A new JAMA study found that patient portal messages jumped 153% in five years. Here’s what the headline isn’t saying — and what healthtech SaaS companies should do with it.
Published June 29, 2026 | Daree Allen Nieves | ReeWrites.com
A study published in JAMA on June 22, 2026 drew on more than 8 billion patient-provider interactions across 2,067 hospitals and 47,100 clinics. The finding that’s bouncing around health IT circles is that patient-authored portal messages rose 153% between 2020 and 2025.
That number deserves more than a headline treatment.
The 153% figure is only half of what the NYU Langone researchers documented. The other half is where the argument lives.
Contents
What the Data Says
The study used Epic Cosmos data, which makes it the largest-ever analysis of EHR-based patient-provider communication. A few notable stats:
- 153% — how much patient-authored portal messages increased from January 2020 to December 2025
- 42 million — the number of active Epic patients who sent a portal or health app message to a clinician in the first quarter of 2025 alone (~30% of the active Epic patient population)
- 17% — the increase in in-person visits over the same period
- 6% — the decrease in phone calls
The last two numbers are more important than the 153% headline.
In-person visits went up, portal messages went up, and phone calls went down. Patients added another channel. The system on the receiving end absorbed it without any structural redesign.
If Patient Access Improved, Why Are Clinicians Still Overwhelmed?
Expanding digital access worked. Patients found the portal, learned to use it, and 42 million of them used it in a single quarter.
However, the infrastructure built to receive them didn’t keep pace:
- Clinical schedules didn’t build in time for inbox management.
- Compensation structures didn’t change to reflect the work.
I know what it feels like to be on the other end of a portal at 11pm with a question nobody is going to answer before morning.
When my late husband George was in treatment, I sat in front of a treatment options dropdown with no plain-language context and no decision support. The secure messaging function was right there on the screen. There was no one on the other end of it who was going to respond before the window for that decision closed.
Clinicians read these messages, then triage and respond to them. But as AJMC‘s noted, that work happens inside a care layer that health systems and payers haven’t built a payment structure around.
Health systems now effectively operate two workflows: one in the exam room and one in the message queue.
- The exam room workflow has decades of design behind it including staffing models, scheduling logic, triage protocols, documentation standards.
- The message queue workflow is largely being improvised in real time.
Patients are in the portal, but no one built the system for receiving them at this volume.
The Athenahealth 2026 Physician Sentiment Survey found that 63% of physicians feel overwhelmed by portal messages. That’s the self-reported perception.
The JAMA study is the structural data behind why: volume doubled, workflows didn’t, and there’s no billing code for the gap between them.
As Dr. Mark Sendak noted in a HIMSS webinar, “AI dies silently because of clinical workflow, not the tech.”
The NYU Langone data is the empirical evidence that the tool is working. The surrounding system wasn’t built to absorb what the tool created.
The Uncompensated Care Layer
There’s a Chronic Care Management (CCM) parallel to consider here. The Chronic Care Management Improvement Act was introduced in April 2026 because chronic care coordination has been a separately billable Medicare service since 2015, and only 4% of eligible patients use it.
The infrastructure to communicate, consent, and coordinate it was never built. Portal messaging is following the same pattern on a faster timeline.
The care layer exists. Clinicians are doing the work. The reimbursement structure hasn’t moved, and the workflows supporting that labor are largely ad hoc.
SeamlessMD’s early data found that 48% of patient questions come in after 6 pm or before 8 am. That’s almost half of the actual message volume landing outside the hours any staffed inbox was designed to absorb. The inbox that’s already uncompensated during business hours is also receiving messages at 11 pm.
Healthtech SaaS products sit inside this. Say that plainly.
Who is and isn’t in “the 153%”
The study also documented persistent digital divide patterns showing the following groups messaged at significantly lower rates than others:
- Rural populations
- People with high social vulnerability index scores
- Men
- Patients who are very young and very old
The 153% represents the people who adopted the portal and use it heavily. It doesn’t represent the patients who couldn’t, didn’t know how, or didn’t have consistent access to the technology required.
This has two implications for health systems and for healthtech companies.
- The patients generating high portal message volume do so because they are engaged, connected, and have the health literacy to navigate a digital communication tool.
The system is struggling to keep up with its best-positioned patients. What that suggests about its capacity for patients who are harder to reach is not a footnote; it’s the next research question. - Digital divide data is often treated as an equity metric and filed accordingly. It should also function as a product design signal.
If rural, high-SVI, and age-extreme patients are messaging at lower rates, the question for every patient engagement company is whether the platform was designed to meet those patients, or designed assuming they’d eventually behave like the patients already generating the 153%.
What Healthtech SaaS Companies Should Do Now
We see the structural failure that healthtech SaaS companies are not responsible for creating, and not positioned to fully resolve on their own. That framing matters, because the alternative framing (“here’s a market opportunity”) misreads the room and misses the actual implication.
Here’s how to implement this:
- Audit what your product is adding to the inbox. If your platform generates portal messages, automated check-ins, or outreach communications, the question is whether you’ve designed those touchpoints to route, triage, or reduce load, or simply add to it. Adding features to an already-overwhelmed communication layer without an education and triage design is not a neutral product decision.
- Design for the message that shouldn’t need to be sent. Most portal messages exist because a patient has a question the onboarding process didn’t answer, the discharge summary didn’t explain, or the care plan didn’t anticipate.
Patient and caregiver education content structured in plain-language and delivered before the question gets sent is a workflow intervention, not a nice-to-have. Every question answered upstream is a message the clinician doesn’t have to read at 9 pm. - Name the clinician’s experience in your product and content strategy. The physicians who are 63% overwhelmed are also the ones your platform depends on for adoption.
If your thought leadership, product marketing, and sales conversations don’t account for what your product looks like from inside an already-saturated inbox, that’s a gap in how you’re positioning to buyers who experience that saturation daily. - Take the digital divide data seriously as a design constraint, not a footnote. The patients generating the 153% were already best-positioned to adopt. The patients your platform may underserve aren’t in that number.
Rural access, high SVI, and age-extreme populations require intentional design decisions like plain language, alternative access pathways, caregiver-inclusive communication.
The Part That Doesn’t Make Headlines
The 42 million patients who sent a message to a clinician in Q1 2025 alone trusted that someone would read it. Most of them were right — the messages are getting read. What clinicians are absorbing to make that happen, without a workflow or compensation structure built for that volume, is the part the headline skips.
The access problem got solved. That was the right thing to do, and the utilization data confirms patients wanted it. The workflow problem — what happens on the other end of the send button — is still open.
Healthtech SaaS companies build products that live in that space. The JAMA study is the clearest structural case to date for why the education and workflow layer inside those products isn’t optional infrastructure.
It’s the only part of the system that reduces the inbox before it fills.
Daree Allen Nieves is a B2B ghostwriter and content strategist for Series A/B healthtech companies. She writes patient and caregiver education content, SaaS onboarding sequences, and thought leadership for executives making the argument publicly. Learn more at reewrites.com.
Sources
Long, J.J., McAdams-DeMarco, M.A., Schwartz, M.D., et al. Trends in Patient Portal Messages, Office Visits, and Telephone Encounters. JAMA. Published online June 22, 2026. doi:10.1001/jama.2026.8690
Shaw, M.L. Patient Portal Messages Outpace Office Visits. American Journal of Managed Care (AJMC), June 22, 2026.
“Epic Portal Messages Up 153% Since 2020, Study Finds.” Becker’s Health IT, June 2026.
Athenahealth. 2026 Physician Sentiment Survey. Athenahealth, 2026.
Sendak, Mark, MD. Remarks on clinical AI workflow. HIMSS 2026.
Chronic Care Management Improvement Act. Introduced April 14, 2026, Reps. Suzan DelBene (D-WA) and Mike Kelly (R-PA). Utilization data sourced from ASPE HHS 2022 report, linked in congressional press release.
SeamlessMD. Internal platform data on after-hours patient question volume. Published May 2026.
“Providers Back New Bipartisan Bill Eliminating Medicare Chronic Care Management Cost-Sharing.” Fierce Healthcare, April 15, 2026.
FAQ
Does portal message volume going up mean patients are more engaged — or just more anxious?
Both, probably, and the distinction matters. A well-designed patient education layer reduces the volume of anxiety-driven messages — the “is this normal?” and “what does this symptom mean?” questions that arrive at 10pm — while leaving intact the messages that actually require clinical response. The 153% increase doesn’t distinguish between those two categories. What it confirms is that the current system isn’t designed to sort them efficiently either.
Why aren’t portal messages compensated?
Clinicians are generally compensated for documented, billable encounters — office visits, telehealth appointments, specific procedure codes. Portal messages don’t fit cleanly into that structure. CMS has explored asynchronous communication reimbursement, and some CPT codes exist for e-visits and online digital evaluation, but uptake is inconsistent and most portal message volume falls outside what gets captured.
So clinicians are doing billable-equivalent work — triaging symptoms, adjusting medications, fielding post-discharge questions — through a channel the payment system still doesn’t consistently recognize. The Chronic Care Management program is a useful parallel: it’s been a separately billable Medicare service since 2015, and only 4% of eligible patients currently use it, partly because the communication infrastructure to support consent and coordination was never built around it. Portal messaging is following the same pattern at a faster pace.
What does the digital divide data mean for healthtech product teams?
The patients generating the 153% increase were already the best-positioned to adopt — they had broadband access, health literacy, and enough familiarity with the portal to use it repeatedly. The JAMA study found that rural populations, patients with high social vulnerability index scores, men, and age-extreme patients messaged at significantly lower rates.
For a product team, that split is worth reading as a design signal rather than a demographic footnote. If your platform’s usage data shows the same skew — high engagement from already-engaged patients, low engagement from the populations your health system partner is most accountable for — the question is whether the product was designed for the people using it or for the people who need it. Those aren’t always the same group. Equity in patient engagement isn’t only a values question. It’s a coverage question, and eventually a contract question, for any platform operating inside value-based care arrangements.
Should healthtech SaaS companies be worried about adding to portal fatigue?
Yes, with specificity. Not all additions to the communication layer are equal. A platform that generates automated check-ins without a triage or routing function is adding volume without reducing load. A platform that answers common post-discharge questions before they become portal messages is doing the opposite — intercepting demand before it hits the inbox.
The distinction matters because “we integrate with the portal” and “we reduce what goes into the portal” describe two very different product decisions. In a system where 63% of physicians already report feeling overwhelmed by portal message volume (Athenahealth 2026 Physician Sentiment Survey), the relevant question for any new communication feature isn’t whether it works — it’s whether it adds to the pile or reduces it.