2012-13 Program on Data-Driven Decisions in Healthcare


Healthcare is a central political, economic and social issue of our times. In the healthcare process, decisions are made at every level from the treatment of individual patients to formulation and evaluation of national policies.

In the healthcare process, decisions are made at every level from the treatment of individual patients to formulation and evaluation of national policies. At the same time, data generation is increasing dramatically. Electronic medical records are becoming ubiquitous. Tests produce gigabytes of data, including images and biometric samples. The volumes of data are daunting in themselves; concomitant problems such as confidentiality and data quality exacerbate the challenges in producing usable tools that support principled healthcare decisions.

Research Themes

The SAMSI program on DDDHC addressed issues of mathematical and statistical theory and methodology that will improve evidence-based healthcare decision-making. The goals were to:

  • Strengthen the link between data and decisions, a path that includes major challenges in mathematical modeling and statistical inference.
  • Highlight and increase the role that statistics, applied mathematics and operations research can play in making data-driven healthcare decisions.

The program had two principal—and intersecting—themes:  

Operations Research Modeling (ORM), with particular emphasis on resource allocation. The depth of interest in healthcare in the operations research community is enormous. Meeting programs of INFORMS, the principal society, contain dozens of sessions on healthcare topics.  One expected focus is mathematical model building and use of different forms of simulation (e.g., discrete-event and agent-based simulations) as tools for evaluation of models, operational practices and policies. Other anticipated mathematical emphases include discrete optimization, Markov decision processes, dynamic programming, network modeling and stochastic control.

Comparative Effectiveness Research (CER), which attempts to determine what—ranging from behavioral modification to surgical intervention to medication—works for whom, for which medical problems, and under what circumstances. 


Description of Activities

Workshops: The program began with an Opening Workshop on August 26-29, 2012. The Transition Workshop was held May 9-10, 2013. 

Courses: A graduate-level course was taught at SAMSI during the Fall 2012 semester.

Working Groups: Working groups met—in most cases, weekly—throughout the program to pursue particular research topics articulated in the Opening Workshop, as well as topics identified subsequently by Working Group participants. The Working Groups consisted of SAMSI visitors, postdoctoral fellows, graduate students, faculty and scientists from the Research Triangle area. Remote participation was possible via WebEx. Potential Working Group foci included:

  • ORM: Screening Policies, Management of Transplant Lists, Emergency Response, Scheduling (e.g., of patients, nurses, physicians and facilities such as operating rooms), Inventory Management (e.g., of blood banks), Patient Safety, Service Delivery
  • CER: Computational Issues, including Massive Data, Observational Medical Studies, Case Mix Adjustment, Data Quality, Data Integration, Public Health Surveillance

We also anticipate one or more Working Groups that merge the ORM and CER themes.

Opportunities to Participate

Opportunities to engage in the DDDHC program included

  • Visits to SAMSI for researchers in academia, government and industry, for periods ranging from several weeks to an entire semester or year
  • Postdoctoral Fellowships, which include a second year carrying out research begun during the program year
  • New Researcher Fellowships, for those within six years of receiving their doctorate
  • Visits (semester- or year-long) to SAMSI by graduate students
  • The Working Groups, including remote participation by means of WebEx
  • Workshops 


Click here to download a short description of the program.