RPM Virtual Assistants After CMS Code 99445: Chase Day 2, Not Day 16

Under the 2026 RPM codes, the second reading day is the one that unlocks the device payment.

RPM is Medicare-paid care in which a connected device sends a patient's readings to the practice between visits. Since January 1, 2026, two reading days, not sixteen, unlock the device payment.

That one change rewrites the job of an RPM virtual assistant. For years, the outreach playbook was a rescue mission: find the patients stalled just short of the 16-day line and push them over it. Now the patient near day 16 is already billable. The patient with zero readings is the only one earning nothing.

According to CMS, Medicare has paid physicians and other billing professionals under the Physician Fee Schedule since 1992, and the schedule is rewritten every year. RPM has been reshaped several times inside it. The CY 2021 final rule set the line between device work that non-clinical staff can support and treatment management that clinical staff must deliver. The CY 2026 rule moved the device threshold. It did not move that line.

So here is my prediction. Through 2026 and 2027, RPM programs will shift virtual assistant outreach away from pushing engaged patients to day 16 and toward reactivating silent patients to at least two readings a month. Day 2 is the new finish line.

I'll make the case with three patients, carried through every section: Ruth at 0 reading days, Luis at 2 and Dana at 14. First, what changed. Then, which tasks an assistant may take on. Then, who gets called first, and what could upend all of it in 2027.

What changed in RPM billing on January 1, 2026?

Medicare now pays CPT 99445 for 2-15 days of device readings in a 30-day period, at the same rate as CPT 99454 for 16 or more days.

An analysis of 17 sources shows one consistent pattern: the second reading day now unlocks the device payment, and day 16 no longer earns anything extra. The two device codes are mutually exclusive, so a practice bills one of them per patient in each 30-day window. A patient needs readings on at least two days before either code applies.

To make this usable, I sort patients with a simple decision lens I call the reading-day triage. Every enrolled patient lands in one of three buckets each month:

• 0-1 reading days: no device code can be billed. This is where the revenue leaks.
• 2-15 reading days: the practice bills 99445.
• 16 or more reading days: the practice bills 99454, at the same rate.

Let's carry three patients through the rest of this article. They are illustrative, not client records. Ruth has sent 0 readings this month. Luis has sent 2. Dana has sent 14.

So who gets the first call this month? Ruth. Two readings from her turn an unbillable month into a billable one. Two more readings from Dana change the code on the claim, not the payment. Day 2 = KEY.

A common misconception is that 16 days is still the target. The reality is that Medicare contractor guidance for 2026 describes the new codes as filling reporting gaps for monitoring under 16 days and management under 20 minutes. In other words, the short month was designed to be billable.

Why does the timing matter so much? Because margins were already thin going in. According to CMS, the CY 2025 Physician Fee Schedule final rule, issued November 1, 2024 and effective January 1, 2025, cut the conversion factor by 2.93%. When the base rate shrinks, a patient-month that pays nothing hurts more. It's important to note that the fix here is not more clinical time. It is more patients crossing day 2.

That is where a virtual assistant earns its place. Getting a quiet patient to take a reading is reminder and follow-up work, the kind of outreach an assistant or a dedicated healthcare call center team can own. At HelpSquad, a Clutch 2026 Top Virtual Assistant Company that also provides healthcare BPO services, our VAs are HIPAA-certified through HIPAATraining.com, and full-time staffing runs $8-$13 an hour. The rate alone is not the point. The point is that the lowest-cost hour in an RPM program should be spent where it moves a claim, and that means Ruth's phone number, not Dana's.

In practice, day 16 now changes the code, not the check. The takeaway: the silent patient is your real revenue problem.

In summary, the 2026 codes moved the finish line from day 16 to day 2. Every outreach decision in the sections that follow starts from that single fact.

What will matter most for RPM virtual assistants in the next 12-24 months?

Over the next 12-24 months, the programs that win will get every enrolled patient to a second reading day, while watching whether CMS pulls care management back in-house for 2027.

According to Longyear Health's January 2026 review, CMS set 99445 at the same rate as 99454 because practice expense is not related to the number of transmission days. That rationale is the engine behind my forecast. Here are the three signals I'm watching most closely.

Let's run our three patients through these signals. Ruth is the first signal in action: two readings turn her month around. Luis is the second: his clinician's short call is what turns his month into management revenue. Dana is the third, in a quieter way. If her monthly clinical time must someday come from a direct employee, the practice's staffing plan changes, but her device month does not.

What would change my forecast? I'd watch for three things:

• A future fee schedule that ends rate parity between 99445 and 99454.
• A CY 2027 final rule that drops or softens the direct-supervision and direct-employee language.
• Outcome-based payment spreading beyond a single CMS test.

It's important to note that the device-lane work in this article looks less exposed to the 2027 proposal than care management does. That is my reading of the evidence, not a ruling, so I'd keep contract terms flexible until CMS finalizes the rule.

What most buyers miss is that day counts may not be the long-run currency at all. CMS's ACCESS model, a 10-year test that began July 5, 2026, pays an Outcome-Aligned Payment, and full payment depends on measurable improvement against each patient's own baseline. If that logic spreads, getting Ruth to day 2 becomes the floor, not the goal.

Which RPM tasks can a virtual assistant handle, and which need clinical staff?

A virtual assistant can own reminders, device troubleshooting, scheduling and billing prep. Treatment-management minutes, reading review and any clinical judgment still belong to licensed clinical staff.

This is where the simple "chase day 2" answer starts to strain. Before 2026, the pressure point was obvious: a patient who fell short of 16 days earned the practice nothing for the device that month. The new codes removed that cliff on the device side. They did not remove the clinical work on the management side.

According to CMS, its CY 2021 Physician Fee Schedule final rule, issued December 1, 2020, drew the line that still matters today. Auxiliary personnel, including contracted employees who are not clinical staff, may furnish the device setup and supply services (99453 and 99454) under general supervision. The interactive communication behind treatment management (99457 and 99458) means, at a minimum, a real-time, synchronous, two-way audio exchange.

The 2026 rule then added 99470, which allows billing for the first 10 minutes of treatment management. Before that, 99457 demanded its full threshold of clinical staff time before anything could be billed. A lower bar for management is still a clinical bar.

So the work splits into two lanes. I call them the device lane and the care lane:

The device lane has one more rule worth knowing. 99445 requires an FDA-cleared connected device and is billed once per 30-day period. The assistant's job is to get that device transmitting, not to collect numbers over the phone.

Let's go back to our three patients. Ruth's reminder call is pure device lane, right up until she mentions that her numbers have been running high. At that moment the assistant stops, notes what she said and routes it to a nurse. Dana is the opposite case. At 14 days she needs no device push at all. What she may need is her monthly time with a clinician, and no assistant can supply that.

A common misconception is that an assistant's outreach minutes help a patient reach a management threshold. They don't. Those minutes are support time, not treatment management.

Who should sit in the device lane, then? Day counting and code flagging are closer to revenue-cycle work than to reception. HelpSquad was named a Clutch 2026 Top Medical Billing Company, and our screen is strict: less than 3% of applicants clear every dimension, and those are the ones a practice meets. For a smaller program, part-time engagements start at 20 hours per week and are quoted individually, which fits a monthly reactivation push better than a full-time seat.

In practice, the assistant fills the device lane and books the care lane. What this means: clinical minutes stay scarce, so protect them.

In summary, the 2026 codes widened the device lane without touching the care lane. Put the assistant in the first, keep clinicians in the second, and write the escalation rule before the first call is made.

Who should an RPM virtual assistant call first each month?

Call the silent patients first: anyone at 0 or 1 reading days, early in the 30-day window. Patients with two or more reading days drop to the bottom of the device list.

Why this order? Because it follows the payment logic. According to CMS, Physician Fee Schedule payments are built from relative value units for physician work, practice expense and malpractice, adjusted for geography and multiplied by a conversion factor. The device codes are practice-expense payments, and CMS reasoned that practice expense does not rise with the number of transmission days. That is why day 2 and day 16 pay alike.

CMS also has a history of reworking RPM when uptake lags. RPM did not take off after the 2015 push to the extent CMS had hoped, and that shortfall prompted the newer codes. The 2026 codes read as the next attempt to get more patients billable. The day-2 threshold is the lever.

To turn that into a daily routine, I'd recommend a simple rule I call the call-first ladder:

1. 0 reading days (Ruth): call first, early in the window. Two readings turn her month from unbillable to billable.
2. 1 reading day: call next. One more reading day crosses the line.
3. Billable, but no clinician call booked (Luis): the device side is done, so the assistant's job shifts to booking the clinician's live call.
4. Close to day 16 (Dana): no device call at all. Only care-lane scheduling if her clinician needs time with her.

Why book Luis's clinician call at all? According to a 2026 RPM billing guide, 99470 effectively covers 10-19 minutes of management a month, requires at least one real-time interactive communication by phone or video, and pays ~$26. The assistant can't deliver those minutes. The assistant can make sure the call happens.

What does the daily workflow look like?

1. Pull each patient's transmission-day count from the RPM platform every morning. Software built for Medicare RPM in 2026 is designed around 99453, 99454, 99457 and 99458, plus the newer 99445 and 99470, and counts transmission days inside each billing window.
2. Sort the list with the reading-day triage.
3. Work the call-first ladder from the top.
4. Use a short script: confirm the device is charged, paired and placed correctly, then ask the patient to take a reading.
5. Stop and escalate to clinical staff the moment a patient mentions a symptom.
6. Log every attempt and outcome so the billing team can see who crossed day 2.

Contrary to popular belief, more calls to engaged patients don't raise device revenue. Dana's 16th day changes nothing on the claim.

How will you know the shift is working? I'd track three things each month: how many patients are still at 0 or 1 reading days at mid-window, the share of enrolled patients at two or more days, and clinician calls booked versus completed. It's worth noting that the most persuasive proof, a before-and-after view of your outreach mix, only exists if you start logging it now.

If you're comparing virtual medical assistant companies or healthcare BPO firms for this work, ask three questions. Can the assistants work from your platform's day count? Do they follow a written escalation rule? Who answers when a patient calls back? That last one matters more than it looks. HelpSquad has been founder-operated since 2015, and we operate across six countries so your patients are never sent to voicemail. A callback that lands in voicemail is a reading that never happens.

In practice, the ladder puts the assistant's first hour where the payment actually changes. The takeaway: work from zero readings up to two, not from fourteen up to sixteen.

In summary, call Ruth first, book Luis's clinician, and leave Dana's device alone. That is the whole day-2 strategy in one line.

What should an RPM program do before the next 30-day window opens?

Re-sort the patient list by reading days, point the assistant at everyone below two, and keep treatment-management minutes with clinical staff. That order matches how Medicare now pays.

My forecast rests on one reframed fact: a patient's second reading day now earns the same device payment as her sixteenth. Everything else follows from it. The two device codes can't be stacked, so extra readings from an engaged patient add nothing to the claim. The fee schedule pays for practice expense, not effort, so the leak sits with patients who never transmit.

It's worth noting how short this promise is. According to CMS, each annual Physician Fee Schedule final rule governs payment for services furnished on or after January 1 of that year. Parity is a one-year rule until it is renewed. The clinical boundary set in the 2021 rule is older and sturdier.

Here is the part I'd underline. An assistant's value is no longer measured in minutes logged. It is measured in patients moved from zero to two.

So start with Ruth. Her number is already on your list, and her month is still open.