A listing of the efforts of ORC Faculty and students hoping to make a positive impact on the current COVID- 19 pandemic.
Dimitris Bertsimas and ORC students - website: covidanalytics.io
This website represents the effort of a group of about 20 graduate students in the Operations Research Center at MIT under the guidance of Professor Bertsimas. Their overarching objective is to rapidly develop, and deliver tools for use by hospitals, government officials, and healthcare institutions in the United States to combat the spread of COVID-19.
Richard Larson - MIT BLOSSOMS https://blossoms.mit.edu and focusing on STEM education for at-home high school students
1. BLOSSOMS @HOME for Students @HOME
BLOSSOMS @Home is for Students @Home, especially high school students who want to explore STEM-related problems.
Interactive video lessons that deal with real-world problems, asking questions of you along the way, followed by some new challenge questions. We ask that students write up their answers in a brief report, and email it to us. We will send our comments! A gold star is possible! Have fun, and learn a lot!
2. MIT Blossoms @MIT_BLOSSOMS
BLOSSOMS to be on KMTP TV 33, an independent broadcast TV station in the Bay Area, Cal. According to station management, given the current crisis, low-income households that do not have internet access are likely to have a television, the new broadcasts extending to all.
People: Richard C. Larson, Professor, Post-Tenure, MIT IDSS, M; Elizabeth Murray, BLOSSOMS Project Manager; Tara Connelly, BLOSSOMS staff, Jerold Gelfand, BLOSSOMS videographer.
3. The 2019-nCoV coronavirus: Are there two routes to infection?
The new virus may be expelled from a human via respiratory and intestinal means. OR/MS Today (story dated Feb, 18, 2020) https://pubsonline.informs.org/do/10.1287/orms.2020.02.01/full/.
Retsef Levi, Vivek Farias and Simon Johnson with ORC, MBAn and EMBA students https://www.covidalliance.com/
Andreea, Celia and Josh, El-Ghali (MBAin), Deeksha Sinha, Andy Zheng, Andreea Georgescu, Jackie Baek, Tianyi Peng, Patricio Araneda
Multiple faculty and over nine ORC PhDs are working as part of the COVID-19 Policy Alliance (https://www.covidalliance.com/), a team over 70 people including MIT faculty and graduate and undergraduate students from multiple programs. The alliance collaborate with major industry and government partners, such Tableau, Claritis, Amazon and federal and state emergency responders and governments. Thus far, the alliance team has accomplished the following:
- Two public webinars and multiple media publications that highlight policy recommendations, particularly around the importance of managing high risk patients and high risk clusters (sites and micro-geographies) to reduce fatalities and control the number of patients who require acute or critical hospital care, thus avoiding scenarios of hospital system crash.
- Interactive risk analytics tools that map high risk sites, high risk counties and high risk zip codes based on known COVID-19 risk factors (https://www.covidalliance.com/interactive-data-tools).
- A load balancing optimizer API that can support state level triage and allocation of testing and hospital resources, such as ICU. The tool is already integrated to the New Hampshire emergency management system.
- Two projects on the ground in MA and NH, collaborating with local industry and government stakeholders to build supporting systems to senior living facilities, including, training, workforce support, testing, PPE supply and infection control.
- An airlift effort to bring PPEs from China (expected arrival of first shipment in this coming Friday, April 10).
- Research insights related to the COVID-19 outbreaks in other countries with implications to the US.
- Engaging with manufacturing companies to build local production capacity of PPE.
- An advanced predictive demand (COVID-19 cases, hospitalization and deaths) model that supports a range of outbreak management decisions.
Andrew Lo - https://projectalpha.mit.edu/research/covid19
Financing Vaccines for Global Health Security (with Jonathan T. Vu, Benjamin K. Kaplan, Shomesh Chaudhuri, and Monique K. Mansoura)
Recent outbreaks of infectious pathogens such as Zika, Ebola, and COVID-19 have underscored the need for the dependable availability of vaccines against emerging infectious diseases (EIDs). The cost and risk of R&D programs and uniquely unpredictable demand for EID vaccines have discouraged vaccine developers, and government and nonprofit agencies have been unable to provide timely or sufficient incentives for their development and sustained supply. We analyze the economic returns of a portfolio of EID vaccine assets, and find that under realistic financing assumptions, the expected returns are significantly negative, implying that the private sector is unlikely to address this need without public-sector intervention. We have sized the financing deficit for this portfolio and propose several potential solutions, including price increases, enhanced public-private partnerships, and subscription models through which individuals would pay annual fees to obtain access to a portfolio of vaccines in the event of an outbreak.
Bayesian Adaptive Clinical Trials for Anti-Infective Therapeutics during Epidemic Outbreak (with Qingyang Xu, Shomesh Chaudhuri, and Danying Xiao)
Despite growing concern over the imminent pandemic of Coronavirus Disease 2019 (COVID-19), standard clinical trials of anti-infective therapeutics for COVID-19 are unlikely to receive FDA approval within the course of the outbreak. We apply a Bayesian patient-centered adaptive model—which minimizes the expected harm of false positives and false negatives—to optimize the clinical trial development path during an epidemic outbreak. When the epidemic is more infectious, the Bayesian-optimal sample size in the clinical trial is lower and the optimal false-positive rate or threshold p-value is higher. For particularly infectious and deadly diseases (, the optimal can be as high as 45.2%. Our results illustrate the importance of adapting the clinical trial design and the evaluation process to the infectivity and severity of the epidemic.
Estimated Probabilities of Success of Clinical Trials for Vaccines and Other Therapeutics for Infectious Diseases (with Chi Heem Wong and Kien Wei Siah)
We estimate the probabilities of success (PoS) of clinical trials for vaccines and other infectious disease therapeutics using 32,626 unique triplets of trial identification number, drug, and disease between January 1, 2000 and January 7, 2020, which yields 1,886 vaccine development programs and 5,222 drug development programs targeting infectious diseases. The overall estimated PoS for a vaccine program is 40.6%, and 23.0% for an infectious disease therapeutic. Among vaccines, the most successful programs have been against bacterial infection (49.5%), hepatitis (49.1%), and respiratory illnesses (42.5%), while the least successful have been against influenza (37.4%), vector-borne diseases (29.0%), and other viruses (37.7%). The three infectious diseases with the highest PoS estimates for non-vaccine therapeutics are rotavirus (76.2%), Japanese encephalitis (67.6%), and rabies (66.7%). Viruses involved in recent outbreaks—Middle Eastern Respiratory Syndrome (MERS), Severe Acute Respiratory Syndrome (SARS), Zika, and COVID-19—have had a combined total of only 15 development programs initiated over the past twenty years, and no approved therapies to date.
Georgia Perakis and her team of 8 current students from the ORC and one alumni: https://mitsloan.mit.edu/faculty/directory/georgia-perakis
Tamar Cohen-Hillel, Lennart Baardman, Jiong Wei Lua, Raphaelle Dupont for the first project (predictive)
Amine Bennouna, Divya Singhvi, Omar Skali-Lami, Yiannis Spantidakis, Leann Thayaparan in the second project (prescriptive)
We are extending already developed methods in our past research (that combine ideas from ML,Optimization and Econometrics), in order to estimate the probability of disease (COVID-19) transmission between different groups of individuals. The goal is to estimate the probability of certain groups of individuals getting infected at a certain time, given pre-existing health conditions, location, social interactions, and initial infection circumstances.
Optimizing Government Interventions for COVID-19
We are analyzing data from across the world to assess the impact of different government strategies (social distancing, stay-in orders, complete lockdown), in order to recommend optimal government interventions that minimizes deaths due to COVID-19 without severely impacting the economy.
David Simchi Levi - A team working on “the impact of the pandemic on supply chains.” Some of the work is described in https://dsl.mit.edu/.