Is the placebo effect just people imagining they feel better?

No. Neuroimaging and biochemical studies show that expectation and conditioning trigger real neurotransmitter release, including opioids, dopamine and endocannabinoids, producing measurable physiological changes.

How large is the placebo effect in pain trials?

In pain trials, placebo responses often account for 30 to 50 percent of the total treatment effect, with effect sizes comparable to many active analgesics in some meta-analyses.

Does the placebo effect work for objective outcomes like tumour shrinkage?

Placebo effects are strongest for subjective outcomes such as pain, nausea, fatigue and mood. They rarely produce measurable changes in objective markers like tumour size, infection clearance or blood glucose.

Resources - RESEARCH LITERACY

How big is the placebo effect really

The placebo effect is a measurable neurobiological response driven by expectation and conditioning, not a single fixed magnitude, and its size varies widely across conditions, outcomes and study designs. It is not a single number but a distribution that depends on what is being measured, how the trial is designed and the psychological context of the intervention. Understanding this variability is essential for interpreting any clinical research.

What the placebo effect actually is

A placebo response occurs when a person experiences a real physiological or symptomatic change after receiving an inert intervention, solely because they expect benefit or have been conditioned to associate the ritual of treatment with relief. Neuroimaging and microdialysis studies demonstrate that verbal suggestions and conditioning cues activate specific brain regions, including the prefrontal cortex, anterior cingulate and periaqueductal grey, triggering the release of endogenous opioids, dopamine, endocannabinoids and cholecystokinin. These pathways overlap with those targeted by many active drugs, which is why placebo responses can mimic pharmacological effects in pain, Parkinsonian motor function, depression and immune modulation.

The magnitude of this response is not constant. It is shaped by the strength of the expectation, the credibility of the clinician, the invasiveness of the sham procedure, the natural history of the condition, regression to the mean and statistical artefacts such as measurement error. Disentangling the true neurobiological placebo component from these other factors requires a three-arm design comparing active treatment, placebo and no-treatment control, a design that is rare in published literature.

How magnitude differs by outcome type

Subjective, patient-reported outcomes show the largest placebo effects. Pain is the most studied domain. Meta-analyses of placebo-controlled analgesic trials consistently find that the placebo arm captures 30 to 50 percent of the total effect size seen with active medication, and in some headache and neuropathic pain trials the placebo response equals or exceeds the drug response. Nausea, fatigue, depressed mood, insomnia and dyspnoea also show robust placebo responses, often with standardised mean differences between 0.3 and 0.6 versus no treatment.

Objective physiological markers tell a different story. Placebo interventions rarely alter tumour dimensions, viral load, HbA1c, bone mineral density or ejection fraction in a clinically meaningful way. A landmark Cochrane review comparing placebo with no-treatment controls across hundreds of trials concluded that placebo had no significant effect on binary or continuous objective outcomes, while small but statistically significant effects persisted for continuous subjective outcomes. This distinction matters: a therapy that only moves subjective scores may still be valuable for quality of life, but it should not be equated with disease modification.

Condition-specific variability

Even within subjective domains, the placebo effect size fluctuates by condition. In irritable bowel syndrome, open-label placebo trials report response rates of 40 to 60 percent when patients are told they are receiving a placebo but given a persuasive rationale. In major depressive disorder, the placebo response rate in antidepressant trials has risen over the past three decades, now often exceeding 40 percent, complicating drug development. In Parkinson disease, placebo can trigger striatal dopamine release measurable on PET imaging, producing transient motor improvement. In contrast, conditions with hard endpoints such as bacterial infection, hypertension or lipid lowering show negligible placebo effects on the primary biomarker.

The route of administration also modulates magnitude. Sham procedures involving needles produce larger effects than sham tablets; sham surgical procedures produce larger effects still. Colour, size, branding, price and the warmth of the clinician-patient interaction all act as expectation amplifiers. These contextual factors mean that the placebo effect in a tightly blinded regulatory trial may underestimate the effect seen in open-label clinical practice, where expectancy is maximised.

Trial design artefacts that inflate or deflate the number

Many published placebo effect sizes are inflated because they conflate the true placebo response with the natural history of the condition, regression to the mean and the Hawthorne effect. A two-arm trial comparing drug versus placebo cannot separate these components. Only a three-arm trial with a no-treatment or waitlist control can isolate the placebo-specific effect. When such designs are used, the pure placebo effect is often smaller than the raw placebo-arm improvement suggests.

Conversely, trial designs that minimise expectancy, such as balanced placebo designs or hidden administration paradigms, reveal that a substantial portion of drug effect can disappear when the recipient is unaware they received treatment. This demonstrates that the therapeutic ritual itself contributes meaningfully to outcomes even for active compounds, blurring the line between drug effect and context effect.

Neurobiological mechanisms and their limits

Endogenous opioid release mediates placebo analgesia and can be blocked by naloxone. Dopaminergic pathways in the striatum mediate placebo responses in Parkinson disease and reward learning. Endocannabinoid signalling contributes to placebo analgesia in certain conditioning paradigms. Cholecystokinin antagonism can enhance placebo analgesia, suggesting an anti-placebo system. These mechanisms are real, reproducible in experimental settings and dose-dependent on expectation strength.

However, these pathways have physiological ceilings. Endogenous opioid release cannot match the receptor occupancy of large-amount exogenous opioids. Dopamine release in the striatum is transient and does not reverse neurodegeneration. The immune modulation seen in conditioned immunosuppression models in animals is modest compared with pharmacological immunosuppressants. The placebo effect works within the bounds of the body's own regulatory capacity; it does not override pathophysiology.

Common misunderstandings

A frequent error is treating the placebo effect as a single constant that can be subtracted from any treatment effect to reveal the true drug effect. In reality, drug and placebo effects interact. Expectation can amplify drug efficacy, and drugs can enhance expectancy. The assumption of additivity is rarely justified. Another misunderstanding is equating placebo response with deception. Open-label placebo studies show that transparent placebo administration, accompanied by a credible rationale, can still produce significant symptom relief, challenging the notion that deception is necessary.

A third error is assuming that a large placebo response means the condition is not real or the treatment is ineffective. High placebo responsiveness reflects the condition's sensitivity to top-down modulation, not the absence of pathology. Conditions like chronic pain, functional gastrointestinal disorders and depression are genuinely debilitating and highly modulated by central nervous system processing, making them placebo-responsive by nature.

Who the evidence applies to and who it does not

Placebo responsiveness varies between individuals. Genetic polymorphisms in COMT, OPRM1 and FAAH genes explain a modest fraction of variance in placebo analgesia. Personality traits such as optimism, suggestibility and trait anxiety correlate with response magnitude. Prior positive treatment experiences condition stronger future responses. However, these predictors are probabilistic, not deterministic, and cannot reliably identify placebo responders in clinical practice.

The evidence base is heavily weighted toward adult populations in high-income countries, with limited data on children, older adults with multimorbidity or diverse cultural contexts. Expectation formation and illness beliefs are culturally shaped, so effect sizes derived from Western trial populations may not generalise. Most mechanistic data come from healthy volunteers in experimental pain models, not from patients with chronic disease, limiting direct clinical extrapolation.

How to judge claims about the placebo effect

When a claim cites a specific percentage for the placebo effect, ask: what condition, what outcome, what control group, what blinding integrity and what population. A claim that the placebo effect is 30 percent is meaningless without these qualifiers. Look for three-arm trials or meta-analyses that separate placebo from natural history. Distinguish between subjective and objective endpoints. Recognise that the placebo effect is not a nuisance variable to be eliminated but a neurobiological phenomenon that can be understood, measured and, where appropriate, ethically harnessed alongside active treatment.

In research literacy, the placebo effect teaches that context is a biological variable. The ritual of care, the words of the clinician, the sensory cues of treatment and the patient's prior learning all shape physiology. Any intervention, whether a peptide, a light device, a cold plunge or a supplement, is delivered inside a context that modulates its effect. Rigorous evidence grades the specific contribution of the active ingredient above and beyond that context. That is the standard Vivera Education applies across every topic.

This article explains the evidence on placebo magnitude and mechanisms. It does not provide medical advice or suggest that any condition can be treated with placebo alone.

Frequently asked

Is the placebo effect just people imagining they feel better?: No. Neuroimaging and biochemical studies show that expectation and conditioning trigger real neurotransmitter release, including opioids, dopamine and endocannabinoids, producing measurable physiological changes.
How large is the placebo effect in pain trials?: In pain trials, placebo responses often account for 30 to 50 percent of the total treatment effect, with effect sizes comparable to many active analgesics in some meta-analyses.
Does the placebo effect work for objective outcomes like tumour shrinkage?: Placebo effects are strongest for subjective outcomes such as pain, nausea, fatigue and mood. They rarely produce measurable changes in objective markers like tumour size, infection clearance or blood glucose.