Placebo means a treatment that lacks a specific therapeutic effect. For a drug it is an agent that lacks any pharmacological effect. For a procedure it is an intervention that lacks any specific anatomical or physiological effect. Placebo treatment is not the same as no treatment. Normal treatment is the sum of natural history, placebo effect, and medical treatment.
Adequately defining placebo is challenging. Any treatment or intervention can have both specific effects and non-specific effects.
- The specific effects are due to known physiological mechanisms.
- The non-specific effects are a paradox and are due to unknown mechanisms. The paradox is resolved if the unknown mechanisms of the nonspecific effects become known.
The term placebo response is what the individual reports after receiving a placebo, and this usually can't be separated from confounding factors such as natural history because of the lack of a "no-treatment" arm. The placebo effect is any any treatment effect that is not due to the specific, intended effects of a treatment. There have been several attempts to create consensus definitions, but the terms placebo effect and response are often used interchangeably.
The total effect of an intervention may be a combination of both its physiological known specific effects, and unknown non-specific placebo effect. An absolute placebo effect is when the intervention has no known therapeutic physiological effect or an intervention that is designed to simulate medical therapy but doesn't have a specific therapeutic effect. The opposite of a placebo effect is a nocebo effect, where harm occurs for non-specific unknown reasons.
Placebo analgesia is the reduction or disappearance of pain, through the placebo effect, when a placebo is given to a patient who is told that it is a painkiller.
Comparison with Other Effects
The placebo effect is a specific phenomenon that is sometimes confused with other effects.
Natural history: some conditions naturally improve over time because of healing or some other factor. The improvement in pain is a property of the disorder that the patient has not the placebo effect. These two effects are distinguished through comparing the outcomes of patients taking placebo and the outcomes of patients having no treatment ("no treatment group").
Regression to the mean: Patients with chronic pain may be more likely to present to a doctor and join a trial when their pain is on the more severe end of what they have been experiencing to date. Conversely, patients who are doing well compared to their average are less likely to present to doctors and enrol in trials. There is therefore an overrepresentation of patients who are in a severe period of their naturally fluctuating pain levels. Measurements of pain and function may not represent their normal state, but the state of their flare. Regression to the mean refers to the statistical phenomenon that on average their pain is likely to improve to their previous average level of pain, regardless of treatment. This is not a placebo effect, but simply reflects normal fluctuations in pain. This effect can be controlled through using appropriate selection criteria.
Hawthorne effect: When individuals are being observed there is a change in performance simply because they know they are being studied. This is a conscious or subconscious change.
Symptom Ambiguity and Biases: This can be controlled by using objective physiological measurements.
Cointervention: It is important to rule out the effects of co-intervention. For example the simple act of mechanically inserting a needle during inject may in itself produce analgesia.
Behavioural Conditioning: This theory views the placebo response as a conditioned response to features of the treatment setting such as the doctor's style of dress, equipment, medication. Relief occurs because of past experiences of having relief from going to the doctor.
In the acquisition phase, the neutral stimulus (e.g. an environment stimulus or inert drug), is paired with the unconditioned stimulus (e.g. active drug). After one or several pairings the neutral stimulus becomes the conditioned stimulus. In the evocation phase, the conditioned stimulus is able to mimic the effects of the unconditioned stimulus.
This mechanism is hard to apply to chronic pain. If anything, patients with chronic pain are likely to have had repeated failures, and so would be conditioned not to respond. Viewing it from this negative viewpoint can be an explanation for nocebo responses and "placebo sag.'
Response Expectancy Theory: This views the placebo response as being due to an expectation from the patient that a treatment will relieve their pain. The placebo effect has been found to be enhanced through credibility of the doctor, therapeutic setting, treatment, administrative setting; and the nature of the interaction between the patient and doctor.
The effect of expectancy is illustrated by the open/hidden drug paradigm. Identical concentrations of active drugs are administrated in either an open or hidden manner. In the hidden manner the patient is unaware of the timing of administration, for example a computer administers it automatically.
Meaning Model: This model views that part the patients symptoms may be amplified due to fear. It looks at what factors are needed to address fear and thereby maximise the placebo response:
- The patient must feel listened to and receive a valid coherent explanation for their illness
- The patient must feel care and compassion from the doctor
- The patient must feel empowered.
By having these three factors, any fears are reduced and pain and function may improve simply as a result of that. Delivering a treatment with confidence and conviction can also also allay fear.
Cognitive Dissonance Theory: The patient avoids holding two psychologically inconsistent beliefs: that the treatment would work and only very sick people don't improve. To reduce this dissonance the patient alters their perception of their symptoms.
The evidence is compelling that the placebo effect is due to biochemical effects, and is not a purely psychological phenomenon.
There are a variety of brain centres (dorsolateral prefrontal cortex, anterior cingulate cortex, amygdala, and periaqueductal grey) that amplify or inhibit incoming pain signals. Descending signals can bend sent to various spinal segments and cause diffuse inhibition. This is called the diffuse noxious inhibitory control (DNIC). The role of the DNIC is to suppress less important ascending information so that the nociceptive information is highlighted, and this is called centre-surround inhibition (see Nociception)
Suppression occurs with global stimulation of brainstem nuclei, causing inhibitory signals through descending pathways to all segmental levels of the spinal cord. Opioids work by globally inhibiting the descending pathways. The nociceptive information thus becomes obscured by the now uninhibited background noise.
There is compelling evidence that endogenous opioids play an important role in the placebo effect through the descending pain modulating circuit. There are several streams of evidence for this.
- The placebo effect has been shown to be reversed by naloxone in several studies.
- The placebo effect is enhanced by cholecystokinin antagonists. (cholecystokinin has anti-opioid action).
- Placebo analgesia can mimic the respiratory depression side effect seen with exogenous opioids, and this side effect is reversible with naloxone.
- With positron emission tomography (PET) there is evidence of activation of the descending pain-modulating circuit.
- A reduction in beta-adrenergic sympathetic activity during placebo analgesia is blocked by naloxone.
There is also some evidence that catecholamines, dopamine, and cortisol are involved.
A commonly quoted figure is that 35% is the standard incidence of placebo response rates, or otherwise stated as one third of patients will exhibit a placebo effect. However this figure, published in 1955, was the average placebo responses in 15 papers, with the range being 15-58%, and it also didn't take into account natural history or regression to the mean. In fact, subsequent studies has found that the incidence of placebo response varies between 0 and 100% depending on the disease, environment, investigator, and other factors. The "constant" 35% figure is a myth.
Furthermore, any given individual can be both a placebo responder and a placebo non-responder under different settings. There is no such thing as a personality or psychological trait that causes individuals to consistently respond to placebos.
In patients with chronic spinal pain, there is no difference in the psychological profiles of patients who have true-positive responses to placebo-controlled diagnostic blocks and those who have placebo responses. There is also no difference between patients who have successful outcomes after successful treatment with radiofrequency neurotomy with those who do not.
Essentially every pain intervention has a placebo component.
In placebo controlled trials in nociceptive and idiopathic pain the magnitude of the placebo effect is highly variable. The effect size is large on average with an effect size above 0.8. In systematic reviews, analgesics tend to barely reach clinical significance for effect compared to placebo.
In the context of surgical interventions the placebo treatment is called the "sham" treatment. In the study of orthopaedic surgery for the treatment of acute and chronic pain conditions, there have only been 7 sham controlled trials. All but one of these 7 trials found no difference between the intervention and sham. 
The singular positive trial was for vertebroplasty and stands in contrast to the three other negative vertebroplasty sham controlled trials. It was also industry funded and only looked at pain reduction in at 8 weeks (44% vs 21% NRS <4/10 at 14 days). A cheeky systematic review of orthopaedic surgery as a field found that sham surgery was as effective as orthopaedic surgery for reducing pain and disability.
The placebo effect also occurs with less invasive procedures, and the placebo intervention is again called sham.
For example, several sham controlled studies have been done for lumbar facet joints injections. Intra-articular lumbar facet joint corticosteroid injection has been shown to have no greater effect than intra-articular saline injection. Corticosteroid, local anaesthetic, and a mixture are all equally effective when injected into lumbar facet joints. Facet joint injections were also found to be equal to facet nerve blocks.
The study of interventions has tended to move towards using categorical data as a primary outcome rather than average response rate, as using average response rates can hide the positive response in a subgroup. In this lumbar transforaminal corticosteroid injection was found to be superior to a variety of sham interventions.
For zygapophysial joint pain, when patients are adequately selected using controlled blocks, radiofrequency neurotomy has been found to be superior to sham, however the Nath study did not report any data from which categorical outcomes could be calculated. In the lumbar spine, studies that have not followed the evidence based Spine Intervention Society guidelines for selection and/or technique haven't found consistent outcomes. (See Lumbar Zygapophysial Joint Precision Treatment)
Diagnostic procedures are also susceptible to the placebo effect. Patients who think they are having diagnostic zygapophysial joint blocks have a significant false positive rate with the subcutaneous injection of saline. Stellate ganglion blocks using normal saline and intravenous infusions of normal saline can relieve CRPS symptoms.
Minimising The Placebo Response
In general, trials should attempt to minimise the placebo response to optimise drug-placebo differences. This ensures that the efficacy of the drug in question can be properly evaluated. This is a scientific, ethical, and regulator requirement when testing new compounds in order to prevent overestimating the potency of the target drug.
The randomised controlled trial has been the standard study design to address non-specific effects. However this is based on the assumption that the placebo effects in the placebo group are identical to the placebo effects in the drug group, and that they both combine in an additive manner. However this is not necessarily always the case.
There have been many variants of this design over the years to try and optimise drug-placebo differences, all with various problems that either increase, obscure, or even ignore the placebo response.
Enrichment or Multi-dosing Trials: In this strategy, more patients are randomised to the active treatment group. However this resulted in even higher placebo responses compared to trials with equal numbers of patients in both the drug and placebo groups. It turns out that having an increased chance of being in a treatment group increases the magnitude of the placebo response.
Placebo Responder Identification Trials: In this strategy, placebo responders are identified at an early stage through a placebo run-in period or by repeated treatment phases with re-randomisation. This study design can't prevent placebo responses at a later phase of the trial.
Crossover Trials: Here different treatment periods are applied to the same patients. However carry-over effects occur. It is difficult to adequately interpret the results due to intra-individual variability of the placebo response. Furthermore, treatments in the first phase can generate behavioural conditioning effects in the second phase.
Another risk is that crossover designs can lead to un-blinding due to perceived differences in side effects. This risk can be mitigated by using an active placebo, one that mimics the side effects of the drug under investigation, however these are difficult to develop. Without active placebos it can't be ruled out that any difference in effect is due to expectations generated within the patient as a result of experiencing side effects. On the other hand, while active placebos may reduce the risk of false-positive detection of a difference between groups, they can increase the placebo response due to the side effects.
Drug Responder Identification Trials: Instead of identifying placebo responders, the trials aims to identify drug responders in the run-in phase or by pre-selecting patients who were previously exposed to a similar drug. This method increases the drug-placebo difference, but it limits the external validity of the results.
No-Treatment Groups: No treatment groups help to disentangle the effects arising from the natural history of the condition, i.e. the fact that spontaneous remission and fluctuation of symptoms may occur simply through the natural disease process. Some trials use "treatment as usual" groups but this is difficult to standardise and monitor.
Comparative Effectiveness Research Trial: In this design there is no placebo group. The treatment under investigation is compared to another active treatment. This design arises out of ethical concerns where it is viewed that all patients should receive the best treatment available.
This trial design has several flaws. Comparing a new drug to an active comparator results in findings of higher efficacy than when the drug is compared to placebo. This sort of design increases the placebo response without being able to control for it. It is also tricky to select a fair comparator for the treatment under investigation. Furthermore, there are certain problems that arise with statistical analysis of 'non-inferiority,' leading to needing a four-four larger sample size compared to classical RCTs.
Patient Reported Outcomes: Patient reported outcomes are standard as part of pain research. Unfortunately placebo responses are higher in outcomes based on questionnaires than in outcomes based on clinical biomarkers. This is not available for pain and function
General Measures: General measures used to limit the placebo response are standardising for symptom severity, avoiding physician selection bias during recruitment, controlling for centre effects, controlling for patient adherence, and controlling for and ensuring effective blinding.
Maximising The Placebo Response
Unlike with research trials, the placebo effect can be enlisted in clinical practice to maximise the response in an ethical way through practice style and behaviour of the doctor.
Sustained Partnership: Having a long term relationship with patients can lead to better outcomes. The ethical attitudes that enhance the placebo response rate are expressing interest in the total person, providing longitudinal care, being adaptive to the patient's idiosyncrasies, and avoiding cookbook medicine. From the patients side they must feel like they are in a caring, sensitive, and empathetic environment, and also view the doctor as reliable and trustworthy. Share decision making is important.
Mastery: The patient should feel empowered, and change from feeling like a passive entity that is totally dependent on the doctor, to feeling like they are in control of their health. They should be encouraged to express their ideas, concerns, and expectations.
Story: This is the narrative that is weaved by the doctor in a way that the patient can understand. Management is often easier when there is a cohesive and coherent narrative. Patients get less analgesic response if they are unaware that they are receiving an analgesic. They do better if they are expressly told that they are receiving strong painkillers. It is vital that patients understand how the analgesic works, what they are for, and how to use them properly.
Placebo should not be used as sole therapy. Otherwise it can delay seeking more effective treatments, it can reduce the future response of active therapy, it can add to treatment cost, it can result in a nocebo effect, and it can be a constant reminder of illness.
- Benedetti. Mechanisms of placebo and placebo-related effects across diseases and treatments. 2008 (closed access)
- Enck et al. The placebo response in medicine: minimize, maximize or personalize? 2013 (closed access)
- Study Design
- Peck C, Coleman G. Implications of placebo theory for clinical research and practice in pain management. Theor Med. 1991 Sep;12(3):247-70. doi: 10.1007/BF00489609. PMID: 1721730.
- Brody H. The placebo response. Recent research and implications for family medicine. J Fam Pract. 2000 Jul;49(7):649-54. PMID: 10923577.
- Enck P, Bingel U, Schedlowski M, Rief W. The placebo response in medicine: minimize, maximize or personalize? Nat Rev Drug Discov. 2013 Mar;12(3):191-204. doi: 10.1038/nrd3923. PMID: 23449306.
- Amanzio M, Pollo A, Maggi G, Benedetti F. Response variability to analgesics: a role for non-specific activation of endogenous opioids. Pain. 2001 Feb 15;90(3):205-215. doi: 10.1016/S0304-3959(00)00486-3. PMID: 11207392.
- Petrovic P, Kalso E, Petersson KM, Ingvar M. Placebo and opioid analgesia-- imaging a shared neuronal network. Science. 2002 Mar 1;295(5560):1737-40. doi:10.1126/science.1067176. Epub 2002 Feb 7. PMID: 11834781.
- BEECHER HK. The powerful placebo. J Am Med Assoc. 1955 Dec 24;159(17):1602-6. doi: 10.1001/jama.1955.02960340022006. PMID: 13271123.
- Wall PD. The placebo effect: an unpopular topic. Pain. 1992 Oct;51(1):1-3. doi: 10.1016/0304-3959(92)90002-S. PMID: 1454391.
- Voudouris NJ, Peck CL, Coleman G. Conditioned response models of placebo phenomena: further support. Pain. 1989 Jul;38(1):109-116. doi: 10.1016/0304-3959(89)90080-8. PMID: 2780058.
- Lord SM, Barnsley L, Wallis BJ, Bogduk N. Chronic cervical zygapophysial joint pain after whiplash. A placebo-controlled prevalence study. Spine (Phila Pa 1976). 1996 Aug 1;21(15):1737-44; discussion 1744-5. doi: 10.1097/00007632-199608010-00005. PMID: 8855458.
- Wallis BJ, Lord SM, Bogduk N. Resolution of psychological distress of whiplash patients following treatment by radiofrequency neurotomy: a randomised, double-blind, placebo-controlled trial. Pain. 1997 Oct;73(1):15-22. doi: 10.1016/s0304-3959(97)00060-2. PMID: 9414052.
- Vase L, Riley JL 3rd, Price DD. A comparison of placebo effects in clinical analgesic trials versus studies of placebo analgesia. Pain. 2002 Oct;99(3):443-452. doi: 10.1016/S0304-3959(02)00205-1. PMID: 12406519.
- Vase L, Petersen GL, Riley JL 3rd, Price DD. Factors contributing to large analgesic effects in placebo mechanism studies conducted between 2002 and 2007. Pain. 2009 Sep;145(1-2):36-44. doi: 10.1016/j.pain.2009.04.008. Epub 2009 Jun 25. PMID: 19559529.
- Quessy SN, Rowbotham MC. Placebo response in neuropathic pain trials. Pain. 2008 Sep 15;138(3):479-483. doi: 10.1016/j.pain.2008.06.024. Epub 2008 Aug 15. PMID: 18706762.
- Buchbinder R, Osborne RH, Ebeling PR, Wark JD, Mitchell P, Wriedt C, Graves S, Staples MP, Murphy B. A randomized trial of vertebroplasty for painful osteoporotic vertebral fractures. N Engl J Med. 2009 Aug 6;361(6):557-68. doi: 10.1056/NEJMoa0900429. PMID: 19657121.
- Bradley JD, Heilman DK, Katz BP, Gsell P, Wallick JE, Brandt KD. Tidal irrigation as treatment for knee osteoarthritis: a sham-controlled, randomized, double-blinded evaluation. Arthritis Rheum. 2002 Jan;46(1):100-8. doi: 10.1002/1529-0131(200201)46:1<100::aid-art10037>3.0.co;2-v. PMID: 11817581.
- Kallmes DF, Comstock BA, Heagerty PJ, Turner JA, Wilson DJ, Diamond TH, Edwards R, Gray LA, Stout L, Owen S, Hollingworth W, Ghdoke B, Annesley-Williams DJ, Ralston SH, Jarvik JG. A randomized trial of vertebroplasty for osteoporotic spinal fractures. N Engl J Med. 2009 Aug 6;361(6):569-79. doi: 10.1056/NEJMoa0900563. Erratum in: N Engl J Med. 2012 Mar 8;366(10):970. PMID: 19657122; PMCID: PMC2930487.
- 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 Jul 11;347(2):81-8. doi: 10.1056/NEJMoa013259. PMID: 12110735.
- Firanescu CE, de Vries J, Lodder P, Venmans A, Schoemaker MC, Smeets AJ, Donga E, Juttmann JR, Klazen CAH, Elgersma OEH, Jansen FH, Tielbeek AV, Boukrab I, Schonenberg K, van Rooij WJJ, Hirsch JA, Lohle PNM. Vertebroplasty versus sham procedure for painful acute osteoporotic vertebral compression fractures (VERTOS IV): randomised sham controlled clinical trial. BMJ. 2018 May 9;361:k1551. doi: 10.1136/bmj.k1551. Erratum in: BMJ. 2018 Jul 4;362:k2937. Smeet AJ [corrected to Smeets AJ]. PMID: 29743284; PMCID: PMC5941218.
- Clark W, Bird P, Gonski P, Diamond TH, Smerdely P, McNeil HP, Schlaphoff G, Bryant C, Barnes E, Gebski V. Safety and efficacy of vertebroplasty for acute painful osteoporotic fractures (VAPOUR): a multicentre, randomised, double-blind, placebo-controlled trial. Lancet. 2016 Oct 1;388(10052):1408-1416. doi: 10.1016/S0140-6736(16)31341-1. Epub 2016 Aug 17. Erratum in: Lancet. 2017 Feb 11;389(10069):602. PMID: 27544377.
- Moradi A, Pasdar P, Mehrad-Majd H, Ebrahimzadeh MH. Clinical Outcomes of Open versus Arthroscopic Surgery for Lateral Epicondylitis, Evidence from a Systematic Review. Arch Bone Jt Surg. 2019 Mar;7(2):91-104. PMID: 31211187; PMCID: PMC6510924.
- Louw A, Diener I, Fernández-de-Las-Peñas C, Puentedura EJ. Sham Surgery in Orthopedics: A Systematic Review of the Literature. Pain Med. 2017 Apr 1;18(4):736-750. doi: 10.1093/pm/pnw164. PMID: 27402957.
- Carette S, Marcoux S, Truchon R, Grondin C, Gagnon J, Allard Y, Latulippe M. A controlled trial of corticosteroid injections into facet joints for chronic low back pain. N Engl J Med. 1991 Oct 3;325(14):1002-7. doi: 10.1056/NEJM199110033251405. PMID: 1832209.
- Lilius G, Laasonen EM, Myllynen P, Harilainen A, Grönlund G. Lumbar facet joint syndrome. A randomised clinical trial. J Bone Joint Surg Br. 1989 Aug;71(4):681-4. doi: 10.1302/0301-620X.71B4.2527856. PMID: 2527856.
- Marks RC, Houston T, Thulbourne T. Facet joint injection and facet nerve block: a randomised comparison in 86 patients with chronic low back pain. Pain. 1992 Jun;49(3):325-328. doi: 10.1016/0304-3959(92)90239-8. PMID: 1408298.
- Ghahreman A, Ferch R, Bogduk N. The efficacy of transforaminal injection of steroids for the treatment of lumbar radicular pain. Pain Med. 2010 Aug;11(8):1149-68. doi: 10.1111/j.1526-4637.2010.00908.x. PMID: 20704666.
- Nath S, Nath CA, Pettersson K. Percutaneous lumbar zygapophysial (Facet) joint neurotomy using radiofrequency current, in the management of chronic low back pain: a randomized double-blind trial. Spine (Phila Pa 1976). 2008 May 20;33(12):1291-7; discussion 1298. doi: 10.1097/BRS.0b013e31817329f0. PMID: 18496338.
- Lord SM, Barnsley L, Wallis BJ, McDonald GJ, Bogduk N. Percutaneous radio-frequency neurotomy for chronic cervical zygapophyseal-joint pain. N Engl J Med. 1996 Dec 5;335(23):1721-6. doi: 10.1056/NEJM199612053352302. PMID: 8929263.
- Moussa WM, Khedr W. Percutaneous radiofrequency facet capsule denervation as an alternative target in lumbar facet syndrome. Clin Neurol Neurosurg. 2016 Nov;150:96-104. doi: 10.1016/j.clineuro.2016.09.004. Epub 2016 Sep 5. PMID: 27618781.
- Schneider BJ, Doan L, Maes MK, Martinez KR, Gonzalez Cota A, Bogduk N; Standards Division of the Spine Intervention Society. Systematic Review of the Effectiveness of Lumbar Medial Branch Thermal Radiofrequency Neurotomy, Stratified for Diagnostic Methods and Procedural Technique. Pain Med. 2020 Jun 1;21(6):1122-1141. doi: 10.1093/pm/pnz349. PMID: 32040149.
- Schwarzer AC, Wang SC, Bogduk N, McNaught PJ, Laurent R. Prevalence and clinical features of lumbar zygapophysial joint pain: a study in an Australian population with chronic low back pain. Ann Rheum Dis. 1995 Feb;54(2):100-6. doi: 10.1136/ard.54.2.100. PMID: 7702395; PMCID: PMC1005530.
- Price DD, Long S, Wilsey B, Rafii A. Analysis of peak magnitude and duration of analgesia produced by local anesthetics injected into sympathetic ganglia of complex regional pain syndrome patients. Clin J Pain. 1998 Sep;14(3):216-26. doi: 10.1097/00002508-199809000-00008. PMID: 9758071.
- Verdugo RJ, Ochoa JL. Abnormal movements in complex regional pain syndrome: assessment of their nature. Muscle Nerve. 2000 Feb;23(2):198-205. doi: 10.1002/(sici)1097-4598(200002)23:2<198::aid-mus9>3.0.co;2-4. PMID: 10639611.