Informs Annual Meeting 2017

MC01

INFORMS Houston – 2017

Monday, 1:30 - 3:00PM

9 - Causal Analyses for the Impact of Homecare Services on Patient Discharge Disposition Biplab Bhattacharya, University at Buffalo, 80 Sterling Ave, Upper,

Buffalo, NY, 14216, United States, biplabsu@buffalo.edu, Sabrina Casucci, Yuan Zhou, Li Lin, Alexander Nikolaev

MC01

Improving health outcomes for low-income adult Medicaid beneficiaries is challenging due to high rates of chronic disease comorbidities, limited mobility, and inconsistent access to care. We demonstrate how causal inference modeling can derive actionable insights from a home healthcare claims dataset. This innovative understanding of the impact of chronic disease comorbidities and service utilization on patient health outcomes can inform healthcare policies that improve the effectivity of home health care services. 10 - Using Nash Equilibrium Mathematical Model to Explore the Impact of Slots Competition on Profitability of Airlines Yuan Feng, National Tsing Hua University, Dept of Industrial Engineering and Management, Hsinchu, 30073, Taiwan, erinfeng830624@gmail.com, Yu-Ching Lee Having the data of known number of airlines, passenger capacity, load factor, operating cost and average fare income, we use Nash equilibrium model to find an solution, which could be used to analyze the profitability of the airlines. We derive the Nash equilibrium formulation from the KKT condition of the profit maximizing problems for every airline. We assume that the airline will not withdraw from any market. The analysis under two conditions for satisfying the market demand and the total number of flights shows that the routes will concentrate on the higher lucrative routes. Furthermore, by reducing the total slots in a certain range will enhance the average passenger load factor and the profits. 11 - Developing a Location Model to the Planning of Sustainable Perinatal Care Network in Korea Hoon Jang, Science and Technology Policy Institute, 370 Sicheng-daero, Sejong, 30147, Korea, Republic of, hoonjang@stepi.re.kr, Kyosang Hwang, Taeho Lee, Taesik Lee The reformation of rural perinatal service systems is an important public health task in Korea. For this, the government set up a support program; locating hospitals is one of key strategies to cover populations. Motivated this, we proposed the location model addressing how to design a rollout plan to maximally serve the demand. The novelty of our model lies in the incorporation of the uncertainty of future demand and individuals’ hospital choice behavior. To build this model, we combine a discrete choice model with a location model, and the model is formulated by using the robust optimization framework. With the case study using a real dataset, we provide insights on how to design a better rollout plan. 12 - Statistical Modeling and Analysis of Chronic Disease Progression using Electronic Health Record (EHR) Data Rema Padman, Professor of Management Science & Healthcare IS, Carnegie Mellon University, 5000 Forbes Avenue, Pittsburgh, PA, 15213, United States, rpadman@cmu.edu, Vijaya Priya Rama Vijayasarathy Improved risk stratification of chronic disease patients may enable early identification of disease progression and allow appropriate interventions to mitigate preventable adverse outcomes and improve care delivery. Leveraging 22- years of EHR data and two clinical markers of chronic kidney disease, we develop and compare data-driven, statistical models to identify and profile groups of patients following four distinct trajectories of disease progression and their statistically significant covariates, potentially enabling better management of the patients. 13 - Supporting the Definition and Analysis of Cervical Cancer Public Health Policies Ivan Mura, Visiting Professor, Universidad de Los Andes, Carrera 1 Este #19A-40, Bogota, 111711, Colombia, i.mura@uniandes.edu.co, Karen Daniela Angulo, Maria F. Cortés, Daniel Felipe Otero, Raha Akhavan-Tabatabaei Cervical cancer is a global health issue, and developing countries bear the highest death rates. Though, cervical cancer etiology is well understood and prevention is possible. Both vaccines and effective screening tests exist, which can significantly reduce incidence. While designing cost-effective public health policies is of paramount importance in countries with limited resource availability, it requires an ability to generate reliable long-term predictions about the dynamics of a population health state, a task that entails significant complexity. This poster presents a simulation-based computational tool that can assist decision-makers in shaping national public health policies.

310A Decision Analysis, Game Theory, Homeland Security, and Disaster Management, Part I Sponsored: Decision Analysis Sponsored Session Chair: Jun Zhuang, University at Buffalo, SUNY, Buffalo, NY, 14260, United States, jzhuang@buffalo.edu 1 - System Defensibility in Reality Alexander Gutfraind, Uptake Technologies, 600 W. Chicago St, A system is termed defensible if modest investment of resources can significantly improve the outcome to the defender. Defensibility in reality is made quite complex when e.g. the more valuable assets are also more costly to defend. We report on how these realities affect defensibility and optimal defense strategies. 2 - Individual Terrorism Risk Assessment: A Signal Detection Theory Framework Richard S.John, University of Southern California, Dept of Psychology MC 1061, Los Angeles, CA, 90089-1061, United States, richardj@usc.edu Increasingly analytics are utilized to assess individual terrorism threat. Such “risk- based” approaches are designed to bin individuals in order to apply differential levels of scrutiny based on individual characteristics related to terrorism threat. A Signal Detection Theory framework is proposed, and sensitivity analysis is conducted to determine conditions in which such threat assessments are useful. The viability of risk-based approaches depends heavily on priors, indicator diagnosticity, and the relative cost of false-positives and false negatives. Risk- based approaches are severely limited for highly adaptive adversaries who can effectively diminish indicator diagnosticity. 3 - An Overview of Uncertainty-tolerant Decision Support for Cybersecurity Cyber system defenders face the challenging task of continually protecting critical assets and information from a variety of malicious attackers. Defenders typically function within resource constraints, while attackers operate at relatively low costs. Thus, design and development of resilient cyber systems that support mission goals under attack, while accounting for the dynamics between attackers and defenders, is an important research problem. This talk will highlight decision support challenges under uncertainty within non-cooperative cybersecurity settings. 4 - Robust Resource Allocation to Interdependent Networks under Multiple Disruption Scenarios Hannah Lobban, University of Oklahoma, 202 W. Boyd St, Room 124, Norman, OK, 73069, United States, hannahrose@ou.edu In recent years, the threat of intentional attacks on infrastructure networks, such as water, electrical, communication, and transportation, has grown. Due to interdependency in the networks, a disruption could have far-reaching effects on the networks’ ability to meet demand. Common disruption scenarios include degree- and capacity-based attacks; however, attackers targeting a network may give additional weight to components with less quantitative (e.g., religious, governmental) significance. This work attempts to determine the allocation of defensive resources that accounts for all these factors while minimizing both costs and the unmet demand in the disrupted network. 5 - Estimating Effectiveness of Investment, Optimal Resource Allocation, and Predictive Risk Analytics for Fire Protection Vineet Madasseri Payyappalli, University at Buffalo, SUNY, Buffalo, NY, United States, vineetma@buffalo.edu, Adam Behrendt, Jun Zhuang Fire hazards are an everyday phenomenon, and the estimated total cost of fire was $329 billion in 2011. Using the large amount of data available from National Fire Protection Association (NFPA) and National Fire Incident Response System (NFIRS), we create empirical and theoretical models to estimate the effectiveness of investment and to formulate optimal resource allocation strategies. Our results show that fire losses have decreased exponentially in investment with high R2 values (~0.8) and also show potential under- and overspending in fire protection for certain years/regions. The research will be of use to policymakers and analysts in fire protection, and will help in mitigating fire losses. Suite 775, Chicago, IL, 60654, United States, agutfraind.research@gmail.com, Vicki Bier Samrat Chatterjee, Pacific Northwest National Laboratory, Richland, WA, United States, samrat.chatterjee@pnnl.gov

201