasiveness of a therapeutic procedure is surgery, opening the body’s surface by cutting the skin, and removing diseased tissue. Surgical interventions carry a risk, ever under the utmost controlled (e.g. sterile) conditions, to be harmful by itself: wounds can infect, can resist healing, can leave scars, irrespective of whether the surgery was successful or not. This is one reason why – in comparison to drug trials – sham-controlled surgical trials are rarely done, and did not start before the mid-20th century (7), much later than the first placebo-controlled drug trials.
The oldest example, according to (7), illustrates how sham (placebo) surgery is done: Some patients with psychosis underwent cingulate cortex isolation, i.e. surgical removal of a part of this brain area after opening the skull, while during the sham procedure in another few patients, only the skin was opened and a fraction of the skull bone was removed. Because this type of intervention for psychiatric disorders is no longer performed nowadays (for very good reasons), a more recent example may illustrate the purpose and procedure of sham surgery much better.
Moseley et al. (8) conducted a sham-controlled orthopedic surgery study between 1995 and 1998, after noting that knee osteoarthritis surgery is a questionable therapy with limited clinical validity but with “more than 650,000 such procedures … performed each year at a cost of roughly $5,000 each” in the US alone, amounting to annual costs in the range of 3 billion USD.
In painful knee osteoarthritis, standard operation procedure (SOP) was (and still is in many instances) the arthroscopic (minimal) opening of the knee joint space, removal (debridement) of rough material from the bones, and flushing the space with fluid. In the Moseley study, 180 such patients were assigned to one of three groups: receiving the SOP therapy (group A), flushing of the joint space only (B), and surgical incision of the skin but no further manipulation of the joint (C). Patients agreed to be randomized and were not informed what was done in their case, under anesthesia. Blinding the treating doctor was achieved by not disclosing what was done in the individual patients, and none of the study participants was aware of the procedure the patient had received. Blinding the single surgeon that performed all procedures was, of course, impossible, but the surgeon had no influence on the decision on what had to be performed but operated according to a pre-defined plan that was provided immediately prior to surgery and was not involved in the post-operative management of the patient. Post-operative symptoms were monitored for 2 years without disclosing the group assignment.
After two years, as well as at the many assessment points in-between, there were no differences in pain and mobility in the patients that received either of the two treatments (A,B) or no treatment (C) at all; all patients reported minor improvement in pain that was of no clinical relevance, but rather attributable to symptom change over time, or to the belief of the patient to have received effective therapy. While the study is often cited as an example of the “power of placebo” (9), for surgeons it instead underlined the ineffectiveness of a commonly used therapeutic procedure, and the data justify its termination. Following the Moseley study, clinical guidelines in orthopedics (10) have not included the procedure any longer.
Why some state “that sham surgery … was as effective as actual surgery in reducing pain and improving disability” (6) is unclear since with drug treatment the interpretation of similar data would force to conclude the opposite, surgery as un-effective as a sham (11). Systematic reviews and meta-analyses (6,9,12) of nearly 100 sham-controlled surgical and similar intervention trials found no difference between sham and surgery in nearly half of them, while in the other half, the superiority of surgery was small but consistent. Most scientists and many surgeons would agree that more sham-controlled trials are requested for all novel surgical therapies, e.g. also in cardiology and with implantable devices (13).
This is part 4 of a series covering “placebo” provided by Paul Enck and Sibylle Klosterhalfen from the Tübingen University Hospital. Continuous updates on placebo research can be found at www.jips.online.
- Kaptchuk TJ (2011). A brief history of the evolution of methods to control of observer biases in tests of treatments. JLL Bulletin: Commentaries on the history of treatment evaluation (www.jameslindlibrary.org). [Brief history]
- Wartolowska K, Beard D, Carr A. Blinding in trials of interventional procedures is possible and worthwhile. Version 2. F1000Res. 2017;6:1663.
- Streitberger K, Kleinhenz J. Introducing a placebo needle into acupuncture research. The Lancet. 1998; 352:364–5.
- Enck P, Klosterhalfen S, Zipfel S. Acupuncture, psyche and the placebo response. Auton Neurosci. 2010 ;157(1-2):68-73.
- Chae Y, Lee YS, Enck P. How placebo needles differ from placebo pills. Frontiers in Psychiatry 9:243; doi: 10.3389/fpsyt.2018.00243 (in press)
- Jonas WB, Crawford C, Colloca L, Kaptchuk TJ, Moseley B, Miller FG, Kriston L, Linde K, Meissner K. To what extent are surgery and invasive procedures effective beyond a placebo response? A systematic review with meta-analysis of randomised, sham controlled trials. BMJ Open. 2015;5(12):e009655.
- Wartolowska KA, Beard DJ, Carr AJ. The use of placebos in controlled trials of surgical interventions: a brief history. J R Soc Med. 2018;111(5):177-182.
- Moseley JB, O’Malley K, Petersen NJ, Menke TJ, Brody BA, Kuykendall DH, Hollingsworth JC, Ashton CM, Wray NP. A controlled trial of arthroscopic surgery for osteoarthritis of the knee. N Engl J Med. 2002;347(2):81-8.
- Louw A, Diener I, Fernández-de-Las-Peñas C, Puentedura EJ. Sham surgery in orthopedics: A systematic review of the literature. Pain Med. 2017;18(4):736-750.
- Siemieniuk RAC, Harris IA, Agoritsas T, Poolman RW, Brignardello-Petersen R, Van de Velde S, et al. Arthroscopic surgery for degenerative knee arthritis and meniscal tears: a clinical practice guideline. BMJ. 2017;357:j1982
- Enck P. A matter of perspective: Sham surgery as effective as surgery, or surgery as uneffective as sham? Pain Med. 2018; doi: 10.1093/pm/pny127. [Epub ahead of print]
- Smith JA, Carr AJ. The magnitude and temporal changes of response in the placebo arm of surgical randomized controlled trials: a systematic review and meta-analysis. Trials. 2016;17(1):589.
- Byrne RA, Capodanno D, Mahfoud F, Fajadet J, Windecker S, Jüni P, et al. Evaluating the importance of sham controlled trials in the investigation of medical devices in interventional cardiology. EuroIntervention. 2018; doi: 10.4244/EIJ-D-18-00481. [Epub ahead of print]