Designing eCOA Technology Patients Can Actually Use

By John Oncea, Chief Editor, Clinical Tech Leader

Tinatin Jashi helped me look at why clinical trial technology tends to break down not inside any one system, but at the handoffs between functions. One thread from that conversation deserved its own space: what happens when those upstream breakdowns finally reach the patient. That’s where eCOA lives, and it’s where Jashi’s background gives her a vantage point almost nobody else in this industry has.

Jashi has worked as a study coordinator, CRA, clinical project manager, medical writer, and regulatory affairs specialist. But it’s her experience as a cognitive debriefing specialist that sets this conversation apart. Cognitive debriefing, testing whether patients actually understand a questionnaire the way it was intended, is a niche skill set, and it gives Jashi a specific, evidence-based view of where patient-facing technology goes wrong: usually well before the patient ever opens the app.

Technology Usually Isn’t The Problem

Ask Jashi how much of the trouble with eCOA usability and data quality traces back to the tool itself, and her answer might surprise sponsors who assume a UX redesign will fix things. “A significant proportion of eCOA and patient-facing technology issues are not caused by the technology itself, but by upstream workflow, communication, and design problems,” she told me. “The tool may function as intended, but the patient experience may still be poor if the study team has not properly considered who will use it, under what circumstances, and with what level of support.”

That distinction matters more than it might sound. A well-engineered app with a clean interface can still fail patients if it wasn’t built around who those patients actually are, their diagnosis, their treatment burden, their language, their fatigue. In other words, the eCOA inherits the sins of decisions made long before anyone tested a build.

Remembering Who’s On The Other End Of The Screen

It’s easy, sitting in a vendor demo or a sponsor planning meeting, to lose sight of what a study is actually asking of patients. Jashi doesn’t let the conversation drift there. “We already ask a great deal from patients,” she said. “They may be dealing with a serious diagnosis, treatment-related fatigue, emotional stress, uncertainty about prognosis, and the practical burden of frequent study visits.” On top of that, they’re handed devices, diaries, reminders, and questionnaires that aren’t always intuitive, properly translated, or validated for the population actually using them.

This isn’t an abstract concern for Jashi; it’s the specific mechanism by which good intentions produce bad data. A patient who doesn’t understand a question doesn’t usually stop and ask for clarification. They guess, they skip, or they disengage. Data quality erodes quietly, and it often looks in the dataset like noncompliance rather than what it actually is: a design failure.

Why Linguistic And Cultural Adaptation Isn’t A Checkbox

If there’s one place where Jashi pushes back hardest against how the industry treats eCOA, it’s here. Usability testing, linguistic validation, and cultural adaptation of questionnaires get treated, too often, as a translation vendor’s line item rather than a scientific safeguard. Jashi sees it differently. “These directly affect patient understanding, data quality, compliance, and the credibility of the final outputs,” she said. “A questionnaire or digital tool may be scientifically appropriate in one language or setting, but if it is not properly adapted and validated for the target population, the patient may misunderstand the question, complete it inconsistently, or disengage from the process.”

That risk isn’t evenly distributed. It concentrates on exactly the populations where data quality matters most and where patients have the least reserve to push through a confusing tool. “This becomes especially important in vulnerable or older patient populations, oncology studies, chronic disease studies, psychiatric indications, or any setting where the patient’s physical or emotional capacity may be limited,” Jashi said. A digital tool that reads as simple to a vendor or sponsor sitting in an office can feel exhausting to a patient managing pain, anxiety, or the aftereffects of chemotherapy.

Judging eCOA By More Than Whether It Collects Data

The industry tends to evaluate patient-facing technology on a fairly narrow question: did it capture the data point? Jashi thinks that’s the wrong bar. “Patient-facing technology should be judged not only by whether it collects data, but by whether it does so respectfully and realistically,” she said. “True patient-centered technology should reduce friction, support participation, and protect data integrity without forgetting the human being behind the data point.”

That reframing has real operational consequences. An eCOA tool judged purely on completion rates might look successful while quietly generating unreliable responses from patients who didn’t understand what they were being asked. A tool judged on whether patients could engage with it accurately and without added burden tells a very different, and more useful, story about the data underneath it.

Where This Connects Back Upstream

None of this happens in isolation from the workflow issues I covered in the companion piece to this article. Jashi has been clear that eCOA problems are frequently symptoms of decisions made earlier, how a protocol defined its assessments, how consistently a site was trained, how clearly a protocol amendment was communicated down the chain before it reached the patient. Fix the upstream handoffs, and the patient-facing layer has a real chance of working as intended. Leave them broken, and no amount of interface polish will compensate.

What Better Looks Like

Jashi’s prescription isn’t complicated, even if it requires discipline the industry doesn’t always have: build in usability testing and linguistic and cultural validation as early, non-negotiable steps, not late-stage checkpoints. Involve people who understand the target patient population, not just the protocol, in questionnaire design. And resist the temptation to treat a completed electronic diary entry as proof that a patient understood what they were completing.

The throughline across both conversations I’ve had with Jashi is consistency, not coincidence: whether she’s talking about sponsor-CRO-site handoffs or a patient filling out a diary at 9 p.m. after a treatment session, her point is the same. Technology doesn’t fail because it’s poorly engineered. It fails because it was built without enough attention to the person on the other end of it, and in clinical trials, that person is usually not the one in the room when the decisions get made.