The Problem of Placebo Responses in Clinical Trials and How To Combat Them

Randomised controlled, double-blind trials are considered gold-standard study designs. In these studies, the active intervention is often compared to a placebo condition. Ideally, in pharmaceutical and nutraceutical/herbal trials, these placebos are matched for appearance, smell, and taste.

A common phenomenon occurring in such gold-standard studies is the ‘placebo response’. This is where participants are placed on a placebo and experience symptomatic improvement. This placebo response is partly due to improvements that occur with the passage of time. However, the placebo response is more than just transient changes that occur over time. Placebo responses are often substantial and can sustain over time. The placebo response is so pervasive and occurs across so many conditions that it would be reasonable to question the validity of a study if a placebo response did not happen!

Unfortunately, a large placebo response can be an important factor contributing to the failure of many clinical trials. In the area of mental health, depression-related trials have placebo responses of around 20 to 30%, with some trials where there is a 50% improvement in symptoms.

There has been significant work in trying to understand the placebo response, and researchers have tried to identify strategies to reduce or control for it. The goal is that if the placebo response is reduced, it will help to determine the true effects of the active treatment. If we think about it in basic terms, if there was a placebo response of 1% improvement, then a 5% improvement in the active treatment means that it is 5 times more effective. However, if there is a 20% improvement in the placebo group, to achieve a 5 times better outcome, the active treatment would need 100% improvement!

Several factors may affect the placebo response. Some of these include:

• Treatment expectancies. Research confirms that people with more positive expectations about a treatment generally experience larger placebo responses.
• Positive researcher-participant interactions. In some trials, there is regular interaction with researchers, who typically try to ensure pleasant interactions. Some people may also complete tasks that are novel and interesting. This can affect the placebo response, particularly in people who may be experiencing mood disturbances and have limited social interactions.
• Behavioural changes during the study. If people participate in a trial to help improve their physical or mental condition, they may become more inclined to improve their overall health. This may lead to dietary, lifestyle, social, and environmental changes that positively affect their wellbeing. It is also common for many participants to start a new treatment during the study, which will majorly affect the results.

To help combat the placebo response, several strategies could be utilised in a clinical trial. Some options include:

• Identifying and utilising biomarkers that provide a measure of clinical improvement. Although biomarkers (e.g., from blood, saliva, urine, and other physiological measures) are also subject to placebo responses, they seem to be associated with a lower placebo response.
• Using self-report and assessor-rated instruments. Overall, assessor-rated instruments seem to be associated with a lower placebo response.
• Using placebo run-in phases. This is where everyone is initially placed on the placebo and then randomised to take either the placebo or active treatment. However, research indicates this strategy’s ability to reduce the placebo response is inconsistent.
• Withdrawing placebo-responders. In some studies, people who experience a significant improvement in their symptoms during the placebo/ lead-in phase are withdrawn from treatment.
• Utilising longer follow-up periods. Sometimes, having a longer treatment period can result in the effects of the placebo diminishing over time.
• Measuring and controlling for expectancies. Assessments can be administered to assess for treatment expectancies at baseline. This information can be utilised to help control for the placebo response.
• Conducting personalised intervention trials. It is unlikely that a treatment will be effective for everyone experiencing a specific condition. Therefore, identifying and recruiting participants who may experience larger treatment effects can be a useful strategy. For example, in a depression trial, it seems logical to assume that people with high inflammation are more likely to experience greater benefit from a treatment that has significant anti-inflammatory effects.
• Utilsing sensitive outcome measures. Some outcome measures may be reliable and valid but lack sensitivity in detecting change. Therefore, utilising measures that are more sensitive to change may help delineate treatment-related effects. For example, it is commonplace for many questionnaires to ask about symptoms over the last 4 weeks. Many people struggle with accurately responding to such questions and may actually give responses based on how they feel now (or over the last few days) rather than the average symptom severity over the last 4 weeks.
• Utilising experimental procedures that can be utilised in controlled environments. Often, in animal studies, a treatment may have shown promise. However, when the human trial is conducted, disappointment results. This may be partly because animal trials are conducted in controlled environments. However, this is certainly not the case in human trials. Life gets in the way and is rarely constant. Even though we can’t place people in a bubble during the trial period, some outcome measures can be administered in controlled settings. For example, if you are interested in identifying whether a treatment affects the stress response, people could be exposed to an experimental stressor in a controlled environment, and their stress response could be measured before, during and after this stressor.

Although this list is not exhaustive, I have outlined a selection of options that researchers have available to help reduce the placebo response and identify the true effects of an active treatment.

If you are interested in conducting a clinical trial on a herbal or nutritional ingredient/ product, do not hesitate to contact us.