Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

MC59

strategies for kidney transplant recipients. We develop a model calibrated with an econometric study of patient data from a national registry to simulate the long- term course of these patients, and derive patient outcomes based on the characteristics of patient population. We study the role of patient heterogeneity on health outcomes, and the relative risk of steroid withdrawal versus steroid maintenance. 2 - Comparison of Post-transplantation Diabetes Mellitus Features and Outcomes in Kidney, Liver, and Heart Transplantation Patients Vidit Munshi, Harvard University, Cambridge, MA, United States, Soroush Saghafian, Curtiss Cook, Harini Chakkera Corticosteroids and immunosuppressive agents prevent graft rejection in transplantation patients, but also bring significantly increased risk for development of impaired glucose tolerance and post-transplantation diabetes mellitus. While this tradeoff is well-known, little work exists characterizing similarities and differences in these tradeoffs across different organ transplantations, particularly from the same hospital system. We present empirical work comparing characteristics and outcomes across kidney, liver and heart transplantation patients and propose a method for modeling these patients to assess interventions and determine optimal management strategies. 3 - Strategic Decisions of Transplant Centers under Competition Sait Tunc, University of Chicago, IL, United States, Burhaneddin Sandikci, Bekir Tanriover Transplant centers in the United States are regularly evaluated for their transplant-related performance. Considering the consequences of their performance outcomes, centers strategically accept/reject offered organs. We study the equilibrium behavior of a heterogeneous set of centers in a competitive market and investigate the effect of their decisions on social welfare. Chair: Eric Gourdin, France Telecom, Orange Labs, CORE/TPN, France Telecom, Orange Labs, CORE/TPN, Issy les Moulineaux Cedex 9, 92794, France Co-Chair: Ivana Ljubic, ESSEC, France 1 - Implications of Zero Rating Programs on Consumer Welfare: A Model Based Study Jialin Song, University of Illinois at Urbana Champaign, 104 S. Mathews Avenue, Urbana, IL, 61801, United States, Qiong Wang Mobile Service Providers (MSPs), such as AT&T, have introduced several zero- rating programs, allowing Content Providers (CPs) to pay for data usage on behalf of customers. However, some worry that zero-rating behavior may hurt the welfare of customers of a MSP’s existing data plans. We develop a linear integer programming model and a dynamic programming solution procedure to analyze MSP’s optimal offering of data plans and the impact of zero rating. We carry out numerical experiments to examine the implications of zero-rating programs with respect to the clustering algorithms to aggregate subscribers into groups, data demand and willingness to pay associated with each group and their correlation. 2 - Clique Interdiction Attacks in Networks Ivana Ljubic, ESSEC Business School of Paris, 3 Av. Bernard Hirsch B.P. 50105, 3rd floor, office 315, Cergy, 95021, France We consider a leader-follower game where the leader removes a fixed number of vertices, and the follower finds the maximum clique over the remaining network. The leader searches for the vertex removal policy that minimizes the largest remaining clique. The problem has applications in the surveillance of communication networks. We provide results concerning the problem complexity and polyhedra, combinatorial lower bounds and safe reduction techniques. We implement a useful tool for analyzing the resilience of (social) networks with respect to vertex-interdiction attacks. Networks with up to 3M edges can be analyzed and solved to optimality within short computing times. 3 - A Flexible Natural Formulation for the Network Design Problem with Vulnerability Constraints Okan Arslan, HEC Montreal, Montreal, QC, Canada, Ola Jabali, Gilbert Laporte Given a graph, a set of origin-destination (OD) pairs with communication requirements and an integer k, the network design problem with vulnerability constraints (NDPVC) is to identify a subgraph with the minimum total edge costs such that between each OD pair, there exist a hop-constrained primary path, and a hop-constrained backup path after any k-1 edges of the graph fail. We develop a natural formulation based on the notion of length-bounded cuts. Experimental results show that for single edge failures, our formulation increases the number of solved benchmark instances from 61% to over 95%. Our formulation also efficiently solves the NDPVC for multi-edge failures (k=3). n MC61 West Bldg 102C Telecommunication Networks Optimization Sponsored: Telecommunications and Network Analytics Sponsored Session

n MC59 West Bldg 102A

Healthcare Operations Sponsored: Health Applications Sponsored Session Chair: Hessam Bavafa, The Wisconsin School of Business 1 - Who Is an Efficient and Effective Physician? Evidence from Emergence Medicine Raha Imanirad, Harvard Business School, Boston, MA, 02472, United States, Soroush Saghafian, Stephen J. Traub Improving the performance of the healthcare sector requires a deep understanding of the efficiency and effectiveness of care delivered by providers. Despite recent advances, fair and scientific methods of measuring efficiency and effectiveness of care delivery have proven elusive. In this study, we use evidence from care delivered by emergency physicians in conjunction with Data Envelopment Analysis (DEA), and shed light on scientific metrics that can gauge performance in terms of efficiency and effectiveness. We then carry out a Tobit analysis to identify factors related to patient, physician and peer physician characteristics that are associated with higher levels of physician performance. 2 - Process Innovation in the Pharmaceutical Industry: Evidence from the Manufacturing of Generics Dimitrios Andritsos, HEC Paris, Departement MOSI, 1 Rue De La Liberation, Jouy-en-Josas, 78351, France, Ivan Lugovoi, Claire Senot We explore the economic effects of process innovation on pharmaceutical manufacturing. Through a collaboration with expert patent attorneys we construct a unique dataset that: i) evaluates process innovation through a detailed observation of pharmaceutical manufacturers’ portfolios of process patents and ii) measures key qualitative dimensions of process innovation, such as novelty, strength of protection and locus of application. 3 - Elective Admission and Patient Discharge Policies in Hospital Environments Vanitha Virudachalam, The Wharton School, 3730 Walnut Street, 500 Jon M. Huntsman Hall, Philadelphia, PA, 19104, United States, Sergei Savin, Hessam Bavafa, Lerzan E. Ormeci We consider the problem of managing the profitability for a hospital faced with a particular reimbursement structure. We assume that the hospital can utilize two main levers: patient admissions, through the size and composition of its portfolio of elective procedures, and early discharges, which modify the patient length-of- stay distribution and relieve pressure on hospital resources. For each type of procedure, we find the optimal admission rate and discharge threshold. We characterize these values under different reimbursement structures and cost structures. 4 - Customized Optimal Diabetes Screening Policies in Resource- restricted Settings Michael Hahsler, Southern Methodist University, Bobby B. Lyle School of Engineering, P.O Box 750123, Dallas, TX, 75275, United States, Hossein Kamalzadeh, Vishal Ahuja, Michael Bowen About 10% of the US adult population has diabetes, and almost 40% are at risk of developing diabetes. Guidelines suggest population-based screening, however, in resource-restricted settings (e.g., for safety-net providers), prioritizing whom to screen and when is essential. We combine analytics methods (hidden Markov models and predictive analytics) with partially observable Markov decision process models to derive the optimal screening policy customized to the characteristics of the provider’s subpopulation. We will present results for electronic health record data from the Parkland Health & Hospital System. n MC60 West Bldg 102B Modeling and Optimization in Organ Transplantation Sponsored: Health Applications Sponsored Session Chair: Murat Kurt, Bristol-Myers Squibb, 3401 Princeton Pike, Lawrence Township, NJ, 08648, United States Co-Chair: David L. Kaufman, University of Michigan-Dearborn, Dearborn, MI, 48126, United States 1 - Investigating Steroid Withdrawal Strategies for Kidney Transplant Recipients Yann Ferrand, Clemson University, Clemson, SC, United States, Christina M. Kelton, Vibha Desai, Teresa M. Cavanaugh, Jaime Caro, Jens W. Goebel, Pamela C. Heaton We study long-term complications associated with various steroid withdrawal

206