2016 INFORMS Annual Meeting Program

WE25

INFORMS Nashville – 2016

WE23

2 - Chemotherapy Treatment Scheduling In Public Health Juan-Carlos Ferrer, Professor, Pontificia Universidad Catolica de Chile, Santiago, Chile, jferrer@ing.puc.cl

108-MCC Surveillance of Diseases with Big Data Sponsored: Health Applications Sponsored Session Chair: Qingpeng Zhang, City University of Hong Kong, 83 Tat Chee Ave, Kowloon Tong, TX, 00000, Hong Kong, qingpeng.zhang@cityu.edu.hk 1 - Patient-specific Depression Monitoring By Selective Sensing Ying Lin, University of Washington, linyeliana.ie@gmail.com, Shuai Huang, Shan Liu Development of personalized health surveillance is enabled by sensing and information technologies. Scaling it up to large scale depression population needs a seamless combination of data analysis and sensing strategy design. We developed a selective sensing method to capture individual depression progressions from acquired health data and optimally allocate sensing resources to high-risk individuals by exploiting the similarities of their progression trajectories. The proposed method can lead to efficient and cost-effective monitoring of depression population. 2 - An Optimization Approach To Concussion Management Gian Gabriel Garcia, University of Michigan, garciagg@umich.edu, Mariel Sofia Lavieri We apply data-driven optimization to improve concussion diagnosis for athletes suspected of concussion and apply dynamic programming to solve the sequential decision-making problem of optimal return-to-play management for athletes who have concussions. 3 - Biosurveillance Of Climate Sensitive Mosquito-borne Diseases Using Online Social Media Kusha Nezafati, University of Texas, Dallas, TX, United States, Kusha.Nezafati@utdallas.edu, Yulia Gel, L. Leticia Ramirez Ramirez Chikungunya is a mosquito-borne virus that is transmitted by the same type of mosquito as dengue and zika. Chikungunya is relatively well documented in Asia, Africa, and the Indian subcontinent. In 2014 the first confirmed case of chikungunya virus has been reported in the Americas, and since then its spread has been attracting a lot of attention from the health care professionals. However, the data on chikungunya still remain relatively scarce which makes forecasting of its epidemiological curve a very challenging task. In this talk we discuss predictive utility and limitations of online social media, particularly, Google trend, as a proxy for unavailable data on chikungunya. 4 - Semantic Social Network Analysis Of Online Health Communities Ronghua Xu, City University of Hong Kong, ronghuaxu2-c@my.cityu.edu.hk, Qingpeng Zhang The understanding of how people use online health communities/groups (OHCs) to discuss health-related topics is critical to the the effective use of social media to provide social support. In this research, we collected a comprehensive dataset of mental health related OHC in China, and developed a set of methods to model and analyze the information spread and semantic patterns of users’ discussions. The results unveiled the unique topological features and semantic patterns of mental health related OHCs, with managerial insights of how to utilize OHCs to provide social support to complement conventional offline approaches. 109-MCC Operations Research in Healthcare Management in Chile Sponsored: Health Applications Sponsored Session Chair: Jorge Vera, Professor, Pontificia Universidad Catolica de Chile, Vicuna Mackenna 4860, Macul, Santiago, 7820436, Chile, jvera@ing.puc.cl 1 - Physiotherapy Treatment Appointments Scheduling Using An MDP-based System Sergio Maturana, Pontificia Universidad Catolica de Chile, smaturan@ing.puc.cl, Ignacio Lazo Scheduling physiotherapy treatment appointments in a hospital faced with a very high demand is complex. The current system in a Chilean hospital results in many patients waiting long times before their treatments can begin. This hospital has three types of specialists: a physiatrist and two types of therapist. Before the therapy can begin, patients must see the physiatrist, who indicates the appropriate treatment. We propose a scheduling system, based on a Markov Decision Process (MDP), which determines how to assign the patients to the physiatrist and how to distribute the patient’s sessions within a planning horizon in order to reduce waiting times and assure that sessions are evenly spread. WE24

This research addresses a real scheduling problem for chemotherapy patients at a Chilean public Hospital. We divide the problem into two subproblems, scheduling patients on an infinite time horizon and daily patient scheduling. The benefits of both stages are evaluated for a real case in the Hospital s Chemotherapy Unit using simulation in the first stage and solving the model to optimality in the second one. We evaluate potential opportunities for efficiency through a sensitivity analysis of key resources. 3 - A Hierarchical Solution Approach For Bed Capacity Planning Under Uncertainty In The Healthcare Service Industry Ana Celeste Batista, PhD Candidate, Pontificia Unversidad Catolica de Chile, Santiago, Chile, abatista@uc.cl, Jorge Vera Effective capacity planning under uncertainty ensures robust and consistent plans in time. The health sector is a service system of great relevance to consider better methods for inter-temporal decisions, since poor planning affects directly the welfare of people. This work presents a hierarchical multistage model applied to beds planning. We propose a solution method based on the formulation and solved using stochastic optimization. The problem is to determine the availability of beds that minimizes overall patient welfare loss as a function of waiting time. From the solution we propose policies allowing better decisions in different planning horizons. Chair: Tayo Fabusuyi, University of Michigan and Carnegie Mellon University, 5520 Baywood Street, Floor #3, Pittsburgh, PA, 15206, United States, Fabusuyi@umich.edu 1 - Minimizing Customer Waiting Time In A Unit-load Warehouse Mahmut Tutam, PhD Student, UARK, 1359 N Leverett Avenue, Apt 31, Fayetteville, AR, 72703, United States, mtutam@uark.edu, John A White Although the number of delivery trucks arriving at a unit-load warehouse per unit time may well be Poisson distributed, the time required to perform rectilinear round-trip travel between the dock and a uniformly distributed point in the storage region is not exponentially distributed. However, it can be approximated using a k-Erlang distribution. Using the Method of Moments, the value of k is determined. The analysis includes one or more docks distributed along one wall of a rectangular-shaped warehouse. The dimensions of the warehouse that minimize customer waiting time are determined for a given storage area. 2 - The Mode Most Traveled: Parking Implications And Policy Responses Tayo Fabusuyi, University of Michigan and Carnegie Mellon University, 5520 Baywood Street, Floor #3, Pittsburgh, PA, 15206, United States, Fabusuyi@umich.edu, Robert Hampshire, Zhen Qian Driving to work alone continues to be the travel mode most utilized by commuters. Using the US Census public use microdata sample (PUMS) dataset from the Pacific region of the U.S., we examine why this trend persists by generating travel mode profiles for representative individuals. In addition to the profiles, we examine the marginal effects on travel mode choice for selected explanatory variables at their representative values. The empirical exercise provides insights on the influence policy measures may have in shaping individuals’ travel mode preferences and how this will impact on demand for parking spaces. WE25 110A-MCC Logistics V Contributed Session

491