For many Medicare patients, telehealth meant a phone call

The week

On April 30, 2020, the Centers for Medicare & Medicaid Services issued a second interim final rule on COVID-19 flexibilities. CMS is the federal agency that runs Medicare and sets the payment rules many other insurers follow. An interim final rule is a regulation that takes effect immediately, before the usual period of public comment, a tool agencies use in emergencies.

Where Medicare telehealth stood

The rule landed on a system that had been rebuilt in six weeks. Before 2020, Medicare paid for telehealth only in narrow circumstances set by statute. The patient generally had to be in a rural area and at a medical facility such as a clinic, not at home. Since 2019 Medicare had also paid small amounts for short “virtual check-ins” by phone or other technology, but those were brief contacts, not visits.

That changed in stages. In early March, Congress gave HHS authority to waive the telehealth restrictions during the emergency. On March 17, CMS used it to let beneficiaries receive telehealth anywhere, including at home. The same day, the HHS Office for Civil Rights said it would not penalize providers for good-faith use of everyday communication tools for telehealth. The CARES Act, signed on March 27, widened the waiver authority and let rural health clinics and federally qualified health centers deliver telehealth. A first CMS interim final rule at the end of March had already agreed to pay for certain services delivered by telephone alone.

The patient the rule was written for

Take an 80-year-old patient in a small town that spring. Her doctor’s office had closed its waiting room and was now offering “video visits.” She had a flip phone, no home internet worth the name, and no grandchild nearby to set anything up. For her, the promise of telehealth meant nothing unless the doctor would simply pick up the phone. Across the country, providers were telling regulators that patients like her were common.

What the April 30 rule did

Among other changes, the rule raised Medicare payment for telephone evaluation and management visits to levels equivalent to similar office and outpatient visits. “Evaluation and management” is the billing category for the core of medicine: a clinician talking with a patient, assessing the problem and deciding what to do. According to CMS, the change moved payment for these calls from a range of about $14 to $41 to about $46 to $110. The increase was retroactive to March 1, 2020, so practices that had already been calling patients could be paid fairly for that work.

The rule also broadened the range of services that could be furnished by audio-only telephone, meaning by voice alone, with no video. CMS said the list now included many behavioral health and patient education services. It waived the video requirement for certain telephone visits and added them to the Medicare telehealth list. And CMS said it would add services to that list on a quicker, subregulatory basis during the emergency. The telehealth list is the official roster of services Medicare will pay for when delivered remotely. “Subregulatory” means CMS could update it through guidance rather than through a full new rulemaking, which normally takes months.

Other parts of the same rule shaped who could use the phone. Physical therapists, occupational therapists and speech-language pathologists could now furnish Medicare telehealth. Hospitals could bill for certain services that their practitioners delivered remotely to registered outpatients, including patients at home. Rural health clinics and federally qualified health centers could be paid as distant sites for telehealth, as the CARES Act provided.

Why the phone

The reason for all this was practical. Providers had reported that many Medicare patients could not use video visits because they lacked devices, broadband or familiarity with the technology. Paying for video alone would have paid for telehealth mainly for patients who were already easiest to reach.

The connectivity problem was being attacked from other directions too. That month the Federal Communications Commission began awarding grants from a $200 million COVID-19 Telehealth Program, funded by the CARES Act, to help providers buy devices and connectivity. Those grants helped providers, not patients’ homes. A patient without broadband still had no broadband.

Medicare’s choice also mattered beyond Medicare. State Medicaid programs and commercial insurers set their own telehealth rules, but many looked to CMS for direction. A federal decision to pay for phone visits like office visits gave states and plans a reference point, even though each moved on its own terms.

The timing was also unusual. Interim final rules take effect first and invite comment afterwards, so organizations were redesigning workflows around a rule whose final shape, and end date, nobody knew.

For hospital and practice leaders, the rule turned the humble telephone from an informal stopgap into a channel with real revenue attached. That brought an obligation along with the money: to run phone care with the same discipline as any other service.

Telehealth in 2020 is usually remembered as video. For a large share of Medicare patients, it was a phone call. The April 30 rule recognized that by paying for audio-only visits at levels comparable to in-person care. Health systems then had to treat the telephone as a channel in its own right. That raised three design questions for patient experience, digital and operations leaders: who the phone channel served, how phone visits fitted clinical and administrative workflows, and which controls a low-friction channel required.

Those questions sound operational, and they are. They are also strategic. How an organization answered them determined whether its most vulnerable patients could reach care, whether its clinicians could work sustainably, and whether its billing would hold up when auditors arrived later.

The three questions come after a short primer on the channels and the billing machinery, because the design choices only make sense against them. They are followed by how to measure the channel, a contrast between two hypothetical practices, what the channel means for revenue, risk and reach, and a checklist.

Video, phone, portal and the record behind them

Telehealth is simply health care delivered at a distance using technology. In 2020 that mostly meant three things. Video visits connect clinician and patient through a camera and screen, using an app or a web link. Audio-only visits are telephone calls in which the clinician assesses and advises the patient by voice. And asynchronous care, such as messages through a patient portal, lets patient and clinician exchange information without being connected at the same time.

Behind every one of these is the electronic health record, or EHR. It is the software system where a health organization schedules appointments, documents care, places orders for tests and prescriptions, and generates the bills sent to insurers. If the EHR does not know a visit happened, or records it incorrectly, the visit is effectively invisible. It cannot be analyzed, audited or properly billed.

Codes, and why the phone was different

Billing depends on codes. Every visit is described to an insurer using standard codes for the service delivered and, increasingly, for how it was delivered. Medicare decides which codes it will pay for, at what rate and under what conditions. Before 2020 Medicare’s telehealth rules were narrow, and telephone calls were paid at lower rates than office visits, if at all. The emergency rules changed that quickly, and the April 30 rule extended the change to the phone.

The telephone visit codes were time-based. Three tiers covered calls of roughly 5 to 10, 11 to 20 and 21 to 30 minutes of medical discussion. That design made the call’s length a billing fact, and therefore something the record had to support. A clinician’s estimate written hours later is weaker evidence than a start and end time captured by the phone system.

Medicare also had two smaller categories that are easy to confuse with telehealth. Brief virtual check-ins, introduced in 2019, paid for short contacts to decide whether a patient needed a visit. E-visits paid for portal exchanges over a period of days. Neither was a visit in the full sense. A health system that wanted clean data needed to keep the three categories distinct in the EHR, because they carried different rules, rates and documentation.

During the emergency, CMS also told practitioners to bill telehealth with the place of service that would have applied in person, and to add a modifier marking the service as telehealth. A modifier is a short code appended to a claim line to add detail. Modifiers would later become the main way Medicare distinguished audio-only care.

Who sets the rules for which patient

The April 30 rule applied to traditional Medicare. Medicare Advantage plans, run by private insurers, had their own flexibility to cover telehealth. Medicaid telehealth policy is set state by state, within federal limits. Commercial coverage depended on state law and each plan’s choices. A single clinic could therefore have four sets of phone rules running at once, one per payer type. That alone was a reason to record modality as structured data: without it, nobody could tell which calls were billable to whom.

What the rule did not change

The rule changed Medicare payment. It did not change several other regimes that a phone visit touches. HIPAA still applied, softened only by the enforcement relief the Office for Civil Rights had announced. State licensure rules still governed who could treat a patient in which state, and waivers there were state by state and temporary. Prescribing controlled substances remotely fell under separate DEA rules, with their own emergency exceptions. Malpractice coverage for telehealth depended on each policy. A health system that read the April 30 rule as permission for everything by phone would have missed each of these.

Banks face the same choice when customers cannot reach a branch. One that offered only a slick mobile app would leave behind every customer without a smartphone. A sensible one also staffs its phone lines properly, gives them the same systems as the branch, and tracks who uses which channel.

Who the phone channel served

Video visits assume a smartphone or computer with a camera, enough broadband, software that works and the confidence to use it. Many older adults, rural residents and low-income patients lacked one or more of those. Patients with limited English proficiency faced an extra barrier when interpreters could not easily join a video session. Some patients with disabilities found video harder rather than easier. For these groups, the telephone was the only version of telehealth that worked.

An equity decision, measured

That made audio-only an equity decision. A health system offering only video would, in effect, have offered telehealth mainly to patients who were already easier to reach. Leaders needed data on which patients used which channel, broken down by age, language, payer and geography. Many organizations could not produce it, because visit modality was recorded inconsistently or only in free text.

Modality simply means how the visit happened: in person, video or phone. Recording it “in free text” means a clinician typed something like “spoke w/ pt by phone” into a note, which a person can read but a report cannot reliably count. Adding a required modality field to every virtual visit, a drop-down menu that must be filled in, was a small change that made the analysis possible.

Channel data also exposed missed demand. Patients who never completed a scheduled video visit often reappeared as no-shows or emergency department visits. Offering a phone fallback at the point of scheduling, and tracking how often it was used, showed where video was failing patients.

For a board or executive team, that data answers questions they care about. It shows whether the organization is reaching the communities it says it serves. It also shows where patients are falling through the cracks and turning up later, sicker and at higher cost, in the emergency department. Without structured modality data, those questions get anecdotes instead of answers.

Not every barrier is the same barrier

The patients who relied on the phone did so for different reasons, and the reasons call for different designs. A patient with no broadband but a working smartphone may manage video on a cellular connection with a little coaching. A patient with a basic phone and no data plan cannot. Some low-income patients relied on prepaid phones with limited minutes, so a long hold or a missed call-back cost them something real. Scheduling a call at a set time, and calling the patient rather than asking them to call in, protected those minutes.

Hearing loss runs the other way. For many older patients, a phone call is harder than video, because they lose the lip-reading and visual cues that help them follow speech. Deaf and hard-of-hearing patients may need relay services or video with captioning. A phone-first design that ignored them would trade one access gap for another. The intake question should ask which channel works for this patient. Whether they have video is only part of the answer.

Language access had a legal frame as well as a practical one. Section 1557 of the Affordable Care Act already required health programs receiving federal financial assistance to take reasonable steps to provide meaningful access to people with limited English proficiency. A phone channel that could not add an interpreter mid-call was an access problem and a compliance exposure at once.

Caregivers and proxies on the line

Many older patients did not take the call alone. An adult child, a spouse or a paid caregiver often held the phone, relayed questions or managed the medication list. HIPAA generally allows a provider to share relevant information with family members involved in a patient’s care when the patient agrees or does not object. On a call, that agreement needs to be asked for and recorded. The EHR should also hold any formal proxy or authorized representative, so staff can see who may speak for the patient without guessing. A phone channel designed for the patient alone will meet many calls where the patient is not the only voice.

Different organizations, different phone patients

The mix of patients who needed the phone varied by type of organization. A federally qualified health center, serving many uninsured and Medicaid patients, was likely to see heavy phone use across age groups, driven by cost of data and devices. A rural hospital’s clinics saw it among older Medicare patients with weak broadband. An urban academic medical center saw it concentrated in particular clinics, such as geriatrics, and among patients needing interpreters. Behavioral health practices found that some patients preferred the phone even when video was available, because it felt less exposing. Each needed the same data, but the design priorities differed.

How phone visits fitted the workflow

A phone visit needs the same scaffolding as any other visit: scheduling, registration, consent, interpreter access, documentation in the EHR, correct coding, and a route to orders, prescriptions and follow-up. Many practices first ran phone visits outside normal workflows. Clinicians called from personal phones and documented afterwards. That produced gaps in records and caller ID problems when patients ignored unknown numbers. Adding an interpreter was awkward or impossible.

A visit type, not a favor

The better approach treated audio-only as a scheduled visit type in the EHR. Telephony showed a recognizable number, supported three-way calls with interpreters and logged the time and modality for billing. Small design choices made a visible difference. A reminder text the day before, a call from a known number and a stated time window all raised completion rates. Front-desk staff needed scripts for converting a video appointment to a phone visit when a patient struggled with the technology.

The visit type carried its own required fields. A practical set covered modality, start and end time, who was on the call, the patient’s verbal consent, the patient’s location and the reason video was not used. Location mattered for two reasons. A clinician generally needs a license valid in the state where the patient is, and many states only temporarily waived that rule in 2020. Location also determines where emergency help should go if a call reveals a crisis.

The telephony underneath

The phone system is where much of the design work lives. A clinic can give clinicians a softphone, an app on a computer or mobile phone that places calls through the organization’s phone system. Calls then show the clinic’s number, can be logged automatically and can bring in an interpreter line with one action. Some EHRs and contact center platforms can link a call to the patient’s chart so the time is captured without retyping.

Caller ID had a further twist. Carriers and phone apps increasingly label unfamiliar outbound numbers as possible spam, and from mid-2021 the FCC required large carriers to implement a call authentication framework known as STIR/SHAKEN. Health systems that placed thousands of outbound calls from new numbers sometimes found them flagged. Registering those numbers with carriers’ analytics providers and keeping a small, stable set of outbound numbers protected answer rates.

Follow-up, capacity and the clinician’s day

Follow-up needed equal care. A phone visit that ends with a lab order or a new prescription depends on the patient receiving instructions they cannot see on a screen. After-visit summaries had to go by mail, text or portal, in the patient’s language, with a number to call back.

Capacity planning changed as well. Phone visits shifted load from exam rooms to telephony and scheduling. Call queues, outbound line capacity and the number of interpreters available by phone now set how many patients a clinic could see. Operations leaders who tracked abandoned calls, interpreter wait times and visits started late could see those limits forming. Clinics that set aside protected blocks for phone visits, rather than squeezing calls between in-person appointments, ran closer to schedule and kept patients waiting less.

Clinician experience mattered too. Many clinicians found back-to-back calls more tiring than they expected, without the cues of a room, and scheduling templates had to allow time for documentation between calls. A scheduling template is the pattern of slots that the EHR offers for booking, for example fifteen-minute in-person slots in the morning. Building phone blocks with realistic gaps into those templates is a modest technical change with a large effect on burnout.

Sequencing the build

The order of work matters, because each step depends on the one before. A sensible first week defines the visit type, its required fields and the consent language, with compliance review. The second week configures telephony: outbound numbers, softphones, interpreter conferencing and call logging. The third week trains front-desk staff and clinicians and opens protected scheduling blocks. Measurement starts at once, but the first claim audit waits until there is a month of visits to sample. Doing these in the reverse order, as many practices did by necessity, means reconstructing the record afterwards.

Hospital clinics and independent practices

The workflow also depended on how the organization billed. Physician practices billed Medicare under the physician fee schedule. Hospital outpatient clinics often bill in two parts, a professional claim and a facility claim. The April 30 rule let hospitals bill for certain services their staff delivered remotely to registered outpatients, including at home. Using that option meant registering the patient as a hospital outpatient for the encounter, with the paperwork that implies. Independent practices had a simpler path and less back-office capacity to manage it. Health systems with both kinds of clinic often needed two versions of the same phone workflow.

Which controls a low-friction channel required

Lower friction brings risk in four areas, each with its own control: clinical, financial, identity and privacy.

Clinical: knowing when the phone is not enough

Some conditions cannot be assessed without seeing the patient. Clinicians need written criteria for when to escalate to video or in-person care, and the EHR should make escalation a single step. Typical triggers include new or worsening symptoms that need examination, signs of acute illness, and any situation where the clinician is unsure the patient understood. Tracking how often phone visits led to an in-person visit within a short period shows whether the criteria are being used.

Financial: documentation that supports the claim

Audio-only services are easier to bill without adequate documentation, which attracts audit and fraud scrutiny. Documentation standards and periodic claim audits address that. A claim audit is an internal review of a sample of bills sent to insurers, checking that the clinical record supports what was billed, before an outside auditor does the same.

The audit should check a few specific things. It should confirm that the note records modality and time consistent with the code billed, and that the patient’s consent is documented. It should check that the content of the note supports the level of service. Outlier reports help target the sample: clinicians who bill only the highest time tier, or who bill unusually many calls in a day. Those are patterns an outside reviewer is likely to look for too.

Identity: confirming who is on the line

On a call, confirming that the person answering is the patient, or an authorized representative, needs a defined procedure rather than a clinician’s judgment in the moment. Two identifiers, checked the same way every time, are a reasonable baseline. Inbound calls need the same care. Callers who claim to be patients and ask to change an address, a pharmacy or a phone number on file are a known route for fraud and privacy breaches. Changes of contact details should trigger a verification step and a notice to the old contact.

Privacy: at the patient’s end and the clinician’s

Patients at home may be overheard, and staff calling from home must protect what they say and write. Asking at the start whether the patient can speak freely, and offering to call back, protects sensitive conversations. Voicemails should carry the minimum needed to get a call back. Clinicians using personal mobile phones leave patient numbers and call histories on devices the organization does not manage. A softphone tied to the organization’s system keeps those records where they can be protected and retained.

The security rules that apply depend on the technology. A call over a traditional landline is treated differently under HIPAA from a call carried over internet-based phone services or apps, which handle the conversation as electronic data. Most modern clinic phone systems are internet-based. Security and compliance teams should confirm which category their systems fall into and whether a business associate agreement is needed with the telephony vendor.

The evidence a reviewer will ask for

An auditor or payer reviewing phone visits will usually ask for a short list of documents. The first is the written policy defining the audio-only visit type and its required fields. The second is the escalation criteria and the identity procedure. The third is a sample of notes matched to claims, with call logs showing start and end times. The fourth is the organization’s own audit results and what it did about any errors, including refunds of overpayments where they were found. An organization that can hand these over in a day is in a different position from one that must reconstruct them from memory.

Controls that keep the channel adaptable

These controls also make the channel adaptable. Payment rules for audio-only care were temporary, and they were likely to diverge across Medicare, Medicaid and commercial payers. A health system that ran phone care as a distinct, measured service could adjust which services it offered by phone as rules changed. One that had absorbed phone calls informally into clinicians’ days could not even see what it was doing.

Measuring the phone channel

A channel that is not measured cannot be managed, and the phone is easy to leave unmeasured because it feels so ordinary. A short set of measures, reviewed monthly by the people who run ambulatory operations, turns it into something leaders can see and steer.

The first group of measures is about access. It covers how many visits were delivered by phone, and to whom, broken down by age, language, payer and geography. It also covers how often a scheduled video visit converted to a phone visit, and how often a patient missed a virtual visit entirely. Those numbers show whether the channel is reaching the people it was meant for, and where video is failing them unnoticed.

The second group is about operations: calls abandoned before a patient reached anyone, waits for an interpreter to join, and phone visits that started late. These are the early warnings that capacity, rather than clinical demand, is setting the limit on care. Outbound answer rates belong here too, because a falling answer rate can signal a caller ID or spam-labeling problem before anyone reports it.

The third group is about quality and integrity. It tracks how often a phone visit led to escalation to video or in-person care within a short period. It tracks what share of sampled audio-only claims had documentation that fully supported what was billed. And it tracks whether patients complained about identity or privacy on calls.

Definitions before dashboards

The measures only work if the definitions are fixed first. A “phone visit” must mean the audio-only visit type, not every call a clinician makes. A “conversion” must be counted when a scheduled video visit is completed by phone, with the reason recorded. Without agreed definitions, two clinics can report the same reality as very different numbers, and the board sees noise.

No new technology is needed for any of these measures if the visit type, modality and duration are captured as structured data from the start. All of them become guesswork if they are not. The unglamorous decision to add a required field was therefore one of the most important choices a health system made about telehealth in 2020.

A final point on interpretation. Phone use that is high among older or rural patients is not a failure of the video program; it may be the channel working exactly as intended. The worrying signal is the opposite pattern: groups that use neither video nor phone, and reappear later in the emergency department.

Two rural practices, one rule

Suppose two primary care groups serving similar rural counties both saw their in-person visits collapse in March 2020, and both welcomed the April 30 rule.

The first group let clinicians handle it themselves. Doctors called patients from their mobile phones between other tasks and wrote notes at night. Many patients did not answer calls from unknown numbers. Interpreters were rarely used. Within a few months, the group could not say how many phone visits it had delivered, to whom, or how long they had lasted. When its billing team prepared claims, it had to reconstruct details from notes, and some visits were never billed at all. When the group later needed to show auditors why particular visits were done by phone, the records were thin.

The second group spent two weeks setting things up. It created an audio-only visit type in its EHR with required fields for modality and duration. It bought a telephony service that displayed the clinic’s main number and could add an interpreter to the call. It wrote a one-page script for front-desk staff, a short escalation guide for clinicians, and a two-identifier check at the start of every call. It set aside two afternoon blocks a day for phone visits.

Six months later, the second group could show its board how many older and non-English-speaking patients it had reached by phone. It could show how often phone visits led to in-person follow-up, and how its documentation held up in a quarterly internal audit. When payment rules shifted, it could see which services depended on the phone and plan accordingly. Both groups had the same rule and the same patients. Only one had a channel it could manage.

What the two weeks cost

The second group’s setup was modest but not free. It needed an EHR analyst to build the visit type and fields, and a few days of a telephony vendor’s time. It also needed a compliance review of the scripts and consent language, and training for front-desk staff and clinicians. It also gave up some flexibility: clinicians could no longer squeeze an unscheduled call into a gap and bill it later without the required fields. The first group saved that effort in April and spent more of it later, reconstructing records and answering auditors.

The same choice in other settings

A community health center faced the same fork with different constraints. Its patients were more likely to have prepaid phones and to need interpreters, so scheduled call times and interpreter conferencing mattered more than caller ID alone. A behavioral health practice found the phone suited many therapy sessions and needed stronger privacy checks at the start of each call. A hospital-owned clinic had to decide whether to register phone patients as hospital outpatients under the new rule, which brought revenue and paperwork together. The design steps were the same in each case; the priorities were not.

Revenue, risk and reach

The audio-only story touches every concern on an executive’s list. Revenue: phone visits carried real payment, but only if they were scheduled, documented and coded correctly. Risk: poorly documented audio-only claims are an obvious target for auditors, and clinical decisions made without seeing a patient need clear escalation rules. Patients and community: the phone channel reached the people least able to use video, which is often exactly the population a health system’s mission and community commitments emphasize. Staff: clinicians’ wellbeing depended on schedules designed for phone work rather than calls squeezed between appointments. Regulators and reputation: organizations that could explain, with data, why and how they used the phone were in a far stronger position when oversight bodies came asking.

Who owns the channel

There is a governance point here as well. Decisions about the phone channel were often made at the level of individual clinics or even individual clinicians, because the phone did not look like a technology project. Yet those decisions touched revenue cycle, compliance, patient experience, interpreter services and IT all at once. Someone senior needed to own the channel as a whole, with authority to set the visit type, the telephony standard, the documentation rules and the measures. Where nobody did, each department solved its own piece, and the gaps between the pieces were where patients and payments got lost.

A workable split gives the overall channel to the leader of ambulatory operations or patient access. IT owns the telephony and the EHR build. Revenue cycle owns coding and claim audits. Compliance owns documentation standards and privacy. Language services owns interpreter access. The channel owner chairs a short monthly review of the measures with each of them.

Payer mix and the financial case

The financial value of the phone channel depended on who paid. For traditional Medicare, the April 30 rule set clear rates. Medicaid rates and rules varied by state, and community health centers were paid under their own methods. Commercial plans followed state mandates and their own policies. Organizations in risk-based contracts, such as accountable care organizations, valued the phone differently again. For them, a call that kept a patient out of the emergency department mattered more than the visit payment itself. A finance team modeling the channel needed revenue by payer and the avoided costs that value-based contracts reward, not visit counts alone.

The case against paying for calls

The strongest objection to this argument is clinical. A phone call omits the physical examination and much of what a clinician sees. Paying for it like an office visit could encourage substitution of cheaper, weaker care, and make some diagnoses later and harder. There is a financial objection too: payment parity for a lower-cost service invites overuse and fraud.

Both objections have force, and they are the reason for the controls above. Yet the alternative for many patients in 2020 was no visit at all, not an office visit. The equity data showed who those patients were. Escalation criteria, documentation that records why video was not used, and audits of billing patterns answer the quality and fraud concerns without withdrawing the channel. Where an organization cannot run those controls, the objection wins, and it should limit which services it offers by phone.

A channel built on purpose

A well-run phone channel has:

• A distinct audio-only visit type in the EHR, with modality, duration, consent, patient location and the reason video was not used captured as structured fields.
• Telephony that shows a recognizable, registered number, supports interpreter conferencing and logs call details.
• A phone fallback offered at scheduling, with its use tracked.
• Written escalation criteria and a standard identity check for every call.
• Protected scheduling blocks with time for documentation.
• Regular internal audits of a sample of audio-only claims, targeted with outlier reports.
• A named owner of the whole channel, with a monthly review of agreed measures.

The usual ways it goes wrong:

• Treating phone calls as informal favors rather than visits, so they never appear in the data.
• Recording modality only in free-text notes.
• Calling from personal phones, which hides caller ID and leaves no log.
• Sending follow-up instructions only through a portal the patient never uses.
• Designing for patients without broadband and forgetting patients who cannot hear well on the phone.
• Assuming temporary payment rules would last forever, or end everywhere at once.

Where a phone channel leaks

1. Can we report, today, how many visits were in person, by video and by phone, broken down by age, language, payer and geography?
2. Is audio-only a defined visit type in our EHR, with required structured fields?
3. What number do patients see when we call them, and can we add an interpreter without hanging up?
4. What are our written criteria for escalating a phone visit to video or in person?
5. When did we last audit a sample of audio-only claims, and what did we find?

With hindsight

Audio-only care proved durable for some services and contested for others. Through its rulemaking for 2022, CMS made audio-only delivery permanently available in certain circumstances for mental and behavioral health services furnished to patients at home. Many states also required Medicaid programs or commercial insurers to cover audio-only services, each on its own terms.

The 2022 rule and its conditions

The 2022 physician fee schedule set the template. It limited audio-only to mental health services, and only where the patient was not capable of, or did not consent to, video. It required a new modifier on claims for audio-only services, so the channel became visible in Medicare data. It required an in-person visit within six months before starting mental health telehealth at home, and generally every 12 months after, with exceptions. Rural health clinics and federally qualified health centers could also bill audio-only mental health visits on similar terms.

Temporary rules, repeatedly extended

For most other services, audio-only coverage depended on temporary extensions Congress passed after the public health emergency ended in May 2023. Those extensions came a few months or a year at a time, which made multi-year planning hard. By October 2026, CMS’s telehealth guidance stated that, under current law, Medicare beneficiaries could receive telehealth anywhere and audio-only services at home through December 31, 2027. From January 1, 2028, absent further action, most telehealth would again require the patient to be at a medical facility in a rural area, with behavioral health the main exception.

A permanent definition with a condition

CMS also went further on its own authority. In the 2025 physician fee schedule, finalized in November 2024, it changed its definition of telehealth technology, with no end date, to include audio-only for telehealth services furnished to patients at home. The condition is that the practitioner can offer video but the patient cannot use it or does not consent to it. The physician fee schedule is Medicare’s annual rulebook for what it pays clinicians and on what terms, so this was a lasting change rather than an emergency measure. Its practical reach still depends on which services the statute allows at home, which is why the 2028 date matters.

That framing matches the argument here: the phone is a patient-chosen channel, and the record has to show why it was used.

Oversight followed the money

Oversight bodies kept pressing for stronger documentation of how and why telehealth was delivered. In August 2022 the HHS Office of Inspector General published an analysis of telehealth billing in the first pandemic year. Using seven measures of high-risk billing, it flagged 1,714 of roughly 742,000 providers who billed telehealth, receiving about $127.7 million in fee-for-service payments. It stressed that the flags did not by themselves prove fraud. It recommended that CMS strengthen monitoring and targeted oversight of telehealth, and educate providers on billing.

Privacy guidance caught up too. In June 2022 the HHS Office for Civil Rights published guidance on audio-only telehealth under HIPAA. It distinguished calls over traditional landlines from calls carried by electronic technologies such as internet-based phone services and apps, which fall under the Security Rule. The emergency enforcement relief for telehealth tools ended with the public health emergency in May 2023, after a 90-day transition. Health systems still running clinician calls through unmanaged apps then had a clear compliance gap.

What held up

Two things stand out. First, the patchwork itself is the reality. Medicare, state Medicaid programs and commercial insurers moved on different timetables and with different conditions. A health system has to know, payer by payer, which phone services it can offer and be paid for. That is only possible if phone visits are recorded as a distinct, structured service.

Second, the condition in the 2025 rule turns good documentation from a nice-to-have into a requirement. A clinician must be able to offer video, and the record must show that the patient could not use it or chose not to. Organizations that built a measured phone channel in 2020 already had most of what that requires. Those that let phone calls drift informally into clinicians’ days had to build it later, under more scrutiny.

Less certain is how much the phone will matter after 2027. If the statutory flexibilities lapse as scheduled, audio-only care at home narrows to behavioral health for Medicare, while some states keep broader rules for Medicaid and commercial plans. The organizations best placed for either outcome are those that can already see, by payer and by service, how much of their care runs over the phone.