Electronic patient-reported outcomes — ePRO — have been part of clinical research for long enough that the technology is no longer the story. What has taken longer to understand is how to use it well: which study types benefit most, where common problems arise, and what separates high-quality remote data collection from a frustrating experience for participants and study teams alike.
The core advantage: data collected closer to life
The fundamental appeal of ePRO is ecological validity. Paper diaries and recall-based questionnaires capture what participants remember, which is a poor proxy for what actually happened — particularly for dietary intake, symptoms, and daily behaviour. ePRO tools that prompt participants in real time, or close to it, capture data that is more accurate, more granular, and more useful.
This matters most for outcomes that vary day to day: dietary patterns, sleep, energy, mood, gastrointestinal symptoms, physical activity. For these outcomes, a questionnaire completed at a clinic visit weeks after the period of interest is measuring a reconstruction, not a record. ePRO changes what is measurable.
The completeness advantage is not just theoretical. A retrospective analysis of 474 patients completing patient-reported outcome measures found that, among those who completed any part of their questionnaire, 95.0% of electronic respondents completed it in its entirety compared with 75.2% of paper respondents — a statistically significant gap (p < 0.001) that held for both pre- and post-assessment timepoints. The study was in a surgical population rather than a nutritional or observational one, but the underlying mechanism — electronic formats catching missed items and preventing partial submission in a way paper cannot — applies just as directly to a dietary symptom diary as to a post-surgical recovery questionnaire.
Where implementation goes wrong
Choosing the platform before the study design. The most common implementation mistake is selecting an ePRO platform early — often driven by organisational familiarity or vendor relationships — and then designing the data collection around its constraints. The right sequence is the reverse: define exactly what data you need to capture, at what frequency, with what validation logic, and then evaluate platforms against those requirements.
Underestimating the burden of frequent prompts. Daily prompts feel lightweight at the design stage. For participants, they accumulate quickly. A study requiring five minutes of responses every morning for twelve weeks is asking for over six hours of participant time — which, across a full study population, is a significant commitment. Burden should be estimated explicitly and justified, not assumed to be acceptable.
Poor onboarding and technical support. Participant dropout in ePRO studies is often concentrated in the first week. People who struggle with the interface, receive confusing instructions, or encounter technical problems early are unlikely to persist. Investment in clear onboarding, a simple troubleshooting process, and responsive support for technical issues pays back disproportionately in retention.
A worked example. A 16-week dietary intervention study asks participants to log meals via an app three times daily. Week one retention looks fine — 92% of participants log at least one entry. By week three, daily logging has dropped to 61%, concentrated among participants over 55 and those who reported difficulty during onboarding in a brief post-enrolment survey. Reviewing completion data weekly rather than only at study close makes this visible while it is still fixable: a short phone call to the affected group, offering a walkthrough of the app's least intuitive feature, recovers a meaningful share of them before the pattern hardens into permanent non-response. Waiting until database lock to look at completion rates means finding out about this problem only when it is too late to do anything but caveat the analysis.
Failing to monitor completion rates in real time. ePRO generates completion data as the study runs. Not using it for proactive participant management is a missed opportunity. Participants with declining completion rates can often be re-engaged with a brief, personalised contact. By the time the data are locked, it is too late.
What works well
Short, focused instruments. The temptation to include additional questions because they can be asked without adding visit burden should be resisted. Each additional item increases completion time and reduces completion rate. Instruments should include what is necessary for the study question — no more.
Validated questionnaires where they exist. For common outcomes in nutritional research — dietary intake, gastrointestinal symptoms, quality of life, fatigue — validated instruments exist. Using them enables comparison with published evidence and gives confidence that the instrument measures what it claims to. Developing novel instruments where validated alternatives exist adds unnecessary burden.
Branching logic and personalisation. ePRO platforms that adapt the questionnaire based on previous answers — skipping irrelevant items, following up on positive responses — produce better data and a better participant experience than rigid linear forms. The investment in setting up branching logic is usually recovered quickly in completion rates.
Transparency about data use. Participants who understand how their data will be used, who will access it, and what it contributes to the research are more engaged and less likely to provide careless responses. Informed consent is the start of that conversation, not the end.
The evidential standard
Regulators and HTA bodies increasingly review ePRO data as core evidence rather than supplementary information. For studies where ePRO data will feed into regulatory or HTA submissions, instruments need to demonstrably meet standards of validity, reliability, and responsiveness — and this should be addressed explicitly in the study design and analysis plan, not treated as an afterthought once the data are collected.
ePRO done well produces data that could not have been collected any other way. Getting there requires thoughtfulness about design, realism about participant burden, and sustained attention to the participant experience throughout the study.