Knowledge that Transforms

Racism at the individual and societal levels has been identified as a cause of disparity in healthcare outcomes in the United States but reducing disparity has been slow. This study disentangles the relative effects of two types of racism on inpatient service delivery and hospital mortality: technology bias and implicit bias. Drawing on clinical data from intensive care unit (ICU) patients with sepsis, we use propensity score matching to balance groups of White and nonwhite patients and run a causal mediation analysis to test our model, which links patient race to hospital mortality through two mediating variables related to service delivery: (1) discrepancies in blood oxygen saturation measurements due to technology bias embedded in the medical device (i.e., pulse oximeter) and (2) administration of supplemental oxygen, which could be impacted by clinicians’ implicit bias. We first replicate prior findings that (a) higher discrepancies between oximeter readings and laboratory tests for nonwhite patients compared to Whites and (b) a higher discrepancy lowers the likelihood of patients receiving supplemental oxygen during the ICU stay. We make a unique contribution by finding that nonwhite patients with sepsis have a 79% higher risk of hospital mortality in the ICU compared to Whites, with nearly half of the racial disparity in mortality stemming from technology bias and less than a fifth from clinicians’ implicit bias. Eliminating these two biases would help save thousands of lives annually among racial/ethnic minorities with sepsis in the United States. Our findings indicate that technology bias exerts a greater negative impact on mortality than does implicit bias, highlighting the importance of device approval standards and clinicians' ability to customize decision criteria for supplemental oxygen.

Keeping up with competitors’ prices is one of the top operational challenges in pricing. However, competitive interactions in revenue management have not received much research attention in the past due to their complexity. We conduct a series of laboratory experiments to investigate the dynamic pricing behavior of two capacity-constrained firms under competition. Our experiments initially control for strategic interactions between the sellers and then allow for them. To gain a broader understanding, we also manipulate demand uncertainty and the expected market size. Our results confirm the dependency of dynamic pricing decisions on the competitor's behavior. We find that the theory is much more forgiving—in the sense that it predicts a lower level of competition among the sellers—than what we actually observe in the laboratory. The modeling literature indicates that the seller with the lower capacity has a competitive advantage, but our results reveal the opposite. Further, there is potential for high-capacity sellers to benefit from competition. Sellers tend to underprice (resp., overprice) their units at the beginning (resp., end) of a selling season. Also, competition lasts longer than the theory predicts, and customers benefit from the biases of the competing sellers. The higher-capacity seller following the best-response policy is not harmed due to the biases of the competitor. However, the lower-capacity seller's performance is greatly influenced by the competitor's degree of rationality.

Several low- and middle-income countries’ emergency transportation systems (ETSs) do not have a centralized emergency number. Instead, they have many independent ambulance providers, each with a small number of ambulances. As a result, ETSs in these contexts lack coordination and ambulances. Using a free-entry equilibrium model, we show that in such decentralized systems, the probability that any given call can be served by at least one ambulance, that is, its coverage, is at most 71.54%, regardless of the ETS’s profitability. We examine three business models that can address the ETS’s lack of coordination and ambulances: (i) a competitor-only business model, where an entrepreneur enters the ETS and acquires ambulances to compete with existing providers; (ii) a platform business model, where an entrepreneur coordinates existing providers; and (iii) an innovative platform-plus business model, where an entrepreneur combines (i) and (ii): setting-up a platform and acquiring platform-owned ambulances. We also examine a government-run platform that takes no commissions from providers. Using a game-theoretic approach, we find that it is optimal for all platform models to incentivize all providers to join. However, only the government-run platform may incentivize providers to acquire additional ambulances. Furthermore, a government-run platform offers higher coverage than a platform-plus only when the platform’s power to coordinate ambulance providers is moderate. Our results can help entrepreneurs and policymakers in LMICs navigate various tradeoffs in improving their countries’ ETS.

While government regulation or company policy can be used to curtail discrimination at work, it is hard to regulate away negative experiences like microaggression or perceptions of discrimination against minority employees in hiring and promotion. Using data from interviews with minority ethnic staff at a UK university, I present evidence of microaggression and minority employees feeling excluded and posit that perceptions of discriminatory policy engender negative perceptions of the organization. I further show the link between employee engagement and organizational performance and propose that minority employees’ negative experiences and perceptions lower their job and organizational engagement and, eventually, impact organizational performance. I offer a solution in the form of an enterprise-wide continuous improvement program that would directly improve organizational performance by improving business processes and indirectly by improving minority employees’ experience, perceptions, and engagement.

Ensuring universal health coverage is one of the top priorities under the United Nations’ Sustainable Development Goals (SDGs). The World Health Organization recognizes the critical role of the health workforce in promoting this mission. However, developing countries throughout the world often experience shortages in the health workforce, especially during pandemic outbreaks such as coronavirus disease 2019. Governments and humanitarian organizations worldwide have supported health workforce development through development aid, but many questions remain about its effectiveness. Therefore, this study investigates whether development aid is effective toward health workforce growth, in general, and when and where the aid is effective within various segments of the health workforce, in particular. Furthermore, the study explores the role of government behavior in recipient countries and the introduction of SDG-3 on aid effectiveness for the health workforce. Analyzing data pertaining to 134 developing countries using fixed-, moderating-, threshold-effects regression models and event study methods, we derive the following results: First, all forms of development aid contribute positively to the health workforce from a “universal” perspective. Second, we establish that aid can be an effective catalyst for “physicians and clinical officials” and the “total health workforce” only when the current workforce density of the corresponding workforce segment exceeds a certain threshold. Thus, countries with severe shortages in the total health workforce as well as physicians and clinical officials, have difficulty reaping the dividends of health personnel development aid. Fortunately, we find a breakthrough from the perspective of “nurses and midwives” wherein the development aid is shown to be always effective for this leading segment of the health workforce. Finally, it is shown that increased government health expenditure, government effectiveness of recipient countries, and the introduction of SDG-3.c significantly enhance aid effectiveness. Our study directly benefits funding managers with development aid allocation and implicitly enhances donor recognition and confidence in health personnel development aid, thereby supporting the universal health coverage goals of the United Nations and the World Health Organization.

As the scale of renewable energy on the demand side continues to grow, a new demand response program (DRP), the virtual power plant (VPP), pays a rebate to the consumer and purchases his rooftop solar electricity. This electricity is dispatched to areas of electricity shortages, which is vital in balancing electricity loads and avoiding energy waste. However, the new DRP differs from the traditional DRP in terms of paying rebates. The traditional DRP requires consumers to respond above a specified baseline to receive a rebate. In contrast, in the new DRP, consumers can receive a rebate even if their response is small (i.e., no baseline). There is no research that addresses this divergence. Therefore, we construct a supply chain consisting of a consumer and a VPP for this novel DRP and analyze the strategic interactions between them in two situations where a baseline is not set/set. Our results show room for improvement in the DRP that dispatches electricity generated from rooftop solar. If the consumer generates more electricity from rooftop solar, the baseline, rather than frustrating the consumer, motivates the consumer to be more responsive. While the baseline also discourages the demand response process, it reduces the emissions generated from the consumer’s over-responsiveness. Additionally, our study suggests that a baseline, whether it facilitates or hinders the DRP, could benefit the VPP, depending on the business model of the VPP. In addition, we add several extensions to demonstrate the robustness of our model.

Referral programs have received increasing attention as a way to enhance project fundraising in crowdfunding. To improve operational efficiency and performance, many crowdfunding platforms allow users to exert more effort (referral effort) when making referrals. In this paper, we study the impact of users’ referral efforts on referral outcomes, using a unique data set from a reward-based crowdfunding platform. We analyze three types of referral effort beyond merely sharing the referral link, namely, writing the referral message, matching the referral recipients’ pledges, and supporting their own referrals. Our results suggest that users’ referral efforts can help improve the outcome of referrals, as these efforts may be perceived as positive signals of quality, thus reducing referral recipients’ uncertainty about the crowdfunding project. We further explore the heterogeneity in this observed impact, and find that matching recipients’ pledges have a greater impact than the other two types of referral effort. Furthermore, the impact of referral effort is attenuated under conditions of low uncertainty, wherein project quality can be inferred from observable features, such as projects launched by large creators, selected by editors, or referred by experienced backers. Our analyses also indicate that referring users’ efforts can collectively improve the fundraising performance of projects that receive referrals. We discuss the implications of our findings for platform managers, policy makers, and users in crowdfunding marketplaces.

Choosing the right development team is crucial for companies. We investigate how project collaboration requirements and uncertainty levels affect the choice between “specialists” with higher levels of task-specific abilities and “generalists” with higher levels of collaboration skills. We also examine how these factors affect optimal team incentives. In addition to performance-enhancing helping, we consider another type of collaboration that has received less attention in the incentive literature: information-sharing, which can reduce uncertainty and lead to more compatible design decisions. In the case of helping, we show that if uncertainty is high then specialists might be preferred in order to reduce risk exposure—even if their task-specific abilities are only slightly better. Conversely, if information-sharing can significantly reduce uncertainty then generalists may be favored even if their task-specific abilities are much lower. Our study also reveals that task and collaboration incentives can be either complements or substitutes depending on the type of collaboration and level of project uncertainty: in projects that benefit from helping, firms will always substitute task incentives for collaboration incentives when selecting a team of specialists (rather than a team of generalists), yet this need not be the case with information-sharing. In such projects, it can be optimal to offer higher task incentives and also higher collaboration incentives to a team of specialists (than to a team of generalists) even though specialists’ collaboration skills are relatively lower.

We build on the national culture, complex adaptive systems, and operations complexity literature to develop hypotheses about the moderating role of national culture on the negative relationship between dynamic operations culture and operational complexity. Monthly performance data were collected from 179 plants in a multinational corporation's global plant network that spans all national culture clusters. We use random effects time series modeling to incorporate the transitory nature of this relationship and control unobserved heterogeneity over time. The results indicate that the relationship between dynamic operations complexity and operational performance is moderated by every dimension of national culture except gender egalitarianism and institutional collectivism. We contribute to the literature on complexity by introducing the dynamic operations complexity construct and showing the moderating effect of national culture. We contribute to the national culture literature by comprehensively examining the role of national culture supporting the national culture divergence thesis. Managerial implications relate to locating new plants in regions whose national culture is supportive of key goals and allocating tasks and products accordingly.

Malaria remains a significant public health concern in several developing countries. Tropical weather, coupled with poor environmental and socioeconomic conditions, increases mosquito activity and disease transmission in impoverished nations. The most effective strategy to prevent and control malaria is the use of long-lasting insecticidal nets, or LLINs. Ensuring that LLINs are effectively and equitably distributed to those who need them most is crucial yet challenging, especially since financial and health resources are scarce in endemic countries. Through a real-world case study conducted in the Brazilian Amazon, which accounts for 99% of malaria cases in Brazil, we propose a Malaria Vulnerability Index (MVI). This composite index encompasses epidemiological, socioeconomic, and environmental factors. Using the MVI, we developed a prioritization-based location-allocation model to maximize the benefits of targeting the most vulnerable municipalities for malaria interventions. Our mathematical model relies on discrete coverage levels, which makes the allocation of LLINs more flexible while ensuring community-wide protection. We also explore the theoretical relationship between our proposed model and classical demand covering problems. Finally, we discuss some approaches to delivering more equitable solutions by minimizing the number of underserved areas. Our results include comparisons between various malaria interventions and provide practical insights. In particular, we show that we can drastically reduce the number of underserved areas without compromising the effectiveness of the LLINs allocation through lexicographic optimization. The results also reveal that with investment levels up to 50% of the ideal, we can fully protect the endemic area after 2 years of successive interventions.