Telehealth After COVID: What Stuck and What Didn't

In March 2020, hospitals were telling patients to stay home. The recommendation made sense from an infection-control standpoint, but it created an immediate problem: people with chronic conditions still needed medication management. People with diabetes still needed check-ins. People in therapy still needed their sessions. The infrastructure to handle those encounters remotely was not ready. The policy barriers were still in place. And then, within a few weeks, the Centers for Medicare and Medicaid Services issued emergency waivers that temporarily lifted restrictions on telehealth reimbursement, expanded eligible provider types, and allowed care to be delivered to patients at home rather than only in designated originating sites. Clinicians who had never thought much about video platforms started using whatever they had. Patients who would never have tried a video call had no other option.

That two-year experiment ended, at least in its emergency form. And the question I find more interesting than "did telehealth work?" is: which pieces of what we learned are still in use, and why?

The short version is that telehealth adoption did not collapse when the acute phase of the pandemic ended, but it did not stay at its peak either. It stratified. Some types of care settled into a permanent hybrid pattern where video is now a normal option. Others returned almost entirely to in-person. The split is not random. It maps fairly cleanly onto what video can and cannot do, and onto which patient populations found the access shift genuinely useful rather than just necessary.

Mental health is the clearest example of what stuck. Video-based therapy and psychiatry maintained high utilization rates long after pandemic restrictions ended, which makes sense once you think about what the physical waiting room was doing before COVID. It was a barrier. Traveling to an office, sitting in a waiting room where you might run into someone you know, and then trying to enter a therapeutic headspace during a rushed appointment slot, all of that friction was not neutral. When it went away, a lot of people found it easier to engage consistently. The therapist's face on a screen, the patient in their own home, turned out to work well enough for most talk-based sessions. The missing piece, the in-person reading of body language and presence, matters more for some modalities than others. For many patients doing cognitive behavioral work, it mattered less than the friction had.

Chronic disease management followed a similar pattern. Medication check-ins, diabetes reviews, blood pressure follow-ups, these encounters were always largely conversational. The physician asks questions, the patient reports symptoms or readings, and adjustments are made. Video changed the medium without changing the substance of most of those exchanges. What it did change was access. A patient managing a chronic condition who lives an hour from their specialist can now have a monthly video check-in and an in-person visit every six months, rather than a monthly drive. That trade made sense before COVID. The policy environment just did not allow it.

Primary care for low-complexity concerns, colds, rash reviews, medication refills, also found a durable if smaller telehealth footprint. Not everything, but enough that most large health systems appear to be maintaining virtual care infrastructure as a permanent capability rather than dismantling it. Gartner has noted that health system investment in virtual care platforms has shifted from emergency response spending to planned operational infrastructure, though I would hedge any specific figures they cite in that direction. The pattern I find credible, even without exact numbers, is that the question changed: it is no longer "should we have telehealth?" and is instead "which visit types should default to telehealth and which should default to in-person?"

The sharper question is what did not stick, and why.

Physical examination is the obvious limit. Musculoskeletal complaints, new skin lesions, abdominal pain, and anything requiring palpation, auscultation, or direct assessment still require the patient to be physically present. The clinical judgment that happens in those encounters depends on information that a video call cannot transmit. This is not a technology gap that remote patient monitoring will fully close. Some of it maybe, but a physician pressing on an abdomen or feeling a lymph node is still doing something that a camera cannot replicate. The early optimism in 2020 that we could build a fully remote primary care practice always ran into this wall, and that optimism has mostly receded.

The equity story is the part I think about most, because it is uncomfortable. Telehealth was supposed to expand access. In some populations it clearly did. Mental health patients who would never have sought in-person care because of stigma, time, or geography found a format that worked. But the rural broadband gap did not close between 2020 and now. Older patients in low-income households who lacked devices or who struggled with the technical requirements of a video visit were pushed toward a format that was actually harder for them to access than the in-person alternative. The waiver expansion was designed for a hospital emergency, not for a health equity strategy, and the access assumptions built into it were assumptions about populations who already had broadband and smartphones. The patients who most needed expanded access were often the ones least able to use the expanded modality.

The CMS waiver story is still unresolved in an important way. The emergency waivers have been extended repeatedly through congressional action, but as of my writing the permanent policy status of telehealth reimbursement is still not settled. Congress has kept pushing the expiration date forward, and the debate over which waiver provisions should be made permanent is real and ongoing. This matters for IS researchers and health IT practitioners equally: the infrastructure investments health systems are making in virtual care platforms depend on whether the reimbursement model that makes them viable will persist. Building for telehealth permanence under temporary policy is its own kind of institutional risk, and it is a risk that the CMS waivers did not eliminate. They just deferred it.

My honest take is that the COVID experiment answered one question clearly. Telehealth is viable for a substantial portion of what primary and specialty care does. It is not a replacement for in-person medicine, and it was never going to be. The question was always whether the regulatory and reimbursement environment would evolve to match what was clinically feasible. The pandemic forced that evolution on an emergency timeline. What remains to be seen is how much of the forced experiment becomes permanent design.