Jan 14

Placebos for irritable bowel syndrome

Journal reference: Kaptchuk TJ, Friedlander E, Kelley JM, et al. Placebos without deception: a randomized controlled trial in irritable bowel syndrome. PLoS ONE 2010; 5(12): e15591 [1]

Link: http://dx.doi.org/10.1371/journal.pone.0015591

Evidence cookie says...

The positive results widely reported for placebo pills in consenting patients with irritable bowel syndrome are likely invalid due to serious methodological limitations in the study.

This study does not have any direct bearing on the practice of community medicine by Australian general practitioners.

Placebo therapies in consenting patients should be avoided; concealed or deceptive placebo therapies are unethical.

Note: see below for more details.

More details:


Article details


Study design:

open-label randomised controlled trial


Study aim:

to test whether open-label placebo (non-deceptive and non-concealed administration) is superior to no-treatment control with matched patient-provider interactions in the treatment of irritable bowel syndrome (IBS)


Methods summary:

  • single academic medical centre
  • August 2009 to April 2010
  • randomised to two groups
    • placebo pill twice daily
    • no-treatment
  • three week trial
  • all visits were in the context of a “warm supportive patient-practitioner relationship”

Recruitment:

  • advertisements for “a novel mind-body management study of IBS” in newspapers and fliers and from referrals from healthcare professionals
  • potential participants screened by telephone
  • were told they would receive “either placebo (inert) pills, which were like sugar pills which had been shown to have self-healing properties” or no treatment.

Eligibility:

  • adults, ≥ 18 years of age
  • meeting the Rome III criteria for IBS
  • IBS symptom severity score (IBS-SSS) ≥ 150
  • diagnosis of IBS
    • based on typical symptoms and exclusion of alarm symptoms
    • confirmed by a specialist gastroenterologist (Lembo AJ, last author) or nurse practitioner experienced in functional bowel disorders (Friedlander E, second author)

Exclusion:

  • patients with unexplained alarm features:
    • weight loss > 10% body weight
    • fevers
    • blood in stools
  • family history of colon cancer
  • inflammatory bowel disease
  • history of pelvic floor dyssynergia
  • need for manual manoeuvres to achieve bowel movement
  • history of surgery of the colon
  • abdominal surgery within 60 days
  • any surgery within 30 days
  • pregnancy or breastfeeding
  • history of laxative abuse
  • “other medical conditions (e.g., neurological disorders, metabolic disorders, or other significant disease)”
  • “pretreatment laboratory or ECG findings believed to impair their ability to participate in the study were also excluded”

Interventions:

  • prior to randomisation:
    • patients met with either the gastroenterologist or nurse practitioner who explained the following four “discussion points”:
      • “the placebo effect is powerful”
      • “the body can automatically respond to taking placebo pills like Pavlov’s dogs who salivated when they heard a bell”
      • “a positive attitude helps but is not necessary”
      • “taking the pills faithfully is critical”
  • placebo pill group
    • given placebo pills marked as “placebo pills” in a typical presentation medicine bottle
    • at the midpoint (day 11) visit, patients received a short reminder regarding the “four discussion points”
  • no-treatment group
    • at the point of randomisation, patients who were allocated to the no-treatment group were “reminded of the importance of the control arm”
    • at the midpoint visit, patients  were “encouraged and thanked for helping make a successful study”

Assessments:

  • primary outcome measure:
    • IBS Global Improvement Scale (IBS-GIS)
    • this is a self assessed questionnaire
    • 7-point Likert scale
  • other measures:
    • various self assessed IBS required scales
  • at end of study:
    • short qualitative open-ended questionnaire

Statistical analysis:


Results summary:

  • 92 patients screened and 80 eligible patients randomised
    • 37 into open-label placebo → 6 patients did not complete study (1 adverse event + 4 drop outs)
    • 43 into no-treatment → 4 patients did not complete study (3 drop outs)
  • missing outcome data for 13 patients (16%) at midpoint and 10 (13%) patients at endpoint

Primary outcome:

  • IBS-GIS (mean ± standard deviation) at 3 weeks:
    • placebo pill: 5.2 ± 1.5, vs.
    • no-treatment: 3.9 ± 1.36, P = 0.002

Qualitative check out questionnaire:

  • placebo pill group, “What did you think was in the placebo pills?”
    • of the 29 (of 31) who responded, 5 “did not know” and 1 “possible test medication”, i.e., ~ 20% were not clearly aware that the pill was an inert placebo
  • no-treatment group, “Were you disappointed to be in the treatment as usual arm?”
    • of the 38 (of 39) who responded, 9 (~ 24%) said “yes” or “a little”.

Study conclusion:

Placebos administered without deception may be an effective treatment for IBS. Further research is warranted in IBS, and perhaps other conditions, to elucidate whether physicians can benefit patients using placebos consistent with informed consent.


Participants:

  • single academic centre
  • responded to advertisements
  • adults, ≥ 18 years of age
  • meeting the Rome III criteria for IBS
  • IBS symptom severity score (IBS-SSS) ≥ 150
  • diagnosis of IBS
    • based on typical symptoms and exclusion of alarm symptoms
    • confirmed by a specialist gastroenterologist (Lembo AJ, last author) or nurse practitioner experienced in functional bowel disorders (Friedlander E, second author)
  • substantial list of exclusions
  • mean age 46 years, SD 18 years
  • most patients were white women

Differences in baseline characteristics between the two study groups:

  • diarrhoea predominant IBS: no-treatment group (37%) > placebo pills group (27%)
  • constipation predominant IBS: no-treatment group (33%) < placebo pills group (44%)
  • upper GI symptoms: no-treatment group (42%) > placebo pills group (30%)
  • taking medications for IBS: no-treatment group (35%) < placebo pills group (54%)
  • taking antidepressants: no-treatment group (16%) < placebo pills group (24%)

Methodological weaknesses

There are very substantial methodological problems with this study.

  • firstly, the sample size is small and the study duration is very short (3 weeks)

It is likely that there is a substantial recruitment bias given that the advertisement stated that it was “a novel mind-body management study of IBS”:

  • it is likely that the study recruited patients who would respond particularly positively to unconventional therapy
  • moreover, the exclusion criteria seems to contain a completely arbitrary component of “other medical conditions (… other significant disease)”
  • this in itself is a serious threat to the external validity of the study

The study wanted to test the administration of a placebo pill in a non-deceptive and non-concealed manner but:

  • strong positive statements such as “the placebo effect is powerful” and “placebo pills … have been shown in rigorous clinical testing to produce significant mind-body self-healing processes” are misleading and deceptive insofar that it is inconsistent with best evidence [2]
  • of the participants in the placebo pill group who answered the qualitative check out questionnaire, 6 of 29 (20%) were not clearly aware that the placebo pill contained only an inert substance
  • the study is arguably not intentionally valid; the study was not adequately non-deceptive or non-concealed in the handling of the placebo pills

Despite randomisation, there are a number of notable and clinically significant differences between the groups:

  • IBS types were different:
    • no-treatment group patients were more likely to have diarrhoea predominant disease
    • placebo pill group patients were the opposite, they were more likely to have constipation predominant disease
  • placebo pill group patients were much more likely to be taking medications for IBS and were more likely to be taking antidepressants
    • the placebo pill group may be expected to have a better response to a supportive patient-practitioner relationship
  • the placebo pill group had a greater proportion of participants who did not complete the study (16% vs 9%)
    • this is likely to bias results towards being more favourable to the placebo pill group
  • these are all problematic biases and threats to internal validity

Differences between treatment received between groups apart from the studied intervention:

  • although the authors claim that the no-treatment group received “the same frequency and duration of contact time and the content of the interaction was very similar”, this is clearly not the case
    • the placebo pill group received a structured reminder about the discussion points of placebo treatments in the review midway in the study
    • the no-treatment group did not receive a similar alternative structured discussion in their midway review
  • almost a quarter of the participants in the no-treatment group who answered the check out questionnaire were disappointed, or a little disappointed to have been in that group

The trial as described is focussed on the placebo pill group and neither the patients or physicians are blinded:

  • the Hawthorne effect is likely (participant improvement simply due to being in a study)
  • the observer-expectancy effect is likely (researcher unconsciously influencing the participants in the study)
  • the Rosenthal effect is possible (participants perform better due to greater expectation placed upon them)
  • all are serious threats to study validity

Self reported measures:

  • there is a question of the construct validity of using a self-reported questionnaire; it may be that the scale simply measured social desirability
  • the participants may report improvements in symptoms as a reward to the researchers to whom they developed a warm and supportive relationship when in fact there has been no clinical improvement

Methodological strengths

  • novel pilot study; attempted to study and measure placebo treatment as compared no-treatment
  • randomised design
  • power calculation made a priori
  • measured potential confounding factors

Biases and conflicts of interests

  • study was partially funded by the National Center for Complementary and Alternative Medicine
  • study was partially funded by a gift from the Bernard Osher Foundation; this has an interest in integrative medicine (complementary and alternative medicine)
  • Kaptchuk TJ (primary author) is a doctor of Oriental medicine and has an interest in complementary and alternative medicine.  He has extensively studied the placebo effect (and is thus a beneficiary if positive results are reported).

Clinical relevance to primary health care

The study authors acknowledge that this study is small and a “proof-of-principle” pilot study, but nevertheless suggest in their conclusion that placebos given without deception may be an effective treatment in irritable bowel syndrome (IBS). The study was widely reported in the popular news media with little critical review or analysis.

However, the study is beset with numerous methodological problems that seriously threatens the validity of (and in this author’s opinion, invalidates) the results. There are:

  • substantial recruitment biases
  • substantial differences in the baseline characteristics of the groups after randomisation
  • probable substantial differences in the treatment of the two groups other than in the studied intervention
  • differences in drop out between the two groups
  • note: see section on “methodological weaknesses” above for elaboration

The use of a self-reported questionnaire as the primary outcome measure raises the spectre that the investigators were measuring social desirability rather than objective improvements in IBS symptoms. Furthermore, it can be argued that despite the investigators’ intention, placebo pill treatment was neither free from deception (positive statements for placebo treatments given to the participants were exaggerated and cannot be objectively supported by the evidence base), or concealment (about 20% of participants in the placebo pill group at the end of the trial were not clearly aware that the pill contained only a therapeutically inert substance).

The most recent update on the Cochrane Review (systematic review and meta-analysis) on placebo interventions concluded that they do not have important clinical effects in general [2]; and although placebo interventions can influence patient-reported outcomes it was difficult to distinguish this from biased reporting.

This study does not have any direct bearing on the practice of community medicine by Australian general practitioners. Placebo therapies in consenting patients should be avoided; concealed or deceptive placebo therapies are unethical.

References

  1. Kaptchuk TJ, Friedlander E, Kelley JM, et al. Placebos without deception: a randomized controlled trial in irritable bowel syndrome. PLoS ONE 2010; 5(12): e15591
  2. Hróbjartsson A, Gøtzsche PC. Placebo interventions for all clinical conditions. Cochrane Database of Systematic Reviews 2010, Issue 1. Art. No.: CD003974
Print Friendly, PDF & Email

Leave a Reply

%d bloggers like this: