Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

SA12

errors. We find that when congestion increases, physicians prevent an increase in wrongful discharges by lowering the threshold for hospital admission. This leads to a surge in avoidable hospitalizations and creates “false demand for hospital beds when ED physicians should be protecting this constrained resource. Introducing a second gatekeeping stage - to which physicians can pass patients if they are unable to make an accurate referral decision - can mitigate this effect. 2 - Inpatient Overflow: An Approximate Dynamic Programming Approach Pengyi Shi, Purdue University, 403 W. State St, Krannert School of Management, Kran 472, West Lafayette, IN, 47907, United States, Jim Dai Due to the inherent variations in arrivals and discharges, hospital managers may assign a patient to a non-primary inpatient ward, especially when the patient has boarded in the Emergency Department (ED) for several hours. Such overflow practice helps alleviate ward congestion but may compromise quality of care. We formulate this overflow decision problem as a Markov decision process (MDP) within a multi-class, multi-pool queueing network setting. To overcome the curse-of-dimensionality, we develop an approximate dynamic programming (ADP) algorithm, where we use a novel combination of a fluid control model and an integrated single-pool model to guide the choice of the basis functions. 3 - Impact of Health Information Technology Enabled Coordination on Inpatient Length of Stay Temidayo Adepoju, Boston University, Boston, MA, 02215, United States, Helen Jin, Anita Tucker, Rebecca Lara, Chris Manasseh Preparing patients for discharge is a complex process that involves the aggregate effort of multiple care providers. We examine how health IT-enabled coordination in discharge process can improve inpatient length of stay (LOS) through the use of an electronic tool called a pre-discharge order (PDO). A pre-discharge order is a tool in the EHR to facilitate discharge process. When completed, it appears as an electronic signal in a patient’s chart to alert care providers of any discharge barriers a patient may have. We find that inpatient LOS is reduced when discharge process is coordinated using the pre-discharge order, particularly for patients with complex discharge placement. 4 - Capacity Pooling in Hospitals: The Hidden Consequences of Off-service Placement Hummy Song, The Wharton School, University of Pennsylvania, Philadelphia, PA, United States, Anita L. Tucker, Ryan Graue, Sarah Moravick, Julius J. Yang Given a highly variable patient census at the service level yet a fixed allocation of inpatient beds to services, a significant portion of admitted patients become “off- service” patients. These patients are physically located in a bed that belongs to a different service (e.g., general surgery) while still being cared for by a physician of the service (e.g., cardiac medicine). We examine the tradeoffs and consequences of assigning incoming patients to an off-service bed as opposed to an on-service bed. n SA14 North Bldg 126C Innovative Service Models Sponsored: Manufacturing & Service Oper Mgmt/Service Operations Sponsored Session Chair: Guillaume Roels, INSEAD, Fontainebleau, 77305, France 1 - Information, Subsidies or Surge? An Experimental Approach to Drivers Relocation Guangwen Kong, MN, United States, Yinghao Zhang, Zhongzhong Jiang Relocating drivers to different geographical regions to reduce the mismatch between the demand and supply is important for on-demand ride hailing platform. We consider a variety of policies to incentivize drivers for relocation, including sharing information about supply and demand, suggestions for relocation, relocation cost subsidies, surge pricing, and their combinations. We examine the effectiveness of those policies to relocate drivers in a series of controlled lab experiments. 2 - Examining Links Between Service Design and Customer Performance in High Anxiety Settings Michelle A. Shell, Harvard Business School, 1 Upland Woods Circle, Unit 406, Norwood, MA, 02062, United States, Ryan Buell Many service settings are rife with anxiety, yet the impact of anxiety on service relationships is not well understood, nor is it often factored into service design. Through a series of lab and field experiments, conducted in the high-anxiety domain of financial services, we document the negative effects of anxiety on customer performance. We further demonstrate how providing customers with access to human contact when they’re feeling anxious can improve customers’ willingness to engage, elevate their satisfaction with their own decisions, and engender greater trust in the company.

n SA12 North Bldg 126A

Joint Session OR Frontiers/Practice Curated: Novel Applications of Operations Research Emerging Topic: OR Frontiers Emerging Topic Session Chair: Srinivas Bollapragada, GE Global Research Center, Niskayuna, NY, 12309, United States 1 - Optimal Scheduling of Field Resources for Power Plant Outages Rajeev Namboothiri, GE Global Research, John F. Welch Technology Centre, Plot # 122, EPIP Phase 2,, Bangalore, 560066, India, Srinivas Bollapragada, Babu Narayanan, Guillaume Camard, Ahmed Khattab Power plants around the world require field resources to carry out maintenance outages. A host of complex constraints restrict the eligibility and availability of a field resource to perform a maintenance task. We developed a Field Resource Scheduling algorithm minimizing the total cost associated with field services while meeting all the real-world constraints. Using exhaustive preprocessing, an intelligently designed cost function, and a novel approach that iteratively solves a series of linear programs, we achieve the optimal solution within a few minutes. GE Power has been using this tool in the Middle East Africa region, resulting in operational cost savings of millions of dollars. 2 - Optimizing Maintenance Operations at GE Wind Sites Rajesh Tyagi, GE Global Research Center, 1 Research Circle, Niskayuna, NY, 12309, United States, Nitish Umang GE Renewable, a leading producer of wind turbines, also sells service agreements to guarantee their operating availability at wind sites. These sites have up to 200 turbines, with outstanding maintenance work for a subset of turbines at any given time. We present a Digital Plan of the Day (DPOD) system that has been developed to optimize the maintenance operations to determine which turbines will have maintenance performed that day and develops a schedule for each crew. DPOD has so far been implemented at over 50 wind farms in North America. 3 - Rail Network Operations Optimization Srinivas Bollapragada, Chief Scientist, GE Global Research Center, 1 Research Circle, K1-4A50A, Niskayuna, NY, 12309, United States Each of the seven major railroads (called Class 1 railroads) in North America owns over 20,000 miles of track and runs over 1000 trains per day. Moving freight on such complex networks requires meticulous planning and execution. Multiple opportunities exist to realize large savings from improving railroad operations. For example, increasing the average speed of freight trains by one mile per hour saves a Class 1 railroad around $200 million per year. Improving terminal (rail yard) operations and decreasing fuel consumed by locomotives also result in similar sized savings. We will describe some of our work on developing algorithms to improve operations of rail networks. 4 - Rail Network Operations Optimization Srinivas Bollapragada, PhD, Chief Scientist, GE Global Research Center, Niskayuna, NY, USA. Each of the seven major railroads (called Class 1 railroads) in North America owns over 20,000 miles of track and runs over 1000 trains per day. Moving freight on such complex networks requires meticulous planning and execution. Multiple opportunities exist to realize large savings from improving railroad operations. For example, increasing the average speed of freight trains by one mile per hour saves a Class 1 railroad around $200 million per year. Improving terminal (rail yard) operations and decreasing fuel consumed by locomotives also result in similar sized savings. We will describe some of our work on developing algorithms to improve operations of rail networks. n SA13 North Bldg 126B Hospital Operations Sponsored: Manufacturing & Service Oper Mgmt/Healthcare Operations Sponsored Session Chair: Anita L. Tucker, Boston University, Boston, MA, 02215, United States 1 - Gatekeeping Under Congestion: An Empirical Study of Referral Errors in the Emergency Department Michael Freeman, INSEAD, 1 Ayer Rajah Avenue, Singapore 138676, Singapore, Stefan Scholtes, Susan Robinson Using data from an ED, we study the effect of congestion on the accuracy of gatekeeping decisions (hospital admission or discharge home) and the effectiveness of a second gatekeeping stage (a clinical decision unit) in reducing

6