Public and professional attitudes towards deceptive and open-label placebo in functional neurological disorder and wider neurological practice

Abstract

Background

Placebo effects are powerful and have been suggested to be particularly relevant in certain neurological conditions, including functional neurological disorder (FND).

Methods

A survey on attitudes towards and current practice of deceptive placebo treatments and ethical alternatives, notably positive suggestion, trust and open-label placebo was performed among health professionals and lay people with and without neurological diagnoses.

Results

116 healthcare professionals and 631 lay people (176 FND, 332 with other neurological diagnoses, 61 with medical diagnoses, 62 healthy controls) completed the survey.

71% of lay people but only 46% of healthcare professionals were in favour of deceptive placebo treatments. Among lay people, healthy individuals were most in favour (87%), and people with FND were least in favour (62%). All groups were sceptical towards open-label placebo, yet neurologists were most open to this practice.

Placebo was considered more effective for functional than non-functional disorders by healthcare professionals, but not by patients. Healthcare professionals reported only rarely using placebo in clinical practice, and if so, mainly in the diagnosis or treatment of FND.

Conclusions

This is the first survey on opinions and current practice of placebo treatments in neurological practice. The results show a mixed picture, with deceptive placebos being perceived as effective and acceptable by most lay people (though strongly opposed by some, particularly by some patients with FND) and mostly considered more negatively by healthcare professionals. Ethically acceptable alternatives of harnessing the power of placebo without deception were considered with scepticism by all respondents, but least so by neurologists.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Placebo effects are genuine, potent and share pathophysiological mechanisms with some neurological conditions, including functional neurological disorder, indicating a potential special relevance.

WHAT THIS STUDY ADDS

• This is the first survey in neurological illness on opinions and clinical practice regarding placebo treatments. It shows an openness towards deceptive placebo treatments in most lay people, but less so in people with functional neurological disorder and healthcare professionals. Importantly, a significant minority strongly opposes deceptive placebo even when presented as a potentially curative intervention. There is widespread scepticism towards open-label placebo.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• When debating policies surrounding deceptive placebo treatments, given their associated risks, those in strong opposition deserve particular consideration.

• A stronger focus, both in terms of education and clinical practice, on ethically viable alternative ways of harnessing the power of placebo, such as open-label placebo, enhancing treatment effects through positive reinforcement and strengthening trust in doctors is less contentious, but currently viewed with scepticism by lay people and most health professionals.

Introduction

The placebo effect, the improvement of symptoms following administration of inactive substances or sham interventions, is an intriguing phenomenon. Placebos affect the neurotransmitters and central nervous system (CNS) regions modulated by equivalent active drugs or interventions, demonstrated among others in analgesia,¹,³ antidepressant⁴ and Parkinson’s disease placebos.^{5 6} Placebo effects are thus real and potent. Thousands of double-blind randomised controlled trials (RCT) placebo arms corroborate this.

Several mechanisms have been proposed: (a) prefrontal cortex involvement indicating a ‘top-down’ mechanism^{2 7 8}; (b) dopaminergic reward system involvement⁹; (c) conditioned responses to treatments^{10 11} and (d) Bayesian framework with sensory input and prior probability, for example, strong belief in symptom improvement, shaping final perception.

Given these mechanisms, deception is not essential for placebo responses. In open-label placebo, patients are explicitly told that the treatment is a placebo without active ingredient. Several studies suggest that open-label placebo is as effective as deceptive placebo.¹²,¹⁶ Apparently essential factors are convincing explanations of placebo’s power, coupled with positive suggestion. One could view open-label placebo as willingness to change one’s beliefs, helped by positive suggestion and conditioned responses to the ritual of taking medication.

Could neurological conditions be especially susceptible to placebo given both their origins lie in the CNS? Particularly in functional neurological disorder (FND), potential placebo responses and, more controversially, placebo treatments, remain a topic of discussion. Some diagnostic criteria for FND include placebo responses.¹⁷,²⁰ Numerous cases report symptom resolution through placebo responses.¹⁹ In clinical practice, FND symptoms sometimes resolve with unrelated drugs, treatments or non-physiological manoeuvres. FND and placebo/nocebo effects share common physiological bases, including prefrontal cortex involvement,²¹,²³ and Bayesian priors notably beliefs and expectations,²⁴ suggesting a link.^{25 26} Among other neurological conditions, placebo effects in Parkinson’s disease and migraine have also been highlighted, with particularly large placebo effects in double-blind RCTs.²⁷

Despite their obvious power, there is still no consensus about how placebos might be deployed in clinical practice and whether it may be unethical to withhold them. Previous surveys enquired about placebo treatments in general medicine, but not in neurology nor in FND.

Here, we report the results of a survey evaluating the opinion of (a) people with FND, (b) people with other, non-FND neurological conditions, (c) neurologists and (d) FND health professional experts, compared with people with non-neurological medical conditions, healthy individuals and other healthcare professionals. Additionally, we enquired about ethically acceptable alternative ways of harnessing the placebo effect, notably (a) open-label placebo, (b) positive suggestion (direct or indirect verbal or nonverbal messages that aim to influence the receiver in a positive manner) and (c) the influence of trust in one’s physician. Finally, we enquired about the current use of placebo in healthcare professionals’ practice.

Methods

The survey was conducted through an online questionnaire, with optional paper format. The questions included questions about the adequacy/inadequacy of deceptive and open-label placebo treatments in clinical practice, their perceived effectiveness and risks, and their use as diagnostic tools and in a clinical vignette. Additional questions enquired about the influence and components of trust in one’s doctor. To allow direct comparison, the healthcare professionals’ questionnaire was largely identical but additionally enquired about their use of placebos in clinical practice. The questionnaires are available in online supplemental figures 1 and 2. An introduction explained deceptive placebo. The question style was inspired by and in part adapted from previously published surveys.²⁸,³² Questions were answered on a Likert scale, with space for optional comments. We included the terms ‘organic’ and ‘psychogenic’ to ensure all healthcare professionals knew what was referred to, reflecting the diversity of current terminology in FND and related conditions.

Questionnaires were anonymous. Data were collected using REDCap^{33 34} and analysed with STATA (StataCorp. 2013. Stata Statistical Software: Release 13. TX: StataCorp LP).

Respondents: The survey was disseminated using the following organisations/databases: patient organisations: FND Hope, FND Action, Dystonia UK London Branch; research databases: National Institute of Health Research BioResource (neurology and FND patients), Queen Square Movement Disorders Centre Movement Disorders Research Registry, UCL Institute of Cognitive Neurosciences healthy subject database; professional organisations: Association of British Neurologists and trainees, UK Functional Neurological Society; other means: ClinicalTrials.gov (lay people only), first author’s research study participants, colleagues and acquaintances.

Exclusion criteria

Since misunderstanding the concept of placebo made all answers uninterpretable, the last question verified understanding: A placebo pill (select all that apply) (a) is an inactive treatment that never improves health issues, (b) is an inactive treatment that can improve health issues, (c) works because of the substance it contains, (d) works because patients believe it will improve their health issues (online supplemental figure 1). Anyone not answering this question or selecting an incorrect answer (a or c) was excluded, unless their additional comments demonstrated good understanding. Duplicate entries were excluded.

Statistical analysis

The main point of interest for each question was the overall attitude, whether respondents were overall in favour or against that specific matter. Thus, for the quantitative statistical analyses, the ‘mostly and strongly agree’ responses were combined into an ‘agree’ response, with the equivalent for disagree responses, unless otherwise stated.

For each question, Pearson’s χ² tests (Fisher’s exact test if non-normality) established whether there was a significant difference in the agree versus disagree responses between lay people and healthcare professionals.

For the subgroup analyses, Pearson’s χ² tests (Fisher’s exact test if non-normality) on the agree versus disagree responses were performed for each question for the following comparisons: FND patients versus non-FND patients; FND experts versus non-experts; patients with other neurological conditions versus patients with non-neurological medical conditions; patients with Parkinson’s disease, multiple sclerosis, epilepsy or headache disorders, respectively, versus patients with other neurological, non-FND conditions; healthy individuals vs patients; neurologists versus psychiatrists/psychologists; neurologists versus non-neurology doctors (excluding FND experts).

Cramer’s V estimated effect sizes. All tests’ significance level was 0.05, two-tailed.

The McNemar test (or its exact version) was performed for pairwise comparisons of the agree versus disagree responses to two questions.

Patient and public involvement

This survey was conducted following witnessed patient experiences with standard treatments and alternative methods, which according to current knowledge rely largely on placebo effects. Patient organisations disseminated the survey and aim to share the results. Patients and healthy controls participated.

Results

Respondents

Out of 757 respondents, 100 were excluded: 10 because they did not indicate their diagnosis/condition/healthy status and 90 because they misunderstood the concept of placebo (see above): 84 patients (35 FND, 43 other neurological, 6 medical condition), 2 healthy controls and 4 healthcare professionals (3 neurologists, 1 doctor in training). After exclusion, there were a total of 112 healthcare professionals and 545 non-healthcare professionals, henceforth referred to as lay people. The lay group comprised 141 FND patients, 289 other neurological patients (patients with neurological, non-FND conditions), 55 medical patients (non-neurological and non-FND) and 60 healthy controls. Table 1 summarises their characteristics.

Table 1. Survey respondents’ characteristics (after exclusion).

Lay people (N=545)				Healthcare professionals (N=112)
FND (N=141)	Other neurology (N=289)	Medical (N=55)	Healthy (N=60)	Doctor (N=91)	Other healthcare professionals (N=21)
Male: female: prefer not to say
35:102:4	125:161:3	20:34:1	22:38:0	54:36:1	5:16:0
Age (mean*)				Years of experience Mean (SD)
50.7	57.2	49.7	45.6	As specialist: 12.3y (11.4) As doctor in training: 5.6y (3.1y)
Main diagnoses				Specialty/profession
Functional movement disorders: 103 Functional seizure: 31 Other FND: 13	Parkinson: 79 Multiple sclerosis: 56 Migraine: 38 Non-migraine headache: 8 Epilepsy: 33 Action tremor: 24 Dystonia: 21 Other movement disorder: 13 Neuropathy: 21 Myasthenia gravis: 8 Other neuro Dx: 15			Neurology: 45 Psychiatry: 13 General medicine: 9 Surgery: 6 Other doctors: 6 In specialist training: 12	Psychologist: 8 Physiotherapist: 3 Other healthcare professionals: 10
				Place of work
				UK: 49 Europe (non-UK): 32 Other place of work: 10	UK: 16 Europe (non-UK): 2 Other place of work: 3
				FND expert
				33%	19%
				Seeing many patients with FND
				76%	76%

Other neurology=patients with neurological, non-FND conditions, Medical=patients with medical, non-neurological and non-FND conditions, Healthy=healthy controls. The main diagnoses are not mutually exclusive. Medical conditions were also present in 29% of patients with neurological and 28% of patients with FND. However, if they had either of the former, then they were included in the other neurology or FND group and not the medical group.

Other neuro Dx=other neurological diagnoses: 4 cerebellar ataxia, 4 progressive supranuclear palsy, 3 myopathy, 2 Huntington’s disease, 2 tics. Other doctors=2 paediatricians, 1 radiologist, 3 not specified. Other healthcare professionals: 2 nurses, 2 speech and language therapists, 1 medical student, 1 psychology student, 1 therapies assistant, 3 not specified. Other place of work: for doctors 2 North America, 2 Asia, 6 not specified; for other healthcare professionals: 1 Asia, 2 Oceania.

^*Age was given in age intervals, these mean values are thus an approximation.

Significant differences were found mainly between lay people and healthcare professionals and not between the subgroups. Thus, for easier oversight, only the lay people versus healthcare professionals’ responses are plotted alongside the questions—with subsequent clarification in the subgroup analyses if they differed significantly.

Deceptive placebo treatment

Lay people, particularly healthy individuals, were generally in favour of deceptive placebo use in clinical practice (71% agreed). Healthcare professionals were more conservative (46% agreeing) and only 38% considered its use ethically acceptable in organic disorders (figure 1). Repeating the first question at the end, after consideration of different aspects did not lead to significant change in any group (data available on request).

Figure 1. Overall attitude, perceived risks and effectiveness of deceptive placebo. The percentage of respondents mostly or strongly agreeing or disagreeing to each item are inserted in the stacked bar charts. The white dotted line indicates 50%. V=Cramer’s V providing the effect size. *The terms ‘organic’ and ‘psychogenic’ were included to ensure all respondents knew what was referred to, even though these are not terms that we personally use. FND, functional neurological disorder.

There was significant concern among lay people, and even more so healthcare professionals, that deceptive placebo use would induce loss of trust in doctors and medicine (lay: 57% agreed, professionals: 73%) (figure 1).

38% of non-FND patients believed that placebo treatments could improve their own symptoms. Healthcare professionals considered they could be effective in 31% of non-FND patients (figure 1). The corresponding answers for FND patients are discussed in the FND section below.

64% of healthcare professionals never used deceptive placebo, 5% used it once, 20% between 2–10 times. The most common application was as a diagnostic tool to distinguish FND from non-FND. Online supplemental table 1 summarises healthcare professionals’ estimated use of deceptive placebo in their clinical practice, the circumstances in which it was applied and the provided explanations.

Open-label placebo treatment

None of the groups thought that open-label placebo treatments were as effective as deceptive placebo treatments. Lay people preferred deceptive over open-label placebo treatments. Healthcare professionals were more cautious (figure 2).

Figure 2. Attitude to open-label versus deceptive placebo. The percentage of respondents mostly or strongly agreeing or disagreeing to each item are inserted in the stacked bar charts. The white dotted line indicates 50%. V=Cramer’s V providing the effect size.

Lay people were asked at what point they would want to be told if given a placebo treatment. While 17% would never want any type of placebo, 59% chose a deceptive and 19% an open-label version (table 2).

Table 2. Lay people’s preferred type of placebo use.

(4e) Imagine your doctor giving you a placebo treatment for a medical problem. When would you want your doctor to tell you that the treatment is a placebo? (select all that apply)
If it has worked and I am no longer taking it	36%
If it hasn’t worked	30%
If it is working and I am still taking it	25%
Before starting (open-label placebo)	19%
Irrelevant—I would never want my doctor to give me a placebo, neither open-label nor deceptive	17%
Never	12%
Other	2%
No options chosen	2%

545 lay respondents. All values are given as percentages of respondents. Three respondents who selected ‘other’, suggested patients consenting in general to being given placebo treatments, without the doctor subsequently specifying when using it.

One question enquired about a potential middle way between open-label and deceptive placebo. Two-thirds of lay people and healthcare professionals considered placebo use acceptable when accompanied by a vague statement such as: “This treatment sometimes works really well against the type of problems you have, so let’s give it a try” (figure 2). Lay people were similarly in favour of this type of placebo as they were of deceptive placebo (question 1, figure 1). Significantly more healthcare professionals, on the other hand, agreed to this type of placebo use compared with deceptive placebo use (question 1, figure 1) (McNemar test: lay: χ²(1)=4.64, p=0.03, professional: χ²(1)=12.6, p=0.004). Such vague explanations were also the healthcare professionals’ preferred option when having used deceptive placebo treatments or if they were to ever do so (online supplemental table 1).

Online supplemental table 1 summarises healthcare professionals’ estimated use of open-label placebo in their clinical practice and the circumstances in which it was applied. Two-thirds of healthcare professionals never used open-label placebo.

The influence of trust in one’s doctor

All groups believed that trust in one’s doctor had a beneficial effect on treatment outcome, yet a trusted doctor’s positive suggestion was not believed to be as effective as placebo (figure 3).

Figure 3. The influence of trust in one’s doctor on placebo effects. The percentage of respondents mostly or strongly agreeing or disagreeing to each item are inserted in the stacked bar charts. The white dotted line indicates 50%. V=Cramer’s V. Several respondents added an optional comment to question 8, stating that such a suggestion would give the impression of dismissing the symptoms as fake/imaginary or insignificant.

Nevertheless, significantly more respondents agreed that a trusted doctor’s positive suggestion was as effective as deceptive placebo, than open-label placebo being as effective as deceptive placebo (McNemar test: lay: χ²(1)=37.0, p<0.0001, professional: χ²(1)=6.7 p=0.009) (question 4d figure 2 and question 8 figure 3).

Online supplemental table 2 summarises the most important factors for this survey’s respondents to fully trust a doctor, with the top four being: (1) explaining the diagnosis and treatment, (2) listening to the patient, (3) viewing the patient as a whole person, not an isolated medical problem and (4) giving sufficient time. Note that these were considered far more important than, for example, performing many investigations.

Functional neurological disorder

Several aspects were enquired about: (a) placebo treatments in FND, (b) placebo as a diagnostic tool for FND, (c) perceived effectiveness of placebo treatments in FND, (d) attitudes of people with FND throughout this survey and (e) attitudes of FND experts throughout this survey.

Placebo treatments in FND

Lay people were significantly more in favour than healthcare professionals of using deceptive placebo treatments in FND in general and in a functional paraplegia clinical vignette. Among lay people, healthy individuals showed most approval, FND patients least approval, with a relatively high number of FND patients strongly disagreeing (22% in general and 28% for the clinical vignette). Healthcare professionals reflected the rather conservative opinions of people with FND (figure 4). There was more scepticism among lay people towards open-label placebo in the same situations, whereas professionals favoured open-label equally to deceptive placebo (McNemar test deceptive vs open-label: general question: lay: χ²(1)=10.8, p=0.001, professional: χ²(1)=1.1 p=0.29; clinical vignette: lay: χ²(1)=22.5, p<0.0001, professional: χ²(1)=0.02 p=0.88) (figure 4).

Figure 4. Attitude towards deceptive and open-label placebo treatments in FND. The percentage of respondents mostly or strongly agreeing or disagreeing to each item are inserted in the stacked bar charts. The white dotted line indicates 50%. V=Cramer’s V. FND = functional neurological disorder.

Deceptive placebo treatment was considered ethically acceptable if there were no better treatment options in functional disorders by 55% healthcare professionals, yet by only 38% in non-functional disorders, reflecting a large and significant contrast (Pearson’s χ² test: χ²(1)=45.2, p<0.001, V=0.64) (figure 1).

In the rare instances when deceptive or open-label placebo was used in clinical practice, this was primarily in FND, for treatment or diagnosis (online supplemental table 1).

Placebo as a diagnostic tool for FND

There was overall agreement to deceptive placebo use as a diagnostic tool in FND (figure 4).

Perceived effectiveness of placebo treatments in FND

36% of people with FND thought that placebo treatments could improve their own symptoms, which is not significantly different from non-FND patients (38%) (Pearson’s χ²: χ²(1)=0.5, p=0.48, V=−0.04). However, 27% of people with FND strongly disagreed (compared with 19% non-FND patients, Pearson’s χ²: χ²(1)=3.6, p=0.059, V=−0.09) (figure 1).

Healthcare professionals considered placebo would be effective in 55% of patients with functional disorders, significantly more than their estimate of 31% in non-FND disorders (Wilcoxon matched-pairs signed-rank test: Z=−8.49, p<0.0001, r=0.80) (figure 1).

Attitudes of people with FND

As detailed above and in figure 4, people with FND had more conservative views than other lay people on deceptive placebo treatments specifically for FND. Throughout the rest of this survey, FND patients were slightly more conservative than non-FND patients with regards to deceptive placebo treatments, even when they were overall in favour or evenly split.

1. The general question about placebo use failed to reach statistical significance: 62% of FND patients were generally in favour of doctors using deceptive placebo treatments, compared with 72% of non-FND patients (question 1: Pearson’s χ²: χ²(1)=2.6, p=0.10, V=−0.08), with 16% being strongly opposed (compared with 12% of non-FND patients) (Pearson’s χ²: χ²(1)=1.9, p=0.16, V=−0.06).

2. A higher proportion felt that patients would lose trust if deceptive placebos were used (question 3a: 66% of FND patients agreed, vs 53% of non-FND patients, Pearson’s χ²: χ²(1)=4.9, p=0.028, V=0.10).

3. A higher proportion felt that patients should be told very clearly from the beginning if a treatment is a placebo (question 4a: 46% of FND, vs 35% of non-FND patients, Pearson’s χ²: χ²(1)=4.9, p=0.026, V=0.10).

The other questions in this survey did not significantly differ between people with FND and patients without FND. Among others, they demonstrated similarly sceptical attitudes towards open-label placebo treatments.

FND experts’ attitudes

Throughout this survey, only one question was answered significantly differently between health professional FND experts and non-experts. Fewer FND experts considered deceptive placebo treatments ethically acceptable in purely functional disorders (38%) than non-experts (63%) (χ² test: χ2(1)=5.2, p=0.02, V=-0.22) (figure 1). Hence, FND experts did not consider it ethically more acceptable in functional compared with non-functional disorders (38% in both).

Subgroup analyses

For the following additional group comparisons, all survey questions were analysed except for the questions allowing multiple answers (questions 4e, 9 and 13 supplementary questions—online supplemental figures 1 and 2). Data and statistical analyses are available on request.

Other neurology patients versus non-neurology medical patients

Patients with other neurological (non-FND) conditions did not significantly differ in any answer compared with patients with non-neurological medical conditions.

Parkinson’s disease, headache disorders, epilepsy or multiple sclerosis versus other neurological conditions

Patients with none of the above neurological conditions significantly differed in their opinions compared with patients with other neurological (non-FND) conditions. The only exception was that even more people with Parkinson’s disease (70%) than other neurological patients (56%) thought that if placebos were to be used, doctors should not tell the patient that it is a placebo (question 4b: χ² test χ2(1)=4.5, p=0.034, V=−0.13).

Healthy individuals versus patients

Healthy individuals generally favoured deceptive placebo treatments significantly more than patients: (87% thought they should be used by doctors in general, compared with 69% of patients (question 1: χ² test χ2(1)=5.9, p=0.015, V=0.11); and 78% thought they should be used in a clinical vignette of a functional paraplegia, compared with 60% of patients (question 10a: χ² test χ2(1)=5.7, p=0.017, V=0.11).

More healthy individuals (67%) than patients (38%) thought that a placebo could improve their own symptoms (question 2: χ² test χ2(1)=11.8, p=0.001, V=0.16) (figure 1).

Answers to all other questions did not significantly differ between healthy individuals and patients.

Neurologists versus other healthcare professionals

Neurologists versus psychiatrists/psychologists

Only one question was answered significantly differently: No psychiatrist/psychologist, but 31% neurologists mostly agreed that open-label placebo treatments were as effective as deceptive placebo treatments (question 4d: Fisher’s exact test: p=0.002).

Neurologists versus non-neurology doctors (excluding FND experts)

In order for FND expertise not to confound the picture, the 25 non-FND-expert neurologists were compared with the 36 non-FND-expert non-neurology doctors.

Neurologists significantly differed from non-neurology doctors in only three questions: Significantly more neurologists (36%) than non-neurology doctors (14%) thought that open-label was as effective as deceptive placebo (question 4d: Pearson’s χ²: χ²(1)=5.4, p=0.02, V=0.31).

Significantly fewer neurologists (44%) thought it ethically acceptable to use deceptive placebo treatments in functional disorders than non-neurology doctors: (72%) (question 14b Pearson’s χ²: χ²(1)=4.9, p=0.03, V=−0.28); or that doctors should give deceptive placebo in the functional paraplegia clinical vignette (Neurologists: 36%, non-neurology doctors: 64%) (question 10a: Pearson’s χ²: χ2(1)=5.2, p=0.02, V=−0.29).

Comments

Respondents could add optional comments to all questions and at the end of the survey. Figure 5 provides examples that reflect recurring comments surrounding the main issues.

Figure 5. Selection of lay peoples’ comments. FND = functional neurological disorder.

Discussion

This is the first patient and healthcare professional survey on attitudes to and clinical use of deceptive and open-label placebo treatments in neurological practice, including in FND.

The results show a mixed picture. Deceptive placebo was deemed rather effective by all groups, but risky, primarily through the undermining of trust. Most lay people (non-professionals) were overall in favour of its use in clinical practice, while healthcare professionals tended to be rather opposed, or evenly split and hence more conservative than lay people.

Neurologists differed from non-neurological doctors by being more conservative towards deceptive placebo treatments in FND and more positive about the potential effectiveness of open-label placebo.

Other (non-FND) neurology patients did not differ in their answers compared with non-neurology patients. Furthermore, neither patients with Parkinson’s disease, headache disorders, epilepsy nor multiple sclerosis significantly differed from patients with other neurological (non-FND) conditions.

People with FND were generally rather in favour of deceptive placebo use, yet evenly split and thus relatively less in favour when explicitly asked about deceptive placebo treatment for FND. They were overall more conservative than the other lay groups, with a higher proportion being strongly opposed, even when presented with the idea that a deceptive placebo treatment could lead to full recovery. They shared the other groups’ scepticism towards open-label placebo.

While people with FND estimated placebo’s effectiveness on themselves similarly to non-FND patients, professionals considered them more effective in functional disorders. The professionals’ opinion reflects a meta-analysis, primarily involving hypnosis and functional seizure induction through suggestion, which identified increased responsiveness to verbal suggestion in FND, with higher symptom-specific than general suggestibility.³⁵ Another study found that people with FND were no more susceptible to placebo effects per se. Nevertheless, FND is not caused by structural or molecular damage that imposes unsurpassable limits to the degree of possible recovery. This changeability of FND symptoms means that if a placebo effect is to occur, it can be of a much larger magnitude than in non-functional disorders.³⁶

More professionals, but not FND experts, considered deceptive placebo treatments in the absence of better treatment options ethically acceptable for functional compared with non-functional disorders. In the rare instances in which placebo treatments had been used by healthcare professionals, this was primarily for FND, either for diagnosis or treatment. Our data show some strong negative opinions regarding deceptive placebos, particularly among people with FND (figure 5). Thus, placebo treatments were thought to be most effective, most ethically acceptable and in practice most commonly used for the group of people most strongly opposed to it. This should alert any clinician and at the very least calls for great caution in the use of deceptive placebo in clinical practice, particularly among a group of patients who are often already subject to stigma and misunderstanding. As one patient commented: “Patients with functional disorders have usually been treated poorly or dismissed by multiple doctors. They need to be able to trust medicine again. Providing any type of treatment without full explanations is wrong and may cause more distrust and anxiety about doctors.”

Could the right wording legitimise placebo use? There was overall support for the use of placebo, accompanied by vague, semiopen, semideceptive explanations, such as: “We don’t quite know how it works/this type of treatment sometimes works really well in your condition.” Arguably, such statements are true. Yet it can be argued that this remains a deceptive placebo, given the implicit understanding between physicians and patients that tablet/injection/infusion ‘treatments’ involve pharmacological substances. Furthermore, 19% lay people and 10% healthcare professionals strongly opposed such a use.

As in many controversies, different interpretations are possible. Is overall agreement or strong opposition most important? Because of deceptive placebos’ risks, we believe those in strong opposition deserve particular consideration (figure 5). A patient feeling deceived and losing trust in the medical profession will have devastating long-term effects. After an initial honeymoon period, placebo use might gradually erode trust, leading to patients doubting all treatments, thereby not only diminishing placebo effects but also inducing lessebo effects—decreased efficacy of active treatments because patients think they might be placebos.^{35 36}

All groups in our survey were sceptical about open-label placebo treatments. Across the whole lay group, if they were to receive placebo treatments, only 19% opted for open-label administration and 59% preferred deceptive forms. This scepticism, despite the experimental evidence for the effectiveness of open-label placebo, suggests the need for education.

It is worthwhile considering what generates placebo effects. Core, intertwined elements are belief; trust; previous experience; learnt associations and reassurance decreasing anxiety. ‘Indeed, a placebo is the whole ritual of the therapeutic act’.³⁷ It may be more beneficial, ethical and ultimately cost-effective to boost existing treatments through an additional placebo effect by focusing on establishing a trusting doctor–patient relationship.

Throughout most of its history, medicine was based on natural recovery and placebo effects.³⁸ As reflected in Voltaire’s quote:

The art of medicine consists in distracting the patient while nature cures him.

Yet centuries later, placebo effects are still integral to medicine: Postoperative pain relief is significantly stronger with open compared with hidden morphine administration; and pain reappearance is more severe with open compared with hidden discontinuation.³⁹ An ethically acceptable way physicians can harness the power of placebo is by actively sharing their knowledge and honest belief about treatments’ effectiveness. This can boost effectiveness through additional placebo effects without involving any deception.

Our study has several limitations. The data indicate opinions only. The low response rate may well reflect response bias, for example, selecting out those who strongly oppose or support use of placebo. Explanations, phrasing and questions can also bias responses, despite our efforts to avoid such bias. Misunderstanding placebo treatments was mitigated by an initial definition and by later double-checking understanding, leading to 90 respondents being excluded. The subgroup analyses must be interpreted with caution, as they included many comparisons and relatively small numbers. People’s self-declared diagnoses could not be verified.

In conclusion, deceptive placebo treatments are considered effective by many lay people and health professionals, but with a sizeable proportion strongly opposing their use, particularly people with FND. Ethically acceptable alternative ways of harnessing placebo were viewed with scepticism, but with further research combined with education, may be a way for people with neurological conditions to benefit from the placebo mechanism without the personal and ethical perils of deception.

The funders had no involvement in any part of the study.

Data availability statement

Data are available on reasonable request.