The placebo response is a major contributing factor to late-stage clinical trial failures in many therapeutic areas, and Placebell's innovative predictive method addresses this complex challenge. Placebell combines machine learning-based algorithms with individual patient psychology and personality data to predict the individual placebo response in a patient before the first dose of a drug is administered. This predictive tool reduces variability caused by the placebo response in patients, all with zero statistical or operational risk to a study.
"These results highlight the impact and applicability of Tools4Patient's approach in clinical development and allowed the sponsor to make thorough conclusions about the study results," says Alvaro Pereira, Ph.D., CSO of Tools4Patient and co-author of the abstract. "With our Placebell product, for the first time, we are now able to gain insight into placebo response in drug treated as well as placebo treated patients."
In the study, Tools4Patient applied a Placebell chronic pain model that had been built by meta-analysis that included individual data from 211 placebo-treated patients from several studies conducted by Tools4Patient. The model was pre-specified and used to predict placebo response in all trial patients receiving an intra-articular injection of a novel compound in the sponsored Phase 2 RCT. Results were compared for five different endpoints in the trial: average pain score, three components of the WOMAC battery focusing on pain (primary efficacy endpoint), joint stiffness and physical function, and the patient global assessment score. Using the same Placebell model, equivalent performance was achieved for all endpoints, as a reduction in data variability between 20 and 35 percent was demonstrated in both placebo-treated patients and all patients.
The model, or the comparison between predicted placebo response and observed placebo response, was highly statistically significant for every endpoint, indicating it is valid to apply the Placebell model to make these predictions. The results confirm the applicability of Placebell in multi-center, sponsored randomized controlled trials and demonstrate that model performance is maintained under these conditions.
"Reducing drug development risk related to the placebo response is a real game changer, as previous attempts to address this issue may provide some benefit but have not completely solved the problem," says Dominique Demolle, Ph.D., CEO of Tools4Patient. "The feedback we are receiving from industry collaborators about this breakthrough is re-affirming. We continue to work with international partners on clinical studies in other therapeutic areas, and the publication of results will follow at a steady pace in 2021."
The OA data supports Placebell as a platform solution in multiple indications, including recent data analyzed for chronic pain, neurology and ophthalmology. The Placebell portfolio is supported by Tools4Patient's partnerships, which include industry stakeholders and key opinion leaders. Beyond this, Tools4Patient is developing tools to predict other key aspects of data variability that jeopardize data interpretation in clinical studies, such as variability between clinical sites and geographies, trial drop-out rate and lack of patient compliance during trials. The company's expansion plans, including a subsidiary in the U.S., will be supported by a capital increase later this year.