The most effective feature in patient engagement software is not in the software.
In national survey data, 87 percent of individuals whose healthcare provider encouraged them to use the patient portal accessed it at least once in the past year. Among those who were not encouraged, 57 percent did. A thirty point swing, produced by a person saying a sentence in a consultation room.
Hold that number while we walk a cohort through the funnel, because almost every product decision below competes with it, and most of them lose.
Take 1,000 patients who have been offered access.
1,000 offered → the offer is nearly universal now
Access is no longer the constraint. As of 2024, 99 percent of US hospitals let patients view their medical records, 96 percent allow download, and 84 percent support transmitting them elsewhere.
So if your product thesis rests on giving patients access to their records, it is not a thesis. It is table stakes that federal rules already delivered. The interesting question moved downstream years ago.
1,000 → 610 activate
In a study of 250,345 adults aged 50 and over with chronic conditions, 61 percent activated their portal account.
Notice who that population is. Older adults managing chronic illness are the group with the most reason to engage and the most contact with the system. If four in ten of them never activate, a general population product should not be modelling anything better.
Activation is where the friction is stupidest and most fixable. Identity proofing, an access code that expires, a password policy written for an enterprise system, a flow that assumes a desktop. None of it is clinical. All of it is where the cohort thins.
610 → 540 open a session
In that same cohort, 54 percent logged at least one session. So roughly one in nine of the people who completed activation never came back to use the thing they had just signed up for.
That gap is the clearest signal in the whole funnel. These people wanted in. They completed a tedious registration. Then the product gave them no reason to return within whatever window their intent survived.
Usually that is a timing failure rather than a feature failure. Activation happens at discharge or after a visit, and the first genuinely useful moment, a result arriving, a bill, an appointment, comes days or weeks later with nothing bridging the gap.
540 → the number nobody publishes
Here the public data thins, and that is itself worth noticing.
Activation and first login are measured everywhere because they are reportable. Sustained use is measured far less consistently, which means most organisations running patient engagement software cannot tell you what happens after month one, and many quietly report activation as though it were engagement.
What is known is what brings people back. The dominant drivers of repeated access are viewing test results and reading clinical notes. Not messaging, not education content, not gamification. People return for information about themselves that they cannot get anywhere else.
The design implication. Build the funnel around results delivery and make everything else secondary. A patient app whose home screen leads with wellness content instead of the result the patient is waiting for has misread why anybody opened it.
Who falls out, and why it is a design problem
The drop off is not evenly distributed, and this is where most patient app development goes wrong by optimising for the people already using it.
Portal use is consistently lower among older adults, people who do not speak English as a first language, and patients in more deprived areas. A 2024 analysis of 499,098 patients across two London teaching hospital systems found activation highest in the 31 to 40 age band at around 71 percent and lowest among the very elderly at around 24 percent, with a clear socioeconomic gradient running through the middle. The pattern holds across health systems and countries, which tells you it is structural rather than local.
Two things follow for anyone building this.
Your analytics are biased. Usage data comes only from people who got through activation, so the product optimises itself toward the cohort with the fewest barriers and away from the patients with the highest clinical need. That is a measurement problem before it is an equity problem.
Language and reading level are features. Not localisation as a phase two item. The population that drops out disproportionately includes people for whom your default interface is genuinely harder to use, and this is squarely the remit of UI and UX design rather than a translation budget.
What organisations built instead
Here is the uncomfortable comparison. In 2024, 99 percent of hospitals had the foundational capability to let patients view records, but only 45 percent had advanced capabilities such as letting patients contribute their own data.
The foundational things were mandated. The advanced things were not. Investment followed regulation almost exactly, which is a reasonable description of how most patient engagement roadmaps get written.
Meanwhile inpatient access to apps rose from 68 percent in 2021 to 80 percent in 2024, having peaked at 83 percent in 2023, and outpatient access from 62 to 75 percent. Capability is broadly present. The gap is not what got built. It is that being buildable and being used are separate problems, and the second one has never had a deadline attached.
What actually moves the number
Four things, ordered by measured effect rather than by how interesting they are to build.
1. Clinical encouragement. The thirty point swing. It costs nothing and it belongs in the workflow, not the marketing plan. That means the product has to make it trivially easy for a clinician or front desk staff member to hand someone in, at the moment of the visit, in seconds. If enrolling a patient takes a staff member longer than a sentence, it will not happen.
2. Activation that survives a phone. Every step you remove from identity proofing and first login converts directly. This is unglamorous frontend development work and it has a better return than almost any feature.
3. A reason to return inside the intent window. Results arriving, and arriving fast, with a notification the patient actually receives. That depends on the integration underneath, which is why custom EMR and EHR development and the data plumbing behind it decide the engagement outcome more than the interface does.
4. Measuring past first login. You cannot fix a drop off you do not track. Cohort retention at 30, 90, and 180 days, segmented by age, language, and deprivation, is the instrumentation most patient engagement software is missing.
Where Woltrio fits
Woltrio builds patient facing products and the systems underneath them, which in this category are the same project.
Most patient app briefs that arrive are feature lists. The more useful first conversation is about the funnel: what activation currently converts at, what happens between activation and first session, and whether anybody is measuring month three. That usually reduces scope rather than expanding it, because the highest return work is almost always earlier in the funnel than the roadmap assumes.
Where the shape of the product is genuinely unsettled, a scoped MVP measured against activation and thirty day return will tell you more than a longer specification. And where staff workflow is the lever, as it usually is, workflow automation around enrolment is often worth more than anything shipped in the app itself.
Start with a scoped discovery from Woltrio.
