In 2019, I met with dozens of colleagues to discuss incorporating telemedicine into our practice. Almost all engaged politely with me. But afterward, very few acted. The subtext was clear: "We’re not interested." Their sentiment was quite typical. At the time, telemedicine accounted for a tiny fraction of U.S. clinical encounters.

In March 2020, this changed suddenly. In response to the pandemic, and enabled by relaxed regulations and expanded reimbursement, our practice — like countless others — quickly pivoted to telemedicine to preserve care access, protect our patients, sustain our workforce and maintain revenue. Soon, we were providing more telemedicine visits in a single day than we had the entire year before.

By July 2020, telemedicine accounted for roughly half of our Department of Medicine's outpatient visits. Nationally, telemedicine reached nearly one-in-six visits, leading healthcare pundits to declare it the "new normal" and proclaim that "the telemedicine genie is out of the bottle." Our Jetsons-like future was quickly taking shape.

Fast forward to today. With stay-at-home orders having long expired and Americans back to their usual routines, telemedicine visits have dropped to just 5% of all visits nationally, primarily concentrated in mental health. Though this far exceeds 2019 levels, it falls way short of predictions and raises two key questions. Why is the telemedicine genie mostly back in the bottle? And will it remain there?

Telemedicine does not always fit our physical bodies.

Experienced physicians can gauge their patients' health status by looking at them. And they can often diagnose them based on medical history alone. Consequently, video and in-person visits typically lead to the same diagnoses.

Yet physicians must examine or test certain patients in person. Early in the pandemic, physicians stretched the bounds of who they were willing to manage remotely, given the need for physical distancing. Now they are defining which conditions are “tele-amenable” more stringently.

This helps explain why telemedicine use varies widely across specialties. For example, a core reason dermatologists provide almost all in-person visits is that comprehensive skin exams require good lighting and sometimes the ability to magnify, touch and biopsy skin lesions. Similarly, lung doctors often opt for in-person visits to obtain concurrent pulmonary function tests. Conversely, mental health services account for more than half of all telehealth visits, mainly because physical exams and diagnostic tests are rarely required.

Many physicians do not like practicing online.

Physicians may find many aspects of telemedicine appealing. They may use it to extend their reach, expand care access, improve patient experiences, see their patients' home environments, work remotely and better tailor their schedules. Select physicians even choose to practice telemedicine full time.

Still, physicians generally prefer in-person care. For one, simple in-person tasks can be much harder for them over video visits, such as informing patients the clinic is running late, checking blood pressure, or arranging tests they would have performed in the clinic. Additionally, physicians must often provide tech support when patients cannot connect to video or enable their microphones. Combined, telemedicine can slow physicians down and saddle them with more after-hours work.

Patients like telemedicine less than many expected they would.

I recently saw a 28-year-old social media software engineer in the clinic. At the end of the visit, I suggested he follow up in three months over video. He said he'd instead return to see me in person. I was shocked.

Many patients prefer telemedicine because it reduces the time and effort required to receive care. Yet, most — even those in Gen Z — favor in-person care, perhaps because it is more familiar. Or maybe, in our increasingly atomized, screen-based world, people would rather see their caregivers in real life.

In-person care is the default orientation.

Healthcare systems and individual practices are optimized to provide in-person care. They must reconfigure their teams, spaces and workflows for telemedicine.

Yet healthcare is a complex, adaptive system that "pulls strongly towards inertia." Today, physical distancing is a thing of the past. Telemedicine generates less revenue than in-person care, and its long-term reimbursement is uncertain. Therefore, healthcare systems and independent practices lack strong enough incentives to motivate change.

The future does not progress in a straight line.

Early in the pandemic, all types of services appeared to be permanently shifting online. But it is difficult to predict how new technologies alter our behaviors. Today, online shopping is losing ground to brick-and-mortar stores, remote work is relocating back to the office, and home exercise is moving back to gyms. Healthcare’s shift back to in-person care is no different.

Still, Amara's law — that "we tend to overestimate the impact of a new technology in the short run, but we underestimate it in the long run" —  suggests telemedicine will ultimately become a core healthcare feature. But first, it needs to be redesigned.

Many practices implemented telemedicine on the fly, simply retrofitting it to their typical in-person care models. They must now reconsider and revamp telemedicine to fit better into their overall services and appeal more to patients and physicians.

For one, practices must expand their view of telemedicine beyond solely video visits — which do not fundamentally increase productivity — to include other asynchronous (e.g., self-service and messaging), synchronous (e.g., chat and phone) and remote monitoring modalities. They may learn from virtual-first upstarts, which, lacking legacy baggage, intentionally design care from the (virtual) ground up to meet patients' needs more effectively and efficiently.

Additionally, they should reframe telemedicine as a complement to — rather than a substitute for — in-person care. In other words, it is not either telemedicine or in-person care. It is both telemedicine and in-person care, depending on the individual's needs and preferences at the given time. They must build "operating systems" to help orchestrate this.

The pandemic unleashed the telemedicine genie from the bottle. But we have already lost many of these gains. Eventually, with the right vision, design and incentives, we can redesign telemedicine to fully emerge and play a pivotal role in improving healthcare.

About the Author: Spencer Dorn, MD, MPH, MHA is a gastroenterologist, professor and vice chair of medicine for care innovation at the University of North Carolina.