On 1 October 2025, the US federal government shut down, and with it lapsed the temporary rules that had let most Medicare patients see a clinician by video from their own homes since the pandemic. For six weeks, practices faced a choice: keep doing telehealth visits and risk not being paid, or cancel them. CMS told its contractors to hold claims for affected services. One large category was carved out of that hold: behavioral and mental health care, which Congress had already given a more permanent footing.
When the shutdown ended on 12 November, the flexibilities were restored, retroactively, only until 30 January 2026.
That episode tells you most of what you need to know about telemedicine after the boom. The general-purpose video visit is living on short-term extensions. The part that has become permanent, in law and in patient behaviour, is mental health.
The boom settled into a niche, and the niche is large
It is easy to say telemedicine “faded” after 2020. It is more accurate to say it sorted itself.
FAIR Health, which tracks private insurance claims across the United States, reported that telehealth accounted for about 5.1% of medical claim lines in January and February 2025. That is far below the pandemic peak and far above the pre-pandemic baseline. The more revealing number is what those visits are for. In the first quarter of 2025, mental health conditions were the leading telehealth diagnosis category nationally and in every region, accounting for between 58% and 62% of patients with a telehealth claim month to month. Psychotherapy was the second most common category of telehealth service overall and ranked first in the Midwest and Northeast.
So the question “who kept using telemedicine?” has a fairly clear answer. People getting talk therapy and psychiatric follow-up kept using it. So did clinicians who provide those services.
Why mental health stuck and much else didn’t
This is less about technology than about the shape of the clinical encounter.
A psychotherapy session is a conversation. It rarely needs a stethoscope, a blood draw or a hand on an abdomen. Moving it to video changes very little clinically, while removing travel, childcare arrangements, time off work and, for some patients, the discomfort of being seen walking into a mental health clinic.
Primary care and specialty follow-up are different. Many visits end with a test, an examination or a procedure, which means a video visit often becomes a prelude to an in-person one rather than a replacement. For clinicians, that can feel like doing the visit twice. For patients, it can mean two co-payments and two sets of arrangements for what used to be one trip.
There is also a supply-side story. Many therapists and psychiatrists have built practices that are partly or entirely remote. They can see patients across a whole state, fit sessions into gaps and run a practice without renting as much space. Once a workforce is organised around a modality, it becomes sticky in a way that a hospital outpatient department simply offering a video option is not.
The uncomfortable trade-off
Here is the tension policymakers keep tiptoeing around. The strongest case for telehealth has always been access for people who struggle to reach care: rural patients, people with disabilities, older adults without transport. The strongest pattern of actual use looks rather different, weighted heavily toward behavioural health, and video care in particular favours people who already have devices, private space and reliable broadband.
That does not make the use less valuable. Mental health care delivered is mental health care delivered. But it complicates the political argument. Rural access is the story used to justify extensions. Urban and suburban therapy is a large part of what those extensions actually pay for.
Payers see a second tension. Easier access can mean more visits. Some of that is unmet need finally being met. Some of it may be care that would not have happened otherwise and did not need to. Nobody has a clean way to tell the two apart in claims data, which is why every extension debate returns to cost and fraud.
What a sensible settlement might look like
The shutdown showed the cost of treating a durable mode of care as a permanent emergency. A more stable settlement would probably:
- Make telehealth policy clinical rather than calendar-based. Decide which services work well remotely and cover them permanently, rather than renewing everything for a few months at a time.
- Keep audio-only options where video fails. Some older and lower-income patients rely on the phone, and dropping it hits them first.
- Tie in-person requirements to clinical need. Blanket rules requiring a periodic face-to-face visit can protect quality in some cases and simply add friction in others.
- Measure outcomes, not just volume. Whether remote therapy patients stay in treatment and improve matters more than how many visits were billed.
The mechanism: telehealth sorts by the shape of the visit
It helps to think of a clinical encounter as having three possible components: talking, examining and doing. Talking includes history, counselling, explanation and shared decisions. Examining means using hands, instruments and eyes up close. Doing means drawing blood, giving injections, taking images or performing procedures.
Telehealth handles talking well, examining poorly and doing not at all. The share of a specialty’s visits that are mostly talking is a good predictor of how much of its care stayed virtual. Psychotherapy is almost entirely talking. Medication management in psychiatry is mostly talking. Routine follow-up of a stable chronic condition can be mostly talking, if the tests were done beforehand. A new abdominal pain, a suspicious skin lesion or a child with a fever usually needs examining.
This framing explains something that frustrated early advocates. They expected telehealth to reshape primary care. But primary care is a mix of all three components, and the proportion is unpredictable before the visit starts. A patient booked for a medication review mentions chest pain. A video call about a rash reveals something that needs a closer look. For a practice, the uncertainty itself is costly, because an unplanned conversion to an in-person visit disrupts the schedule. A common response is to reserve telehealth for specific, predictable visit types rather than offer it across the board.
A worked scenario: two patients, one policy
Consider two Medicare patients during the six-week lapse.
The first is a woman in her seventies living in a small town, two hours’ drive from her psychiatrist. She has been having monthly video appointments for depression and medication management for three years. Because behavioural and mental health services had a more durable legal footing, her care was less exposed to the lapse, although the rules on in-person requirements remained a live question for her clinicians.
The second is a man of similar age in the same town with heart failure. His cardiology practice had been doing some follow-ups by video to review symptoms, weights and medication changes between in-person visits. During the lapse, the practice had to decide whether to continue those visits, with uncertain payment, or convert them to in-person appointments that would require a long drive or a family member taking time off work. Some practices kept going and absorbed the risk. Others cancelled.
The policy treated these two patients differently not because of any assessment of clinical need, but because of the legislative history of mental health telehealth. That is the core problem with calendar-based policy. It creates distinctions that no clinician would design.
The strongest counterargument
Sceptics of permanent telehealth expansion have serious points. Early in the pandemic, rapid expansion was accompanied by some fraudulent billing and aggressive marketing, particularly by companies with no ongoing relationship with the patient. There are also legitimate clinical concerns: remote prescribing of controlled substances, missed physical findings and fragmentation when patients use a telehealth company that does not share records with their regular doctor. And budget scorers have tended to treat permanent telehealth coverage as adding cost, because more convenient care is used more.
This argument partly holds. Short extensions give Congress repeated opportunities to tighten rules and respond to evidence. Some guardrails, such as requiring an established relationship for certain services or in-person contact before prescribing some medicines, may be reasonable.
But short extensions have their own costs. Practices cannot plan. Health systems hesitate to invest in virtual care infrastructure that might become unpayable in a few months. Patients cannot be sure their care will continue. And the lapse showed that the most vulnerable patients, older, rural and less mobile, are the ones most exposed when the rules change suddenly. Guardrails and permanence are not opposites. A permanent framework can include guardrails and still give people something to rely on.
Second-order effects to watch
Workforce geography. If mental health telehealth remains stable while other telehealth stays uncertain, more clinicians may build fully remote practices in behavioural health. That could improve access in underserved areas, or concentrate therapists serving patients who can pay privately.
Hybrid models. The most durable arrangements may combine a local in-person clinician with remote specialists, for example a primary care nurse examining the patient while a distant specialist joins by video. These models need payment rules that recognise both participants.
Audio-only care. For patients without video capability, the phone is often the only realistic option. Whether audio-only services remain covered will shape who benefits from virtual care.
Cross-border licensing. In large federal systems, clinicians often need a licence in the state where the patient is located. Licensing compacts and reciprocity will affect how far remote care can reach.
What this means for different players
Legislators and ministries should move from time-limited extensions to service-based rules: which services, for which patients, with what safeguards, on a permanent basis, reviewed periodically on evidence.
Health system leaders should design virtual care around visit types that genuinely suit it, rather than offering telehealth as a generic option. Clear triage reduces the frustrating conversions to in-person visits that make clinicians sceptical.
Clinicians can help by documenting when telehealth avoided a missed appointment, a hospitalisation or a lost patient. Those cases are the evidence policymakers need and rarely see.
Payers should measure continuity and outcomes, not just volume and cost. A therapy patient who stays in treatment because sessions are remote is a different result from a one-off consultation with an unknown provider.
What to watch next
The immediate test is 30 January 2026, when the current extension runs out. Watch whether Congress chooses another short extension, a multi-year one or a permanent framework. Watch too whether any new rules distinguish between services by clinical suitability, or simply carry the pandemic-era list forward again. And watch the claims data: if telehealth volumes outside mental health fall further during periods of uncertainty, that tells us how much of the remaining use is fragile rather than settled.
The takeaway
Telemedicine did not fail after the pandemic, and it did not take over. It found the part of medicine it fits best, which turned out to be mental health, and settled there. The policy framework is still behaving as though the whole thing is a temporary experiment, renewing it in short increments and occasionally letting it lapse.
Patients who have built their care around a weekly video session deserve a more reliable answer than “until the end of January.”
In your practice or system, which services have stayed virtual, and which quietly went back to in-person? We’d like to hear your experience.
