2016 INFORMS Annual Meeting Program

TC23

INFORMS Nashville – 2016

2 - Incentive Programs For Reducing Readmissions When Patient Care Is Co-produced Dimitrios Andritsos, HEC Paris, andritsos@hec.fr, Christopher S Tang To compare the effectiveness of three different hospital reimbursement schemes (i.e., Fee-for-Service, Pay-for-Performance and Bundled Payment) in reducing readmissions, we develop a “health co-production” model in which the patient’s readmission is “jointly controlled” by the efforts exerted by both the hospital and the patient. 3 - Reference Pricing For Healthcare Services Shima Nassiri, University of Washington, shiman@uw.edu, Hamed Mamani, Elodie Adida The traditional payment system between an insurer and hospitals does not incentivize hospitals to limit their prices and patient to choose less expensive providers, hence contributing to high insurer costs. Reference pricing (RP) has been proposed as a way to better align incentives and control costs. Under RP, the patient may be responsible for part of the cost if they select a high-price hospital. We propose a model to analyze the RP payment scheme that incorporates an insurer, competing hospitals, and patients with the goal of understanding how RP compares with the current payment system. 4 - Role Of Payment Models In The Value And Adoption Of Health-information Exchanges Mehmet U Ayvaci, University of Texas-Dallas, 800 W Campbell Rd SM33, Richardson, TX, United States, mehmet.ayvaci@utdallas.edu, Huseyin Cavusoglu, Srinivasan Raghunathan We study the interrelationships among the payment model, the providers’ incentives to exchange health information (HIE), and the value of HIEs in terms of improving quality or reducing costs. In the context of a stylized healthcare setting, we examine the fee-for-service, performance-, and episode-based payment contracts that induce socially optimal care levels and HIE adoption. Our findings suggest that as payment models evolve over time, there is a real need to reevaluate the value of HIE adoption and the government policies that induce providers to adopt HIEs. TC22 107B-MCC Dealing with Uncertainty in Hospital Operations Sponsored: Health Applications Sponsored Session Chair: Song-Hee Kim, USC Marshall School of Business, Bridge Hall 307A, 3670 Trousdale Pkwy, Los Angeles, CA, 90089, United States, songheek@marshall.usc.edu Co-Chair: Tinglong Dai, Johns Hopkins University, 100 International Dr, Baltimore, MD, 21202, United States, dai@jhu.edu 1 - Time-driven Activity Based Costing Of Coronary Artery Bypass Grafting Across National Boundaries To Identify Improvement Opportunities Feryal Erhun, University of Cambridge, f.erhun@jbs.cam.ac.uk Coronary artery bypass graft (CABG) surgery is a well-established, commonly performed treatment for coronary artery disease—a disease that affects over 10% of US adults and is a major cause of morbidity and mortality. In 2005, the mean cost for a CABG procedure among Medicare beneficiaries in the USA was $32,201±$23,059. The same operation reportedly costs less than $2,000 to produce in India. The goals of this study are to (1) identify the difference in the costs incurred to perform CABG surgery by three Joint Commission accredited hospitals with reputations for high quality and efficiency and (2) characterize the opportunity to reduce the cost of performing CABG surgery. 2 - Clinical Ambiguity And Conflicts Of Interest In Interventional Cardiology Decision-making Tinglong Dai, Assistant Professor, Johns Hopkins University, 100 International Drive, Baltimore, MD, 21202, United States, dai@jhu.edu, Xiaofang Wang, Chao-Wei Hwang, Chao-Wei Hwang Cardiovascular disease is the leading cause of death in the United States, and coronary artery disease (CAD) is the major underlying culprit. Percutaneous coronary intervention (PCI) has proven to be beneficial to patients with acute coronary syndrome, yet its benefit to stable CAD patients is more nuanced. Indeed, unnecessary PCI procedures for stable CAD patients have contributed to wasteful health spending and, in certain cases, patient harm. In this paper, we model both clinical ambiguity and conflicts of interest in interventional cardiology decision-making. Among other results, we show the PCI usage may be non- monotonic in the conflict-of-interest level.

3 - The Value And Price Of Flexibility In Robust Assignment Of Patients To Radiation Therapy Machines Philip Allen Mar, Dept. of MIE, University of Toronto, 5 King’s College Road, Toronto, ON, M5S 3G8, Canada, philip.mar@mail.utoronto.ca, Timothy Chan In a radiation cancer therapy program, radiation therapy machines are allocated to treat particular types of cancers to form a fixed network that acts as a guideline for the hospital when assigning patients to machines for treatment. We study the operational efficiency of this system from a manufacturing process flexibility viewpoint. Furthermore, we use robust optimization to prescribe new allocation and assignment guidelines which are robust against deviations from the optimal assignment, and against capacity uncertainty. 4 - Maximizing Intervention Effectiveness Through Robust Optimization Rong Qing Brian Han, Marshall School of Business, University of Southern California, Los Angeles, CA, United States, rongqing.han.2019@marshall.usc.edu, Vishal Gupta, Song-Hee Kim In medicine and social science, practitioners often seek to implement interventions that have previously been proven effective via randomized control trials (RCT). Typically, practitioners cannot access the raw data of the RCT, but do have summary statistics from published papers. We propose a novel robust optimization framework to identify a small, targeted group of candidates for the intervention to maximize effectiveness based on these summary statistics. Using data from a large urban hospital, we show that our method often outperforms conventional methods, especially when the target and RCT populations differ substantially. Chair: Lawrence Wein, Stanford University, 655 Knight Way, Stanford, CA, 94305, United States, lwein@stanford.edu 1 - Personalized Medicine Dimitris Bertsimas, MIT, dbertsim@mit.edu We use a) Electronic Medical Records from 1.5 million patients over 15 years from the Boston Medical Center and 200 thousand cancer patients from Dana Farber and b) state of the art as well as new machine learning algorithms to propose an algorithmic theory of personalized medicine for several human diseases. We discuss the overall vision, results and possible impact. 2 - Data Uncertainty In Cost-effectiveness Analyses Of Medical Innovations Joel Goh, Harvard University, jgoh@hbs.edu, Mohsen Bayati, Stefanos Zenios, Sundeep Singh, David W Moore Cost-effectiveness studies of medical innovations often suffer from data inadequacy. When Markov chains are used as a modeling framework for such studies, this data inadequacy can manifest itself as imprecision in the elements of the transition matrix. We study how to compute maximal and minimal values of the chain as these uncertain transition parameters jointly vary within a given uncertainty set. We show that these problems are computationally tractable if the uncertainty set has a row-wise structure but generally intractable otherwise. We apply our model to assess the cost-effectiveness of fecal immunochemical testing (FIT), a new screening method for colorectal cancer. 3 - New Models For Fecal Microbiota Transplantations Lawrence Wein, Stanford University, lwein@stanford.edu, Abbas Kazerouni A nonprofit organization, OpenBiome, has created a public stool bank to facilitate fecal microbiota transplantation, which is an effective treatment for Clostridium difficile infection and is being investigated as a treatment for other microbiota- associated diseases. We discuss two problems: optimizing OpenBiome’s operations, and using pooled stools to improve the efficacy in clinical trials against microbiota-associated chronic diseases such as ulcerative colitis. 4 - Designing Strategic National Stockpile – A Two-stage Robust Optimization Approach Peter Yun Zhang, Massachusetts Institute of Technology, Cambridge, MA, United States, pyzhang@mit.edu, Nikolaos Trichakis, David Simchi-Levi We present a model that captures two sets of decisions a supply chain network designer faces: placement of inventory in preparation for demand uncertainty, and resource allocation after the uncertain events unfold. We show optimality and tractability results for problem structure that arises from designing the Strategic National Stockpile. TC23 108-MCC New Models in Health Care Sponsored: Health Applications Sponsored Session

309