Text & media
Reading Trials,
Reading Endpoints.
An interactive guide to clinical trial appraisal — from understanding endpoint hierarchies, to spotting them in the wild, to judging whether a trial actually delivers on what it promises.
The Endpoint Explorer.
Endpoints are the questions a trial is designed to answer. Click any node to learn what it measures, why it matters, and what to watch out for when appraising it.
Sort the Endpoints.
Drag each endpoint into the correct category. Hard endpoints measure events that matter to patients directly. Surrogates predict those outcomes indirectly. Composites bundle multiple events. PROs come from the patient.
Appraise a Trial.
A simulated trial summary modeled on CONSORT reporting standards. Click each item in the checklist on the right to highlight the relevant passage and read the appraisal rationale.
Drug X versus Standard Care in Adults with Condition Y: A Randomized, Double-Blind, Placebo-Controlled Trial
We conducted a randomized, double-blind, placebo-controlled, parallel-group trial across 84 sites in 12 countries. Eligible adults were assigned in a 1:1 ratio to receive Drug X or matching placebo using centralized, computer-generated permuted-block randomization stratified by site and disease severity. Participants, investigators, outcome assessors, and the data analysis team remained blinded to treatment allocation throughout the study. The primary endpoint was time to first occurrence of a major clinical event (a composite of disease-related hospitalization or all-cause mortality). Key secondary endpoints — tested in a pre-specified hierarchical sequence — included all-cause mortality, change in biomarker Z at 24 weeks, and patient-reported quality of life (EQ-5D). The study was powered at 90% to detect a 25% relative risk reduction in the primary endpoint, requiring 720 events at a two-sided alpha of 0.05.
Of 4,210 patients screened, 3,488 underwent randomization and all were included in the intention-to-treat analysis. Baseline characteristics were well balanced between the two arms. Over a median follow-up of 32 months, the primary endpoint occurred in 412 patients (23.6%) receiving Drug X versus 528 (30.3%) receiving placebo (hazard ratio 0.74, 95% CI 0.65–0.84, p<0.001). Driven principally by a reduction in hospitalizations; all-cause mortality did not differ significantly between groups (HR 0.91, 95% CI 0.78–1.06). Per the pre-specified hierarchical testing procedure, formal inference on subsequent secondary endpoints was conducted only after the primary endpoint reached significance.
Drug X significantly reduced the composite of hospitalization or death compared with placebo, although the effect on mortality alone was not statistically significant. The trial enrolled a relatively young, predominantly male population at specialized centers, which may limit external validity.