INFORMS Open Forum

Invitation to participate in a clinical trial prediction study

  • 1.  Invitation to participate in a clinical trial prediction study

    Posted 11 days ago
    Edited by Pavel Atanasov 11 days ago

    Dear INFORMS Colleagues,

    We invite you to participate as judgmental forecasters in a prediction tournament focused on clinical trial development. The study is supported by two grants from the National Science foundation. It is structured as a horse race between machine models, led by PI Sauleh Siddiqui at American University, and human crowdsourcing methods, led by CO-PIs Pavel Atanasov and Regina Joseph. The tournament starts next week, on September 15. 

    Our focus is on biomedical innovation and clinical development: predicting which treatments will proceed toward approval in clinical trials. Forecasters will be asked to estimate the probability of clinical trial progression of various treatments for COVID-19, oncology, neurology and cardiology. Forecasters may choose to consult statistical models or use purely judgmental techniques. Both experts and non-experts are welcome. The only requirement is that forecasters are at least 18 years old.

    As a forecaster, your participation will help advance the study of prediction, including reference class forecasting and probabilistic reasoning. You will also receive information, training and feedback that may help you improve your prediction skills. Finally, the most active and accurate forecasters may win prizes of up to $1,500. The total prize purse is over $14,000. The first tournament ran in 2020-2021 and featured a similar structure & prizes. We plan to present the initial results of the first tournament at the INFORMS Annual Meeting this October.

    Forecaster involvement: The first tournament starts on September 15 and ends in February 2022. We will feature approximately 50 forecasting questions, of which 35 will open in September and 15 in October. To be eligible for prizes, forecasters must answer at least 25 questions. During their first sessions, forecasters will undergo a short battery of online questionnaires, receive forecasting training, and make their first forecasts. Forecasters will then log in occasionally to submit forecasts on new questions, and to update their previous estimates.

    Sign up here: http://www.pytho.io/human-forest

    If you think colleagues or acquaintances may be interested, please forward this message to them.

    Regards,

    Pavel Atanasov, Regina Joseph & Sauleh Siddiqui



    ------------------------------
    Dr. Pavel Atanasov
    Co-Founder, Pytho LLC
    Penn & Good Judgment Project Alum
    ------------------------------