Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

TE84

4 - An Alternative Method for Measuring Performance Outcomes of Hospital Acquired Infections Christine Pitocco, Research Professor, Stony Brook University, 333 Harriman Hall, Stony Brook, NY, 11794, United States, Herbert F. Lewis, Jonathan Liu Many hospital acquired infections (HAI) are preventable however it is estimated 2 million patients will acquire one. In this paper, we utilize NYS health data to measure the performance of hospitals in four categories of HAIs. We demonstrate the current method used by NYS is flawed and we propose and alternative method based on standard statistical methods. 5 - Characterizing Obstructive Sleep Apnea with Association Rule Mining and Sequence Analysis Rupesh Agrawal, Research Assistant, Oklahoma State University, Tulsa, OK, 74136, United States, Dursun Delen, Bruce Benjamin A healthy individual goes through five stages of sleep in one standard sequence from wake (W) stage to rapid eye movement (REM) stage and three non-REM stages (NREM1, NREM2, NREM3). Obstructive sleep apnea (OSA) is highly correlated with several life-changing health manifestations such as atrial fibrillation, obesity, hypertension, etc. Pathological changes due to OSA and the severity of OSA (mild, moderate & severe) cause disturbances to the sleep pattern transition from one stage to another. This study aims to use association mining and sequence analysis to discover underlying patterns in sleep stage transitions. 6 - Advanced Analytics in Healthcare Stephen J. Stoyan, Director, Business Analytics & Strategy, Abbott, 100 Abbort Park Road, Chicago, IL, 60064, United States Digital platforms, global demands, and customers have evolved in today’s healthcare industry, where innovation, connecting with customers, and servicing the supply chain require higher levels of precision and execution. Abbott is a global leader in healthcare and advanced analytics is becoming an integral part of business operations and sales effectiveness. We present analytically tuned tools that are providing information that discover new opportunities, allow for better predictions, and provide business value. Chair: Jens Brunner, University of Augsburg, Universitatsstra e 16, Faculty of Business and Economics, Augsburg, 86159, Germany 1 - Master Surgical Scheduling and Elective Surgeries Scheduling M’Hallah Rym, Professor, Kuwait University, P.O. Box 5969, Safat, Kuwait, 13060, Kuwait This paper proposes a stochastic model that determines a two-week master surgical schedule for a leading hospital. It determines the optimal expected number of surgeries per surgery type to plan for every two-week planning horizon. The model accounts for the random duration of both surgery times and post-surgery length of stays and the availability of post-surgery beds. It maximizes the expected operating theater’s throughput. The model is solved via a sample average approximation with ranking and selection. Its application illustrates the potential savings it induces, the results’ sensitivity to the length of the planning horizon and to the pooling of post-surgery beds. 2 - A Two Stage Stochastic Mixed Integer Programming Approach for Short Term Nurse Schedule Adjustments Osman Tuncay Aydas, Assistant Professor of Supply Chain Management, Oakland University, 275 Varner Drive, Elliot Hall 308, Rochester, MI, 48309, United States, Anthony Ross, Buket Aydas We study short-term staffing adjustments of nurse schedules for the upcoming 4- hour shift. Developed two-stage stochastic programming model minimizes total nurse staffing costs, and the cost of adjustments to the original schedules developed in the medium-term planning phase, while ensuring adequate coverage of nursing demand. Our findings indicate, short-term stochastic schedule adjustments can deliver cost savings of up to 20%. As the number of unscheduled admissions increase there is a greater need for scheduling flexibility. We demonstrate the significance of part-time nurses, which alleviate under- and over-staffing risks and provide scheduling flexibility to the nurse managers. 3 - Differentiating the Types of Inter-hospital Transfer Helps Hospital Choice Decisions Lei (Raymond) Fan, Student, University of Houston, 4800 Calhoun Rd, Houston, TX, 77004, United States, Ming Zhao, Xiaosong David Peng Inter-hospital patient transfer adds complexity to healthcare operations. We study inter-hospital transfers at the patient level. We develop a methodology to classify the transfers, and for each transfer type, we investigate the effects of transfer decisions on patient care performance. The results may be used by healthcare providers to develop better inter-hospital transfer guidelines. n TE83 Hyatt, Remington Practice- Health Care V Contributed Session

4 - Does Announcing the Visit Matter? An Empirical Examination in US Nursing Homes Sehwon Kang, PhD Student, University of Minnesota, Carlson School of Management, Minneapolis, MN, 55455, United States, Rachna Shah, Kevin Linderman Recently, healthcare accrediting agencies changed their on¡-site accreditation survey visits from announced to unannounced. In this study, using panel data from more than 3,000 US nursing homes over 4 years, we empirically examine the impact of this two regimes in terms of improving and sustaining quality performance. 5 - Leveling of Intensive Care Unit Utilization by Efficient Surgery Scheduling Jens Brunner, Prof. Dr., University of Augsburg, Universitaetsstrasse 16, Faculty of Business and Economics, Augsburg, 86159, Germany, Steffen Heider The intensive care unit (ICU) is an expensive hospital resource serving both emergency and elective patients. The vast majority of ICU patients arrive from the operating theatre. In order to increase efficiency, we propose a mixed-integer programming model to optimize surgery scheduling with the goal of balancing the expected occupancy of patients in the ICU. The model is applied to two years of data from our cooperating hospital. We show that a more balanced utilization is possible without changing the master surgery schedule and that days without elective surgery have a negative effect when leveling the utilization at the ICU. n TE84 Hyatt, Russell Practice- Disaster and Disruption Management I Contributed Session Chair: Duygu Pamukcu, Middle East Technical University, Ind., Ankara, 06400, Turkey 1 - A Multistage Robust Model for Blood Transshipment Problem in Disasters Jie Deng, Lecturer, Jiangnan University, School of Business,Jiangnan University, Wuxi, China This research introduces a robust multi-echelon multi-stage model that can assist in blood transshipment problem with uncertain demand and supply in disasters, the proposed model considers service level-based policy in the blood scheduling process to minimize the unsatisfied demand and the time span between blood production in regional blood centers and consumption in demand zones for reducing blood shortage and wastage. 2 - A Hypercube Queuing Equilibrium Approach to Dynamic Volunteer Firefighter Planning under Probabilistic Service Disruptions Han Liu, University of Illinois, Urbana-Champaign, 722 Foley Ave, Champaign, IL, 61820, United States, CHAO LEI, Yanfeng Ouyang We propose a multi-period hypercube queuing equilibrium model for the dynamic volunteer firefighter planning problem with multiple rescue rounds, with the consideration of that rescue services are under the risk of being disrupted. The applicability of the proposed model is tested on an empirical case study in Baoding. Numerical results show that more rescue resources should be dispatched in the first rescue round if the disaster deteriorates fast. 3 - Minimization of the Social Impact of an Earthquake in Lima Metropolitana and Callao Optimizing the Speed of Response with Mathematical Models Renzo Benavente, Pontificia Universidad Catolica del Peru, Universitaria 1801, Lima, LIMA 32, Peru, Jonatán Edward Rojas Polo This research shows opportunities for improvement in the response plan after an earthquake for the distribution of humanitarian aid. The proposed methodology begins with a clustering of the geographic area of Lima Metropolitana and Callao in 9 groups with a main warehouse in each. Then, the amount and location of temporary warehouses that will supply the affected population is calculated. Finally, a linear programming model is executed for the distribution of humanitarian aid goods from the central warehouse to the temporary ones. With this, it is possible to reduce the response time and make it more dynamic for the population.

409