In-person peer support for critical care survivors: The ICU REcovery Solutions cO-Led through surVivor Engagement...

Peer support is a promising intervention to mitigate post-ICU disability, however there is a paucity of rigorously designed studies.

The objective of this study was to establish feasibility of an in-person, co-designed, peer-support model.

Prospective, randomised, adaptive, single-centre pilot trial with blinded outcome assessment, conducted at a university-affiliated hospital in Melbourne, Australia. Intensive care unit survivors (and their nominated caregiver, where survivor and caregiver are referred to as a dyad), >18 years of age, able to speak and understand English and participate in phone surveys, were eligible. Participants were randomised to the peer-support model (six sessions, fortnightly) or usual care (no follow-up or targeted information). Two sequential models were piloted: 1. Early (2-3 weeks post hospital discharge) 2. Later (4-6 weeks post hospital discharge). Primary outcome was feasibility of implementation measured by recruitment, intervention attendance, and outcome completion. Secondary outcomes included post-traumatic stress and social support.

Of the 231 eligible patients, 80 participants were recruited. In the early model we recruited 38 participants (28 patients, 10 carers; 18 singles, 10 dyads), with an average (standard deviation) age of 60 (18) years; 55 % were female. Twenty-two participants (58 %) were randomised to intervention. Participants in the early intervention model attended a median (interquartile range) of 0 (0-1) sessions (total 24 sessions), with 53% (n = 20) completing the main secondary outcome of interest (Impact of Event Scale) at the baseline and 37 % (n = 14) at the follow-up. For the later model we recruited 42 participants (32 patients, 10 carers; 22 singles, 10 dyads), with an average (standard deviation) age of 60.4 (15.4) years; 50 % were female. Twenty-one participants (50 %) were randomised to intervention. The later intervention model attended a median (interquartile range) of 1 (0-5) sessions (total: 44 sessions), with the main secondary outcome impact of events scale (IES-R) completed by 41 (98 %) participants at baseline and 29 (69 %) at follow-up.

In this pilot trial, a peer-support model that required in-person attendance delivered in a later posthospital phase of recovery appeared more feasible than an early model. Further research should investigate alternative modes of intervention delivery to improve feasibility (ACTRN12621000737831).