INFORMS 2021 Program Book

INFORMS Anaheim 2021

TE22

2 - Implicit Racial Bias in Healthcare Scheduling Delays Michele Samorani, Santa Clara University, Leavey School Of Business, Santa Clara University, Santa Clara, CA, 95053, United States, Nan Liu, Shannon Harris, Haibing Lu We show that the traditional objective of minimizing patients’ waiting time and provider overtime leads to scheduling the patients at higher risk of no-show farther into the future than other patients. The reason is that by doing so, the clinic increases the show probability of the patients that are more likely to show up while decreasing the show probability of the patients that are less likely to show up. This strategy will consequently reduce the variability in number of shows, and ultimately decrease the schedule cost. However, because no-show probabilities are often correlated with race, this scheduling strategy results in unintended racial disparities in terms of access to care. 3 - Adaptive Design of Personalized Dose-finding Clinical Trials Saeid Delshad, Clemson University, Clemson, SC, 29630, United States, Amin Khademi A key step toward personalized medicine is redesigning dose-finding clinical trials and finding the right therapeutic dose for each patient type. This work studies a problem of fully response-adaptive Bayesian design of Phase II dose-finding trials with patient information, where the decision maker seeks to identify the right dose for each patient type by minimizing the overall expected variance of the target dose over all patient types. We formulate this problem by a SDP and exploit its properties. Since the optimal solution is intractable, we propose an approximate policy by an adaptation of a one-step look-ahead framework. We show the optimality and asymptotic rate of sampling of the proposed policy for a case with homogeneous patients and two doses. We also adapt other policies such as posterior adaptive sampling and test their performance against our proposed policy. TE22 CC Room 204B In Person: Healthcare Operations Management General Session Chair: Alex Mills, Baruch College, City University of New York, New York, NY, 10010-5585, United States Co-Chair: Masoud Kamalahmadi, University of Miami, Bloomington, IN, 47405-1701, United States 1 - Optimization of Pediatric Vaccines Distribution Network Configuration under Uncertainty Zahra Azadi, University of Miami Herbert Business School, Coral Gables, FL, 33158, United States, Sandra D. Eksioglu, Harry Neil Geismar Millions of young people are not immunized in low- and middle-income countries because of low vaccine availability resulting from inefficiencies in cold supply chains. We create supply chain network design and distribution models to address the unique characteristics and challenges facing vaccine supply chains in these countries. The models capture the uncertainties of demand and the resulting impacts on immunization, the unique challenges of vaccine administration, the interactions between technological improvements of vaccines and immunizations, and the trade-offs between immunization coverage rates and available resources. The objective is to maximize the percentage of fully immunized children and the vaccine availability in clinics. We tested the model using Niger’s Expanded Program on Immunization, which is sponsored by the World Health Organization. 2 - Telehealth Expansion and Patient Demand in Acute Care Ozden Engin Cakici, American University, Washington, DC, 20016, United States, Alex Mills Many healthcare providers have recently expanded telehealth services where patients can see a doctor online. In theory, increasing capacity of telehealth services should expand the provider’s panel size because telehealth is more convenient, but the drawback of telehealth is that it may require a follow up visit for a physical exam. We model the patient’s strategic choice for acute care among balking, telehealth, and walk-in office visit using a game theoretic model, as a function of the provider’s capacity allocation to telehealth. We find that too much expansion of telehealth can decrease the provider’s panel size, and we discuss policy implications that emerge from this result. 3 - Planning Follow-up Capacity to Reduce Hospital Readmissions Alex Mills, Baruch College, City University of New York, New York, NY, 10010-5585, United States, Jonathan Eugene Helm Many hospitals now schedule patients for follow-up care at the time of discharge, in an effort to reduce readmissions. Using analytical models and a case study with real hospital data, we show that careful capacity planning in a post-discharge follow-up program can increase the number of patients with timely follow-up appointments by about 30% without increasing the cost of the program.

TE20 CC Room 203B In Person: Emerging Topics in Agricultural Supply Chains General Session

Chair: Somya Singhvi, MIT, Cambridge, MA, 02139-4230, United States 1 - Yield Improvement Through Smallholder Farmer Certifications: Price and Profit Implications Utku Serhatli, Nova School of Business and Economics, Lisbon, Portugal, Guillaume Roels In agricultural-intensive economies, manufacturers often help smallholder farmers improve their yields through training and certification programs. However, and perhaps paradoxically, some farmers feel that these programs can lower their profit, in part due to a decrease in commodity prices. Using a Cournot model, we show that a) certification programs can push prices down, which may indeed decrease profits of some farmers, b) the objectives of minimizing market prices and protecting farmer well-being might be conflicting and c) certifying low- cost farmers performs well in terms of both individual and aggregate farmer well-being. 2 - Restricting Mobile Data Can Accelerate Digital Development Evidence From a Smartphone Experiment Kamalini Ramdas, London Business School, Regent’s Park, London, NW1 4SA, United King dom, Alp Sungu This paper identifies a significant and heretofore unnoticed barrier to digital development: data shortages. Low-income smartphone users in Mumbai who were randomly assigned to a data plan with daily usage caps increased late-plan access of WhatsApp invites to health camps, increased attendance at these camps, and reduced social media checking; without compromising sleep or subjective well-being. Our novel smartphone usage tracking app reveals why. Absent usage caps, participants binge on YouTube and social media, resulting in subsequent data shortages. Consequently, access to information significantly reduces later in a data plan. Participants with low self-control and high fear of missing out are more likely to prefer this data-saving mechanism, even at a higher price. Data caps present a non-obvious and inherently cost-saving path to alleviating poverty. 3 - Designing Payment Models for the Poor Sasa Zorc, University of Virginia, Charlottesville, VA, United States, Bhavani Shanker Uppari Some life-improving technologies for the poor are unaffordable to them, as their limited liquidity puts the purchase costs out of reach. Thus, business models have emerged where the consumers pay a fraction of price upfront to acquire the technology, and make a series of payments for continued access, at the end of which the ownership of the technology may be gained by the consumers. This offers flexibility to sometimes pay low/no amounts (alleviating cash constraints), and at the same time, disciplines consumers by remotely turning off the technology when they lag behind on payments (reducing default risk). Using the optimal contracting approach, we investigate the payment mechanisms that balance flexibility, discipline, and ownership incentives. Several implementable features that improve both the firm’s profits and the consumers’ welfare emerge from our analysis. TE21 CC Room 204A In Person: Equity and Social Justice in Health Care Operations General Session Chair: Michele Samorani, Santa Clara University, Santa Clara, CA, 95053, United States Co-Chair: Amin Khademi, Clemson University, Clemson, SC, 29634, United States 1 - Fair Allocation Decisions in Multi-stakeholder Healthcare Scenarios David Rea, Lehigh University, Bethlehem, PA, 45221-0211, United States, Leonardo Lozano, Craig Froehle Healthcare is rife with difficult multi-stakeholder tradeoffs. Decisions have direct implications for the well-being of patients, providers, and healthcare systems. In such scenarios, inter-stakeholder fairness is a natural concern. Importantly, stakeholder groups are not monoliths. Individual patients and providers differ in their needs, preferences, and expectations. Simultaneously management of these intra-stakeholder and inter-stakeholder tradeoffs is further complicated by the discrete nature of healthcare allocation decisions. This research proposes a framework for incorporating fairness into algorithmic objectives. The generalizability of the framework is shown through examples from teleradiology and inter-hospital transport.

143