What are the COA types?
A clinical outcome assessment (COA) is a measure that "may be influenced by human choices, judgment, or motivation," which is what separates it from a biomarker. There are four types, sorted by who does the rating. eCOA means any of them captured electronically; ePRO is the patient-reported kind on a phone, tablet or web form.
| Type | Who rates | Example | How it is captured |
|---|---|---|---|
| Patient-reported outcome (PRO) | The patient | Pain or low mood | Paper, computerized questionnaire forms, or an interview recorded exactly as spoken |
| Clinician-reported outcome (ClinRO) | A trained member of the investigator team | Rating of findings from a physical exam or interview | The clinician's rating under the COA's definition |
| Observer-reported outcome (ObsRO) | Someone other than the patient, with no professional training needed | A parent, spouse or caregiver's observations | The observer's report |
| Performance outcome (PerfO) | No rater judgment; the patient performs a defined task | Distance walked in 6 minutes; symbols matched to a key in a fixed time | The task is scored in a specified way |
Sources: Walton et al., Value Health 2015; FDA PFDD guidance series.
Moving a questionnaire from paper to electronic
FDA's PRO guidance lists "changing an instrument from paper to electronic format" among the changes that can alter how patients answer, and says additional qualitative work "may be adequate depending on the type of modification made." The ISPOR ePRO Task Force ties the evidence to how much the migration changes the measure's "content, meaning, or interpretation":
| Level of modification | Evidence the task force recommends |
|---|---|
| Minor | Cognitive debriefing and usability testing only may be sufficient |
| More substantive | Equivalence testing |
| Substantial | Full psychometric testing |
A 2023 ISPOR update takes an "evidence-based approach to determine whether new research is needed." Where enough comparability evidence already exists and the migration follows best practices, it says further testing across modes is unnecessary, including when a trial mixes modes such as bring-your-own-device (BYOD) designs.
Sources: FDA PRO guidance, 2009; Coons et al., ISPOR ePRO Task Force, Value Health 2009; O'Donohoe et al., ISPOR, Value Health 2023.
BYOD vs. provisioned devices
With BYOD, participants answer on their own phones; with a provisioned device, the sponsor hands out a study phone or tablet. A crossover study from the PRO Consortium and eCOA Consortium gave 64 people with COPD 15 days on each approach and found the scores equivalent.
| Result (Hudgens et al. 2022) | BYOD | Provisioned device |
|---|---|---|
| Score agreement between the two | Equivalent: ICC 0.863 to 0.908, within 10% of total score (TOST) | |
| Across 3 screen sizes | Equivalent: ICC 0.972 to 0.989 | |
| Weekly compliance | 89.7% to 100% | 76.9% to 100% |
This is one study in one condition, so check comparability evidence for your own instrument and population before relying on it.
What FDA reviewers look for
FDA's 2009 PRO guidance (final, docket FDA-2006-D-0362) describes how the agency evaluates PRO instruments used to support labeling claims. Points that matter most for an app or eCOA build:
- Content validity first. It is "the extent to which the instrument measures the concept of interest," and good reliability or other validity "will not overcome problems with content validity." FDA says to establish it before other measurement properties.
- Recall period. Sponsors should justify it; items that ask patients to recall over a long period, compare with an earlier time or average over time "are likely to undermine content validity."
- Reliability, other validity and ability to detect change are reviewed as measurement properties.
- Investigator control of ePRO source data. The investigator must keep "the ability to control and provide access to the records," and the protocol should say how ePRO source data are maintained.
- Audit trail and security. Avoid sending data from the device "without an electronic audit trail," avoid source control by the sponsor alone, and plan system, security and cyber-security controls, consistent with FDA's computerized-systems and Part 11 guidance.
FDA's patient-focused drug development (PFDD) series adds detail. Guidance 3, on selecting, developing or modifying fit-for-purpose COAs, was finalized in October 2025. Guidance 4, on turning COAs into endpoints, is still listed as a draft and covers "methodologies, standards, and technologies" for COA data collection. This page is a summary, not regulatory advice.
Sources: FDA PRO guidance, 2009; FDA PFDD guidance series.
From our work
Our principal, Mike Hittle, created and led the PHENOMS eCOA platform for neurology.
Related resources
Frequently asked questions
Does moving a questionnaire from paper to an app need revalidation?
It depends on how much the change affects content, meaning or interpretation. The ISPOR ePRO Task Force says cognitive debriefing and usability testing may be enough for minor changes, equivalence testing for more substantive ones, and full psychometric testing for substantial ones. A 2023 ISPOR update says no new testing is needed when enough comparability evidence already exists and best practices are followed.
Is BYOD as good as a provisioned device for ePRO?
In a crossover study of 64 people with COPD, scores on participants' own phones and on provisioned devices were equivalent (ICC 0.863 to 0.908), and weekly compliance was 89.7% to 100% with BYOD vs. 76.9% to 100% with provisioned devices. It is one study in one condition.
What is the difference between a PRO and a PerfO?
In a PRO, the patient is the rater and answers questions directly. In a PerfO, the patient performs a defined task, such as walking for 6 minutes, and the result is scored without rater judgment.
Scope it in 30 minutes
Book a scoping call and we'll name the smallest engagement that moves you forward. Most teams start with the fixed-scope Initial Audit.
Book a 30-min scoping callPrefer to write? Send a short note.