2016 INFORMS Annual Meeting Program

MB24

INFORMS Nashville – 2016

3 - Risk-adaptive E-triage In Emergency Medicine: A Prospective Analysis Scott R Levin, Johns Hopkins School of Medicine, slevin33@jhmi.edu, Matthew Toerper, Diego A. Martinez, Heather Gardener, Eric Hamrock, Sean Barnes Unprecedented levels of crowding and consequential delays in care have intensified the need for accurate triage in emergency departments (ED). For the majority of ED patients, the projected clinical course at presentation not is obvious. Almost half of adult ED visits nationally are triaged to emergency severity index (ESI) Level 3; the ambiguous midpoint of a 5-Level algorithm standard in the US. The objective of our electronic (e) triage tool is to improve differentiation of ED patients by enabling data-driven prognostication of risk of critical events and illness severity. The tool, prospectively evaluated at multiple sites, demonstrates improved detection of critically ill patients. 4 - An Evolutionary Computation Approach For Optimizing Multi-level Widespread adoption of electronic health records and objectives for meaningful use have increased opportunities for data-driven applications in medicine and healthcare. Optimally specifying multi-level patient data—which can be defined at varying levels of granularity—for predictive modeling is a challenge that must be addressed. We present a general evolutionary computational framework to optimally specify multi-level data to predict individual patient outcomes. We evaluate its performance in predicting critical events for emergency department patients across five populations. 5 - Control System For Electronic Triage In The Emergency Department: Integrating The User Into Development Loop Diego A. Martinez, Scott R. Levin, Johns Hopkins University School of Medicine, Baltimore, MD, dmart101@jhmi.edu The potential for machine learning systems to improve via exchange of informa- tion with knowledgeable users has yet to be explored in much detail. In a pilot study in an emergency department of a large hospital, nurses were presented with triage level predictions, and they were able to provide feedback through a real-time communication system. The types of some of this feedback seem prom- ising for assimilation of clinical gestalt by machine learning systems. The results show that to benefit from clinical gestalt; machine learning systems must be able to absorb information in a graceful manner and provide clear explanations of their predictions. 109-MCC Strategy and Uncertainty Invited: Strategy Science Invited Session Chair: Hart Posen, University of Wisconsin, University, Milwaukee, WI, 4, United States, hposen@bus.wisc.edu 1 - High On Innovation: The Impact Of Liberalization Policies On Creative Outcomes Laurina Zhang, Ivey Business School, Western University, London, ON, Canada, lzhang@ivey.uwo.ca, Keyvan Vakili We investigate the impact of two social liberalization policies and one anti- liberalization policy on innovation. We find that liberalization policies increase state-level patenting while the anti-liberalization policy reduces patenting. Liberalization policies increase incumbent inventors’ patenting rate and the rate of entrance into inventorship. The policies do not impact average innovation quality but patents filed after liberalization are more likely to be built upon novel technological recombinations and cite more recent prior art. The findings highlight the impact of the social context on the rate and direction of innovation. 2 - Seeding The S-curve? The Role Of Early Adopters In Diffusion Christian Catalini, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA, 00, United States, catalini@mit.edu, Catherine Tucker In October 2014, all 4,494 undergraduates at MIT were given access to Bitcoin. As a unique feature of the experiment, students who would generally adopt only mature and established technologies were placed into an early-adopter condition: suddenly they had to decide to either learn more about Bitcoin and try to use it, to bet on its volatile future by holding it, or to simply cash out and convert it into US dollars. In this paper, we explore the students’ response to the digital currency, and in particular how randomly delaying different types of students relative to their peers affected their adoption decision. Our results point to a novel mechanism through which early-adopters may influence diffusion. MB24 Data To Predict Individual Patient Outcomes Sean Barnes, Univ of Maryland-College Park, sbarnes@rhsmith.umd.edu, Suchi Saria, Scott R Levin

3 - The AQ Model Of Probabilistic Judgment And Patterns Of Risk And Return Ulrik W. Nash, University of Southern Denmark, Odense, Denmark, uwn@sod.dias.sdu.dk We have long known that uncertainty about the world is crucial for understanding profit. Moreover, there are reasons to suspect that differences in the degree of uncertainty that firms perceive about the same situation may be a fundamental cause of their performance heterogeneity. Here I introduce the AQ model of probabilistic judgment and use it to predict the flow of money between firms in factor markets. Heterogeneous distributions of profit that capture observed patterns of risk and return summarize these flows. 4 - The Impact Of Learning And Overconfidence On Entrepreneurial Entry And Exit Hart E Posen, University of Wisconsin-Madison, Madison, WI, 53705, United States, hposen@wisc.edu, John Chen, David Croson, Daniel Elfenbein Research examining entrepreneurial entry and performance highlights the phenomena of excess entry and delayed exit. We develop a computational model wherein agents learn from experience both pre- and post-entry making endogenous entry and exit decisions. The model suggests excess entry and delayed exit result from a common process — entrepreneurs’ ongoing efforts to learn about their prospects and act according to their updated information. One interesting result is that a population of unbiased entrants exhibits beliefs that overestimate their true success probabilities, providing a rational explanation for empirical patterns typically explained by individuals’ biases. Chair: Yael S Grushka-Cockayne, Darden School of Business, Charlottesville, VA, United States, GrushkaY@darden.virginia.edu 1 - Multifarious Project Management Methodologies Vered Holzmann, Tel Aviv University, veredhz@post.tau.ac.il, Yael S Grushka-Cockayne, Hamutal Weisz, Daniel Zitter In order for a project manager to deliver an effective and efficient solution to the customer’s needs, an adaptable methodology for the planning and execution of the project is to be adopted. Following the paradigm that “one size does not fit all”, meaning each project has different characteristics that should be taken into consideration when selecting the appropriate management method for a project, this study suggests the exploitation of several methodologies in a project to effectively and efficiently delivery of a successful product. The conceptual framework is based on an integration of the waterfall, agile, and TOC methods to Peter D Simonson, North Dakota State University, Fargo, ND, United States, psimonson@mac.com, Joseph Szmerekovsky For a project manager, planning for uncertainty is a staple of their jobs and education. But the uncertainty associated with a catastrophic event presents difficulties not easily controlled with traditional methods of risk management. This dissertation proposes to bring and modify the concept of a project schedule as a bounded “Alphorn of Uncertainty” to the problem of how to reduce the risk of a catastrophic event wreaking havoc on a project and, by extension. The dissertation will present new mathematical models underpinning the methods proposed to reduce risk as well as simulations to demonstrate the accuracy of those models. be applied in complex projects derived from specific attributes. 2 - Limiting Financial Risk From Catastrophic Events In Project Management. MB25 110A-MCC Project Management Methodologies Invited: Project Management and Scheduling Invited Session

156