June 6, 2019


Daniel Levine:               Leading a startup team. Whether you’re delivering a sugar rush, stocking coffee, or getting a regular delivery of snacks, Office Depot has solutions that fit every startup culture, from getting those first business cards and stationery, to ordering fleece pullovers with your new logo. To learn how Office Depot and the California Technology Council have partnered to bring you savings on all of these startup essentials and more, go to

Daniel Levine:               I’m Daniel Levine, and this is The Bio Report. The placebo response in clinical trials can derail a promising experimental therapy that might benefit patients. In cases where trials rely on subjective endpoints or patient-reported outcomes, the placebo response can be more pronounced. Tools4Patient has developed a means of identifying patients who are likely to be placebo responders, and allow trial sponsors to take steps to account for that in clinical studies. We spoke to Erica Smith, vice president of business development at Tools4Patient about the placebo response, the consequence this phenomenon has on drug development, and how the company is seeking to address this problem. Erica, thanks for joining us.

Erica Smith:                  Thanks so much for having me.

Daniel Levine:               We’re going to talk about clinical trials, the placebo effect and your company, Tools4Patients. Let’s start with the placebo effect in trials though. How problematic is this for companies developing new therapies? How often does a high placebo response derail a potentially promising therapy?

Erica Smith:                  Absolutely. That’s a great question. We all know, everyone that’s developing drugs recognizes that the placebo response or the phenomenon which patients have a positive response to a dummy or sham treatment is a very serious issue in drug development. Currently double-blind, randomized placebo-controlled trials are the gold standard for evaluating drug efficacy. It’s very difficult for drugs to obtain regulatory approval unless they can show superiority over a placebo treatment.

Erica Smith:                  The complexity of the placebo effect is profound, and it makes it very difficult to actually evaluate true drug efficacy. The effect size which is the size of the drug effect relative to the size of the placebo effect is really the critical issue. We know that the placebo effect, for example, is increasing over time. This is reducing the effect size in diseases like pain, depression and psychiatric disorders.

Erica Smith:                  This reduction in the effect size over time results in increased difficulty to show efficacy even for drugs that are efficacious. It increases clinical trial size. It increases the timeline of clinical drug development and the risk, and all of these things increase the expense of trials. Importantly, it delays patient access to new drugs, and in many therapeutic areas where those drugs are very needed. It’s a very substantial issue.

Erica Smith:                  It’s been estimated in the last 10 years or so in pain particularly that about 90% of drugs have failed late-stage clinical trials in part as a response to the placebo effect. It’s quite a significant issue.

Daniel Levine:               Would that also suggest that a placebo effect could be getting drugs that really aren’t efficacious to market?

Erica Smith:                  The problem is actually the opposite. The placebo effect makes it so difficult to show efficacy that it’s preventing drugs that are efficacious from reaching the market. What it’s doing is it’s obscuring the true efficacy of the drug. Therefore, making it more difficult for companies to actually be able to demonstrate that to the regulatory bodies.

Daniel Levine:               You’ve developed a tool called Placebell that is designed for drug developers to identify placebo responders and on a patient-by-patient basis to determine who might be most likely to have a placebo effect. What makes one patient more likely to have a placebo effect than another?

Erica Smith:                  That’s a really good question. Essentially we need to understand a couple of things about the placebo response. The first is that it’s a psychobiological phenomenon. It involves both the psychology of the patient, so part of that is their personality. It involves the biology of the patient, so we know, for example, using fMRI and PET studies that patients that receive placebo have certain areas in their brain that are activated that can result in release of certain neurotransmitters, for example. There is a true biological response. There are genetic components.

Erica Smith:                  All of these things are components of the individual patient. On a population basis, we know that things like age and gender and patient geography, which can also infer cultural influences will impact the placebo response. There’s all these different factors that taken together can describe which patient may be a high placebo responder. The tool that we’ve developed at Tools4Patient is the culmination of several years, about five years now of research into the placebo effect.

Erica Smith:                  It really aims to identify those factors in each individual patient that gives it a high propensity to be a high placebo responder. Things, like I mentioned, demographics, age and gender, geography, the medical history of the patient. For example, how long they’ve had a specific disease, how many drugs they’re taking for the disease. The intensity of the disease at baseline, so at the start of the clinical trial. Very importantly, we’re one of the first groups to really be able to systematically integrate personality traits into this approach.

Erica Smith:                  All these things are combined and weighted in different ways using our proprietary algorithm to come up with a score that describes each patient’s likelihood of having a strong placebo response in the trial.

Daniel Levine:               How does it work? What’s the process you do for a clinical trial sponsor?

Erica Smith:                  What we’re trying to target with Placebell is actually the variability in data that the placebo effect brings to clinical trial data. One of the issues with the placebo effect is because each patient is unique, each patient has a very unique response to placebo, and this is an issue for both patients in a placebo group as well as patients that are treated with active drug. Some component of the response in patients treated with active drug is actually attributed to the placebo response.

Erica Smith:                  The underlying issue, if every patient responded to placebo in the same way, it would be much easier. The underlying issue is that this is a main driver of variability in clinical trial data. The variability in the data makes it very difficult to distinguish the difference between the placebo group and the drug treated group. The way our tool works, at the outset, the team that developed the tool wanted to make sure it was as easy to execute and had a minimal burden on the clinical trial team and the clinical site staff.

Erica Smith:                  Essentially, all we do is submit the patients to a personality questionnaire. Takes about 35 to 40 minutes to complete. Very straightforward and it assesses multiple different personality traits in the patient. We collect that data from the patients as well as some standard patient data that would be collected in the trial anyway, things like I mentioned, the patient demographics, the patient medical history. These things are combined into our proprietary algorithm that is calibrated on a disease-by-disease basis.

Erica Smith:                  The algorithm then outputs a single number for each patient called the Placebell covariate. That covariate is then used as any baseline covariate could be used in the ultimate statistical analysis. It aims to reduce the variability in the data due to the placebo response, and improve the ability to demonstrate efficacy, and ultimately de-risk the trial and the clinical development process.

Daniel Levine:               Once you identify people who are more likely to respond to a placebo, what do you do with them? Do you exclude them or do you make sure they’re evenly distributed among the different arms of a trial?

Erica Smith:                  That’s a very good question. Really these data can be used in a variety of ways. Any covariate that’s calculated in a clinical trial can be used for any of those things, for patient selection, for patient stratification, and for data adjustment. I’ll talk about each of those individually. Theoretically the data could be used for patient selection or to include or exclude patients in a trial based on the outcome of our assessment.

Erica Smith:                  That may or may not be advised depending on the strategy of the program and the stage of development. There are a couple of issues with that. First, there is the possibility in many diseases that the patients that are high placebo responders may also be high drug responders. The converse is also true. Patients that do not respond well to placebo, often do not respond well to drugs. While you might be tempted to exclude placebo responders as a means to increase the effect size, that might backfire and you actually might decrease the effect size, because you’re losing some of your strong drug responders as well.

Erica Smith:                  The regulators, once you start limiting the population in which you’re testing a drug, the regulators for pivotal trials may impose a limitation on the label of the drug. There may be labeling issues down the line. As well, anytime you exclude any large population, it makes it more difficult to recruit patients for trials. The placebo responders in any given even population could be 50% of the entire population. If you exclude them, it’s going to lengthen the length of the trial, and make it, again, more expensive, and delay access to patients.

Erica Smith:                  The tool that we have developed, Placebell, can be used to adjust data, which is a way to account for some of the high placebo responders that actually allows you to leave them in the trial. As I mentioned, it can also be used for patient stratification. Could be a very viable tool particularly in smaller patient populations or in early stage trials to make sure there are an equal number of placebo responders in each study arm of the trial.

Erica Smith:                  Really the mechanism or the way in which we propose to use these data would be to adjust the data at the end of the trial. It’s essentially a statistical tool to reduce the variance that the placebo response adds to the data, and improve the ability to show treatment efficacy.

Daniel Levine:               Are there indications or classes of drugs where there’s a greater risk of a placebo response?

Erica Smith:                  Absolutely. Absolutely. Classically, one of the therapeutic areas that has been most plagued by the placebo response has been pain. There are obviously many different types of pain, acute pain, chronic pain, arthritis pain, neuropathic pain. Ultimately, pain is one area where the placebo response has made it very difficult to show treatment efficacy. That’s one area that we’ve done quite a bit of work in currently.

Erica Smith:                  Also, in any neurologic diseases, Parkinson’s disease, for example, I mentioned psychiatric disorders, things like schizophrenia and depression, but there are also other diseases. Really anything that has a subjective endpoint or a patient-reported outcome will tend to have a strong placebo response, and anything that has a sensory endpoint, so dryness, for example, dry eye disease, itching, so atopic dermatitis. Things like that, anything where there’s a sensory endpoint that the patients are evaluating the response of that to any sort of treatment is really a good candidate for use of Placebell.

Erica Smith:                  We’ve explored its utility in therapeutic areas and disease states across the board.

Daniel Levine:               There’s a big push towards using patient-reported outcomes right now. Does that become a bigger challenge, or does that necessitate using the tool in a different way?

Erica Smith:                  I think the tool is very amenable to being used for any type of patient-reported outcomes. You’re absolutely correct. We believe that this is going to become more and more common particularly as part of the 21st Century Cures Act, where companies running trials are encouraged to focus and emphasize the endpoints that are most meaningful to patients. Often those endpoints are things like pain, or things like quality of life, or things like ability to carry on your activities of daily living, which are often assessed as a patient-reported outcome.

Erica Smith:                  These are the types of things where Placebell, we think, can really help to minimize data variability. We suspect that this is going to become a bigger, bigger problem as time goes on.

Daniel Levine:               What impact does this tool have on statistical significance?

Erica Smith:                  That’s a really good question. What our data has shown currently in pain is that we can explain about 30% of the variability in clinical trial data due to the placebo response. As I mentioned, that applies to both the active treated group and the placebo treated group. That’s quite a significant amount of reduction in variability. What that means is that it increases clinical trial power. In general, a 30% reduction in variance translates to about a 12% to 14% increase in power.

Erica Smith:                  If you have a trial that’s powered, say, at 0.7, it increases the power to 0.82, 0.84. Another way to look at it is it has the same effect, or it gives the trial the same power as a trial that has about 30% to 40% more patients. What I mean by that is if you have a trial that is 100 patients with a given statistical power, using Placebell makes that 100 patients seem like it’s 130 or 140 patients. Because it increases the power of the trial to that level.

Erica Smith:                  From a statistical perspective, one of the really important components of Placebell is it decreases the risk of type two error. What I mean by that is it decreases the likelihood that a company will reject a drug and not pursue it if that drug is truly efficacious. It does so without increasing the risk of type one error. It does so without actually increasing the risk of accepting a drug that’s not efficacious. That’s very important both to us, because it’s a fairly conservative approach, but it’s also extremely important to the regulatory body.

Erica Smith:                  We are not actually developing a method that could give you a false positive result. We’re just trying to reduce the risk of a false negative.

Daniel Levine:               What has the discussion been with regulators and what have you done to validate the tool?

Erica Smith:                  Again, another great question. We’ve had several discussions with the FDA and the EMA, and both of those conversations are ongoing. I think one common outcome from all the conversations is that the regulatory bodies recognize the crippling effect that the placebo effect has on drug development. It’s truly open to innovative and novel solutions. There are things that companies have done for the past several decades, alter trial designs and patient training, and this tool can be used easily in conjunction with all of those.

Erica Smith:                  The reality is that those things may have had a positive impact, but they haven’t solved the problem yet. The regulators appreciate the need to have a new tool to be able to enable drug development by reducing the placebo effect. That being said, this tool is still relatively new. It’s a very novel approach. It’s never been done before. We’re currently in discussions to try and develop a path towards approval for something that doesn’t neatly fit in to any of the boxes that the regulators have.

Erica Smith:                  For example, it doesn’t fit into the drug development tools qualification program at the FDA. While the FDA was very receptive to the idea, they don’t have a means to give a “stamp of approval.” Where we are right now is we’ve agreed with the agencies that they will evaluate it on a case by case basis, as they would with any covariate that’s included in a clinical trial package, and at part of a pre-IND meeting, and as part of the statistical analysis plan.

Erica Smith:                  That’s essentially the bottom line is it’s just the covariate. A covariate approach is generally considered very low risk. That’s how the regulators are approaching it right now.

Daniel Levine:               Are there companies using the tool in a clinical trial right now?

Erica Smith:                  We do have companies using the tool in trials right now, yes.

Daniel Levine:               In the broader view of the company, where does this fit in, and do you just provide this service or is this part of a broader range of clinical trials consulting you do?

Erica Smith:                  The founders of Tools4Patient set out five years ago to really address multiple different issues that plague the drug development industry by developing predictive tools. Placebell is the first predictive tool that we have developed and are commercializing. We’re making this available to sponsors in two ways. The first is in a collaborative way in areas where the tool has not yet been calibrated. I did mention that the Placebell model is disease specific.

Erica Smith:                  If we haven’t worked in a specific disease, and a company is interested in applying Placebell, we do so in a collaborative way essentially at little to no cost to the sponsor, and in a way that is highly scientific and really engaged at developing a highly predictive tool for that particular indication. In areas where we already have a validation that we can provide this off the shelf, we’re doing this as a service to help sponsors. It’s still done very collaboratively and with a lot of scientific background, and a lot of scientific interaction with our team.

Erica Smith:                  We have been approached by several companies that have had trials that were inconclusive because of the strong placebo response. We have been able to in some cases use some of the advanced mathematical and statistical tools that we’ve developed while developing Placebell to help companies figure out what may have happened in the course of that trial. That’s not always the case. Sometimes everything that’s been done, has been done, but sometimes we can add a little something extra to provide some insight. We’ve engaged with companies like that as well.

Erica Smith:                  I can say there are certainly other tools that are in our pipeline that we’re working on as well. That information is not something we’re disclosing at this point.

Daniel Levine:               Erica Smith, vice president of business development at Tools4Patient. Erica, thanks so much for your time today.

Erica Smith:                  Thank you, Danny. I really appreciate being able to talk to you today. Thanks so much.

Daniel Levine:               Thanks for listening. The Bio Report is a production of The Levine Media Group. To automatically download this podcast each week, subscribe to our RSS feed or through iTunes or other podcast manager. To join our mailing list, go to We’d love to hear from you. If you want to drop us a line or are interested in sponsoring this podcast, send a email to

Daniel Levine:               Special thanks to Jonah Levine who composed our theme music, and the Jonah Levine Collective which performs it.


VP, Business Development

Related content


Video Interview at Partnerships in Clinical Trials, Europe 2017

Dominique Demolle, CEO speaks with Andy Burrows of PCT TV about the early vision and development of Placebell©™...

Read More

Video Interview at Biotech Showcase 2019

Erica Smith, VP of Business Development, discusses the benefits of the Placebell©™ approach in video interview at Biotech...

Read More

Video Presentation – Can Daily Self-Assessment Induce a Learning Effect Mitigating Pain Evaluation Error in Randomized Clinical Trials?

LEARNING EFFECT & EVALUATION ERROR At Tools4Patient, we are working on the prediction of placebo response and especially the placebo response with endpoints assessing the patients’ pain. As you know, these endpoints represent most of the efficacy endpoints in Osteoarthritis Randomized Clinical Trials. Nevertheless, the assessment of…

Read More

The next frontier in clinical research & patient management

We’re proud to be leading the charge into the next era of drug development.
Cognivia helps clinical trials reduce data variability, empower decision-making, and accelerate the launch of new therapies.
Tell us about your clinical trial below and we’ll be in touch.

"*" indicates required fields

This field is for validation purposes and should be left unchanged.