Randomized controlled trial

A randomized controlled trial or RCT is a clinical treatment trial with at least two treatment groups, one receives the treatment being tested and the other group is either allocated to a waiting list or receives a placebo, or sham therapy, or alternative treatment. Participants are allocated to a treatment group randomly rather than being able to choose whether to undergo the treatment being tested or the alternative. Treatment groups are known as arms.

Theory

Well-conducted RCTs or systematic reviews of RCTs are considered the highest standard of treatment evidence.

Types

• Unblinded: participants know whether the treatment is a placebo / sham or the one being tested, which introduces possible bias.

This is particularly important when a trial's success is measured by subjective outcomes like participant or clinician questionnaires rather than independent measurable outcomes like employment rates, physiological measures (e.g., distance walked, heart rate, muscle strength, blood test results, hospital admissions).

• Blinded (single blinded): participants do not know whether the treatment is a placebo / sham or the one being tested, but the researchers / clinicians delivering or assessing the treatment do know. Bias from clinicians may affect results, and may have an effect on patient reports, e.g., positive or negative clinician feedback, and clinician encouragement.

• Double-blinded: participants and clinicians / researchers do not know whether the treatment is a placebo / sham or the one being tested, the researcher analyzing the trial data is also not aware which group received which intervention.

Comparison

Other types of treatment evidence include:

• Case studies: The treatment of one or more patients is described, typically without any original intention for the results to form part of research. These commonly identify treatments that may be worth further investigation, or diagnostic issues. In rare diseases, most evidence is in the form of case studies or expert consensus due to the very small number of patients available for trials.
• Case series: A group of patients are reported on who share the same or similar characteristics, or who receive the same diagnosis or treatments. For example, outcomes of a group of 100 consecutive patients admitted to hospital ICU or outcomes of 12 patients receiving surgery for craniocervical instability.
• Randomized trial: There are two or more treatment arms, which participants are randomly allocated to. Without a control group with a placebo or sham treatment for comparison, it is possible to determine which treatment is most effective but it is not be possible to determine if both are ineffective, harmful, or effective.
• Expert consensus: A selected group of clinicians discuss or agree on diagnostic and/or treatment guidelines based on their personal experiences. Clinicians may be representatives of particular organisations or members of a particular professional organization. Examples include the Italian consensus for multiple chemical sensitivity.
• Historical use or historical evidence: A treatment that has been used for an extended period of time and prior to modern medicine systems may have been assessed as safe and/or effective based on considerable historic anecdotal evidence. This type of evidence commonly applies to certain alternative medicines or traditional treatments.
• Patient surveys: These are typically disregarded in evidence-based medicine, but can provide significant information on the effectiveness of real-world treatments that may differ from the results of RCTs, or on treatments that may be worth further funding RCTs for. Large surveys are commonly conducted by a number of ME/CFS patient groups. In the case of graded exercise therapy, patient surveys consistently reported both very high rates of harm and a lack of effectiveness which was not found by most RCTs.
• Patient evidence: Individual patients are interviewed or asked to provide information based on their personal experiences.

Limitations

It should be noted that the perfect RCT has yet to be undertaken. — BMJ Best Practice Guide

Blinding

Double-blinded studies are typical in medication trials because it is easy to provide placebo medications that look like the intervention being tested.

For some type of treatments, it may not be possible to blind participants to the intervention, for example in psychotherapy, counseling, or behavioral treatments such as cognitive behavioral therapy or graded exercise therapy.

Methodology

RCTs may have a variety of problems with methodology, that can affect the outcomes reported.

Weighting

RCTs are not automatically of equal standing, and systematic reviews assign a weighting to RCTs, which effectively ranks some RCTs as more significant than others.

The GRADE framework is a standardised way developed to compare different RCTs by assessing a number of factors from the trials, in order to provide an independent, systematic review of exercise. Both Cochrane and the systematic reviews created for the UK's NICE guidelines use the GRADE framework.

The largest RCT of ME/CFS treatments to date is the highly controversial PACE trial. The PACE trial had significantly more participants than any other RCTs of graded exercise therapy and cognitive behavioral therapy, for example the FINE trial tested the same treatments in severely ill patients but had far less participants. The FINE trial reported a null effect for the treatments, but the PACE trial reported a moderately positive result. The Cochrane systematic review of exercise as a therapy for CFS assigned a higher weight to the PACE trial results compared to the smaller FINE trial.

Criticism

Several RCTs of the same treatment may produce very different results. For example, the PACE trial and it's sister trial the FINE trial tested the same interventions, one rating CBT and GET as moderately effective and adaptive pacing therapy as ineffective, and the other weighting CBT and GET as ineffective.

RCTs are a type of treatment trial, and are not able to provide the sum of all evidence of the effectiveness of different treatment interventions. Outcomes from RCTs may differ significantly from real world results once a treatment is in use, with some successful RCTs treatments leading to either ineffective or harmful treatment reports.

RCTs require organization, patient recruitment, and funding, which may not be available for some types of treatment or for some illnesses or treatments.

Notable studies

• 2018, The limitations of using randomised controlled trials as a basis for developing treatment guidelines (Abstract)

See also

• Research bias in ME/CFS
• Intimidation and bullying of PACE trial critics
• Research terminology (category)
• Subjective outcome
• PACE trial
• FINE trial