2016 INFORMS Annual Meeting Program

MC35

INFORMS Nashville – 2016

3 - Using Patient-centric Quality Information To Unlock Hidden Health Care Capabilities Guihua Wang, Ross School of Business, University of Michigan, Ann Arbor, MI, 48105, United States, guihuaw@umich.edu Jun Li, Wallace J Hopp We document a wide variation in quality among 188 surgeons at 35 hospitals in New York state that perform mitral valve surgery. Our analysis shows that patients of different demographics and levels of acuity benefit differently from elite surgeons. In this paper, we develop an approach for computing patient- centric information from outcome data and evaluate the potential health benefits from using such information to guide patients to surgeons. We estimate that the total societal benefits from using patient-centric information are comparable to those achievable by enabling the best surgeons to treat 40% more patients under population-average information. 4 - Evidence Of Upcoding In Pay-for-Performance Programs Hamsa Sridhar Bastani, Stanford University, Stanford, CA, United States, hsridhar@stanford.edu, Joel Goh, Mohsen Bayati Medicare has sought to improve patient care by penalizing providers for hospital- acquired infections (HAIs). However, these efforts may be undermined if providers upcode, i.e. mis-report HAIs (possibly unintentionally) to increase reimbursement. Identifying upcoding is challenging due to unobservable confounders. We exploit state-level variations in adverse event regulation and instrumental variables to estimate that over 10,000 infections (nearly 15%) are upcoded each year, resulting in an added cost of $200 million. Our findings suggest that increasing financial penalties alone may not reduce HAI incidence. We make several policy recommendations accordingly. MC35 205A-MCC” Empirical Research in Services Sponsored: Manufacturing & Service Oper Mgmt, Service Operations Sponsored Session Chair: Qiuping Yu, Indiana University, Kelley School of Business, Bloomington, IN, 47405, United States, qiupyu@indiana.edu Co-Chair: Gad Allon, Professor, Kellogg School of Management, 2001 Sheridan Rd, Evanston, IL, 60208, United States, g-allon@kellogg.northwestern.edu 1 - Understanding Customers Retrial In Call Centers: Preference Of Service Quality And Service Speed Kejia Hu, Kellogg School of Management, Northwestern University, k-hu@kellogg.northwestern.edu Gad Allon, Achal Bassamboo In this paper we want to understand retrial by connecting customers decisions with their preferences on service aspects: the speed in service access and the quality in service delivered. We use a dynamic random-coefficient structural model along with observations in a call center to capture customers behavior. Our results suggest different preferences for service speed and service quality exist across different customer segments. Using counterfactual analysis, we suggest two cost-effective strategies to reduce retrial. Our new service system design increases private customer welfare by 8.91% and business customer welfare by 37.6%. 2 - The Reference Effect Of Delay Announcements: A Field Experiment Qiuping Yu, Assistant Professor, Kelley School of Business, Indiana University, Bloomington, IN, 47405, United States, qiupyu@indiana.edu, Gad Allon, Achal Bassamboo We empirically explore whether delay announcements induce the reference effect and customers’ loss aversion in a field experiment approach. We find that customers exhibit loss aversion, whether they are provided with announcements or not. Moreover, providing delay announcements appears to impact customers’ reference points and may reduce customers’ per unit waiting cost compared to the case when announcements are not provided. 3 - When You Work With A Super Man, Will You Also Fly? An Empirical Study Of The Impact Of The Coworkers On Workers’ Performance Tom Tan, Cox School of Business, Southern Methodist University, ttan@smu.edu, Serguei Netessine We examine a large operational data set in a casual restaurant setting to study how coworkers’ sales ability (measured as servers’ sales premium) affects workers’ performance in terms of service speed and service quality. We find that servers react non-linearly to their coworkers’ ability. Our empirical findings imply that managers should mix servers having heterogeneous ability levels during the same shift. Through a counterfactual analysis, we find that considering the inverted-U-shaped peer effects to optimize current servers’ schedules without changing their capacity may increase total sales by 2.7%.

4 - Responsiveness And Learning In The Medical Device Product Recall Process George Ball, Kelley School of Business, Indiana University, gpball@indiana.edu, Rachna Shah Product recalls move through multiple steps in a firm. Deciphering the impact of moving fast or slow can help unravel the complexities associated with product recalls and reveal the impact of learning on future recalls. We investigate recall responsiveness with unique data obtained from the Food and Drug Administration consisting of over 4,000 medical device recalls from 2003 to 2013. We find that moving too quickly to identify root cause and corrective action may hamper learning, and lead to additional future recalls. We also find that the speed at which the firm opens a recall is not associated with future recalls, supporting the view that firms should move quickly to recall risky products. Incentive Design in Marketing and Operations Sponsored: Manufacturing & Service Oper Mgmt, Supply Chain Sponsored Session Chair: Tinglong Dai, Johns Hopkins University, Baltimore, MD, United States, dai@jhu.edu Co-Chair: Kinshuk Jerath, Columbia University, 3022 Broadway, New York, NY, 10027, United States, jerath@columbia.edu This paper examines the management problem of “selling” platforms, i.e., designing appropriate salesforce management and incentives schemes to obtain participation by paying customers. The paper shows that network effects increase not only the mean, but also the variance of the performance metrics used to compensate sales agents. 2 - Salesforce Contracting Under Yield Uncertainty Tinglong Dai, Assistant Professor, Johns Hopkins University, 100 International Drive, Baltimore, MD, 21202, United States, dai@jhu.edu, Kinshuk Jerath We consider a scenario in which a firm hires a salesperson to market a product with uncertainty in both demand and supply. We build a principal-agent model of the above situation and study the optimal structure and timing of the contracts, and obtain a number of interesting results. We find that bonus contracts are optimal in both cases, and the bonus may be higher if the yield is lower. Our paper also provides interesting insights into optimal timing of salesforce contracting. For example, we find that when it is difficult to infer marketing effort from observing the sales outcome, it may be in the best interests of the firm to contract with the salesperson before the inventory information becomes available. 3 - Long-term Versus Short-term Contracting With Effort Shifting Fei Long, Columbia University, New York, NY, United States, FLong18@gsb.columbia.edu, Kinshuk Jerath, Fangruo Chen We investigate multi-period contracts to understand agents’ gaming in terms of effort shifting, and what is a firm’s best response. We model a risk-neutral principal employs a risk-neutral agent with limited liability to exert unobservable effort to increase demand over two periods. We find that the principal may find it optimal to use either a short-term plan paying at each period, or a long-term plan paying at the end that concentrates rewards at a single output level. Under limited liability, the latter one provides larger incentives compared to the former, but it suffers from agents’ effort shifting. We extend to a case when the inventory is limited and find it makes the short-term plan more preferred. 4 - Who Compensates The Sales Agent? Duo Shi, PhD Student, Washington University in St. Louis, KH 401, Olin Business School, 1 Brookings Drive, Saint Louis, MO, 63130, United States, dshi@wustl.edu, Panos Kouvelis We analytically study a value chain consisting of three segments: a manufacturer, a retailer, and a sales agent. Five distinct value-chain structures are considered: an integrated value chain, an integrated distribution channel (the manufacturer and the retailer) compensating the sales agent, non-integrated channels with the manufacturer compensating the sales agent, with the retailer compensating the sales agent, and with joint compensation. We compare the strategic implications across all these value-chain structures. 1 - Salesforce Compensation With Network Effects Hemant Bhargava, University of California, Davis, hemantb@ucdavis.edu, Olivier J Rubel MC36 205B-MCC

194