Logged Out: Why Patient Portals Must Be Designed for the Patients Most Likely to Be Left Behind
Photo by Photo by Georg Pflueger on Unsplash on Unsplash
Patient portals were sold as a simple idea: put the record in the patient’s hands. Give people a way to see test results, request refills, schedule appointments, and message a clinician without waiting on hold. In many health systems, the portal has delivered on that promise—for patients who have the devices, broadband, language skills, time, and confidence to use it. The problem is that a portal designed for the most capable users can become a barrier for everyone else.
Convenience for Some, a Gate for Others
A patient portal assumes a stable email address, a smartphone or computer, reliable internet, a password that can be remembered or reset, English literacy, and the time to navigate a website. Those assumptions are not neutral. They map onto income, age, disability, language, and housing. A patient who moves frequently may lose access after a phone number changes. A patient with limited English proficiency may face a portal that is English-only. An older adult may need a caregiver to log in, but proxy access can be cumbersome or unavailable.
When the portal works, it is genuinely convenient. When it does not, the same patient may be pushed toward overburdened phone lines or told to use a tool that was never designed with them in mind.
The Design Choices That Exclude
Exclusion is rarely the result of a single dramatic failure. It is built from small design choices: complex login requirements, CAPTCHAs that are difficult with low vision, session timeouts that punish slow typists, PDFs that are not mobile-friendly, medical jargon that requires a translator, and interfaces that assume constant high-speed internet. A portal may offer translation for the login page but not for the lab results. It may offer proxy access for parents but not for adult children caring for aging parents.
These are not merely technical defects. They are decisions about who the system expects its users to be. A portal that works only for the digitally confident is not a universal convenience. It is a new front desk with a narrower door.
When the Portal Becomes a Gatekeeper
Portals increasingly control tasks that patients cannot easily complete elsewhere. Appointment reminders, test results, medication refill requests, billing questions, and secure messages may be portal-first or portal-only. If phone lines are understaffed because the health system assumes patients will go online, those who cannot use the portal wait longer. A delayed refill request can become a medication gap. A missed result notification can delay follow-up. A scheduling barrier can turn into a no-show.
Medication safety is a particular concern. A patient who cannot request a refill through the portal may ration pills, skip doses, or arrive at the pharmacy without a valid prescription. A caregiver who cannot access the portal may not see a changed dose. These are equity problems disguised as IT problems.
What Usable Looks Like
A usable portal is not simply a portal with larger fonts. It is part of a multichannel system that includes phone, text, paper, and in-person support. Plain language should be standard, not an optional setting. Translation should cover the tasks patients actually perform, not just the welcome screen. Low-bandwidth and mobile-first design should be tested on the devices patients actually own. Caregiver proxy access should be straightforward and clearly explained. Accessibility should meet recognized standards, and usability testing should include older adults, people with disabilities, and patients with limited English proficiency.
Digital navigators—staff who can help patients set up accounts and complete tasks—can make a difference, but they are not a substitute for simpler design. The goal is not to force every patient online. The goal is to make the portal one reliable option among several.
The Equity Test
Health systems should measure portal success by task completion, not activation alone. They should stratify use by age, language, disability, broadband access, and race and ethnicity where appropriate. If one group is consistently less able to schedule, refill, or view results, the portal is failing that group. Publishing those metrics would make the problem visible and create accountability.
Patient portals can be a real advance. But an advance that widens disparities is not progress. The test is not whether the portal exists. It is whether the patients who need it most can actually get through the door.