TP485: Digital Equity Is Health Equity

Touch Point Podcast
Touch Point Podcast
TP485: Digital Equity Is Health Equity
Loading
/

The Industry Has Stopped Pretending Otherwise.

In 2025, U.S. digital health startups raised $14.2 billion. AI-enabled companies captured 54% of it. The trade press is calling 2026 the year AI moves from scattered pilots to enterprise-scale deployment. Every prediction in every roundup carries one quiet assumption underneath it. The patient on the receiving end can use what’s being built.

The Pew data from January says something different. The income gap on home broadband is 40 points and widening. About one-third of low-income households are smartphone-dependent, on a phone with no broadband, hitting patient portals built desktop-first. 19 million Americans 65 and older have no wireline home internet. 29% of U.S. hospitals offer their patient portal login in English only, even in counties chosen specifically because they have hundreds of thousands of limited-English-proficient residents. The Affordable Connectivity Program expired in June 2024. The Digital Equity Act, $2.75 billion for state digital inclusion programs, was canceled in May 2025.

Two trajectories. One looks like progress in aggregate. The other looks like the patients with the worst health outcomes being structurally locked out of the system that’s being built. Chris Boyer and Reed Smith examine what happens when digital strategy and health equity stop being parallel tracks and become the same problem.

  • Why the 2026 AI investment narrative quietly assumes a digitally capable patient, and what the population data actually shows
  • The smartphone-dependent patient most health systems haven’t internalized, and why portal UX fails them by design
  • Why disparities in patient portal access are widening for low-income, less-educated and 65-plus populations, even as overall use rises
  • What the 2025 cancellation of federal digital equity funding means for health systems whose patient panels actually need the work done
  • Modality mix as the reframe: digital, phone, in-person and printed channels as a portfolio allocated by segment, not a hierarchy everyone migrates toward

The University of Michigan study published in JAMA Network Open in October is the one to anchor on. Researchers looked at 511 hospitals in 51 counties in 17 states where census data showed at least 300,000 LEP residents. 29% of those hospitals offered the patient portal login in English only. 60% offered English plus Spanish. 11% offered three or more languages. In counties specifically chosen because they have hundreds of thousands of patients who don’t speak English at home.

If your most-invested-in digital experience reaches the patients who already had the most options, and barely touches the patients with the worst outcomes, what is your digital strategy actually optimizing for?

Mentions from the Show:

Read the Argument

We talked through the tension on air. Here it is on the page, with the data tightened up and the case made in full.

In 2025, U.S. digital health startups raised $14.2 billion. AI-enabled companies captured 54% of it. Q1 2026 alone saw $7.4 billion in funding, with 60% of the capital concentrated in 19 mega-rounds averaging $108 million apiece. The trade press is calling 2026 the year AI moves from scattered pilots to enterprise-scale deployment. Read any prediction roundup. The language is unanimous. Not “if AI.” Not “when AI.” The conversation has moved to “AI at scale.”

Underneath every prediction sits one quiet assumption. The patient on the receiving end can use what’s being built.

The Pew data from January says something different. The income gap on home broadband adoption is 40 points, and widening. About one-third of low-income households are smartphone-dependent, doing healthcare on a phone with no broadband, hitting patient portals built desktop-first. 19 million Americans 65 and older have no wireline broadband at home. 29% of U.S. hospitals offer their patient portal login in English only, in counties specifically chosen because they have hundreds of thousands of limited-English-proficient residents. The Affordable Connectivity Program expired in June 2024. The Digital Equity Act, $2.75 billion for state digital inclusion programs, was canceled in May 2025.

Two trajectories. One looks like progress in aggregate. The other looks like the patients with the worst health outcomes being structurally locked out of the system that’s being built. The honest question for every health system board in 2026 is not whether AI works. It is who exactly the work is for, and what happens to everyone the work doesn’t reach.

The Investment Trajectory

The 2026 industry consensus is unambiguous. Healthcare captures nearly half of all vertical AI spending. Healthcare AI investment tripled from $450 million in 2024 to a projected $1.5 billion in 2025. The narrative across consultancies, investors and health system leadership is consistent. Ambient documentation, agentic orchestration, AI-driven scheduling, EHR-native AI. All framed as enterprise infrastructure rather than experiments. Cleveland Clinic gets cited as the proof case, $18.3 billion in revenue, twice the growth rate of peers, attributed to corporate-discipline AI scaling.

This is not a hype cycle. It is a capital allocation pattern. And capital allocation patterns at this scale are difficult to redirect.

The patient implicit in every pitch deck has broadband. Owns a smartphone they’re confident using. Reads a login screen in English. Completes multi-step authentication on a five-inch screen without abandoning the flow. Trusts an automated message from an unfamiliar number. None of those assumptions is named in the strategy. Because if they were named, the strategy would have to account for the patients for whom none of them are true.

Banking ran this play between 2010 and 2020. Mobile-first as the default channel. Branch closures. Digital onboarding as the assumed pathway. The industry then spent years quietly rebuilding branch and phone infrastructure for the segments the digital strategy excluded. Healthcare is repeating the arc with higher stakes and a much smaller margin for error. Skipped clinical care is not deferred deposits.

The Population Reality

Income is the largest and most consistent divide. The gap is widening rather than closing. 54% of households earning under $30,000 subscribe to home broadband, compared to 94% of households earning $100,000 or more. The under-$30K rate dropped from 57% in 2023 and 2024 to 54% in 2025. The 40-point gap has held for years.

The smartphone-dependent category is the one most health systems have not internalized. About one-third of low-income households are smartphone-dependent. Compared to 4% at the top. These are the patients most likely to be on Medicaid, most likely to have unmanaged chronic conditions, most likely to be the target of “high-utilizer” outreach programs. Patient portals were built desktop-first, with mobile as an afterthought. The patients least likely to have broadband are hitting an experience that’s hostile to small-screen use.

Older adults are the largest disconnected population by absolute numbers. 19 million Americans 65 and older with no wireline broadband at home. Older seniors, 75 and up, are less connected than the 65-to-74 cohort. This is the population health systems are simultaneously trying to engage on Medicare Advantage retention, remote patient monitoring, post-acute navigation and chronic condition programs.

Race and ethnicity disparities persist after controlling for income and age, particularly in portal access and use. Black and Hispanic adults are offered patient portals at lower rates and access them at lower rates when offered. Asian and Hispanic adults face the additional language access barrier. 25.7 million U.S. adults have limited English proficiency.

The University of Michigan study published in JAMA Network Open in October is the one to land on. Researchers looked at 511 hospitals in 51 counties in 17 states where census data showed at least 300,000 limited-English-proficient residents. 29% of those hospitals offered the patient portal login in English only. 60% offered English plus Spanish. 11% offered three or more languages. In counties specifically chosen because they have hundreds of thousands of patients who do not speak English at home. The digital “front door” most health systems are pouring investment into is locked from the outside for one in three of the patients they’re meant to serve.

What Changed in 2025

This is the variable that makes 2026 different from 2021. The federal scaffolding that used to absorb this problem has been dismantled.

The Affordable Connectivity Program expired in June 2024. The $30-a-month broadband subsidy that 23 million households relied on, gone, with no replacement. The Digital Equity Act, $2.75 billion for state digital inclusion programs, was canceled in May 2025 and is currently in litigation. The Broadband Equity, Access and Deployment program was restructured in 2025 toward technology neutrality rather than fiber-first, with the affordability and digital adoption components weakened. Section 1557 of the Affordable Care Act still requires meaningful language access for federally funded healthcare. The connective tissue around enforcement is weaker than it was.

The framing this raises is structural, not partisan. The work the federal programs were doing did not disappear because the funding did. It got reassigned. To states with less coordination. To nonprofits and libraries running on smaller budgets. And, increasingly, to the health systems whose patient panels actually need the work done.

A health system serving a heavily-Medicaid panel does not have the option to lobby its way out of this. The patients are still on the panel. The function the federal programs were absorbing has been dropped. Either the health system absorbs it, or it goes nowhere. There is not a third option that involves the work continuing without anyone doing it.

The Modality-Mix Reframe

Most health systems treat the digitally capable patient as the strategic target and treat anyone unable to engage digitally as the exception. For systems with significant Medicaid, dual-eligible, rural or LEP populations, the math is closer to the opposite. Digital strategy is a patient-segmentation problem, not a technology-deployment problem.

Johns Hopkins Bloomberg School published a useful frame in early 2025. Digital determinants of health. Broadband, devices, digital literacy and culturally appropriate user experience get treated as social determinants alongside housing, food and transportation. Once you accept that the infrastructure a patient uses to reach care is part of whether the patient reaches care at all, the strategic implication shifts. You don’t optimize for the median patient and call it equity. You design for the panel you actually serve.

A 2025 JMIR review on rural digital health proposed three pillars. Infrastructure, affordability, adoption. A health system that solves only one has solved nothing. Building a beautiful patient portal solves nothing if the patient does not have broadband. Subsidizing a hotspot solves nothing if the patient has not been trained to use it. Training the patient solves nothing if the device is unusable. All three have to be present for any of them to matter.

The actionable reframe is modality mix. Treat digital, phone, in-person and printed channels as a portfolio that gets allocated across patient segments rather than a hierarchy everyone is meant to migrate toward. The segment that engages best by SMS gets reached by SMS. The segment that needs a phone call gets a phone call. The segment that responds to a printed bill insert gets the bill insert. None of these is second-best. Each is appropriate to the segment.

The intervention with the strongest evidence base is also the cheapest. Provider encouragement is the largest single predictor of portal adoption among populations otherwise less likely to enroll. Not technology improvement. Not better UX. The conversation in the exam room. That finding has been replicated in HINTS data across cycles going back years. Most systems are not operationalizing it.

Where the Work Is Already Being Done

The work is being done. It is being done by entities that are not health systems. Schools providing hotspots. Public libraries running digital navigator programs. Electric cooperatives building rural fiber. School-based telehealth networks. Community health workers picking up the digital onboarding function from inside Federally Qualified Health Centers.

Some of this is starting to be formalized. Rural health systems and safety-net hospitals are increasingly providing digital navigators, hotspots and devices as part of care delivery. The function is shifting from optional outreach to required infrastructure. The job titles are still inconsistent. Patient access coordinator. Community health worker. Digital navigator. Different names for the same function. Closing the gap between the digital tool and the patient who cannot use it.

Section 1557 still applies. Enforcement is uneven, but the legal exposure for systems whose digital-only pathways exclude LEP patients is real. The compliance question is not whether language access matters. It is whether the digital implementation of access matters as much as the in-person version. The trajectory of regulatory attention suggests it does.

The Question on the Table

For 15 years, healthcare has talked about health equity as a moral commitment and digital transformation as an operational priority. They have been worked on by different teams, measured against different metrics, funded out of different budgets. In 2026, that separation stops being defensible. The infrastructure the patient uses to reach care is part of whether the patient reaches care at all. Once you accept that, digital equity stops being a parallel track and becomes a clinical operations question.

The strategic question is not abstract. If your most-invested-in digital experience reaches the patients who already had the most options, and barely touches the patients with the worst outcomes, what is your digital strategy actually optimizing for? The federal infrastructure that used to absorb this question is gone. The patients are still there. The function the federal programs were doing has been reassigned by default. To the institutions whose panels need the work done.

The strategy decision has been made. Most boards just have not named it yet. The only question left is whether health system leadership names it deliberately, or accepts the version that happens by default.

This essay accompanies Touch Point episode 485, “Digital Equity Is Health Equity,” with Chris Boyer and Reed Smith. Listen to the full conversation at touchpoint.health.

Find Us Online

Reed Smith

VP @jarrardinc, Advisor @MayoClinic Social Media Network and @SXSW Health & MedTech. @SocialHealthIns and @tpmnetwork founder

Chris Boyer

Principal, @chrisboyer LLC  Using digital marketing, experience and UCD/CX to transform healthcare. Co-host of the touch point podcast and founder of the Touch Point Media network. 

, , , , , , , , , , , , , , , , , , ,

Sign Up to become a touchpoint insider

Sponsors

Become a Sponsor

Subscribe

Subscribe today to stay up to date with Touch Point episodes, special events, and promotions. We promise to never spam you or sell your information.