Is the placebo effect strong enough to use clinically?

How big is the placebo effect—and does it really matter in practice?

The placebo effect is not just a statistical nuisance; it's a clinically meaningful phenomenon. A landmark meta-analysis of 186 randomized trials across dozens of conditions found that, on average, 54% of the overall treatment effect in the active drug group was attributable to contextual effects—meaning the placebo response accounted for more than half of what patients experienced as improvement [7]. In antidepressant trials specifically, the placebo effect size was d=1.69, which means the average person on placebo improved more than 95% of untreated patients, and this placebo response represented 68% of the drug effect [6]. These are not trivial numbers: they show that the ritual of treatment—the pill, the injection, the consultation—can produce real, substantial benefits.

But the effect is not uniform. In a meta-analysis of diabetes trials, placebo treatment actually led to a slight worsening of blood sugar (HbA1c increased by 0.14%), suggesting that for hard, objective biomarkers, the placebo effect may be weak or absent [3]. In contrast, for conditions like functional dyspepsia (stomach pain with no clear physical cause), placebo produced a 42-45% adequate relief rate at 8 weeks—a clinically meaningful response [2]. The takeaway: the placebo effect is strongest for subjective, symptom-based outcomes (pain, nausea, mood) and weaker for objective lab values.

Can placebos work even if patients know they're taking a placebo?

Surprisingly, yes. A systematic review and meta-analysis of 13 trials found that open-label placebos—where patients are told they are receiving a placebo—produced a significant overall effect (standardized mean difference = 0.72) compared to no treatment [5]. This effect was seen across conditions including back pain, irritable bowel syndrome, and menopausal hot flushes. The effect size is moderate to large, meaning that even without deception, the act of taking a placebo can trigger real improvements, likely through expectation and conditioning.

This finding is important for clinical use because it removes the ethical barrier of deception. A doctor could say, 'This is a placebo—it has no active drug, but your brain can still produce real benefits from it,' and patients may still respond. However, the research is still early, and the effect varies by condition and how the placebo is presented [5].

What causes the placebo effect—and can we make it stronger?

The placebo effect is driven by a combination of patient expectations, the quality of the patient-provider relationship, and the ritual of treatment. A large meta-analysis found that contextual effects were higher in trials with blinded outcome assessors and concealed allocation, suggesting that the patient's belief in the treatment—and the care with which it is delivered—matters [7]. In a study of functional dyspepsia, the placebo group's adequate relief rate was 42-45%, which is comparable to some active treatments, and this was achieved simply by giving a placebo pill in a clinical setting [2].

However, attempts to deliberately boost the placebo effect by training study teams have had mixed results. One recent trial found that briefing the study team on placebo mechanisms did not significantly alter pain reduction or treatment expectations, regardless of whether the team was trained to maximize or minimize placebo effects [8]. This suggests that the placebo effect is not easily manipulated by brief interventions—it may require deeper changes in the clinical environment or longer-term patient conditioning.

Interestingly, even the belief that an AI system is adaptive can produce a placebo effect: participants who thought they were receiving AI support performed better on word puzzles, even though the AI was not actually helping [9]. This shows that expectations alone—whether from a doctor, a device, or a label—can drive real performance changes.

When should clinicians actually use the placebo effect?

The evidence suggests that the placebo effect is most useful for conditions where symptoms are subjective and where there is no highly effective, low-side-effect treatment. For example, in functional dyspepsia, pregabalin outperformed placebo (70.6% vs 44.7% adequate relief at 8 weeks), but the placebo response was still substantial and could be a first-line option for some patients [2]. In ulcerative colitis, a placebo response of 34.5% was observed, meaning about one in three patients improved on placebo alone [1].

However, the placebo effect is not a replacement for effective drugs. In the ulcerative colitis trial, the active drug (olamkicept 600 mg) produced a 58.6% response rate, significantly better than placebo [1]. And in NASH (fatty liver disease), while 20-28% of placebo patients showed a 30% reduction in liver fat, the effect was modest and variable [4]. The smart clinical use of the placebo effect is to harness it alongside active treatment—by building positive expectations, improving communication, and reducing anxiety—not to substitute it for proven therapies.

About These Sources

This answer is built on 9 peer-reviewed studies — published from 2009 to 2025, 1 from 2024 or later, 5 in Q1 journals, collectively cited 560 times — selected as the most relevant from 9 studies that passed quality screening, drawn from 48 papers retrieved from a database of over 500 million.

Sources used in this answer

Related research questions