Glossary category
Outcomes and endpoints
Terms describing what a study measures and what the measurement supports.
Terms
17 documents
Table 1. Terms in the outcomes and endpoints category.
| Term | Definition |
|---|---|
| Adverse eventAE | Any untoward occurrence in a participant, whether or not attributed to treatment. Ascertained by spontaneous report in most trials, which underestimates incidence relative to systematic questioning. |
| Anchor outcome | The outcome the Institute designates as primary for an indication across its whole document set, so that estimates for different compounds in that indication are reported on a common measure. |
| Clinical outcomePatient-important outcome | An outcome a person would notice or care about directly: death, an event, a symptom, a function. Distinguished throughout the document set from a surrogate. |
| Composite endpoint | An endpoint counting the first occurrence of any of several component events. Interpretable only when the components are of comparable importance and move in the same direction, and the Institute reports components separately. |
| Discontinuation for adverse events | Cessation of randomised treatment because of an adverse event. Reported by the Institute in the summary-of-findings table alongside efficacy outcomes rather than in a tolerability annex. |
| Effectiveness | The effect of an intervention under ordinary conditions of use. Generally smaller than efficacy, principally because adherence and persistence outside a trial are lower. |
| Efficacy | The effect of an intervention under the conditions of a controlled trial, in a selected population with protocol-driven adherence and follow-up. |
| Exploratory endpoint | An outcome analysed without multiplicity control and without pre-specified inferential status. Hypothesis-generating in every case, including where the result is striking. |
| MACEMajor adverse cardiovascular events | Major adverse cardiovascular events. Conventionally a three-point composite of cardiovascular death, non-fatal myocardial infarction and non-fatal stroke; four- and five-point definitions exist and are not interchangeable. |
| Number needed to harmNNH | The reciprocal of the absolute risk increase for a harm. Reported alongside the number needed to treat wherever both are estimable. |
| Patient-reported outcomePRO | An outcome reported directly by the participant without interpretation by a clinician. At elevated risk of bias where the intervention produces a perceptible effect that unblinds the participant. |
| Placebo-subtracted | An effect expressed as the difference between the change in the active arm and the change in the placebo arm. The only form in which the Institute reports a within-arm change. |
| Primary endpointPrimary outcome | The outcome on which a trial is powered and against which its success is judged. A trial has one primary endpoint, or a small number designated co-primary with an explicit multiplicity plan. |
| Responder analysisCategorical responder threshold | An analysis reporting the proportion of participants reaching a defined threshold. Reported by the Institute alongside the mean because a mean conceals the distribution of response. |
| Secondary endpoint | An outcome measured alongside the primary endpoint. Confirmatory only where it occupies a place in a pre-specified testing hierarchy that was reached. |
| Serious adverse eventSAE | An adverse event that results in death, is life-threatening, requires hospitalisation, produces persistent disability or is a congenital anomaly. A regulatory classification and not a judgement of severity. |
| Surrogate outcomeSurrogate endpoint | A measurement substituted for a clinical outcome on the assumption that a change in one predicts a change in the other. Validation of that assumption requires trial-level evidence across interventions and is available for few surrogates. |