Columbia Business School DRO Healthcare Conference


Please "Save the Date" for the Columbia Business School DRO Healthcare Conference, which is being held in honor of Linda Green on the occasion of her retirement. Join us on Friday, April 24, 2020 in New York at Columbia Business School to discuss research in the healthcare space and to celebrate Linda's career.

Conference location: 

142 Uris Hall
Columbia Business School
3022 Broadway 
New York, NY 10027


  • 8:00 am – 8:45 am  Breakfast
  • 8:45 am – 9:00 am  Opening Remarks
  • 9:00 am – 10:25 am Session 1
    • 9:00 am - 9:35 am Christian Terwiesch
      Connected Healthcare: A Framework and Some Research Results

      Can you anticipate the products and services patients and clients will need before they even ask? A shift in health care delivery from occasional interaction to continuous connection has the potential to do just that. Wearable devices, smart pill bottles, digestible sensors, and other disruptive health technologies are all associated with the promise of improving the quality of care while also making efficient use of resources. This talk will provide a framework on how to think about connected healthcare and overview a handful of research studies in this space.

    • 9:35 am - 10:10 am Timothy Chan
      An Inverse Optimization Approach to Measuring Clinical Pathway Concordance

      Clinical pathways outline standardized processes in the delivery of care for a specific disease. Patient journeys through the healthcare system, though, can deviate substantially from recommended or reference pathways. Given the positive benefits of clinical pathways, it is important to measure the concordance of patient pathways so that variations in health system performance or bottlenecks in the delivery of care can be detected, monitored, and acted upon. This paper proposes the first data-driven inverse optimization approach to measuring pathway concordance in any problem context. Our specific application considers clinical pathway concordance for stage III colon cancer. We apply our novel concordance metric to a real dataset of colon cancer patients from Ontario, Canada and show that it has a statistically significant association with survival. Our methodological approach considers a patient's journey as a walk in a directed graph, where the costs on the arcs are derived by solving an inverse shortest path problem. The inverse optimization model uses two sources of information to find the arc costs: reference pathways developed by a provincial cancer agency (primary) and data from real-world patient-related activity from patients with both positive and negative clinical outcomes (secondary). Thus, our inverse optimization framework extends existing models by including data points of both varying "primacy" and "goodness". Data primacy is addressed through a two-stage approach to imputing the cost vector, while data goodness is addressed by a hybrid objective function that aims to both minimize and maximize suboptimality error for different subsets of input data.

    • 10:10 am - 10:25 am Discussion led by Carri Chan
  • 10:25 am - 10:45 am  Coffee Break
  • 10:45 am – 12:10 pm Session 2
    • 10:45 am - 11:20 am Kamalini Ramdas
      Increasing Patient Engagement Through Shared Medical Appointments

      Through a randomized control trial, we examine the impact of shared medical appointments (SMAs), in which a group of patients with similar chronic conditions meet with a doctor simultaneously, on levels of patient engagement. Relative to traditional one-on-one care models, we study how SMAs affect engagement levels, both during the appointment (such as making eye contact with the physician, engaging in the proceedings, and asking questions) and after (such as complying with prescribed medications in the home). Although SMAs hold obvious promise for improving the efficiency of healthcare, to the extent that they may lead to increased patient engagement, they may result in improved outcomes as well.

    • 11:20 am - 11:55 am Hummy Song
      Speed-Quality Tradeoffs in Home Health: The Effects of Visit Length on Hospital Readmission

      Home health care is an industry that has been experiencing significant growth in the United States. It is viewed as an avenue for achieving reductions in the cost and utilization of expensive downstream health care services. Using a novel dataset on home health care visits, we examine the extent to which there is a speed-quality tradeoff between the length of a post-acute home health visit and the likelihood of hospital readmission. Since unobserved patient health status may influence both the length of a home health visit and the likelihood of hospital readmission, we use an instrumental variable approach for our estimation in addition to controlling for operational, demographic, and patient condition-related characteristics. We find that longer home health visits may contribute to a lower likelihood of hospital readmission in a meaningful way. We also conduct a cost-benefit analysis that suggests that the cost of investing in additional home health capacity may be outweighed by the cost savings arising from fewer hospitalizations. We suggest several approaches that managers could take to attain reductions without incurring significant additional costs. Joint work with Elena Andreyeva and Guy David.

    • 11:55 am - 12:10 pm Discussion led by Fanyin Zheng
  • 12:10 pm - 1:45 pm  Lunch Break
  • 1:45 pm – 3:45 pm Session 3
    • 1:45 pm - 2:20 pm Lawrence M. Wein
      The National Backlog of Untested Sexual Assault Kits: Which Kits to Test and How to Test Them

      Motivated by the debate over how to deal with the huge backlog of untested sexual assault kits in the U.S., we use data from Detroit to construct and analyze a mathematical model that predicts the expected number of hits to the criminal DNA database as a function of the proportion and mix (i.e., stranger vs. nonstranger assaults) of the backlog tested, and carry out a cost-benefit analysis for testing the kits. We find that testing all kits is highly cost effective. Although the backlog of untested sexual assault kits in the U.S. is starting to be addressed, many municipalities are opting for selective testing, where only a subset of the collected samples that are deemed most probative by sexual assault forensic examiners are actually tested. We use data from San Francisco to address a fundamental issue in criminal investigations: how much evidence to collect and process. We find that the number of solved sexual assaults could be increased several-fold by optimizing which samples are collected and testing all collected samples.

    • 2:20 pm - 2:55 pm Nikolaos Trichakis
      Analytics-driven Organ Transplant Allocation: Challenges and Opportunities

      Current organ transplant allocation policies have resulted in inefficient utilization of supply and persistent disparities in transplant access across different waitlisted candidates based on their geographic location, sex, and/or disease. We discuss our collaborations with the United Network for Organ Sharing (UNOS) Research Labs, the UCSF Department of Surgery, and the UT Southwestern Medical Center to improve allocation. We describe a novel optimization scheme that leverages machine learning and simulation techniques to devise allocation policies that could alleviate disparities and allow for a more efficient use of donated organs. We find that our proposed allocation policies could provide substantial mortality reduction (of the order of 20% for end-stage liver disease patients), while providing a more equitable organ access in comparison with UNOS policies. We discuss challenges and opportunities in revising national allocation policies based on our interactions with UNOS policymakers.

    • 2:55 pm - 3:30 pm Margaret Brandeau
      What Should We Do About the Opioid Epidemic? Models to Support Good Decisions

      The US is currently experiencing an epidemic of drug abuse caused by prescription opioids and illegal opioid use, including heroin. In addition to crime and social problems, rising levels of drug abuse have led to a sharp increase in overdose deaths in the US as well as significant outbreaks of infectious diseases such as HIV and hepatitis C. How should we deploy limited public health resources to help solve this complex public health problem? This talk describes models to support decision making regarding the control of drug abuse – and associated diseases such as HIV and hepatitis C – in the US. We conclude with discussion of key areas for further research.

    • 3:30 pm - 3:45 pm Discussion led by Jing Dong
  • 3:45 pm - 4:15 pm  Coffee Break
  • 4:15 pm - 5:15 pm  Panel discussion “Making an impact” led by Carri Chan
    • Margaret Brandeau
    • Linda Green
    • Lawrence M. Wein
  • 5:15 pm - 6:00 pm  Reception