Very superstitious,
Writing’s on the wall,
Very superstitious,
Ladders ’bout to fall,
Thirteen month old baby,
Broke the lookin’ glass
Seven years of bad luck,
The good things in your past
When you believe in things
That you don’t understand,
Then you suffer,
Superstition ain’t the way
—Stevie Wonder, “Superstition,” 1972
Superstition, in particular negative superstition, is the most common form of nocebo in everyday culture. From breaking a mirror, to walking under a ladder or stepping on a crack, superstition attributes cause of a consequent action or event to logically unrelated preceding actions or events. This misattribution is characteristic of how nocebos, and in turn expectations, can shape perceptions and the meanings we make of the experience of subsequent exposures. In the extreme, there is documented evidence that fears from strongly held cultural beliefs can induce the ultimate nocebo effect: death. A team of researchers found that Chinese Americans who believe in Chinese astrology, and “have a combination of disease and birth year which Chinese astrology and medicine considers ill-fated,” die younger than patients with the same disease that were born in a different year. According to the researchers, the “strength of commitment to traditional Chinese culture” correlated with the difference in life span. This surprising phenomenon is evident in several other cultures where the power of a hex or curse can invoke cultural beliefs resulting in unexplained sickness, sleep paralysis, and even sudden death.1 Though you might take these extreme nocebo examples with a grain of salt, you probably have not been above tossing a pinch of salt over your shoulder from time to time.
As we saw in previous chapters, positive expectations can induce neurological changes that enhance the effectiveness of a treatment and lead to the production of placebo effects and amelioration of symptoms. Unfortunately, the inverse is also true. Negative verbal suggestion, associative learning, and contextual cues create negative expectations that are accompanied by neurological changes that can increase the experience of symptoms. This phenomenon, termed the nocebo effect, is the antithesis of the placebo effect. Derived from the Latin nocere, “to harm,” Walter Kennedy coined the term nocebo in 1961 to describe the impact that negative suggestions or information can have on clinical outcomes. In this chapter, I examine the neurological circuitry activated by nocebos, and discuss the factors that promote the negative expectation and transmission of nocebo effects in randomized clinical trials, popular culture, and the clinical encounter. Strategies to mitigate nocebo effects will also be explored.
Not surprisingly, there are similarities between the neurological pathways that mediate placebo and nocebo effects. The expectation of a negative experience (i.e., pain) creates a mental representation of the pending experience by activating the prefrontal cortex (PFC), ACC, and vmPFC. The thalamus and insula, two regions that encode pain, are activated too. When an individual is exposed to the painful stimulus, their experience is proportional to the combined effect of the intensity of the stimulus and precision of the expectation. If the expectation is for less pain, more often than not, less pain is experienced. If the expectation is for more pain, the experience of pain is frequently amplified.2 The worker landing on a seven-inch nail in chapter 2 is a perfect example; observing the nail through his boot, he anticipated and felt extreme pain, even though the nail wasn’t penetrating his foot at all. In this way, negative expectations drive nocebo-induced hyperalgesia or increases in pain.
Nocebo-induced hyperalgesia can be observed through brain imaging in real time. Pain and the expectation of pain induces activation in overlapping brain structures. These structures include the regions of the pain matrix discussed in the earlier chapters along with the thalamus, insula, dorsolateral prefrontal cortex, ACC, and somatosensory cortex. Whereas positive expectations dial down signaling in pain-related regions, negative expectations can enhance this signaling. Nocebo effects are not limited to placebo studies in hyperalgesia; nocebos and their harmful effects are everywhere. To get a sense of real-world nocebo effects, let’s take a look at statins, side effects, and cardiovascular disease.
As a young man, Akira Endo was inspired by Alexander Fleming and the discovery of penicillin. After college in Japan, he spent two years in New York City, where he was struck by how many people suffered from cardiovascular disease, the leading cause of death in the United States. Cholesterol was in large part to blame. Endo reasoned that if mold made molecules that killed bacteria, perhaps they also made molecules that could perturb cholesterol synthesis. On returning to Japan, he screened culture broths from thousands of mold extracts in search of an anticholesterol “antibiotic.”
Nocebo effects are not limited to placebo studies in hyperalgesia; nocebos and their harmful effects are everywhere.
It was the 1970s, and Endo was about to revolutionize cardiovascular disease prevention. Compactin, the molecule he discovered, had potent lipid-lowering effects in animal models, but in his clinical trials in patients with severe familial hypercholesterolemia, an inherited form of high cholesterol, one of the subjects experienced muscular dystrophy. Although this side effect disappeared when the treatment was stopped, it was enough to halt development of compactin. Luckily, Merck, a pharmaceutical company, took up the cause, and soon after another mold extract, lovastatin, gained regulatory approval. In all, seven statins including atorvastatin (Lipitor) and rosuvastatin (Crestor) would do well in clinical trials, and make it to the now-competitive statin market.
Statins are remarkably effective at lowering low-density lipoprotein or the “bad” cholesterol. Patients taking statins have almost a 50 percent lower risk of heart attack, stroke, or the need for bypass surgery.3 Overall, statin users have a 20 percent lower chance of dying from any cause compared to their counterparts who don’t take a statin. But despite this impressive efficacy, statins could never shake the stigma of muscle pain.
While statins can cause muscle pain, arguably most muscle pain and other statin-related adverse effects are not caused by the drug but instead by nocebo effects. To ensure patient autonomy and decision-making power, it is important to ensure that the patient is fully informed about possible negative outcomes of potential treatments. The informed consent procedure in randomized clinical trials requires the disclosure of all the possible side effects associated with a given treatment. Because statins do induce myopathy and rhabdomyolysis, or muscle wasting, in rare cases, this information has to be communicated to patients in clinical trials to allow them the autonomy to decide if the drug is right for them. Incredibly, the side effects observed in the placebo arm of a trial are strikingly similar to those explained and attributed to the active treatment in informed consent. Up to a quarter of the patients in the placebo arm of statin clinical trials drop out because they experience adverse effects.
Information about the potential side effects of statins is not limited to informed consent in clinical trials. It is not even limited to shared decision-making discussions in the clinical setting. Negative information about statins is everywhere. A recent Google search of “statin side effects” produced over ten million hits; many of these sites post unsubstantiated theories about why statins are harmful and cause harmful adverse effects. A common trope is that statins inhibit cholesterol, and as your brain is made of cholesterol, statin use can shrink your brain. As a result of what seems to be a campaign of misinformation, many who might benefit from taking a statin to reduce their risk of cardiovascular disease opt not to.
One of the many studies used to get to the bottom of the statin side effect conundrum was an n-of-1 clinical trial in which patients served as their own controls by taking either the statin or a placebo control for specified time intervals. Between 2016 and 2019, sixty patients who had previously discontinued statin use within two weeks of starting treatment due to side effect complaints were enrolled in the trial.4 These patients each received a month’s worth of atorvastatin, a placebo, or no pills for one-month intervals over a twelve-month period. The order in which they were instructed to take the pills varied between the patients and was randomized from month to month. The patients reported their daily symptom intensity on a scale of 0 (no symptoms) to 100 (worst symptoms) via a smartphone app. If their side effects were too strong, they could stop taking that set of pills for the rest of the month and resume with the next designated set of pills the following month.
At the end of the n-of-1 study, the researchers found that during the no-pill month, the average pain intensity was 8.0; during placebo months, it was 15.4; and during statin months it was 16.3. Therefore taking any pill increased average pain, but there was little statistical difference between the average statin and placebo-related pain intensities. Of the enrolled sixty participants, forty-nine completed the trial, and thirty were able to restart statin treatment without difficulty after the trial. This study not only supports the assertion that statin-related side effects are commonly misattributed to the drug but also underscores the ethical predicament facing clinicians: on the one hand, too little information robs the patient of their autonomy; on the other hand, too much negative information can induce nocebo effects, and lead to poor adherence to or rejection of a treatment that might be of great benefit to the patient.
Statins are not the only drugs plagued by nocebo effects. The suggestibility of side effects is observed in many other conditions. Meta-analyses (a systematic assessment of the results of multiple clinical trials) of side effect suggestibility in patients being treated for migraines or using SSRIs found that the adverse events reported tended to be closely related to those of the active substance in the trials.5 For example, erectile dysfunction (ED) is a commonly cited symptom of beta-blockers, a blood pressure medication.
In a study in which ninety-six male patients were randomized to a beta-blocker, subjects given no information about the treatment had few cases of ED (3.1 percent), those given the name of the treatment, “atenolol,” reported more cases (15.6 percent), but almost a third of the patients who were given the name of the drug and information about side effects reported ED (31.2 percent).6 In the second part of this study, those participants who reported experiencing ED were rerandomized to receive a placebo or Sildenafil, a drug that treats ED. In the study, the placebo and Sildenafil were equally effective at reversing ED. This trial clearly demonstrates the potency of negative as well as positive expectations to influence clinical outcomes. It is possible that anxiety about this particular side effect in the first part of the study contributed to a higher incidence of ED. This link between nocebos and anxiety is emblematic of the potential of negative experience and nocebo effects in the context of clinical care to lead to worse clinical outcomes. But this is just the tip of the iceberg of the pain and suffering that nocebo effects can cause.
Studies have shown that when patients walk into a doctor’s office and experience being seen and heard by a doctor who is present and attentive, their treatment outcomes are far more successful. Components of the treatment encounter, including therapeutic alliance, warmth, and competence, have been shown to augment treatment outcomes. In one of the landmark papers in the placebo literature, participants with IBS were randomized to three groups that re-created an increasing “dosage” of the clinical encounter. The minimal dose was just an assessment or observation of symptoms. Adding the ritual of treatment by a clinician to the observation had a stronger effect. Establishing a positive relationship between the patient and treatment practitioner during treatment by ensuring eye contact, listening to the patient’s symptoms, expressing confidence in the treatment, and emphasizing physical touch was the strongest dose of placebo.7 One can imagine that the absence of these placebo-enhancing elements in the clinical encounter could reduce the quality of the treatment. In other words, the absence of placebo-inducing effects could result in suboptimal care.
With the ethical challenges of creating negative experiences, less has been done to modify and study negative physician encounters. Lauren Howe and colleagues, however, used an allergic reaction paradigm to study practitioner effects.8 In this study, an allergen was administered in a pinprick, followed by a topical inert cream, and then either positive (“that the cream would reduce the reaction and itching”) or negative (“that the cream would increase the reaction and itching”) information was provided by a clinician who demonstrated either high or low warmth and competence. The size of the wheals among the participants who were randomized to the positive framing was significantly smaller than among those given the negative framing. Surprisingly, regardless of expectation, the participants randomized to the high warmth, high competence clinician interaction had the smallest wheal size. In this way, the impact of expectations on allergic reactions was enhanced when the clinician demonstrated warmth and competence, and this effect was nullified by a cold and less competent physician.
Although not commonly framed in nocebogenic terms, implicit bias and racism can seed negative expectations and drive nocebo effects.9 Just as important, the absence of placebogenic factors can also lead to negative clinical outcomes. Substantial evidence suggests that nonwhite patients often receive unfavorable treatment in comparison to their white counterparts.10 This disparate, frequently inferior treatment is characterized by less empathy and time spent with patients, poor communication, and less caring behavior on the part of physicians.11 Other studies found that physicians were less likely to participate in shared medical decision-making or establish rapport with nonwhite patients, and that nonwhite patients were less likely to receive sufficient information.12 In another study, independent, nonbiased raters found that physicians were 23 percent more verbally dominant and 33 percent less engaged in patient-centered communication in medical visits with Black patients as compared to white ones.13
Given the potential for nocebos to cause harmful effects, it is important that clinicians and physicians understand their mechanisms of action, and what steps can be taken to mitigate their effects.14 Generally, experts agree it is important to educate physicians and patients about the potential for nocebo effects in a way that matches the clinical context.15 These experts and other clinician placebo researchers encourage clinicians to help patients explore their concerns and expectations, and learn coping strategies to manage their expectations.16 They suggest contextualizing worrisome information by setting patients’ expectations of the information they will encounter on the internet and describing treatments in a realistic but positive way (i.e., present the proportion of patients who don’t get side effects as opposed to the proportion who do get them).
In the specific encounter between white doctors and Indigenous, Black, or other people of color patients, antiracism training and communication and/or cross-cultural training to enhance clinician skills in caring for patients of other cultures or races is recommended to address the potential for the nocebo effects of racism and bias to negatively impact patient care.17
Excipients are literally the stuff of placebos; they are the nonactive ingredients that allow manufacturers to compound a drug into a comestible pill with a reasonable shelf life. When Merck changed the excipients in the thyroid hormone replacement drug Levothyroxine from mannitol plus citric acid to lactose, it set off a virtual firestorm. Before releasing the new formulation, Merck had conducted a randomized bioequivalence study that found that the biological outcome measures were essentially identical and the new formulation was completely safe. Nonetheless, patients reported significant side effects; from hair loss, headaches, and weight gain, to diarrhea, extreme fatigue, and increased heart rates, the rash of side effects experienced by the tens of thousands of patients with hypothyroidism in France gave rise to more than sixty lawsuits accusing Merck and the French government of a “failure to assist a person in danger.”18 Finally, the minister of solidarity and health asked Merck to bring back the old formulation, and the outcry died down. At the same time, the company introduced the new formulation in other countries, with more deliberate communications to patients about the change.
A similar mass nocebo effect happened a decade earlier in New Zealand. In 2007, GlaxoSmithKline moved the manufacturing of Eltroxin, the only thyroid hormone replacement drug approved for use by the New Zealand government, from Canada to Germany, resulting in some changes to the inert components of the drug’s formulation. Though the new manufacturing process was actually more expensive than the old one, some New Zealanders believed that the switch was a cost-cutting tactic of the pharmaceutical company that put patients at increased risk. A local pharmacist was concerned about the adverse events he observed in patients and reported them to the media. The story of a “small town health professional taking on the ‘medical establishment’” gained widespread media attention, and an almost two thousandfold increase in reported adverse events followed.19 Conspiracy theories proliferated that the company was lying to patients and that the new formula contained genetically modified or toxic agents. Studies at the time showed that the regions with the most media coverage of the switch also tended to have the highest rates of reported adverse events. Some of the adverse events reported were consistent with hypothyroidism, but a great number were nonspecific. After testing, it was determined that the new and old drugs were bioequivalent, and once it was announced that an alternative drug would be made available, the adverse events dwindled significantly. Clearly, nocebo effects transmitted through media can be pervasive, and have widespread public health impacts that severely affect treatment outcomes and adherence.
In the midst of the COVID-19 pandemic, nocebos and their effects are increasingly relevant. Opportunities for nocebo effects to influence behavior and outcomes are rampant.20 Combined with potential preexisting medical mistrust, the stress of deciphering misinformation or conspiracy theories, and the dramatization or exaggeration of potential virus symptoms and vaccine side effects in the media, the scene is set for nocebos to take over. Though it is difficult at this time to assess how many of the COVID-19 vaccine-related side effects, which include headache, fatigue, fever, and sore arms, are potentially influenced by nocebo effects, a recent initial meta-analysis found that there is a high rate of adverse events reported in patients assigned placebo shots.21 Though at this point it would be hard to prove, the impact of media hype about these side effects likely has the downstream effect of creating vaccine hesitancy. While people need to be informed about the potential risks of the vaccine, broadcasting those risks does indeed appear to be exacerbating the nocebo effect.
In the United States, fear and misinformation about the vaccine side effects are prevalent and surveys in 2020 showed that a substantial portion of people were unsure about the COVID-19 vaccine or did not plan to get the vaccine when it became available to them.22 Consequently perhaps, the rates of vaccination are relatively low; the Mayo Clinic reports that in the United States as of 2021, only 67 percent of the eligible population has received a first dose, and an average of about a thousand people die a day from the virus.23 Portugal, on the other hand, boasts one of the highest vaccination rates in the world; with 86 percent of its total population fully vaccinated, and 98 percent of the eligible population fully vaccinated, Portugal was able to lift all COVID restrictions and had a death rate of less than ten per day in 2021.24 As we all know, the COVID-19 landscape is changing rapidly, and it remains to be seen if nocebo or placebo effects will prevail.