From Intensive Care to Recovery: Feasibility of a Randomised Controlled Study Design Evaluating a Cognitive Rehabilitation Intervention for Critically Ill Patients.

The study was designed to test trial procedures, not treatment effect, and no formal hypothesis testing was done. The planned 120 participants were not reached; 83 were randomised. Only 33 of 83 patients completed 6-month follow-up, and 12 died, so retention and follow-up estimates are imprecise. Blinding was not possible, and planned assessor blinding stopped after about 20 participants, which may have introduced assessment bias. Screening was limited to weekday daytime hours and depended on research staff availability, so some admissions were not screened. Patients had to be awake and free from delirium at inclusion, and those with primary neurological disease, dementia, cardiac arrest, severe substance abuse or palliative care status were excluded, limiting generalisability to more unstable ICU populations. ICU discharge cognitive and quality-of-life assessments were often not completed because of fatigue, delirium or competing care transitions.

This study was funded by Novo Nordisk Fonden and Aarhus University. The funders had no role in the design, conduct or reporting of the study. The authors declare no conflicts of interest. OpenAI ChatGPT was used for language editing.

Do not read this as proof that mindfulness or brain training after ICU improves recovery. Patients also received usual care, ICU nurse support and, in some groups, post-ICU contacts, so the contribution of any single component cannot be separated. No formal hypothesis testing was undertaken.