Operational and behavioral interventions to jointly manage telemedicine and in-person channels in epidemic contexts

This thesis investigates a multichannel healthcare system and studies operational and behavioral interventions to manage provider decisions and patient demand for telemedicine and in-person channels in the presence of an infection risk. The first part of the thesis formulates an optimization model that determines optimal capacity allocation and pricing decisions for the channels to maximize the provider's revenue. The main novelty in our model is that the in-person channel has a contagion disutility driven by patients' fear of contracting the disease in the healthcare center. The contagion disutility has two components: a behavioral parameter capturing the contagion awareness of patients and a contagion function. The contagion disutility is increasing in the number of in-person appointments, which makes the patient's choice endogenous. The optimization model offers several interesting results. We prove that if there is a contagion disutility and the telemedicine channel is not offered, the revenue grows logarithmically in potential demand (patient pool). However, the revenue grows linearly in potential demand if the telemedicine channel is offered. Therefore, the telemedicine channel is necessary to mitigate the provider's revenue loss due to contagion disutility. Additionally, when only the in-person channel is offered, patient access (total number of appointments) decreases significantly. We prove that the optimal telemedicine appointment price is strictly less than that of the in-person appointment, and the price gap increases as the patients become more contagion-aware. When we include the physician's sickness due to contracting the disease from patients in our model, we see that a social planner maximizing the welfare allocates more capacity to telemedicine than a revenue-maximizing provider, making physicians better off. The first part of the thesis shows that capacity allocation and pricing decisions that incorporate patients' contagion disutility can mitigate revenue losses experienced by healthcare providers during epidemics. This result continues to hold when some telemedicine appointments may require follow-up in-person appointments if the likelihood of a follow-up appointment is under a certain value. We study a setting where we include the direct cost of sick patients (disease burden) in the provider's/social planner's objective. When the disease burden is sufficiently large, high contagion awareness among patients can increase revenue/welfare. This indicates the importance of modulating patients' contagion awareness levels. Considering the benefits of telemedicine in a context of infection risk, in the second part of the thesis, we test the effectiveness of several operational and behavioral interventions in increasing the use of telemedicine during epidemics. We conduct a series of laboratory/platform experiments, in which each study considers an operational dimension alongside a behavioral one. The operational dimensions are motivated by the optimal capacity allocation and pricing decisions we studied in the first part of the thesis and include appointment prices and wait times. The behavioral dimensions include infection fear and health condition deterioration over time. The novelty of our experiments is that they include both dimensions in an epidemic context and let participants actively choose between the telemedicine and in-person appointment options. Our study is the first to propose and test informational interventions to manage patient demand between channels in an epidemic context via controlled experiments. Our experimental setting also allows us to control for behavioral characteristics such as risk perception and inherent attitudes toward the channels that can impact channel choice. The first operational dimension that we study is a price difference intervention where the price of an in-person appointment is 25% higher than that of a telemedicine appointment. Although we observe a directional increase in the use of telemedicine, the impact is not significant. Insurance considerations, a need for increased saliency, or a larger price difference can be potential explanations. The second operational dimension is the appointment wait time advantage of the telemedicine channel. We approach appointment wait time from two different perspectives: the systems perspective and the patient's perspective. In the systems perspective, due to the provider's capacity/resource allocation policy, telemedicine appointments have shorter appointment wait times. In the patient's perspective, the provider offers similar appointment wait times for the channels. Yet, telemedicine appointments can easily fit into the schedule of patients due to their convenience in eliminating the need to travel to a healthcare facility. That can allow patients to see the doctor sooner. In addition to the operational interventions, we test the effectiveness of several behavioral (informational) interventions. In line with the benefit of increasing patients' contagion awareness in the first part of the thesis, the first set of informational interventions aims to increase the infection risk awareness of patients and nudge them to choose the telemedicine channel. The infection messages differ in their framing (negative or positive) and risk focus (only self-risk or including the risk for others). For the messages regarding the appointment wait time difference, we emphasize that telemedicine appointments allow patients to see the doctor sooner. When there is an appointment wait time difference due to the provider's capacity/resource allocation policy, informational interventions do not further nudge patients toward the telemedicine channel. When the provider offers similar appointment wait times, the informational interventions can be effective in increasing the use of telemedicine. We observe that a negatively framed infection message (emphasizing the infection risk associated with the in-person channel for self and others) and a positively framed convenience message (emphasizing the ease of scheduling for the telemedicine channel) effectively increase telemedicine use. The convenience message has the advantage of being further used in non-epidemic contexts, increasing its generalizability. Overall, the results of the second part of the thesis show that healthcare providers can manage patient demand with easy-to-apply informational messages. Together, the findings of this thesis offer actionable insights for healthcare providers to manage channel decisions and patient demand in multichannel healthcare systems.