A Scholar Studies How Digital Health Can Deliver on Its Promise

Technology promises to improve health care delivery, enhance disease management, and promote satisfaction for patients and providers alike. Yet the reality often falls short. Mitchell Tang, an assistant professor of Health Policy and Management, studies how digital health technologies function in the real world, uncovering subtle challenges they can introduce and identifying potential solutions.

His latest study, published in JAMA Network Open, analyzed 3.6 million patient portal messages from patients to their health care teams to find clues about why people from historically marginalized groups are less likely to receive a response from their doctors. Tang and his coauthors found that writing style accounted for a substantial share of these disparities.

“Whether you start with ‘Dear Dr. Jones affects whether Dr. Jones ends up responding,” he explains. “The way we have designed portal message interfaces and triage processes places a great deal of responsibility on patients to express their needs in writing,” he continues. “Better design could make portals more equitable and more effective at pulling out what the patient actually needs.”

Tang, who spent four years as a management consultant before earning his PhD at Harvard, has also investigated how technology adoption can increase costs for our health system or disrupt clinical workflows. Remote monitoring of high blood pressure and other physiological data has been marketed as a way to improve patient care. Tang’s prior work has shown support for these claims, but has also raised concerns that indiscriminate adoption could substantially increase Medicare spending, underscoring the need for reimbursement policies that encourage high-value use. His analysis of clinician schedules that require “modality switches” between telehealth and in-person appointments in a single day revealed negative impacts on patient care quality.

“My work homes in on the new challenges and questions that these technologies raise for health care systems and policymakers and how they need to adapt in response,” says Tang.

If health care systems can inadvertently amplify disparities through their patient-provider messaging portals, what’s the fix?

Tang: Portals rely on patients’ technology literacy and skill in communicating through the written word. Due to substantial increases in demand, a lot of health care systems have instituted triage systems where a nurse reads and either responds to the message or routes it to a physician or another member of the care team. That can be a sensible way to allocate scarce resources, but it also creates opportunities for things like writing style to bias how messages are interpreted and prioritized.

I think we need ways to reduce reliance on a patient looking at a blank screen and writing prose. What might be more equitable—and also more effective—at eliciting what the patient needs? For example, AI chatbots can ask follow-up questions and prompt more detail. Or maybe triage tools can aid the nurse by removing some of the stylistic noise. Identifying the disparity and finding its mechanisms has been a really interesting opportunity for imagining solutions

What spurred your analysis of clinicians’ experiences switching between in-person visits and telehealth?

Tang: Many of my research projects are inspired by lived experience, either my own or of those around me. I was chatting with a practicing physician—my wife’s brother-in-law. He told me he was bothered by these intermixed days bouncing between telehealth and in-person visits. That conversation got me thinking: How should practices integrate telehealth into their schedules, and what implications do those choices have for patient and provider experience, access improvements, physician capacity? I decided to take a closer look.

What did you learn about providers’ experience of mixing online and in-person care delivery?

Tang: We found that switches from in-person to telemedicine visits create meaningful operational friction, leading to delayed visit starts, less comprehensive visits, and more after-hours work. One potential solution is to schedule dedicated telemedicine blocks rather than asking clinicians to move between the two types of care, but that approach has a clear downside: What happens when there are not enough patients seeking telemedicine during those periods? Appointment slots may go unused, reducing clinical capacity and potentially making it harder for patients to receive timely care. There is an inherent tension here, but once practices are aware of it, they can make more informed decisions as to how to navigate it.

How do you hope to inform health care organizations as they weigh the challenges and trade-offs you’ve identified?

Tang: As an academic, I want to see this work published and broadly disseminated. I also want to make sure it’s in the hands of my local health system and perhaps help translate these ideas into practice. More broadly, I want to highlight important aspects of these new technologies that administrators and policymakers perhaps aren’t thinking as much about. For example, in future work, I plan to study how telemedicine enables clinicians to work from home and what that flexibility could mean for providers, patients, and health care organizations.

You are in the department of Health Policy and Management. Do you see your work more focused on influencing health policy or management?

Tang: Ideally, both! A lot of my work looks at how Medicare should be paying for emerging health technologies in ways that incentivize high-value use and avoid wasteful spending. I’m also interested in what is happening in health care organizations and providing recommendations to people who run them. Ultimately, understanding both improves work on either side.