Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

TD77

regret and expectation-based objectives, and derive various key structural properties of their optimal solutions. Our case study establishes the benefits of the proposed models over current policies. 2 - Precise: Pancreatic Cancer Prioritization and Screening Evaluation Tool Lena Abu-El-Haija, NC, United States, Julie Simmons Ivy, Osman Ozaltin, Walter G. Park Pancreatic cancer (PC) is the fourth cause of cancer death in the US and is projected to be second by 2030. Incidence of PC is low, but the mortality risk is high with 8% five-year survival. Early stage PC can be asymptomatic leading to late diagnosis after it may have metastasized. The key to improving survival is screening. SEER and secondary data were used to develop a model-based framework for evaluating screening policies. The model provides quantitative clarity to identify areas for future data collection to understand the behavior of PC and whom it targets. The model enables the development of personalized screening policies to include race, sex, and age that influence incidence and progression. 3 - Improving Access of Low Back Pain Patients Through Prioritization at a Neurosurgery Clinic Esma S. Gel, Associate Professor, Arizona State University, 699 S. Mill Ave., Tempe, AZ, 85287-8809, United States, Derya Kilinc Low back pain (LBP) is often cited to cause significant health impairments for a large fraction of the population. Studies point to the high variability in treatment approaches and frequent mismatch between patient needs and services offered by providers from different specialties. This mismatch results in critical access problems for patients that truly need critical interventions such as surgery within a reasonable timeframe. We present findings from a a number of ongoing projects to improve access of LBP patients at a neurological surgery clinic. Our analysis points to the importance of prioritizing surgical patients and demonstrates the potential improvements in patient access. 4 - A Latent-Variable Approach to Potential Outcomes for Emergency Department Admission Decisions Gabriel Zayas-Caban, University of Wisconsin, Mechanical Engineering Building, 1513 University Avenue, Madison, WI, 53706-1691, United States, Amy Cochran, Paul Rathouz, Keith Kocher In emergency departments, providers weigh admissions against continued monitoring and treatment often without knowing their health state. We describe a framework to estimate the causal effect of admission decisions on patient outcomes from observational data. Admission decisions are modeled as a decision-making process in which the patient’s health needs is a latent variable. We estimate latent health needs with limited knowledge of the decision process using the potential outcomes framework and assuming potential outcomes are independent from admission decisions within a latent health state. We apply our approach to patient encounters with the University of Michigan Health System. 5 - A Data-driven Partially Observable Markov Decision Process for Optimizing Individualized Surveillance Strategies for Prostate Cancer Weiyu Li, IOE department, University of Michigan, Ann Arbor, MI, United States, Brian T. Denton Active surveillance (AS) is a strategy that involves regular clinical examinations, biomarker tests, and biopsies to monitor patients diagnosed with low-risk prostate cancer. The ideal strategy must strike a balance between the burden of testing and the benefit of early detection of progression to high-risk prostate cancer. We propose a hidden Markov model (HMM) to estimate the progression rate of cancer, and the sensitivity and specificity of the biomarker tests using longitudinal data from a large surveillance study. We use the HMM as the basis for a partially observable Markov decision process (POMDP) and present results for optimal strategies. n TD77 West Bldg 213A Health and Humanitarian Supply Chain Management Sponsored: Public Sector OR Sponsored Session Chair: Pinar Keskinocak, Georgia Institute of Technology, Atlanta, GA, 30332, United States Co-Chair: Seyma Guven-Kocak, Georgia Institute of Technology, Atlanta, GA, 30340, United States Co-Chair: Pelin Pekgun, University of South Carolina, Columbia, SC, 29205, United States 1 - The Roadside Healthcare Facility Location Problem Harwin de Vries, INSEAD, Boulevard de Constance, Fontainebleau, 77210, France, Albert P. M. Wagelmans, Joris Van de Klundert

n TD75 West Bldg 212B Joint Session MAS/DAS Joint Session Chair: Gregory S. Parnell, University of Arkansas, Fayetteville, AR, 72701, United Statesu 1 - Improving Analysis of Alternatives The DoD is required to perform an Analysis of Alternatives (AoAs) to support major program milestone decisions early in the system life cycle. This paper reviews the statutory, defense department, and service guidance for AoAs. One of the major results of the AoA is a trade-off analysis that informs senior decision makers about the technical, cost, schedule risks of the potential new program. We report on a book project to document best practices for trade-off analyses and our research in the Engineering Resilient Systems program to improve the effectiveness and efficiency of AoAs. 2 - Sensor Modeling for Military Applications Randy K. Buchanan, USACE - ERDC, 3909 Halls Ferry Road, Vicksburg, MS, 39180, United States Sensor models capabilities within the DoD were investigated for supporting tradespace exploration and analysis of unmanned aerial systems (UAS) integrated sensor systems and sensor payloads. The research was conducted to support examining the performance trades of various modalities including size, weight, and power to a rotary-wing and fixed wing aircraft. Sensor modeling and simulation supports the development and verification of optimized sensor systems and is required when knowledge of sensor system tradeoffs are deemed necessary, or the development of cutting-edge deployed systems are required. 3 - Integrating the Cost Model with Set-Based Design Randy K. Buchanan, USACE - ERDC, 3909 Halls Ferry Road, Vicksburg, MS, 39180, United States Integrating cost models with set-based design concepts improves decision making capabilities for analysis of alternatives when implementing model-based engineering early in the design process. These principals were applied to a conceptual Army ground vehicle while integrating an engineering model with a cost model. Stakeholder requirements were used to incorporate value into the design tradespace and sensitivity analysis was applied to input parameters to determine effect on cost. The process of integrating SBD into the cost, engineering, and value models generated analytical insights of the design alternatives within the tradespace that provide guidance for future integration efforts. 4 - Improving Personnel Selection through Multi Objective Decision Analysis Christopher Smith, PhD, Air Force Institute of Technology, OH, United States, Joshua Deehr Personnel selection has always and will continue to be a challenging task for the military special operations. They want to select the best out of a number of qualified applicants. What makes a successful candidate and how to compare candidates against each other are some of the difficulties that top tier organizations like the special operations face. Value Focused Thinking places criteria in a hierarchal structure and quantifies the values with criteria measurements, known as a decision model. The selection process can be similar to a college selecting their students. This research used college student entry data and strategic goals as a proxy for special operations applicants and standards. n TD76 West Bldg 212C Joint Session MIF/HAS: Models and Methods for Improving Patient Outcomes Sponsored: Minority Issues Sponsored Session Chair: Gabriel Zayas-Caban, Madison, WI, 53706-1691, United States 1 - Optimal Biomarker Testing and Subject Classification Under Limited Information Seyedehsaloumeh Sadeghzadeh, Virginia Institute of Technology, Blacksburg, VA, 24060, United States, Douglas R. Bish, Ebru Korular Bish Bio-marker testing is essential in public health screening. For many diseases, related bio-markers may have a wide range of concentrations among subjects, particularly among the disease-positive subjects. Furthermore, bio-marker levels may fluctuate based on external and subject-specific factors. These sources of variability can increase the chance of subject misclassification based on a bio- marker test. We study the minimization of the misclassification cost considering Gregory S. Parnell, University of Arkansas, 4207 Bell Engineering Center, Fayetteville, AR, 72701, United States, Edward A. Pohl, Simon Goerger

375