2016 INFORMS Annual Meeting Program

TA23

INFORMS Nashville – 2016

2 - Optimal Integration Of Kidney Exchange Programs With Antibody Reduction Therapy To Increase Successful Transplant In Difficult To Match Recipients Naoru Koizumi, George Mason University, nkoizumi@gmu.edu, Monica Gentili Kidney paired donation (KPD) allows incompatible pairs to exchange kidneys with other incompatible pairs. However, evidence suggests there stills exist barriers to KPD utilization, especially among difficult-to-match transplant candidates and positive actual or virtual crossmatches. We use optimization and simulation analyses to optimally integrate antibody reduction therapy in KPD matching runs to increase the overall number of transplants. The proposed mathematical model matches incompatible pairs taking into consideration the possibility that some of the recipients could undergo a desensitization protocol to improve compatibility with the matched donor. 3 - Position-indexed Formulations For Kidney Exchange Tuomas Sandholm, Carnegie Mellon University, sandholm@cs.cmu.edu, John Dickerson, David Manlove, Benjamin Plaut, James Trimble We address the tractable clearing of kidney exchanges with short cycles and practical (long, but not unbounded) chains. We introduce three compact integer programming formulations with linear programming relaxations that are at least as tight as the previous tightest formulation (which was not compact) for instances in which each donor has a paired patient. Then, on real data from the UNOS US-wide exchange and the NHS UK-wide exchange, as well as on generated data, we show that our models dramatically outperform all prior solvers. We also present models that are more scalable than the state-of-the-art models for failure-aware kidney exchange. TA23 108-MCC Modelling Care and Treatment of Chronic Diseases Sponsored: Health Applications Sponsored Session Chair: Michael W Carter, University of Toronto, 5 King’s College Rd., Toronto, ON, M5S 3G8, Canada, carter@mie.utoronto.ca 1 - Chronic Care Disease Management Through Operations Research & Analytics – Lessons From Ontario Ali Vahit Esensoy, Manager, Strategic Analytics, CCO, Toronto, ON, Canada, AliVahit.Esensoy@cancercare.on.ca, Kiren Handa CCO acts as a key Ontario government advisor on cancer care, renal, palliative and access to care. Part of CCO’s mandate is to drive improvement through developing multi-year system plans, setting standards and guidelines, developing and deploying information systems, and measuring performance. Since 2010, CCO has actively tested and deployed numerous operational research methodologies as part of their advanced analytics work. This session will review the evolution of CCO’s OR practice within the advanced analytics group and discuss successes and challenges of applying OR for system planning and policy decisions within Ontario’s healthcare system. 2 - An Analytics Approach To Dementia Capacity Planning Tannaz Mahootchi, Cancer Care Ontario, Tannaz.Mahootchi@cancercare.on.ca, Azadeh Mostaghel, Ali Vahit Esensoy We use a data-driven approach to identify and project capacity issues in Ontario for the persons living with dementia (PLwD). Evidence suggests that while PLwD prefer to stay at home for as long as possible. Lack of appropriate community-care could lead to hospitalization and residential long-term care (LTC) placements. Using the person-level care trajectory data and evidence from literature, we develop a simulation model to predict the effect of augmented home and community care options on the patient flow and LTC placements for the future dementia incidence cases. 3 - Nurse Scheduling And Risk Analysis Of Hemodialysis Patients Michael W Carter, University of Toronto, carter@mie.utoronto.ca, Mahsa Shateri Kidney failure patients require dialysis treatment three times a week until a suitable donor is found. During dialysis, nurses monitor several patients at once, but when complications occur, a nurse must be available quickly to attend to the problem and restart dialysis. This paper provides a model to determine the minimum staffing levels required in order to deliver safe, effective care.

4 - Two-stage Stochastic Programming For Adaptive Interdisciplinary Pain Management With PIN Transition Models Gazi Md Daud Iqbal, The University of Texas at Arlington, gazimddaud.iqbal@mavs.uta.edu, Jay Michael Rosenberger, Victoria C. P. Chen, Robert Gatchel This research uses a two-stage stochastic programming approach to optimize personal adaptive treatment strategies for pain management. Transition models are represented by Piecewise Linear Networks. A multi-objective mixed integer linear program is developed to optimize treatment strategies for patients based upon on these transition models. TA24 109-MCC Practice-Based Research in Healthcare OM Sponsored: Health Applications Sponsored Session Chair: Jónas Oddur Jónasson, London School of Business, Regent’s Park, NW1 4SA, London, TX, 00000, United Kingdom, jjonasson@london.edu 1 - Staff Planning For Anesthesiologists Staff planning for human resources like anesthesiologists at hospitals takes place sequentially, where planning is done for regular staffing as well as reserve capacity. The staff planners balance the expected overtime, under-utilization costs as well as the cost of keeping staff on reserve. Some of these costs are not explicitly known. We employ a structural estimation model to infer these implicit costs and subsequently find heuristic solution for the medium term staff planning. We apply this approach to staff planning for anesthesiologists. 2 - Separate & Concentrate: Accounting For Patient Complexity In General Hospitals Sandra Sülz, Assistant Professor, Erasmus University Rotterdam, Rotterdam, Netherlands, sulz@bmg.eur.nl, Ludwig M Kuntz, Stefan Scholtes We show that the positive association between patient volume, focus and service quality is worse for complex patients, and that hospitals that route the majority patients in a disease segment to the same department have fewer department allocation errors and better outcomes, particularly for complex patients. These findings suggest a redesign of general hospitals: Separate out routine patients and route them away from general hospitals into high-volume and focused value- adding process clinics and concentrate disease segments in the clinical departments of solution shop hospitals rather than scattering patients across several departments. 3 - Towards An Equitable Allocation Of Organs Among End-stage Liver Disease Patients Mustafa Akan, Associate Professor of Operations Management, Carnegie Mellon University, Tepper School of Business, Posner Hall 381C, Pittsburgh, PA, 15213, United States, akan@andrew.cmu.edu, Ngai-Hang Z Leung, James F. Markmann, Sridhar R Tayur, Heidi Yeh Patients on the waiting list for liver transplants receive priority based on their MELD scores, which reflect the severity of liver disease. Recent studies have shown that Hepatocellular Carcinoma (HCC) patients have significantly higher liver transplant rates than non-HCC patients due to non-individualized MELD calculation of the former. By using clinical data from SRTR we calibrate Markov Models and build a new simulator (MYATLAS). We then compute alternative MELD scores for the HCC patients that are a function of the candidate’s tumor biology and its progression. 4 - Ambulance Emergency Response Optimization In Developing Urban Centers Justin J. Boutilier, University of Toronto, Toronto, ON, Canada, j.boutilier@mail.utoronto.ca, Timothy Chan Time sensitive medical emergencies are a major health concern comprising over 33% of all deaths in low and middle income countries (LMICs). Despite evidence that ambulance services can save lives, poor access and availability of emergency medical care in LMICs continues to be a widespread problem. In this work, we develop a novel ambulance location-routing model, tailored to address the challenges faced by urban areas in LMICs. We use extensive field data from Dhaka, the capital of Bangladesh and one of the most densely populated cities on earth, to develop and validate our modelling framework. Sandeep Rath, University of California-Los Angeles, sandeep.rath.1@anderson.ucla.edu, Kumar Rajaram

240