Informs Annual Meeting 2017

MD50

INFORMS Houston – 2017

MD50

doctors’ appointments made over a five-month period, along with other online information. We propose a random coefficient logit model to characterize consumer heterogeneity in doctor choices, taking into account both numeric and textual user-generated content with text mining techniques. 2 - The Impact of Health Information Exchanges on Physicians’ Referral Patterns Saeede Eftekhari, State University of New York-Buffalo, 304 Jacobs Management Center, Buffalo, NY, 14228, United States, saeedeef@buffalo.edu, Niam Yaraghi, Ram Gopal, Ram Ramesh The objective of this research is to study how Health Information Exchanges (HIE) improves physicians’ referral processes. Mainly, we argue that a primary care physician who has adopted HIE has access to a larger number of specialists than another physician who has not adopted HIE. This enables her to make a better decision in a referral process with regards to selecting an appropriate specialist. Second, we argue that a primary care physician who has adopted HIE tends to refer to the specialists who have also adopted HIE. This leads to better quality of care since HIE renders patient information readily available at the point of care. This study has significant implications for healthcare policy-making. 3 - Supervised Learning Methods for Predicting Healthcare Costs Mohammad Amin Morid, University of Utah, 130 University Village, Salt Lake City, UT, 84108, United States, amin.morid@utah.edu This study conducted a systematic literature review and identified five methods for predicting healthcare costs. We empirically evaluated the predictive performance of each reported approach, as well as other state-of-the-art supervised learning methods. In this comparative analysis, gradient boosting had the best predictive performance overall and for low to medium cost individuals. For high cost individuals, Artificial Neural Network (ANN) and the Ridge regression model, which have not been previously reported for use in healthcare cost prediction, had the highest performance. 4 - Physician and Patient Hybrid Community-based Recommender System for General and Diabetic Preventative Care Pamella Howell, University at Buffalo, Buffalo, NY, United States, pamellah@buffalo.edu, Mohamed A.Abdelhamid An under-explored area of recommender system research is e-health. Technology is essential in the lives of the aging population and millennials. Integrating technology with healthcare services using recommender systems could help providers transition to a Merit-based payment incentive system. We propose a recommender system that will provide general preventative care prescriptions based on context, personal health information, and domain knowledge. 5 - Standardization, Interoperability and the Patient Interest: How EHRS Influence the Hit Marketplace Nicola Mountford, Director, UCD Connected Health, University College Dublin, Dublin, Ireland, nicola.mountford@ucd.ie, Rajiv Kohli, Georgiana Ifrim Encouraged by EHR adoption incentives such as the Health Information Technology for Economic and Clinical Health (HITECH) Act, healthcare providers held bargaining power as buyers of HIT. As EHRs began to promote quality measurement and reimbursement that power shifted toward an oligopoly of EHR vendors. We explore an emerging power shift toward patients that is driven by consumerism and patients’ control of information. We track this emergent power shift and the evolving role of EHR and discuss implications for providers, vendors and patients. 361E Energy Contributed Session Chair: Shigeki Toriumi, Chuo University, Tokyo, Japan, toriumi@ise.chuo-u.ac.jp 1 - Coal Inventory Optimization for Electric Power Utilities Chung-Hsiao Wang, LG&E and KU, 102 Spruce Lane, Louisville, KY, 40207-1701, United States, chunghsiao@hotmail.com, K. Jo Min Electric power utilities that own coal power plants maintain coal inventory to hedge against coal burn and delivery uncertainties. Coal burn uncertainties are caused by fluctuations in load and generation unit availability. Delivery uncertainties are primarily caused by transportation disruptions. In this paper, we develop a mathematical model that considers those root causes to determine optimal coal inventory levels for electric power utilities. MD52

361C Road and Parking Pricing and Coordination Sponsored: TSL, Urban Transportation Sponsored Session Chair: Sayeeda Ayaz, University of Massachusetts-Amherst, 990 North Pleasant Street, Apt B24, Amherst, MA, 01002, United States, sbayaz@umass.edu 1 - A Sensitivity Analysis Approach to Robust Toll Pricing Considering Boundedly Rational Network Users Mahdi Takalloo, University of South Florida, Tampa, FL, 33620, United States, mtakalloo@mail.usf.edu, Changhyun Kwon Considering an ambiguous system of preferences and lack of complete information, decision makers are not perfectly rational, but boundely rational. Under such behavior, users do not necessarily choose the shortest path, and system travel time can be greater than in the case of rational behavior. In this research, we study a robust toll pricing strategy in congested networks under boundedly rational behavior. Using sensitivity analysis approach, our goal is to find a toll vector that minimizes the system travel time of the worst-case tolled boundedly rational user equilibrium. 2 - Cooperative Scheme - An Alternative Approach to Equitable and Pareto-Improving System Optimum Sayeeda Ayaz, University of Massachusetts-Amherst, 990 North Pleasant Street, Apt B24, Amherst, MA, 01002, United States, sbayaz@umass.edu, Song Gao, Hyoshin Park An alternative approach is adopted to achieve equitable and Pareto-improving system optimum (SO) based on cooperation among travelers assisted by defector penalty. Preliminary analysis shows that under mild conditions, a cooperative scheme always exists for a single-origin-destination two-route network to ensure equitable and Pareto-improving SO. A cooperative scheme without financial transactions always exists where all travelers are cooperators due to high defector penalty. A more practical case is also discussed where a certain number of defectors exist. A constrained optimization problem ensuring equity among cooperators is formulated. 3 - A Bi-level Parking-pricing Problem Hossein Fotouhi, George Mason University, 4630 Buckhorn Ridge, Fairfax, VA, 22030, United States, hfotouhi@gmu.edu, Elise Miller-Hooks, S.Bekhor A bi-level, multi-period, parking-pricing problem is formulated with applicability for online reservation systems. In the upper level, multiple parking operators seek to maximize their own profits. User response to these upper-level decisions are obtained from a lower-level, parking-choice problem. 4 - Optimal Pricing for Priced Managed Lanes with Multiple Entrances and Exits Venktesh Pandey, The University of Texas at Austin, Austin, TX, 78712, United States, venktesh@utexas.edu, Stephen D.Boyles Priced managed lanes are being implemented in many cities. Dynamic pricing can be incorporated to achieve several objectives, such as maximizing system level throughput or revenue. We propose and compare different approximate dynamic programming algorithms for solving an optimal distance-based pricing for a managed lane corridor with multiple entrances and exits for a time dependent demand with varying value of time across the travelers. The results are compared against other non-optimal methods used in practice. 361D Information Systems in Healthcare Sponsored: Information Systems Sponsored Session Chair: YeongIn Kim, University of Texas-Dallas, Richardson, TX, 75080, United States, yxk130930@utdallas.edu 1 - The Effect of Online Reviews on Physician Demand: A Structural Model of Patient Choice Yuqian Xu, University of Illinois at Urbana-Champaign, Champaign, IL, United States, yxu@stern.nyu.edu, Mor Armony, Anindya Ghose In this paper, we wish to derive the impact of online information on patient choice of outpatient care doctors. We are especially interested in how operational factors influence demand. To do so, we study a unique data set from one of the leading appointment booking websites in the United States, that contains online MD51

254