Description

Stroke is the second leading cause of death, worldwide. It is also the second largest cause of disability-adjusted-life-years (DALYs) lost after ischaemic heart disease in developing countries, and third largest contributor to DALYs in developed countries (after ischaemic heart disease, low back and neck pain). In 2019, the absolute numbers of new strokes (12.2 million), stroke-related deaths (6.6 million), and people living with stroke (101.5 million) had increased globally from 1990 (70%, 43%, and 85% increases, respectively).

Steady progress has been made in establishing specific management strategies for patients affected by, or at high-risk of, stroke. Unfortunately, only a few acute treatments have been proven to be beneficial: thrombolysis, endovascular thrombectomy, hemicraniectomy, stroke unit care, and aspirin. There is a continued need for interventions that improve outcomes which can be implemented with wide applicability for stroke.

ACT-GLOBAL is an investigator-initiated, multi-factorial, multi-arm, multi-stage, randomised, global adaptive platform trial for stroke, aiming to find the treatment/s associated with the highest chance of improving outcome after stroke. In ACT-GLOBAL multiple questions will be evaluated simultaneously and sequentially as data accrues and can evaluate interactions between different treatment options.

Frequent adaptive analyses are conducted to assess whether a given intervention is superior, inferior, or equivalent either within a domain or for specific populations within the domain. Where it is anticipated that interactions between interventions in different domains may be likely, the statistical models will allow evaluation of such interactions. Each intervention within a domain with prospectively defined and mutually exclusive strata (sub-groups) of participants will be evaluated within the strata, while information from one stratum may be used (via 'borrowing') to contribute to the analysis of the effect of that intervention in other strata. Specific interventions or subgroups within overall populations may be dropped or cease to enrol, based on pre-specified rules.

The adaptive design allows new interventions or domains or both to be introduced. A Response Adaptive Randomisation algorithm may be used to preferentially randomize participants to interventions that appear to be performing better in domains where applicable.

Unlike traditional trial designs which explicitly prohibit co-enrollment of patients into different trials or multiple therapies, or use a factorial design, the adaptive platform design allows investigators to replicate the real-world environment to estimate possible synergy, competitive interference, or safety profiles of complex treatment protocols.