How much of healthcare can we really afford to digitize?
That’s the question worth sitting with before any list of trends. Healthcare is not just another industry waiting to be made more efficient – it’s deeply human-sensitive.
Great quality care depends not only on knowledge, technology, and professional expertise, but on things that are much harder to measure or automate: human connection, trust, attention, judgment, and the sense that someone is actually there when it matters.
That makes the current healthcare trend of digitalization a little more complicated than a story about faster systems and smarter tools. AI models get better by the month. Devices get smaller and more continuous. Data moves faster and across more systems. Regulations chase capabilities instead of the other way around.
Every one of these shifts opens a door and, quietly, narrows another: more data means more insight and more exposure; more automation means more capacity and less friction, but potentially fewer human checks along the way; more connection between systems means better-coordinated care and a wider net for anyone trying to compromise it.
So this isn’t a “here’s what’s new” piece. It’s closer to: what happens to healthcare when more of its work, information, and relationships move into digital systems? Which changes actually improve the day-to-day experience of giving and receiving care? Which ones just move the friction somewhere else? Where can technology support professional judgement without trying to replace it – and where does human connection stay something no software should be designed around?
AI has gone mainstream in clinical work (81% of physicians now use it), and ambient documentation is the change doctors actually feel – less typing, more eye contact, real burnout reduction.
Underneath that sits an interoperability layer (FHIR) that’s finally connecting fragmented records. Care is spreading into the home via telehealth platforms and remote monitoring, though the plumbing for that data often still runs on phone calls and email.
Prior authorization remains healthcare’s most hated process, and automation is chipping into it without removing the hard cases. Patients have more portal access than ever, but coherence hasn’t caught up with access. And every one of these gains widens that attack surface – healthcare remains the most expensive sector for data breaches.
None of it resolves into a clean verdict. Read on for the “why”.
The clinical intelligence layer: AI tools and ambient documentation
AI didn’t arrive as a replacement for judgement – it slipped into the seam of existing work.
The adoption curve here is genuinely striking. In the American Medical Association’s 2026 physician survey, 81% of physicians reported using AI in their practices, more than double the 38% reported in the AMA’s first survey in 2023.
The uses are mostly less cinematic than the headlines suggest: summarizing medical research, documenting visits, creating discharge instructions, supporting clinical decisions, and handling other administrative work. Healthcare AI isn’t arriving as one giant replacement for clinical expertise. It is slipping into individual parts of existing workflows, one task at a time.
The part clinicians actually feel day to day, though, is ambient documentation. Instead of typing or dictating notes, a growing number of clinicians can talk with their patients while an ambient tool listens, transcribes, and drafts the chart. The results, where studied, are hard to wave away:
- A 2025 JAMA Network Open study of 46 clinicians found that ambient scribing was associated with 20.4% less time spent on notes per appointment and 30% less after-hours work time.
- Another multicenter study of 263 clinicians found that 30 days of ambient-scribe use was associated with lower reported burnout, lower cognitive task load, more undivided attention for patients, and less time spent documenting after hours.
The appeal is almost disarmingly simple: the clinician can look at the patient instead of the screen.
The catch is that these systems still require human review, and clinicians know it. The AMA’s 2026 survey found that 88% of physicians had at least some concern about AI-related skill loss, while 86% identified data privacy as a top concern and 88% emphasized the need for strong safety and efficacy validation.
So the more interesting question isn’t whether AI can produce a note – it is what happens to clinical work when producing that note no longer requires the clinician to do it personally. Some of the answer is clearly positive: less clerical work, more attention, fewer hours spent finishing charts after the day is supposed to be over. The longer-term question is whether the parts of clinical judgment being assisted by AI remain actively exercised by the people responsible for the final decision. That tension will follow almost every healthcare technology that comes next.
The plumbing: FHIR, APIs, and health information exchange
This is the least glamorous trend on the list, and arguably the one everything else depends on.
Healthcare has accumulated an enormous number of digital systems, but digitized information is not automatically connected information. A patient can have an EHR at one organization, imaging somewhere else, laboratory results somewhere else again, and a patient-generated data stream coming from a device at home. The fact that each piece exists digitally doesn’t mean the pieces can easily understand one another.
FHIR, or Fast Healthcare Interoperability Resources, is increasingly part of the answer. HL7’s 2025 global State of FHIR survey (82 experts, 52 countries) found:
- 73% said FHIR was mandated or formally advised in their environment
- 78% reported regulation around electronic health-data exchange
- Most respondents said their countries had a core FHIR implementation guide already available or under development
That sounds technical because it is technical. But the consequence is remarkably human. This is the layer that allows an ambient scribe’s note, a lab result, a wearable’s data stream, or a referral to move between systems without someone manually re-typing it into a different one.
It is also why interoperability has stopped being purely an IT concern. ONC’s current SAFER guidance devotes a section specifically to clinician communication – electronic exchange among clinicians, care teams, and patients, particularly during handoffs like discharge and referral. The infrastructure conversation, in other words, keeps circling back to something much less technical: making sure information actually reaches the people who need to act on it.
Which is where the promise and the risk start to resemble each other. Better-connected systems make care easier to coordinate. They also create more connections that need to work, be governed, and be protected – every link is one more thing that can fail or be breached.
Care beyond the building: remote monitoring, telehealth, and home-based care
The care relationship has stopped being defined by the walls of a clinic, but the plumbing hasn’t fully caught up.
Telehealth, remote patient monitoring, connected devices, and home-based care have pushed more of the care relationship into places where the patient actually lives. This has moved well past its pandemic-era moment: 85% of physicians surveyed by the AMA reported currently using telehealth and remote patient monitoring in increasingly used with patients in their own homes.
The technology is particularly relevant to chronic disease, where something can change substantially between two appointments. A blood-pressure reading, weight measurement, oxygen level, glucose reading, or symptom report collected at home can potentially give a care team information that would otherwise not exist until the next visit.
The attraction is obvious. Care can become more continuous without requiring the patient to physically appear in a clinic every time.
But there is another side to that equation. In the AMA survey, only 8% of telehealth-using respondents said they were using RPM technologies with patients in their homes, and among those who were, 76% said the data was usually shared manually, such as by phone or email.
That’s a useful reminder that the healthcare trend of digitalization doesn’t automatically eliminate operational work. Sometimes it creates a new stream of information that someone now has to watch, interpret, route, and respond to.
The same is true at a larger scale. The more care moves into the home, the more people and systems sit between the patient and the person ultimately responsible for their care. Remote care may reduce the distance between a patient and the healthcare system while increasing the number of connections required to keep that system coordinated.
That is not necessarily a flaw in digital care. It is simply a new problem to solve.
The administrative layer: scheduling, referrals, and prior authorization
The process patients rarely see, and the one clinicians have been complaining about for years – with a good reason.
Prior authorization is the clearest example of an administrative process becoming complicated enough to become a clinical problem. The AMA’s latest survey found:
- Physicians complete an average of 40 prior authorizations per week, consuming about 13 hours of physician and staff time
- 94% say prior authorization contributes to burnout
- 95% say it delays access to necessary care
- 26% report it has led to a serious adverse event for a patient in their care
This is exactly the sort of problem that makes healthcare software attractive to vendors that aren’t building clinical tools at all.
Automation can extract information from the medical record, organize it against payer requirements, route requests, track their status, and reduce some of the repetitive work involved in getting a decision.
And health systems are increasingly looking at AI this way. Deloitte’s 2026 global healthcare outlook found that 64% of surveyed health-system executives expected AI to reduce costs by standardizing and automating workflows. Fifty-five percent saw potential in predictive analytics for workforce optimization, while 49% expected value from technology-enabled patient engagement and remote monitoring.
The upside is real: fewer repetitive tasks, less time spent chasing paperwork, fewer delays between “the doctor ordered this” and “the patient can get it.” But automation doesn’t make the underlying judgment problem disappear. It moves part of it into software. The straightforward cases may be easier to process. The difficult cases are still difficult precisely because they don’t fit neatly into a rule. And those are the cases where someone needs to understand the person behind the data rather than simply match the data to a criterion.
The patient-facing layer: portals, booking, navigation, and digital therapeutics
Patients are more connected to the system than ever. Whether the system is connected around the patient is a separate question.
The patient-facing side of digital healthcare has quietly outgrown the old idea of a “portal.” In 2024, 65% of people in the U.S. reported accessing their online medical records or patient portal, according to ONC data. Among people managing a chronic condition, the figure was 67%. More than half of people also had multiple online medical records or portals, while only 7% used an app designed to organize information from different portals into one place.
That last number says a lot. Healthcare has made patient information more accessible, but accessibility is not the same thing as coherence.
A patient may now be able to see a test result, read a clinical note, message a provider, schedule an appointment, upload information, and connect a health app. The problem is that these experiences can still exist across multiple systems, organizations, logins, and interfaces.
The same pattern appears on the provider side. By 2024, 93% of U.S. hospitals reported offering secure messaging with providers, 96% allowed patients to view clinical notes, and 87% supported API-driven app access to medical information. The patient is increasingly wired into the system. Whether the system is wired around the patient is the harder, still-open question.
Digital therapeutics introduce another version of the same challenge. The technology has moved beyond the idea that an app is automatically innovative simply because it sits on a phone. The FDA’s recent discussion of digital therapeutics points to persistent problems around evidence, regulation, reimbursement, and helping patients and providers distinguish medical-grade interventions from general wellness products.
In other words, putting care into software does not remove the need to earn trust. It may make that requirement more important.
The lock on the door: cybersecurity and access controls
The trend that makes every other trend on this list conditional
Healthcare has been one of the most expensive sectors for data breaches for years. IBM’s 2024 Cost of a Data Breach research put the average cost of a healthcare breach at $10.93 million, the highest of any industry in that study.
But the financial cost is only part of the story. A healthcare breach isn’t losing a list of customer email addresses. Medical information is deeply personal, often permanent, and difficult to replace. A patient cannot simply cancel their medical history and request a new one.
And the risk increasingly extends beyond the hospital itself. In 2025, the U.S. Department of Health and Human Services settled a HIPAA ransomware investigation involving Comstar, a healthcare business associate whose breach affected 585,621 individuals. HHS has also continued to pursue ransomware and risk-analysis cases involving business associates, reinforcing the fact that healthcare’s security perimeter now extends well beyond the walls of the provider organization.
That matters precisely because interoperability and digital care depend on the outside connections. The same link that makes care easier to coordinate can become another place sensitive information leaks out. Which is why access controls, identity management, and vendor risk assessment aren’t merely back-office IT concerns anymore. They are part of the quality of care. A ransomware incident that takes a scheduling system offline, interrupts access to records, or delays a procedure is not just a cybersecurity problem. It becomes a patient-care story.
Is this good for the quality of care?
Mostly, yes – the evidence points toward real benefits, but the asterisks matter.
Freeing clinicians from some clerical work can give them more attention for the patient in front of them. The evidence around ambient documentation is increasingly encouraging, with studies associating these tools with lower documentation burden, less after-hours work, and lower reported burnout.
Continuous monitoring can give care teams information between appointments rather than only at them. Patient portals can make records, notes, and medical communication available without requiring a phone call or a trip to the clinic. Automation trims some of the friction that currently slows treatment down. None of that is insignificant.
But every gain here comes bundled with a corresponding demand on human judgment. Someone still needs to review an AI-drafted note. Someone needs to decide whether an automated result makes sense in context. Someone needs to notice when a patient’s situation doesn’t fit the pattern. Someone needs to determine whether a message is routine or actually urgent.
The more automated a system gets, the easier it becomes to mistake less human effort for less human responsibility – and those aren’t the same thing.
The AMA’s 2026 physician survey captures that tension unusually well. Physicians broadly see value in AI, but 88% reported at least some concern about AI-related skill loss, and 85% wanted to be consulted or directly involved in decisions about AI adoption. Convenience and erosion can be the same trend, viewed at different timescales.
What this actually changes day to day
For clinicians, the picture is a mixture of relief and new vigilance. Where these tools are implemented well, some of the administrative burden really does start to ease. A clinician can spend less time documenting and more time looking at the patient. A staff member can spend less time copying information between systems. A referral can move forward without someone having to call three different people to find out where it went.
But removing one task doesn’t necessarily mean removing the work around it. The clinician still has to check an AI-generated note, the staff member still has to monitor an automated workflow, someone still has to correct bad data or decide whether an alert actually matters. In some cases, the work simply changes from doing the task manually to supervising the system that now does it.
The same shift is happening inside the clinical relationship. Early research on ambient documentation is encouraging, with clinicians reporting greater engagement with patients when they aren’t focused on typing. But that benefit depends on the technology staying in the background and remaining understandable to the people using it. If a patient doesn’t know what is being recorded, where the information goes, or what the system is doing with the conversation, the technology can just as easily introduce distance where it was supposed to remove it.
For patients, digitalization is creating a different kind of continuity. Chronic conditions can be monitored between visits, records can be accessed without returning to the clinic, and questions can be sent without spending twenty minutes on hold. Much of this genuinely improves access and responsiveness, particularly for people who need ongoing care rather than a single appointment.
At the same time, that convenience comes with a growing trail of personal information moving through systems, organizations, devices, and vendors that most patients will never see. A patient may experience healthcare as a simple interaction with an app or portal, while behind it sits a much more complicated chain of data exchanges and permissions.
Trust in digital healthcare therefore isn’t only about trusting the clinician or the hospital anymore. It increasingly depends on trusting the infrastructure that connects them, even when most people never see how that infrastructure works.
The throughline
None of this resolves into a clean verdict, and it shouldn’t. Digital health in 2026 is making some parts of care more attentive while making the underlying system more complex. It can give clinicians more time with patients while creating new oversight work. It can make information easier to access while giving organizations more information to protect. It can automate tedious processes while making human supervision more important in the cases that don’t fit neatly into a rule.
The direction of the market reflects that complexity. KLAS’s 2026 global health IT research identifies AI as a major investment priority, but places it alongside EHR and digitalization, infrastructure, cybersecurity, and accelerating cloud adoption. Healthcare isn’t replacing one old technology with one new technology. It is adding layer after layer to an already interconnected system.
And those layers depend on one another. AI needs data. Data needs to move between systems. Systems need to connect the people responsible for acting on that information. Those people still have to make decisions, communicate them, and understand the context behind them. Patients still have to trust the people and systems handling their care.
That may be the more useful way to look at the healthcare trend of digitalization. Not as a race toward a point where technology does more and people do less, but as a gradual redesign of how people, information, and technology work together.
The challenge is making sure that every new efficiency and every new data flow is evaluated together. A faster process isn’t necessarily a better one if it makes the important information harder to understand. A more connected system isn’t necessarily a safer one if nobody knows who is responsible for acting on what it produces.
Healthcare can become much more digital without becoming less human. But that part won’t happen automatically.
