This is a complete, illustrative example of a TrialScope audit — the same structure you receive when you run your own document. The study below is fictional and the findings are for demonstration only; nothing here describes a real trial or product.
ILLUSTRATIVE ONLY
Fictional study, fabricated for demonstration. A real audit is generated from your document, calibrated to your professional role, and is decision-support — not a substitute for professional review.
DOCUMENT · BRIEFING / PROTOCOL SYNOPSIS
NX-204 in Moderate-to-Severe Plaque Psoriasis
Phase III · randomized, double-blind, placebo-controlled · primary endpoint: PASI-75 at Week 16 · reviewer persona: Regulatory Affairs
Overall verdict
VERDICT
Revise before submission
FINDINGS
1 · 3 · 2
critical · major · minor
The document is well organized and the design is appropriate to the indication, but one finding is submission-blocking: the primary estimand is under-specified relative to ICH E9(R1), and the handling of treatment discontinuation is inconsistent between the synopsis and the statistical section. Resolve the critical and major findings before the document goes to a regulator.
Findings
CRITICALStatistics · Estimand
Primary estimand does not specify intercurrent-event handling
The synopsis defines the primary endpoint as PASI-75 at Week 16 but does not state how treatment discontinuation and rescue medication are handled in the primary estimand. Under ICH E9(R1) this is required and materially changes the estimate. A reviewer cannot determine whether a treatment-policy or hypothetical strategy is intended.
From the source documentread-only
“The primary efficacy endpoint is the proportion of participants achieving a 75% reduction in PASI score (PASI-75) at Week 16.”— Synopsis §2.1, Primary Objective
MAJORConsistency
Discontinuation rule conflicts between synopsis and statistical section
The synopsis states participants who discontinue are 'considered non-responders,' while the statistical section describes a mixed-model repeated-measures analysis that implicitly assumes data are missing at random. These two approaches are not reconcilable and will produce different primary results.
From the source documentread-only
“Participants who discontinue study treatment for any reason will be considered non-responders for the primary endpoint.”— Synopsis §2.3, Handling of Withdrawals
“The primary endpoint will be analyzed using a mixed-model repeated-measures (MMRM) approach including all observed post-baseline data.”— Statistical Analysis §9.4
MAJOREndpoints
Secondary endpoints lack multiplicity control
Five secondary endpoints are listed as supporting label claims, but no hierarchical testing or alpha-allocation strategy is described. Without pre-specified multiplicity control, positive secondary results cannot support confirmatory claims.
From the source documentread-only
“Key secondary endpoints include IGA 0/1 response, PASI-90, change in DLQI, itch NRS, and PASI-100 at Week 16.”— Synopsis §2.2, Secondary Objectives
MAJORPopulation
Key inclusion criterion is ambiguous
'Moderate-to-severe' is defined by body-surface-area involvement in one place and by a PASI threshold in another, without stating which governs eligibility. This will create avoidable screening deviations and an inconsistent analysis population.
From the source documentread-only
“Inclusion 4: Moderate-to-severe plaque psoriasis, defined as BSA involvement ≥10%.”— Eligibility §4.1, Inclusion Criteria
“Eligible participants must have moderate-to-severe disease (PASI ≥12) at screening.”— Synopsis §3, Study Population
MINORSafety
AE collection window not stated for the follow-up period
Adverse-event collection is described through end of treatment, but the safety follow-up visit at Week 20 does not specify whether AEs are collected through that visit. Minor, but worth clarifying to avoid gaps in the safety database.
From the source documentread-only
“Adverse events will be collected from the time of informed consent through the end-of-treatment visit.”— Safety §7.2, Adverse Event Reporting
MINORReporting
Trial registration identifier not included
No registry identifier is provided in the synopsis. Not submission-blocking at this stage, but it should be present before any manuscript or disclosure.
HOW THIS WAS CALIBRATED
This audit was tuned for a regulatory affairs reviewer — the same document audited for a biostatistician or medical writer surfaces and prioritizes findings differently. Over time, corrections from verified clinical professionals in each field sharpen how the system reasons for that role. See the for how the calibration loop works.
RUN YOUR OWN
Upload a protocol, CSR, manuscript, or briefing document and get an audit like this in under 90 seconds. Free to use.