2016 INFORMS Annual Meeting Program

370 / 561

INFORMS Nashville – 2016
370
2 - Dialysis Facility Network Design
Michael Klein, McGill University, 1001 Sherbrooke Street West,
Montreal, QC, H3A 1G5, Canada, ,
Vedat Verter, Brian Moses
We study the problem of improving access for patients in a rural area by designing
a network of dialysis facilities. We incorporate the possibility of home dialysis for
which the patients need to travel to a city centre for training.
WA22
107B-MCC
Healthcare Analytics
Sponsored: Health Applications
Sponsored Session
Chair: Zahra Gharibi, Southern Methodist University,
3145 Dyer Street, Suite 372, Dallas, TX, 75205,
1 - Optimal Care Pathways For Lower Back Pain Patients
Danny R Hughes, Harvey L. Neiman Health Policy Institute,
Reston, VA, United States,
Danny R Hughes, George Mason University, Fairfax, VA,
United States, , Jeremy Eckhause,
Katharina Ley Best
Aggressive treatment of lower back pain is frequently cited as low value care that
contributes to rising health care costs with limited effects on patient outcomes. In
order to identify cost-effective early treatment policies for these patients, we
model the physician’s sequential decision problem as a finite-horizon Markov
decision process where the boundary condition is defined as the patient
completing the episode of care. We compare the results from historical claims data
with optimal decisions solved via stochastic dynamic programming to determine
desirable initial treatment strategies.
2 - Effect Of Report Cards On Kidney Transplantation Related
Decision Making
zahra Gharibi, Dept. of Engineering Management, Information,
and Systems Bobby B. Lyle School of Engineering, SMU,
Mehmet U.S. Ayvaci, Michael Hahsler
Report card programs collect and publicize information on patient outcomes as a
means of improving quality. However, it is unclear whether behavioral responses
to such programs improve patient outcomes. We study the report cards as an
incentive mechanism to induce socially-optimal medical decisions in the context
of kidney transplantation. Using a game theoretic framework, we investigate how
performance reporting and flagging for low performance influence
acceptance/rejection decisions for offered kidneys and patient selection by
transplant centers. We also study the implications of new allocation system for
such decisions.
WA23
108-MCC
Operations Research for Public Health: Data-Driven
and Dynamic Decion Making Approaches
Sponsored: Health Applications
Sponsored Session
Chair: Soroush Saghafian, Harvard Univeristy, 79 John F. Kennedy
Street, Mailbox 37, Cambridge, MA, 02138, United States,
1 - Impact Of Breast Density And Supplemental Screening Methods
On Breast Cancer Screening Policies
Mucahit Cevik, University of Wisconsin - Madison, Madison, WI,
United States, Burhaneddin Sandikci
Mammography screening is the golden standard for breast cancer screening, but it
is less accurate for women with dense breasts. Supplemental screening methods
are recently introduced to improve detection accuracy. We study the impact of
supplemental tests through incorporating breast density information in a partially
observable Markov decision process model.
2 - Impact Of Ambiguity on Medications Management Strategies:
An Application To NODAT
Alireza Boloori, Arizona State University, Tempe, AZ,
United States, , Soroush Saghafian,
Harini A. Chakkera, Curtiss B. Cook
Patients after organ transplantations receive high amounts of immunosuppressive
drugs (e.g., tacrolimus) to reduce the risk of organ rejection. However, this
practice has been shown to increase the risk of New-Onset Diabetes After
Transplantation (NODAT). We propose an ambiguous POMDP framework to
generate effective medication management strategies for tacrolimus and insulin.
Our approach increases the patient’s quality of life while reducing the effect of
transition probability estimation errors. We also provide several managerial and
medical implications for policy makers and physicians.
3 - Robust Dynamic Programming For Medical Decision Making
Lauren N. Steimle, University of Michigan, Ann Arbor, MI,
United States, , Brian T Denton
Markov Decision Processes (MDPs) are useful for studying the management of
chronic diseases, which is characterized by a series of treatment decisions under
uncertainty about the future progression of the disease. Dynamic programming
algorithms can be used to determine the optimal treatment policies for these
diseases, but these policies may not be robust to perturbations of the model
parameters. We discuss robust dynamic programming algorithms that provide
protection against variation in the estimates of MDP model parameters. We
present our results in the context of treatment of cardiovascular disease.
4 - Optimal Intervention Strategies For Hypertensive Disorders Of
Pregnancy
Aysegul Demirtas, Arizona State University, Tempe, AZ, United
States, Esma S Gel, Soroush Saghafian,
Dean Coonrod
Hypertensive disorders of pregnancy (HDP) constitute one of the leading causes of
maternal and neonatal mortality and morbidity. We consider the decision problem
of timing and mode of child delivery for women with HDP. We formulate a
discrete-time Markov decision process (MDP) model that minimizes the risks of
maternal and neonatal adverse outcomes, and assess its outcomes with clinical
data by performing probabilistic sensitivity analysis. We also build a robust MDP
model in which the transition probabilities are contained in a controllable
uncertainty set. Our robust MDP approach considers the sensitivity of estimated
transition probabilities while avoiding over-conservative policies.
WA24
109-MCC
Scheduling and Capacity Management in Healthcare
Sponsored: Health Applications
Sponsored Session
Chair: Maya Bam, University of Michigan, 1205 Beal Avenue,
Ann Arbor, MI, 48109, United States,
1 - Scheduling Operating Rooms With Elective And
Emergency Surgeries
Kyung Sung Jung, University of Florida, 364 Stuzin Hall, PO Box
117169, Gainesville, FL, 32611, United States, ,
Michael L Pinedo, Chelliah Sriskandarajah, Vikram Tiwari
Hospital accounted for 30% of total health expenditures. Operating rooms (ORs)
are typically a bottleneck during the entire processes. We solve multiple-OR
scheduling problems with elective and emergency patients. First, we provide
general guidelines for these scheduling problems, and then develop several
scheduling and rescheduling methods for these patients.
2 - Two-stage Robust Optimization Of Multi-stage Care Planning:
Formulation And Computational Challenges
Saba Neyshabouri, George Mason University, Fairfax, VA, 22030,
United States, , Bjorn Berg
We study the problem of scheduling surgeries in a block-booking setting in which
both the surgery duration and length-of-stay (LOS) in the surgical intensive care
unit are subject to uncertainty. We utilize the theory of robust optimization and
propose a novel formulation that captures the complexities in modeling
uncertainty in LOS, which is modeled as a discrete random variable. We propose
an exact solution approach and perform computational experiments to analyze
the quality of the solutions obtained. Computational challenges and future
directions are discussed.
3 - Capacity And Flow Management In Emergency Departments With
A Fast Track
Elham Torabi, University of Cincinnati, 2925 Campus Green Dr.,
Cincinnati, OH, 45221, United States,
Elham Torabi, University of Cincinnati Medical Center, Cincinnati,
OH, United States, , Craig Froehle,
Craig Froehle, Craig Froehle, Christopher Miller
In EDs with a fast track, sub-optimal allocation of patients to capacity results in
under-utilization of the fast track while main ED area is congested. Using data-
mining we identify sub-groups of moderate-acuity patients who can be treated in
the fast track instead of the main ED. We use simulation analysis to find routing
policies that better allocate the identified patients to the two capacity segments.
The proposed routing policy results in less patient waiting and more parity in
utilization of the capacity segments.
WA22