Telehealth’s three policy clocks all ran down in December, so virtual care needed an annual cycle

The week

On November 1, 2024, the Centers for Medicare & Medicaid Services released the final calendar year 2025 Medicare physician fee schedule. The fee schedule is the annual rulebook that sets what Medicare pays clinicians for each service and the conditions attached to that payment, so it shapes, with little public notice, how care is delivered across the country.

For millions of Medicare patients, a video visit with a doctor or therapist had become ordinary. Yet whether such a visit could legally happen, and be paid for, was decided by rules that kept expiring. The first week of November 2024 showed how many separate clocks were ticking at once.

On telehealth, CMS made three notable moves. It permanently allowed audio-only communication, meaning an ordinary phone call, for services delivered to patients at home when the practitioner can offer video but the patient cannot use it or does not consent. It extended virtual direct supervision through 2025, and made it permanent for a narrow set of lower-risk services. Direct supervision is the requirement that a senior clinician be immediately available while another staff member delivers care; the virtual version lets that clinician be present by live audio and video rather than physically down the hall. CMS also let practitioners deliver telehealth from home through 2025 without listing a home address on enrollment.

Those changes mattered, but bigger ones lay elsewhere. The broader flexibilities that let beneficiaries receive most telehealth at home, regardless of where they lived, rested on statute, a law passed by Congress, not on CMS’s own rules. Those were due to expire on December 31, 2024, and CMS could not extend them on its own.

Separately, the Drug Enforcement Administration’s temporary rule allowing controlled substances, medicines with potential for misuse that federal law tightly regulates, to be prescribed by telemedicine without a prior in-person visit was also due to expire at the end of 2024. Two weeks later, DEA and HHS extended it through 2025.

So in a single month, a patient’s virtual care depended on a regulator, a legislature and a law-enforcement agency, each acting on its own timetable.

Telehealth had already been through 90-day emergency renewals, the first permanent mental health benefit and the end of the emergency in 2023, when flexibilities began expiring on separate schedules set by different authorities. That last stage was treated as a transition. By late 2024 it was clearly something else.

The emergency was long over. Yet whether a given virtual visit could happen, and be paid, still depended on three authorities acting on three schedules, and all three had settled into a pattern of running down in December. What looked like a temporary bridge had become the permanent road.

That changes the management question for an executive. A transition is something you get through once. A standing condition is something you build a routine for. Treating telehealth policy as a recurring annual cycle, with a calendar and an owner, was the change in mindset the 2025 fee schedule called for.

Three rulebooks for one video visit

Telehealth means delivering care at a distance, usually by video, sometimes by phone, and sometimes through tools that send readings such as blood pressure from a patient’s home to the clinic. Technically, it is not hard. A clinician and patient need a secure video platform, a device and a connection. The hard part is the rules about when such a visit counts as real care that Medicare will pay for.

Those rules come from three places. CMS controls what it can change by regulation in the annual fee schedule: how services are defined, how clinicians supervise, and enrollment details. Congress controls what the statute restricts: where the patient may be, which practitioners may bill and the in-person visit requirement for mental health. The DEA, with HHS, controls whether controlled substances can be prescribed without an in-person visit.

The arrangement resembles a road trip that needs three permits from three offices: a registration renewed every year on a predictable date, a route permit granted for short stretches at the last minute, and a special permit for certain cargo. If any one lapses, the trip stops, even if the other two are in order.

Every Medicare telehealth dependency falls under exactly one of these three, and each affects a different group of patients. Medicare Advantage, the private plans that deliver Medicare benefits for many beneficiaries, and commercial payer policies add variation, but the three public clocks set the floor. Knowing which clock governs which service is the starting point for every decision that follows.

The chain of systems behind a video link

From the patient’s side, a video visit looks simple: a text message with a link, a few minutes in a virtual waiting room, a conversation. Behind that link sits a chain of systems and people, and each link in the chain can be affected when a rule changes.

It starts with scheduling. The scheduling system decides whether an appointment can be booked as virtual, for which kinds of clinicians and which kinds of patients. Next comes the video platform itself, which must meet health privacy requirements. During and after the visit, the clinician documents it in the electronic health record, the digital chart that holds a patient’s history, and records how the visit happened: video, phone, or in person.

Then comes billing. Each claim to Medicare carries codes that describe the service, where the patient was, where the clinician was and whether the visit was by video or audio only. If the codes are wrong, the claim comes back. Billing systems apply rules that check these codes automatically, and those rules must change whenever the underlying policy changes.

Finally, there are the people: front-desk staff who explain options to patients, nurses who prepare patients for visits, clinicians who decide when in-person care is needed, and coders who review claims. When a flexibility lapses, every one of these steps needs a new instruction at the same moment.

Policy deadlines are therefore operational events as well as legal ones. A change to a single rule can mean reconfiguring scheduling templates, updating billing logic, retraining staff and rebooking patients, all on the same date. The earlier a health system knows which services each deadline touches, the more of that work it can prepare in advance and switch on only if needed.

The CMS clock: annual and mostly clinician-facing

Fee schedule changes arrive once a year, with a proposed rule in summer and a final rule around November 1. That cadence is predictable, and the November 2024 changes mostly concerned how clinicians work.

Virtual direct supervision lets a supervising practitioner be present by real-time audio and video. Staffing models for infusion, remote patient monitoring and some outpatient services had been built on it since 2020. Take a suburban infusion center where nurses administer treatments while a physician supervises remotely across several sites. If virtual supervision lapsed, that model would need a physician physically on each site, which can mean hiring, rescheduling or cutting hours.

Clinicians working from home could keep their home address off public enrollment records, a safety concern for many. A therapist seeing patients from a home office does not want that address listed in a public database.

For patients, these changes are mostly invisible unless they fail. The planning task is internal: an inventory of which workflows rely on each CMS flexibility, its end date and the staffing change needed if it lapses. Because the CMS clock is annual, that inventory can be reviewed each summer when the proposed rule appears, well ahead of the January effective date. Of the three clocks, this is the easiest to manage, precisely because it is the most regular.

The congressional clock: short, late and patient-facing

The statutory flexibilities are the ones patients feel. If they lapse, most Medicare beneficiaries outside rural areas could no longer receive many services at home by telehealth. Physical and occupational therapists, speech-language pathologists and audiologists would lose their ability to bill for telehealth. The in-person visit requirement for mental health telehealth would take effect. Congress had renewed these flexibilities a few months at a time, usually near the deadline and often inside larger spending bills.

That pattern makes this clock difficult. A health system cannot know until the final days of the year whether January’s video appointments will be payable. It must therefore prepare for both outcomes at once, which is more expensive than preparing for either one.

The same clock governed the Acute Hospital Care at Home waiver, which lets hospitals treat selected inpatients at home, with clinicians visiting in person and monitoring remotely. Its statutory extension also ran to the end of 2024. A health system with a hospital-at-home program faced a sharper version of the telehealth problem: if the waiver lapsed, patients already admitted at home would need a plan, and new admissions would stop. Census planning for the last weeks of each extension period, with a defined path back to a bed in the building, is the contingency that matters.

This clock calls for patient-level planning. Scheduling systems can flag appointments that fall after a deadline, by service type and patient location. Patient messages can be drafted for each outcome. In-person capacity can be reserved for conversions.

Patients who rely on audio-only care, often older or without reliable broadband, need particular attention. The permanent CMS change on audio-only does not by itself restore home eligibility if the statutory flexibilities lapse. CMS allowed the phone to be used, but Congress still decides whether the patient’s home counts as an allowed location. Both permits are needed.

On the billing side, claims systems need configurations ready for both outcomes and a rule for holding claims when coverage on the date of service is uncertain. A claims hold means the bill is prepared but not sent until the rules for that date are clear, which avoids submitting claims that would be denied or later have to be repaid.

The DEA clock: narrow but highest-stakes

The third clock affects a smaller group with the most to lose: patients receiving controlled substances through telemedicine, including buprenorphine for opioid use disorder and stimulants for ADHD. Without the temporary rule, many would need an in-person visit before a prescription could be issued or renewed. For patients in treatment for addiction, an interruption is a clinical risk rather than an inconvenience.

The DEA rule applies to every payer, Medicare included. A patient with commercial insurance, or paying cash, is affected just the same. Health systems needed a registry of patients receiving controlled substances via telemedicine, with the date of each patient’s last in-person visit, so they could prioritize in-person appointments if the extension failed.

They also needed to track state rules, which can be stricter than federal ones, and the DEA’s pending permanent framework, including a proposed special registration for telemedicine. A special registration would be a distinct federal permission for clinicians who prescribe these medicines remotely, which would replace the patchwork of temporary extensions if it were ever finalized.

Running the three clocks as one annual cycle

The three clocks interact at year-end, when the fee schedule takes effect, congressional deadlines cluster and DEA extensions have tended to arrive. Staff are on leave at the same time. Holiday staffing, budget closes and patient volume all peak in the weeks that the policy decisions land.

A single policy owner, with a calendar of every telehealth-related expiry by authority and a pre-agreed trigger for each contingency, prevents the last-week scramble. A trigger is a decision made in advance: “If Congress has not acted by this date, we begin rescheduling these patients.” Making the decision in October, calmly, is far better than making it on December 30 with half the team away.

Messages to patients, clinicians and billing staff can then go out in the right order: billing first, clinicians next so they can advise patients, and patients a week or two before any change. The order matters. A patient who hears about a change before their clinician can explain it calls the clinic confused; a clinician who learns before billing is ready gives advice that the system cannot yet support.

Suppose a regional health system runs primary care, behavioral health and rehabilitation services. In its annual cycle, the summer proposed rule triggers a review of supervision-dependent staffing. In early autumn, the policy owner lists every service by governing authority. In October, scheduling flags January appointments for therapists and audiologists whose billing depends on Congress. Behavioral health pulls its controlled-substance registry and identifies patients without a recent in-person visit. By December, every team knows its trigger date and its message. When the extensions arrive, the plan is shelved for next year.

The same system without the cycle looks very different. The first sign of trouble is a news story in mid-December. The compliance office emails department heads asking what services might be affected. Rehabilitation does not know that its therapists’ telehealth billing depends on Congress. Behavioral health has no list of patients on controlled substances by telemedicine and must build one by hand from charts. Scheduling cannot easily pull January appointments by service type and patient location, so staff call patients one by one. Billing keeps sending claims because nobody set a hold. When the extension arrives at the last minute, the scramble stops, but nothing has been learned, and the same scramble repeats at the next deadline.

The difference between those two systems is not money or technology. It is a calendar, an owner and a set of decisions made before they were urgent.

Revenue, patients and the board’s autumn report

The stakes run across the whole enterprise. On revenue, telehealth visits that become unbillable overnight are lost income, and claims sent in error must be repaid. On cost, emergency rescheduling and temporary staffing are expensive when done in a rush. On patients, sudden cancellations fall hardest on those who are older, rural, disabled or in addiction treatment, the people who most need continuity. On regulators and compliance, billing for services that were not covered on the date of service creates audit exposure. On reputation, a health system that leaves patients confused at New Year looks disorganized, even when the cause is in Washington.

There is also a less visible strategic cost. Organizations that treat each deadline as a one-off crisis hesitate to invest in virtual care at all, because they cannot plan around it. Those that treat the uncertainty as a routine annual cycle can keep investing, knowing they have a tested fallback.

The useful report to a board is short: how many patients and how much revenue depend on each of the three clocks, when each next deadline falls, and what the trigger date for each contingency is and whether it has been rehearsed. A board that sees those answers once a year, in the autumn, can ask the right questions before the holidays rather than after a lapse has already happened.

Where year-end telehealth planning breaks

The working version is unglamorous. There is one owner and one calendar. Every telehealth service is mapped to CMS, Congress or the DEA. Scheduling and billing systems can be switched between outcomes without heroics. Patient messages are drafted before they are needed. The controlled-substance registry is kept current all year rather than assembled in a panic.

The failures are predictable. Treating each extension as the last one, and so never building a routine. Assuming that a permanent CMS change, such as audio-only, settles questions that actually belong to Congress. Forgetting that the DEA rule reaches beyond Medicare to every payer. Leaving hospital-at-home patients without a defined path back to the building. And sending patient communications before clinicians and billing staff are ready to support them.

Which clock each service runs on

• For each telehealth service we offer, which authority governs it, CMS, Congress or the DEA, and when does it expire?
• Can our scheduling system show, today, every appointment that falls after the next statutory deadline, by service type and patient location?
• Do we have a current registry of patients receiving controlled substances by telemedicine, with the date of each one’s last in-person visit?
• If our hospital-at-home waiver lapsed, where would currently admitted patients go?
• Are claims holds configured and ready, and what is the trigger to use them?

With hindsight

Congress extended the statutory flexibilities in December 2024 and again in March 2025, each time for months rather than years. All three clocks kept running, much as the December pattern suggested they would. Each extension bought time but not certainty, so the annual cycle quickly became a cycle of several decision points a year.

The flexibilities then lapsed when the federal government shut down on October 1, 2025. Some visits were converted, rescheduled or held for billing until legislation restored them in November. That episode was the real-world test of the contingency plans: organizations with claims holds configured and patient lists ready could act calmly, while others had to improvise. A longer extension followed a further lapse in early 2026.

The DEA clock behaved similarly. The extension through 2025 was followed in early 2026 by a fourth temporary extension through the end of 2026, while the permanent special registration framework remained unfinished. Patients receiving controlled substances by telemedicine stayed dependent on temporary measures for years after the emergency ended.

CMS continued to make incremental changes within its authority in later fee schedules, keeping its clock the most predictable of the three.

Organizations that had mapped services and patients to the governing authority could see which group each lapse touched and act on that group alone. When the statutory flexibilities lapsed during the shutdown, for example, they knew which appointments and which clinicians were affected, and they did not need to disrupt services that rested on CMS or DEA rules.

“Temporary” can be a permanent state. When a capability your organization relies on is renewed a few months at a time by people outside your control, waiting for a final answer gets you nowhere. The sensible response is a routine that makes each deadline ordinary: a map of dependencies, a calendar, an owner and decisions made ahead of time.