Negative lay perceptions of open-label placebo users may pose barriers to intervention adoption

Abstract

Background

Recent research has revealed that the use of specific medical interventions carries with it social stigma. This “intervention stigma” can pose an obstacle to the use and adoption of interventions that may otherwise be effective in managing medical conditions. Open-label placebos (OLPs) have been identified as a potential intervention for a variety of clinical and nonclinical conditions but are viewed with skepticism among lay populations.

Purpose

This online experimental study aimed to quantify intervention stigma associated with the use of OLP interventions for a medical condition within a warmth-competence framework of social perception.

Methods

In an online experiment fielded in the USA (N = 541), we randomly assigned participants to read 1 of 4 vignettes about a patient who is administered an OLP intervention by a physician for chronic back pain. In each vignette, the patient’s belief in and response to the treatment varied. After reading the vignette, participants rated the patient on several characteristics that captured perceptions of warmth and competence.

Results

We found that patients who believed in the OLP intervention or reported improvement after taking it were perceived as less competent and warmer.

Conclusions

Our results suggest that the use of OLP interventions for medical conditions carries intervention stigma. We contend that this stigma poses an obstacle to the adoption of OLP interventions.

Introduction

Researchers have figured out an ethical way to harness the beneficial effects of placebos without deception.¹ By educating people about placebo effects, explaining the science behind them, and highlighting that they can work even without deception, researchers can induce beneficial effects in various clinical and nonclinical contexts.² As research on the effectiveness of placebos without deception—also known as open-label placebos (OLPs)—accumulates, they may soon be considered a viable intervention for a wide range of disorders and conditions.³ However, perceptions of OLP interventions and people who use them may pose a significant barrier to their adoption. In this research, we extend qualitative work on negative perceptions of placebo responders.⁴ Using an experimental approach, we identify how laypersons perceive individuals who believe in and respond to OLPs using a warmth-competence framework of social perception.

OLPs can be effective interventions

Although interest in OLPs is not new, Kaptchuk and colleagues published a seminal paper showing that OLPs can help reduce irritable bowel syndrome symptoms and improve quality of life, spurring a renewed interest in the topic.^5,6 Since then, there has been an accumulation of research demonstrating the beneficial effects of OLPs for a wide range of clinical conditions, such as irritable bowel syndrome,^5,7,8 chronic back pain,^9–11 and migraines.¹² A recent meta-analysis estimated a standardized mean difference of .72 between OLP treatment and no-treatment controls.² In addition, there is evidence that OLPs can be effective for conditions such as acute and chronic emotional distress.^13–15

Lay perceptions of OLP interventions

As evidence supporting the effectiveness of OLPs grows, laypersons’ perspectives on these interventions have become an important consideration. Although sparse, the available evidence paints a mixed picture. On one hand, there is skepticism about the effectiveness of OLPs, with lay conceptualizations of placebos usually including deception as a required element.¹⁶ Consistent with these lay conceptualizations, research directly comparing perceptions of deceptive placebo and OLP treatments finds that OLPs are perceived to be less credible.¹⁷ On the other hand, transparency is viewed favorably in placebo interventions, with research finding that deception is often considered problematic for personal autonomy and the patient–provider relationship.⁴ Nevertheless, research has found that in the context of insomnia and abdominal maladies, participants believe that it is permissible for a physician to prescribe both OLPs and deceptive placebos, with mixed evidence for which is viewed as more permissible.^18,19

Lay perceptions of patients using OLPs

In addition to perceptions of OLP interventions, it is important to investigate the perceptions of individuals who use them,²⁰ particularly because qualitative evidence suggests that patients who respond to placebos are perceived as gullible, foolish, and childish.⁴ Although several factors impact a patient’s choice to seek a specific treatment, social stigmatization is one factor that may hinder the adoption of OLPs.²¹

A stigmatized person possesses a devalued identity because of perceived negative attributes.²² The concept of stigma has been extended to the social perceptions associated with using specific treatments and termed “intervention stigma.”²³ For example, the stigma associated with using palliative care among people who are living with a serious illness can lead to patients being characterized as “giving up,” which may make it less likely that they elect to receive palliative care.²⁴ This kind of intervention-specific stigma, which can result in treatment avoidance, has been identified in a range of other contexts, such as antipsychotic medications²⁵ and methadone as a medication for opioid use disorder.²⁶ Accordingly, it is important to understand whether there are stigmas associated with OLP interventions.

Warmth and competence as basic dimensions of social perception

To understand the perceptions of those who use OLPs, we draw on the stereotype content model, which suggests that warmth and competence are fundamental dimensions of social perception, leading to distinct downstream patterns of cognition, affect, and behavior.²⁷ Warmth comprises traits such as friendliness and helpfulness, while competence consists of traits such as intelligence and efficacy.²⁸ Research has identified that the affective correlates of different warmth-competence perceptions vary.²⁸ For example, people seen as high in competence and warmth elicit feelings of admiration,²⁹ and people who are low in both warmth and competence elicit feelings of contempt.³⁰ Ambivalent perceptions are characterized by a low-high or high-low configuration of warmth-competence perception.²⁸ People seen as having low competence but high warmth—such as elderly and disabled people—elicit feelings of superiority or pity.²⁸ On the other hand, people seen as highly competent but low in warmth elicit feelings of envy and jealousy, resulting in dislike and hostility.²⁸

The current research

We sought to better understand the risk that intervention stigma might pose to the use of OLP interventions for the treatment of medical conditions. In an online experiment, we investigated how belief in and responsiveness to an OLP intervention impacted perceptions of a patient suffering from chronic back pain. We first tested our hypothesis that participants would perceive patients who believed in or responded to an OLP intervention as less competent. Then, we conducted exploratory analyses on perceptions of warmth, following the same analytic plan.

Methods

Participants

On August 13th, 2019, 616 adult participants from the United States were recruited to complete an online study on “Perception of Patients and Treatment Plans for Medical Conditions” in exchange for $1.25 using the online recruitment platform Prolific. Of these 616 participants, 75 were excluded for failing 1 of the 2 attention checks (N = 69), reporting that they were either under the age of 18 or not specifying their age (N = 3), and/or advancing past the vignette portion of the study in fewer than 20 s (N = 3), resulting in a final sample of 541 participants (see Table 1 for sample characteristics).

The demographic composition of the participants in each of our 4 conditions (the details of which are explained in the Procedures section) did not vary in terms of age (P = .80), race (P = .86), or highest level of education completed (P = .23). However, there was a failure in random assignment such that there were significantly different numbers of men and women across our conditions (P = .02). Analyses reported in the ESM demonstrate that controlling for these gender differences does not substantively alter our findings.

Procedure

Participants were randomly assigned to read 1 of 4 vignettes about a patient who receives an OLP intervention for their severe and persistent back pain (full vignettes are available in the ESM). Participants read that the patient is enrolled in a medical study and is randomly assigned to 1 of 3 treatment groups. In all vignettes, participants learn that the selected intervention is a placebo pill containing no active ingredients, taken every day for 2 weeks. The physician in the vignette provides the patient with more information on the placebo effect and emphasizes that the OLP intervention may help reduce their back pain because a key ingredient is the belief that it can work.

At this point, the information participants received differed across 2 key dimensions: (1) Whether the patient believes (vs does not believe) that the OLP intervention will be effective, and (2) whether the patient reports improvement (vs no improvement) at a 2-week follow-up. These 2 dimensions resulted in 4 conditions. It is explicitly stated in every vignette that the patient takes the placebo pills as directed by the experimenter.

After reading the vignette, participants answered a series of Likert-scale questions probing their social perceptions of the patient. Finally, participants completed demographic questions about themselves and were compensated.

Measures

Patient impressions

Participants provided their perceptions of the patient on 12 characteristics: intelligent, open, optimistic, agreeable, logical, religious or spiritual, scientifically literate, knowledgeable, gullible, educated, warm or friendly, and competent. Correlations between these items can be found in Table 2. Except for “gullible” and “open,” these perceptions were measured using 6-point Likert-scale items (1 = Not at all [characteristic], 6 = Very [characteristic]) with the following structure: “How [characteristic] do you think the participant is?”

Gullible

Participants were asked, “How easy do you think it is to persuade the participant to believe something?” (1 = Not at all easy, 6 = Very easy).

Openness

Participants were asked, “How open do you think the participant is to new perspectives and experiences?” (1 = Not at all open, 6 = Very open).

Overview of data analytic approach

First, we conducted exploratory factor analysis to condense our 12 perceptual characteristics into their underlying factors. Then, we tested for differences in social perceptions among our 4 conditions. We explain these steps in more detail below.

Exploratory factor analysis

To better understand the underlying structure of the social perceptions we assessed, we entered the 12 characteristics into an exploratory factor analysis. To determine the number of factors in our solution, we used both the Kaiser rule (number of factors with eigenvalue > 1)³¹ and scree test.³²

Following guidelines from Tabachnick and Fiddel,³³ we used 0.32 as a threshold for acceptable factor loadings, cross-loading, and as an upper bound for factor correlations when employing an orthogonal factor solution. Because Tabachnick and Fiddel³³ do not offer explicit guidelines for item communalities, we followed guidelines from Costello and Osborne³⁴ which suggest that 0.40 is a lower bound for acceptable communality.

Condition differences in social perceptions

Using the factors revealed in our exploratory factor analysis, we performed 2-way ANOVAs to examine the impact of a patient expressing belief (vs disbelief) in the OLP intervention and reporting a positive response (vs no response) to the OLP intervention on participants’ impressions of them. We began by fitting a model with the interaction between our 2 factors and reduced the model to only main effects if the interaction was not significant at the P < .05 level.³⁵

Results

Warmth and competence emerge as core dimensions of perception

Both the Kaiser rule (number of factors with eigenvalue > 1) and the scree test³² (see Figure S1 in the ESM) suggested a 2-factor solution for our 12 perceptual characteristics. We began by fitting a 2-factor solution with an oblimin rotation. After confirming that the correlation between our factors was less than 0.32 (r = −0.11), we proceeded with a varimax rotation. At this point, “gullible” was removed for high cross-loading (−0.40), and although the cross-loading of “religious or spiritual” was slightly below our cut-point of 0.32, its communality (0.30) was well below conventional cutoffs and was therefore removed. With item selection complete, we confirmed that our factors were still uncorrelated (r = −0.04) and retained loadings from the varimax rotation (see Table 3 for varimax rotation loadings; see Table S1 in the ESM for oblimin rotation loadings). This 2-factor varimax solution explained 68% of the total variance in the 10 retained items. We interpreted Factor 1 as capturing competence and Factor 2 as capturing warmth, with both sets of items displaying high reliability (αs > .85). Analyses were conducted on these EFA-derived warmth and competence scores, which were computed by averaging the set of corresponding items for each factor.

OLPs and patient perceptions: belief in intervention and positive response to intervention

Perceptions of competence

There was a main effect of both belief in the OLP intervention and responding to it on competence (see Figure 1). Participants perceived patients who believed the OLP intervention would work as less competent than patients who did not, F(1, 537) = 39.74, P < .001, d = −0.64. Although the effect was smaller, patients who reported a positive response to the OLP intervention were perceived as less competent than patients who did not respond, F(1, 537) = 11.04, P < .001, d = −0.28. The interaction between these 2 factors was not significant (P =.14).

Perceptions of warmth

Similarly, there was a main effect of belief in the intervention and positive response to the intervention on perceived warmth (see Figure 2). Participants perceived patients who believed the OLP intervention would work as warmer than patients who did not, F(1, 537) = 253.98, P < .001, d = 1.85. Once again, the effect of reporting a positive response was smaller, but in the same direction, F(1, 537) = 30.55, P < .001, d = 0.57. The interaction between factors was not significant (P = .24).

Discussion

A burgeoning literature has identified the promise of OLPs for managing various conditions.² However, research on intervention stigma and preliminary evidence on perceptions of placebo responders suggests that stigma associated with the use of OLPs may pose a significant obstacle to their widespread acceptance, adoption, and use—in spite of their effectiveness.^4,23 In this online experimental study, we demonstrated that along two dimensions of universal social perception—warmth and competence—there is intervention stigma associated with the use of OLP interventions.²⁸

Individuals who believe in or respond to OLPs are stigmatized

We identified that both belief in and responsiveness to OLP interventions impact social perceptions of patients’ warmth and competence. Specifically, we found that patients who believed in (or responded to) the OLP intervention were perceived to be lower competence and higher warmth than patients who did not believe in (or respond to) the OLP intervention. Although less extreme, the social perceptions we identified as associated with OLP use are structurally similar to perceptions of groups like the disabled, those with acquired brain injury, and the elderly.^28,36 Importantly, the effects of belief in and response to OLPs were independent which indicates that individuals may be stigmatized for responding to an OLP intervention, even if they express skepticism about it.

These findings are notable within the stereotype content model because this combination of contrasting stereotypes results in “ambivalent prejudice.”²⁷ On one hand, higher perceived warmth in those who believe in and respond to OLPs may result in positivity toward these individuals. However, given the concurrent perception of lower competence, perceivers are likely to only provide paternalistic positivity, viewing them as subordinate in status—even warranting pity.^28,37

OLP intervention stigma poses barriers to widespread adoption

Empirical research examining the behavioral consequents of concurrent low competence and high warmth social perception has found that it is likely to elicit both helping behaviors (active facilitation) and exclusionary or neglecting behaviors (passive harm).³⁷ Following research on other medical interventions that carry stigma, we argue that the negative perceptions associated with OLPs and their behavioral consequents pose several potential obstacles to their widespread adoption, even if there is evidence that OLPs may be an appropriate intervention for certain conditions.^24,26

First, OLP intervention stigma may dissuade individuals from directly expressing interest in OLP interventions to their medical care team or their peers for fear of incurring social cost. Second, given that we identified intervention stigma even in the absence of belief in the intervention, patients who have benefitted from an OLP intervention (skeptical or not) may be less likely to volunteer that information with others. Third, we speculate that the content of OLP intervention stigma lends itself to the undermining of private beliefs about OLPs. For example, we imagine a patient asking themselves: “The people who believe in OLPs are stupid and helpless. What would it say if they worked on me?” This possibility is particularly germane in the context of OLPs because patients’ belief that the intervention will work has been proposed—not without debate—as an important mechanism of action through which OLPs may work.^{1,5,13,38–40} Mediation models reported in the ESM (see Figures S2 and S3) do, indeed, find that the reported condition differences in perceived warmth and competence significantly predicted lower forecasts of treatment efficacy when participants were asked to imagine they were the patient. We note, however, that we were unable to test the influence of plausible alternative explanations (eg, the presence of a confounding third variable such as negative a priori beliefs about OLPs); therefore, these indirect effects are preliminary and should be viewed with caution.

Finally, although our study focused on perceptions of patients, there is reason to believe that intervention stigma can also impact medical providers who choose to prescribe certain treatments.²³ In this case, we suggest that medical providers may choose not to prescribe OLP interventions due to the risk it poses to their perceived competence, which has been shown to impact patient satisfaction and compliance.⁴¹

Limitations

As with any research, there are limitations of this work that should be acknowledged.

Vignette paradigm

Given our use of 4 short vignettes to generate complex social perceptions, there are inevitable bounds on the generalizability of our findings, which we discuss below.

First, participants were offered only cursory information—much less than a patient would receive—about the OLP intervention and the science underpinning it. Although this was an intentional choice, designed to emulate the information-constrained contexts in which social perceptions are often generated, it does limit the generalizability of our findings. For example, if an individual was to provide a detailed rationale for the OLP intervention they were prescribed to their colleague, the colleague’s social perceptions may be different than if the individual just mentioned the treatment in passing. In particular, given their wide use in research on OLPs, it would be valuable to assess how social perceptions around OLPs are altered if the perceiver is also informed with the 4 widely adopted discussion points put forth by Kaptchuk et al.⁵: “(1) the placebo effect is powerful, (2) the body can automatically respond to taking placebo pills like Pavlov’s dogs who salivated when they heard a bell, (3) a positive attitude helps but is not necessary, and (4) taking the pills faithfully is critical.” With that said, we argue that the second scenario—where a treatment is mentioned briefly, in passing, or even second-hand—is an important and common social context where social perceptions are generated.

Second, our choice to use chronic back pain as the ailment in our vignettes may have impacted participants’ social perceptions. It seems likely that OLPs are viewed as more reasonable interventions for some conditions (eg, abdominal pain) and less reasonable for others (eg, terminal cancer). Interestingly, existing work has found that OLPs are generally viewed as acceptable for a range of conditions, including chronic stress, insomnia, and abdominal maladies.^19,42,43 However, future work examining how social perceptions may vary among an even broader range of conditions and treatment contexts would be valuable.

Third, the patients in our vignettes expressed only very simple unidimensional beliefs about the OLP intervention (eg, “the participant believes the physician”). Recent qualitative work suggests that patients taking OLP treatments often have more complex ambivalent feelings about their treatment.¹⁸ We believe that our study’s focus on 2 poles of belief is an informative starting place, but that follow-up studies should examine how more nuanced patient responses to OLP interventions affect social perceptions.

Fourth, we did not collect perceptions of patients who were prescribed closed label placebos for their back pain. As a result, we cannot infer that the patterns we have identified are unique to OLPs (as opposed to placebos more broadly). In fact, the existing qualitative evidence would lead us to believe that closed label placebo responders would likely be perceived negatively as well.⁴ With that said, we believe that our findings about OLPs are uniquely important because of the potential for widespread adoption OLPs have. In the case of closed label placebos, we believe it is likely that the ethical issues associated with their administration poses a larger barrier to adoption than intervention stigma.

Outcome measurement

As described in the Methods section, we assessed perceptions of the patient in the vignette by asking participants to rate them on 12 characteristics. These characteristics cover considerable breadth because our initial intention was to analyze perceptions at the item-level. However, preliminary EFA analysis suggested that warmth and competence emerged as the underlying conceptual dimensions of social perception in our study. In retrospect this was to be expected, considering the large body of work suggesting these 2 factors represent basic dimensions of social perception.^27,28 As a result, the collection of items we used have not been validated for the assessment of warmth and competence. Yet, we do take comfort in the fact that several of the items retained in our factor analysis (“warm or friendly,” “intelligent,” and “competent”) are among those recommended for use by Fiske²⁷ and that those single items correlate highly with the corresponding EFA-derived scale scores we used in our analyses (“warm or friendly”: r = .75, P < .001; “intelligent”: r = .88, P < .001; “competent”: r = .84, P < .001).

Additionally, our data focuses solely on perceptions of patients, but not the clinicians administering the OLP intervention. This omission is notable because perceptions of clinicians have important implications for the acceptance of OLP interventions in the context of daily clinical practice.^26,44 Future research should address this gap.

Conclusions

Our work demonstrates that OLPs carry intervention stigma along fundamental dimensions of social perception. These perceptions and their consequences may pose barriers to the adoption of OLP interventions despite their empirically demonstrated effectiveness.^2,23 Once future research identifies how these perceptions vary in different contexts, it will become critical to better understand how malleable these perceptions are and what approaches are effective in improving them.

Acknowledgments

We would like to extend our gratitude to Amanda Bogen for her help with this project.

Author contributions

Tyrone J. Sgambati (Methodology, Formal Analysis, Writing—Original Draft, Writing—Review & Editing, Visualization), Luana Colloca (Methodology, Writing—Review & Editing), Andrew L. Geers (Methodology, Writing—Review & Editing), and Darwin A. Guevarra (Conceptualization, Methodology, Formal Analysis, Investigation, Writing—Original Draft, Writing—Review & Editing, Visualization, Supervision)

Transparency Statements

1.Study registration. This study was not formally registered.

2.Analytic plan pre-registration. The analysis plan was not formally pre-registered.

3.Analytic code availability. Analytic code used to conduct the analyses presented in this study are available in a public archive: https://osf.io/phznm/?view_only=12e548eba3ba4f6fb5889f30669da54a

4.Materials availability. All materials used to conduct this study are available in a public archive: https://osf.io/phznm/?view_only=12e548eba3ba4f6fb5889f30669da54a

Funding

None declared.

Conflicts of interest

None declared.

Data availability

De-identified data from this study are available in a public archive: https://osf.io/phznm/?view_only=12e548eba3ba4f6fb5889f30669da54a

References