Usage Examples
- The PRO is a key secondary endpoint, so the analysis was pre-specified before unblinding.
- Moving that PRO from paper diary to tablet is a change to the instrument, not just to data capture.
- Site staff cannot complete the PRO on the patient's behalf; that entry has to be queried.
What is Patient-Reported Outcome (PRO)?
Patient-reported outcome (PRO) is an outcome measure captured directly from the patient about their own symptoms, function, or health status, with no interpretation, filtering, or amendment by a clinician, caregiver, or investigator.
Patient-reported outcomes exist because some treatment effects have no external witness. Pain, fatigue, itch, nausea, breathlessness, and whether a patient can get through a workday live only inside the patient's experience. Every layer between that experience and the record, a clinician's paraphrase or a coordinator's transcription of a paper diary, adds interpretation a reviewer cannot audit years later. A PRO deletes that layer.
A patient-reported outcome covers anything the patient reports about their own state: symptom severity, physical and social function, health-related quality of life, treatment tolerability. The boundary is who speaks. A clinician's rating is a clinician-reported outcome, a caregiver's observation is an observer-reported outcome, and a scored task such as a six-minute walk is a performance outcome. None of those are PROs, however patient-centric the programme sounds.
Patient-reported outcomes are applied through commitments made before the first patient enrolls: an instrument matched to the concept and population, an endpoint and analysis written into the protocol, and a capture system built to hold regulated records. 21 CFR 314.126(b)(6) is the anchor: the method of assessment must be well-defined and reliable, and both protocol and report must explain the variables, methods of observation, and criteria used.
Not to be confused with
- Clinician-Reported Outcome (ClinRO)
- a ClinRO is a trained clinician's judgment about the patient, so it carries professional interpretation by design. A PRO carries none; if a clinician edits or grades the answer, it stops being a PRO.
- Observer-Reported Outcome (ObsRO)
- an ObsRO is reported by someone other than the patient or a clinician, typically a parent or caregiver, and is limited to observable behaviour. It cannot capture an internal state such as pain intensity, which only a PRO can.
- Performance Outcome (PerfO)
- a PerfO measures a standardised task the patient performs and someone else scores, such as a walk test or cognitive battery. The patient produces the data but does not report it, so the interpretation layer is still present.
- PRO instrument vs PRO endpoint
- the instrument is the questionnaire and its scoring rules. The endpoint is the derived measure, timepoint, and pre-specified analysis. A validated instrument analysed post hoc does not produce a defensible endpoint.
PRO data intended to support a marketing application inherits obligations from two places: the adequate and well-controlled studies rule and the electronic records rule.
What you must do
- 1Fix the method of assessing patient response so that it is well-defined and reliable before the study starts, rather than reconstructing its properties afterwards21 CFR 314.126(b)(6)
- 2Explain, in both the protocol and the report of results, the variables measured, the methods of observation, and the criteria used to assess response21 CFR 314.126(b)(6)
- 3Treat ePRO data as a regulated electronic record whenever it is created, modified, maintained, archived, retrieved, or transmitted under an FDA records requirement21 CFR 11.1(b)
- 4Validate the ePRO system for accuracy, reliability, consistent intended performance, and the ability to discern invalid or altered records21 CFR 11.10(a)
- 5Retain secure, computer-generated, time-stamped audit trails recording the date and time of every entry or change to a PRO record21 CFR 11.10(e)
Common mistakes
Running the PRO as an exploratory endpoint, then claiming it in labeling
effectiveness rests on adequate and well-controlled studies in which the assessment method was well-defined in advance and explained in the protocol. A PRO analysed after the fact fails that test, and the claim is stripped in review after the data has already been generated.
Modifying a validated instrument without re-establishing its properties
dropping items, shortening the recall period, translating without cognitive debriefing, or migrating paper to screen creates a different instrument. The prior validation evidence attaches to the old version, so the study ends up measuring with something that has no support behind it.
Treating ePRO as an IT rollout rather than a regulated records system
teams deploy a vendor app with no validation package and no audit trail over patient entries. At inspection the data cannot be shown to be unaltered, which puts the endpoint that depends on it in question rather than just the system.
When This Matters
- The PRO is a key secondary endpoint, so the analysis was pre-specified before unblinding.
- Moving that PRO from paper diary to tablet is a change to the instrument, not just to data capture.
- Site staff cannot complete the PRO on the patient's behalf; that entry has to be queried.
Frequently Asked Questions
A PRO measure counts as validated when its method of assessment is well-defined and reliable for the specific concept, population, and claim at issue, which 21 CFR 314.126(b)(6) requires of any response assessment in an adequate and well-controlled study. Sponsors evidence this with documented content validity, reliability, and ability to detect change.
Related Use Cases
- Pharma Use Cases
Cut NDA and sNDA prep time by 60% with AI-assisted drafting and automated readiness checks
- Biotech Use Cases
Compress IND prep from 8-12 weeks to under 3 weeks with AI-assisted drafting and validation
- Clinical Workflows
Track clinical trial regulations across global markets
- Regulatory Affairs Workflows
Cut regulatory intelligence tracking from 10+ hours/week to automated, real-time alerts
