Skip to content

Placebo response and trial design

TL;DR — In a disorder defined and treated by symptom report, the placebo response is evidence about the condition, not a nuisance to be subtracted. Across 73 RCTs and 8,364 placebo-allocated patients the pooled placebo response was 37.5% (95% CI 34.4–40.6) (Ford 2010, PMID 20412064); a later 73-trial synthesis put it at 27.3% (24.3–30.9) for global improvement, 34.4% (31.2–37.8) for abdominal pain and 17.9% (15.2–21.0) for the composite FDA endpoint (Bosman 2021, PMID 33765447). Endpoint stringency, not biology, drives most of that spread: in licensed-drug trials for IBS-C, moving from a 6/12-week to a 9/12-week responder requirement drops the composite placebo rate from 18.9% to 4.3% and the stool response from 30.1% to 7.7% (Barberio 2022, PMID 34425274). The mechanism is at least partly relational rather than pharmacological: in a three-arm randomised trial, adequate relief rose stepwise from 28% (waiting list) to 44% (sham acupuncture alone) to 62% (sham acupuncture plus a warm, attentive practitioner relationship), p<0.001 for trend, with the patient–practitioner relationship "the most robust component" (Kaptchuk 2008, PMID 18390493). And the placebo works when patients are told it is a placebo: open-label placebo beat no-pill control on IBS-SSS improvement (90.6 vs 52.3 points, p=0.038) and was statistically indistinguishable from double-blind placebo (100.3 vs 90.6, p=0.485, d=0.10) in a 262-patient three-arm trial (Lembo 2021, PMID 33605656). The nocebo counterpart is equally large: 32% (26–38) of placebo-treated IBS patients report adverse events (Li 2022, PMID 36606047).

How big, and measured how

Source Scope Endpoint Pooled placebo response
Ford 2010, PMID 20412064 73 RCTs, 8,364 placebo patients Global symptom improvement or cure, or abdominal pain improvement 37.5% (34.4–40.6)
Bosman 2021, PMID 33765447 73 RCTs (70 articles), search 1959–April 2020 Global improvement 27.3% (24.3–30.9)
Bosman 2021, PMID 33765447 same Abdominal pain 34.4% (31.2–37.8)
Bosman 2021, PMID 33765447 same Composite FDA responder 17.9% (15.2–21.0)
Barberio 2022, PMID 34425274 17 IBS-C RCTs of licensed drugs, 4,603 placebo patients FDA composite, ≥6/12 weeks 18.9%
Barberio 2022, PMID 34425274 same FDA abdominal pain, ≥6/12 weeks 34.6%
Barberio 2022, PMID 34425274 same FDA stool response, ≥6/12 weeks 30.1%
Barberio 2022, PMID 34425274 same, ≥9/12 weeks composite / pain / stool 4.3% / 24.5% / 7.7%
Barberio 2022, PMID 34425274 17 IBS-D RCTs, 3,908 placebo patients FDA composite / pain / stool, ≥6/12 weeks 16.2% / 40.2% / 16.2%
Barberio 2022, PMID 34425274 IBS-D, raising pain threshold from ≥30% to ≥40% then ≥50% abdominal pain 40.2% → 34.5% → 23.4%
Li 2022, PMID 36606047 53 RCTs, nocebo Any adverse event on placebo 32% (26–38); headache 9%, nasopharyngitis 7%, abdominal pain 4%

The single most useful number here is the 18.9% → 4.3% collapse when the responder window tightens from 6 to 9 of 12 weeks (Barberio 2022, PMID 34425274). It means that most of the "placebo response" in IBS trials is transient symptom improvement that fails to persist — and that a drug's apparent NNT can be manufactured or destroyed by an endpoint definition, without anything changing in the patients.

What moderates it

Moderator Direction Source
Publication before 2006 Higher placebo response Bosman 2021, PMID 33765447
Conducted in Europe Higher Bosman 2021, PMID 33765447; Ford 2010, PMID 20412064
Parallel (vs crossover) design Higher Bosman 2021, PMID 33765447
Run-in period ≤2 weeks Higher Bosman 2021, PMID 33765447
Dosing three or more times daily Higher Bosman 2021, PMID 33765447
Smaller control group Higher Bosman 2021, PMID 33765447
Physician-reported (vs patient-reported) outcomes Higher Ford 2010, PMID 20412064
Shorter therapy duration Higher Ford 2010, PMID 20412064

Bosman and colleagues turned these into design advice: use a run-in of at least 2 weeks and dose once or twice daily to minimise the placebo response (PMID 33765447). Barberio and colleagues recommend adhering to FDA-recommended endpoints "because these lead to lower placebo response rates," while noting that some should be refined further to separate drug from placebo (PMID 34425274).

What the placebo response is made of

The dose–response of ritual and relationship. Two hundred and sixty-two adults with Rome II IBS and IBS-SSS ≥150 were randomised for three weeks to waiting-list observation, "limited" placebo acupuncture, or "augmented" placebo acupuncture plus a practitioner relationship characterised by warmth, attention and confidence (Kaptchuk 2008, PMID 18390493):

Outcome Waiting list Limited Augmented p (trend)
Global improvement scale (1–7) 3.8 (SD 1.0) 4.3 (1.4) 5.0 (1.3) <0.001
Adequate relief 28% 44% 62% <0.001
Symptom severity score change 30 (63) 42 (67) 82 (89) <0.001
Quality of life change 3.6 (8.1) 4.1 (9.4) 9.3 (14.0) <0.001

All pairwise comparisons between augmented and limited relationship were significant, and results held at six weeks. The authors' conclusion — the components "can be progressively combined in a manner resembling a graded dose escalation" and "the patient–practitioner relationship is the most robust component" — makes the consultation itself an active ingredient with a measurable effect size larger than most IBS drugs.

Open-label placebo. Three trials, consistent direction:

Trial Design Result
Kaptchuk 2010, PMID 21203519 (NCT01010191) 80 patients, 3 weeks, open-label placebo described as "inert substance, like sugar pills… shown in clinical studies to produce significant improvement in IBS symptoms through mind–body self-healing processes" vs no-treatment control with matched provider interaction IBS-GIS 5.2 ± 1.0 vs 4.0 ± 1.1 at 11 days (p<0.001) and 5.0 ± 1.5 vs 3.9 ± 1.3 at 21 days (p=0.002); IBS-SSS p=0.008 and p=0.03; adequate relief p=0.02 and p=0.03; IBS-QoL trend p=0.08
Lembo 2021, PMID 33605656 262 adults, 6 weeks, three arms: open-label placebo (OLP), double-blind placebo (DBP), no-pill control (NPC) IBS-SSS improvement: OLP 90.6 vs NPC 52.3 (p=0.038, d=0.43); OLP 90.6 vs DBP 100.3 (p=0.485, d=0.10) — no significant difference. "Blinding may not be necessary for placebos to be effective"
Nurko 2022, PMID 35099543 (NCT02389998) 30 children/adolescents aged 8–18 with functional abdominal pain or IBS; crossover, 3 weeks OLP vs 3 weeks control Mean daily pain 39.9 (18.9) vs 45.0 (14.7); difference 5.2 (95% CI 0.2–10.1), p=0.03. Rescue hyoscyamine use nearly halved: 2.0 (3.0) vs 3.8 (5.1) pills, difference 1.8 (0.5–3.1)

The Nurko result is the most resistant to dismissal because rescue-medication count is an objective behavioural measure, not a symptom rating.

Who responds to open-label placebo. In a secondary analysis of 210 participants from the Lembo trial, visceral sensitivity (symptom-specific anxiety) predicted response to OLP and to no-pill control in opposite directions — high visceral sensitivity predicted less improvement on NPC and more on OLP — while pain catastrophising was a negative predictor of OLP response. Neither predicted double-blind placebo response. The authors conclude that "different psychological mechanisms may be involved in DBP and OLP interventions" (Ballou 2022, PMID 35412513). Genomic correlates of placebo response have also been examined in this trial programme (Wang 2021, PMID 35295415; Vollert 2022, PMID 35401282).

The instruments the endpoints are built from

Two measurement decisions shape every effect estimate in this knowledge base.

IBS-SSS. The IBS Severity Scoring System, validated in 141 patients and 40 healthy controls, incorporates pain, distension, bowel dysfunction and quality of life/global well-being on a 0–500 scale. Mild, moderate and severe are 75–175, 175–300 and >300; controls score below 75, so patients scoring below 75 "can be considered to be in remission". Scores repeated within 24 hours were highly reproducible, and a change of 50 points reliably indicates improvement (Francis 1997, PMID 9146781). That 50-point figure is the responder threshold used in CARIBS, DOMINO, the FMT trials and the hypnotherapy trials throughout this knowledge base — and ATLANTIS prespecified a different threshold, 35 points, as its minimum clinically important difference, then reported −27.0 (Ford 2023, PMID 37858323). The field does not agree on what counts as improvement on its own most-used scale.

The FDA composite responder endpoint. For IBS-C: ≥30% reduction in worst abdominal pain and an increase of ≥1 complete spontaneous bowel movement per week, both in the same week, for ≥6 of 12 weeks. For IBS-D: pain reduction paired with a stool-consistency criterion. Composites of this kind were introduced to suppress placebo response, and they do — 17.9% pooled versus 27.3% for global improvement (Bosman 2021, PMID 33765447).

What the composite discards. Pooled analysis of the two linaclotide phase 3 IBS-C trials (1,602 patients) asked what happened to patients who did not meet the FDA responder definition. Thirty-four per cent of linaclotide-treated and 17% of placebo-treated patients met it (p<0.0001). Among FDA non-responders at week 12 (Lacy 2014, PMID 24382134):

Measure, among FDA non-responders Linaclotide Placebo
Abdominal pain at least somewhat relieved 63% 48%
Stool frequency at least somewhat improved 62% 46%
At least some global IBS symptom relief 65% 48%
Adequate relief of IBS symptoms 43% 34%
Satisfied with treatment 57% 41%

A drug–placebo gap of 12–17 percentage points persists inside the group the endpoint classifies as failures. The authors' conclusion is that "differing response thresholds and symptom-specific change from baseline should be considered by clinicians for a complete understanding of clinical response" — a manufacturer-affiliated analysis making a methodologically valid point. The same asymmetry explains the peppermint-oil result, where both co-primary regulatory endpoints were missed while continuous secondaries improved (Weerts 2020, PMID 31470006).

The general lesson: stringent composites lower the placebo response and lower the drug response together, and whether the resulting responder difference is the right target depends on an unresolved question about what patients value — which connects directly to the absence of work-function endpoints (quality-of-life-and-stigma).

Nocebo

Thirty-two per cent (26–38) of placebo-treated IBS patients in parallel-design pharmacological RCTs report adverse events, most commonly headache (9%), nasopharyngitis (7%), abdominal pain (4%) and nausea (Li 2022, PMID 36606047). This has direct consequences elsewhere in this knowledge base: 26% of ATLANTIS placebo participants discontinued, 9% for adverse events (Ford 2023, PMID 37858323), and adverse-event NNH estimates for active drugs are computed as differences against a placebo arm that is itself generating symptoms (Busam 2026, PMID 40471839).

Two natural experiments outside drug trials

  1. Labelling. In a 2×2 factorial trial, 269 adolescents received mebeverine or placebo, and independently were told the treatment was "mebeverine or placebo" or "mebeverine". The drug did nothing (23.4% vs 22.0%, OR 1.08, 0.59–1.99, p=0.81); the label roughly doubled treatment success (31.6% vs 14.1%, OR 2.84, 1.52–5.34, p=0.001) (Rexwinkel 2025, PMID 40074185). See antispasmodics-and-peppermint.
  2. Blinded dietary reintroduction. In a 9-week blinded FODMAP reintroduction, the inert glucose control triggered symptom recurrence in 26% of exposures — against fructans at 56% and mannitol at 54% (Van den Houte 2024, PMID 38401741). Roughly a quarter of "identified food triggers" would be reproduced by an inert powder. See dietary-therapy.

What this does to the rest of the evidence base

Where it bites Consequence
Drug effect sizes Placebo rates of 10.1–32.2% across the licensed-drug trials in this knowledge base mean identical drugs can show very different NNTs depending on the comparator arm's behaviour (constipation-predominant-pharmacotherapy, diarrhoea-predominant-pharmacotherapy)
Unblindable interventions Diet and psychological therapy cannot be blinded; their networks show high risk of bias and funnel-plot asymmetry, and their authors say effects are "likely to have been overestimated" (Black 2020, PMID 32276950; Thakur 2025, PMID 41077057)
Failed primary endpoints Peppermint oil missed both co-primary regulatory endpoints against a 34.4% placebo response while improving continuous secondaries (Weerts 2020, PMID 31470006)
Historical comparisons Pre-2006 trials had higher placebo responses (Bosman 2021, PMID 33765447), so older drugs' apparent effects are not comparable to modern ones
Clinical practice An unhurried, warm consultation increased adequate relief from 44% to 62% within a placebo-acupuncture trial (Kaptchuk 2008, PMID 18390493); this demonstrates a contextual effect but is not a direct comparison with an active drug

A note on endpoint design

The FDA composite for IBS-C requires ≥30% reduction in worst abdominal pain and an increase of ≥1 complete spontaneous bowel movement per week, in the same week, for ≥6 of 12 weeks; the IBS-D composite pairs pain reduction with a stool-consistency criterion. These composites were introduced precisely to suppress placebo response, and they succeed (17.9% pooled vs 27.3% for global improvement; Bosman 2021, PMID 33765447). The cost is interpretive: a composite responder analysis discards information about patients who improved on one dimension only, and the relationship between "responder" status and clinically meaningful change has been questioned directly using linaclotide phase 3 data (Lacy 2014, PMID 24382134).

Open questions

  • Should open-label placebo be offered as treatment for refractory IBS? Three positive randomised trials exist (Kaptchuk 2010, PMID 21203519; Lembo 2021, PMID 33605656; Nurko 2022, PMID 35099543), while the guideline registry audited on 2026-09-02 contained no recommendation addressing it.
  • If OLP and double-blind placebo are equivalent (d=0.10; Lembo 2021, PMID 33605656), what exactly does blinding control for in this condition?
  • Can the practitioner-relationship effect be delivered deliberately and reimbursed? The enhanced-relationship arm produced 62% adequate relief versus 28% with waiting list (Kaptchuk 2008, PMID 18390493), but cross-trial comparison with drug effects cannot establish comparative efficacy.
  • What is the right responder threshold? Tightening from 6/12 to 9/12 weeks cuts placebo composite response from 18.9% to 4.3% (Barberio 2022, PMID 34425274) — but also cuts drug response, and no analysis has established which threshold best predicts patient-valued outcomes.
  • Is the placebo response in IBS a property of the condition or of its measurement? Both readings are consistent with the moderator data (Bosman 2021, PMID 33765447; Ford 2010, PMID 20412064).
  • Does high visceral sensitivity really predict opposite responses to OLP and to no treatment (Ballou 2022, PMID 35412513)? A single secondary analysis.
  • Which IBS-SSS change is clinically meaningful — the 50 points validated in 1997 (Francis 1997, PMID 9146781) or the 35 points ATLANTIS prespecified (Ford 2023, PMID 37858323)?
  • Should regulators accept symptom-specific continuous endpoints alongside the composite, given that a 12–17-point drug–placebo gap persists among FDA non-responders (Lacy 2014, PMID 24382134)?
  • How much of the dietary-trigger literature is nocebo? The blinded glucose control triggered 26% of exposures (Van den Houte 2024, PMID 38401741).

References

  1. Ford AC, Moayyedi P. Meta-analysis: factors affecting placebo response rate in the irritable bowel syndrome. Aliment Pharmacol Ther. 2010;32(2):144-58. PMID 20412064
  2. Bosman M, et al. The placebo response rate in pharmacological trials in patients with irritable bowel syndrome: a systematic review and meta-analysis. Lancet Gastroenterol Hepatol. 2021;6(6):459-473. PMID 33765447
  3. Barberio B, Savarino EV, Black CJ, Ford AC. Placebo Response Rates in Trials of Licensed Drugs for Irritable Bowel Syndrome With Constipation or Diarrhea: Meta-analysis. Clin Gastroenterol Hepatol. 2022;20(5):e923-e944. PMID 34425274
  4. Li R, et al. Nocebo response intensity and influencing factors in the randomized clinical trials of irritable bowel syndrome: A systematic review and meta-analysis. Front Med (Lausanne). 2022;9:1018713. PMID 36606047
  5. Kaptchuk TJ, et al. Components of placebo effect: randomised controlled trial in patients with irritable bowel syndrome. BMJ. 2008;336(7651):999-1003. PMID 18390493
  6. Kaptchuk TJ, et al. Placebos without deception: a randomized controlled trial in irritable bowel syndrome. PLoS One. 2010;5(12):e15591. PMID 21203519
  7. Lembo A, et al. Open-label placebo vs double-blind placebo for irritable bowel syndrome: a randomized clinical trial. Pain. 2021;162(9):2428-2435. PMID 33605656
  8. Nurko S, et al. Effect of Open-label Placebo on Children and Adolescents With Functional Abdominal Pain or Irritable Bowel Syndrome: A Randomized Clinical Trial. JAMA Pediatr. 2022;176(4):349-356. PMID 35099543
  9. Ballou S, et al. Psychological Predictors of Response to Open-Label Versus Double-Blind Placebo in a Randomized Controlled Trial in Irritable Bowel Syndrome. Psychosom Med. 2022;84(6):738-746. PMID 35412513
  10. Wang RS, et al. Genomic Effects Associated With Response to Placebo Treatment in a Randomized Trial of Irritable Bowel Syndrome. Front Pain Res (Lausanne). 2021;2:775386. PMID 35295415
  11. Vollert J, et al. Genotypes of Pain and Analgesia in a Randomized Trial of Irritable Bowel Syndrome. Front Psychiatry. 2022;13:842030. PMID 35401282
  12. Lacy BE, et al. Responders vs clinical response: a critical analysis of data from linaclotide phase 3 clinical trials in IBS-C. Neurogastroenterol Motil. 2014;26(3):326-33. PMID 24382134
  13. Rexwinkel R, et al. Mebeverine and the Influence of Labeling in Adolescents With Irritable Bowel Syndrome or Functional Abdominal Pain Not Otherwise Specified. Gastroenterology. 2025;169(1):94-103. PMID 40074185
  14. Van den Houte K, et al. Efficacy and Findings of a Blinded Randomized Reintroduction Phase for the Low FODMAP Diet in Irritable Bowel Syndrome. Gastroenterology. 2024;167(2):333-342. PMID 38401741
  15. Ford AC, et al. Amitriptyline at Low-Dose and Titrated for Irritable Bowel Syndrome as Second-Line Treatment in primary care (ATLANTIS). Lancet. 2023;402(10414):1773-1785. PMID 37858323
  16. Busam JA, et al. The Safety of Pharmacotherapy for Irritable Bowel Syndrome: A Systematic Review and Meta-Analysis. Am J Gastroenterol. 2026;121(3):745-753. PMID 40471839
  17. Black CJ, Thakur ER, Houghton LA, Quigley EMM, Moayyedi P, Ford AC. Efficacy of psychological therapies for irritable bowel syndrome: systematic review and network meta-analysis. Gut. 2020;69(8):1441-1451. PMID 32276950
  18. Thakur ER, Khasawneh M, Moayyedi P, Black CJ, Ford AC. Efficacy of behavioural therapies for irritable bowel syndrome: a systematic review and network meta-analysis. Lancet Gastroenterol Hepatol. 2025;10(12):1075-1088. PMID 41077057
  19. Weerts ZZRM, et al. Efficacy and Safety of Peppermint Oil in a Randomized, Double-Blind Trial of Patients With Irritable Bowel Syndrome. Gastroenterology. 2020;158(1):123-136. PMID 31470006
  20. Francis CY, Morris J, Whorwell PJ. The irritable bowel severity scoring system: a simple method of monitoring irritable bowel syndrome and its progress. Aliment Pharmacol Ther. 1997;11(2):395-402. PMID 9146781