Author:
By the time you meet with a patient to review a lab result, it’s possible they’ve already read it, searched it, and decided what to do next. In 2024, 65% of U.S. adults accessed a patient portal or online record, up from 25% in 2014. Among patients with results in a portal, 58% viewed them before hearing from a provider. Now with the emergence of AI, between a quarter and a third of U.S. adults now use AI chatbots for health information.
Self-serve health care covers the tasks patients now start on their own: booking, refills, reading results, tracking wearable data, ordering home tests, and asking chatbots what it all means. Each task sits next to a clinical decision that still belongs to the care team. A patient can request a refill online, but a provider still decides whether reassessment comes first. For health systems and health tech teams, the design question is where self-service ends and that decision begins.
Convenience and access gaps drive adoption
Some patients choose self-service because it’s faster. In a February to March 2026 KFF poll, 65% of adults who used AI for health information cited getting quick answers as a major reason. Banking and retail set that expectation years ago, and phone queues now feel slow by comparison. Preferring an easier way to book doesn’t mean patients value clinical expertise any less, it just means there’s an appetite for a more efficient process.
Other patients self-serve because conventional care is hard to reach. In early 2025, new-patient waits averaged 31 days across six specialties in 15 large U.S. metro areas, the longest since the survey began in 2004. In a late-2025 Gallup survey, 32% of recent AI health users earning under $24,000 a year used AI because they couldn’t pay for a visit, compared with 2% of those earning $180,000 or more. Some had felt dismissed by a provider (21%), and others were too embarrassed to ask a person (18%).
The same patient can move between these groups. Someone who books online for convenience may turn to a chatbot when the next appointment is weeks away. A single adoption rate hides that difference, and a tool designed only for convenience can leave the access-driven patient with no clear next step.
Where self-service falls short
Access doesn’t guarantee understanding. Among patients who viewed results before talking to a provider, 66% said they understood them well or very well. That fell to 53% among patients with low digital literacy.
AI accuracy doesn’t transfer to the people using it. In a randomized study of 1,298 UK adults, large language models identified the relevant condition in 94.9% of written scenarios on their own. Participants using the same models did so in fewer than 34.5% of cases and chose the right level of care no more often than a control group using search engines and the NHS website.
Follow-up drops off. In the KFF poll, 42% of adults who used AI for physical health information didn’t follow up with a doctor or other provider. The poll can’t show whether skipping follow-up was appropriate, but a falsely reassuring answer can delay diagnosis.
Private-feeling isn’t protected. In the same poll, 77% of adults were concerned about the privacy of medical information shared with AI tools, yet 13% of all adults had uploaded their own medical information to one. HIPAA applies to covered entities and their business associates, and many consumer apps and wearables fall outside it.
Self-care works when a provider stays connected
The strongest outcome evidence comes from models that send patient data back to a provider. In the HOME BP trial, 622 UK adults with poorly controlled hypertension used home monitoring and digital self-management with feedback to both patients and providers. At one year, systolic pressure was 3.4 mmHg lower than with usual care. A meta-analysis of 65 trials found larger effects when telemonitoring or clinical staff support was built in.
Six things you can do to adopt a hybrid model
Providers can move routine logistics online while still maintaining oversight for interpreting results, answering clinical questions, and arranging follow-up.
Make administrative tasks self-serve by default
Booking, intake, payment, and refill requests can move online, with clinical authorization routed to a provider for review. Keep staffed phone and in-person options for patients with low digital literacy or limited internet access.
Put context next to every result
Federal information-blocking rules limit delays in releasing results, but each organization controls the context around them. A plain-language note on what the result is, any trend, and the planned next step targets the comprehension gap. Patients whose provider encouraged portal use understood their early results more often (67% vs. 54%).
Evaluate what patients bring in
In 2024, 73.4% of wearable users were willing to share device data with a provider. Confirm clinically important wearable readings with a validated measurement. For direct-to-consumer results, review the indication and the laboratory before acting on them. For AI output, ask what the patient entered and what the tool advised, then reconcile it with their history and exam results.
Set expectations for AI use and data sharing
Tell patients that many consumer apps fall outside HIPAA, so anything they upload follows the company’s privacy terms. Disclose when your organization uses AI in their care; 72% of U.S. adults say that disclosure is extremely or very important. Before deploying patient-facing AI, ask vendors for evidence from testing with real users, since model accuracy alone can’t predict user performance.
Build escalation into every self-service entry point
A portal result, home monitoring program, or patient-facing AI tool should show a named route to a provider, such as a nurse line or message pool, on the same screen.
Measure whether the patient’s need was resolved
Usage and click-through rates show that patients used a tool, not that they understood the next step or got follow-up. Track abandoned tasks, follow-up after abnormal results, repeat testing, and whether digital-only pathways leave some patients out.
Keeping providers in the picture is important
Patients will keep self-serving because it’s faster and because traditional care is often slow, costly, or hard to reach. Providers can give patients more control without leaving them to interpret findings, coordinate services, and arrange follow-up alone. As physicians, it’s your job to make sure the self-service side of your process handles the routine work and brings patients back when a result, symptom, or question needs real clinical judgment from a trusted provider.
Platforms like Fullscript support providers in maintaining this connection keeping you and your patients connected beyond the visit, while still providing the benefits of AI and automation. Patients can sync wearable data, upload lab results, and share how they’re feeling each month, and it all comes together in one view providers can review.
