Dr. Dominique Demolle, CEO of Tools4Patient, recently presented data at the 16th Annual Scientific Meeting of the International Society for CNS Clinical Trials and Methodology in Washington, DC. The presentation, entitled ““Modeling of Peripheral Neuropathic Pain and Osteoarthritis Placebo Response: Working Towards a Unique Model of the Placebo Response in Chronic Pain?” was authored by Tools4Patient scientist Dr. Samuel Branders along with co-authors Drs. Dominique Demolle, Alvaro Pereira and Erica Smith.
This poster presentation summarized studies aimed at generating a machine learning model that could predict the placebo response in patients with chronic pain resulting from varying etiologies as a first step in generating a method to control for this major confounding factor in clinical trials. The study evaluated placebo-treated patients from two peripheral neuropathic pain (PNP, N=87) studies and two osteoarthritis (OA, N= 48) studies. In these studies, subjects received placebo only as an add-on therapy to their current pain management regimen. The model utilized standard clinical trial baseline data including demographics, medical history and baseline pain intensity along with the results from a proprietary psychological questionnaire. Placebo response was measured as reduction in average pain score at the end of the treatment phase. A predictive model was trained using only PNP patients or PNP+OA patients, and was then applied to either the PNP population alone or the combined PNP+OA population. Results indicated that both models explained approximately 24-30% of variability in patient data irrespective of whether it was derived from the PNP population or the more heterogeneous population of both PNP+OA patients.
These results established a predictive model to understand the placebo response in PNP clinical trial patient populations. Furthermore, the applicability of the model developed in PNP to a mixed population of both PNP and OA patients suggests basic similarities in factors driving the placebo response in chronic pain patients, even when evaluated in different trial designs and pain etiologies. “This type of modeling can be utilized to reduce data variability related to the placebo response in clinical trials and improve study power”, indicates Dr. Alvaro Pereira, CSO.
Tools4Patient is offering its machine learning-based predictive models of the placebo response (Placebell©™) to the biopharmaceutical and medical device industries. Placebell©™ is easily implemented in virtually any industrial clinical study design, and scalable to large, complex trials. This technology is being deployed to clinical trials in pain, neurology, psychiatry and ophthalmology, with applicability to almost any therapeutic area or indication. “Tools4Patient is making great strides in our mission of developing predictive tools that de-risk clinical trials, improve decision making in drug development, and accelerate drugs to market for our ultimate stakeholders – patients”, said Dr. Dominique Demolle, CEO.