Abstract Background and aims Despite numerous trials, the optimal design of stroke aftercare programs remains unclear. The StroCare trial evaluated a structured, transitional and follow-up care program to improve health-related quality of life (HRQoL) up to 2 years after stroke but found no overall intervention effect. This post-hoc analysis explored whether treatment effects differed across patient characteristics. Methods At 12 and 24 months we fitted timepoint-specific linear regression models for PROMIS-10 Global Physical Health T-scores including interaction terms between treatment group and covariates prespecified in the main analysis. Interactions were examined individually and jointly in the intention-to-treat population using multiple imputation with pooled estimates for missing data, analogous to the main analysis. Results Pre-existing care level was strongly associated with worse HRQoL at 12 (ß=-3.8; p0.0001) and 24 months (ß=-3.8; p0.0001). Single-interaction models indicated a stronger attenuation of this negative association in the intervention group at 12 months (group x care level ß=2.4; p=0.03) and a similarly directed but weaker trend at 24 months (ß=2.0; p=0.066). In the full interaction model, this interaction was weaker (12 months: ß=2.2; p=0.058; 24 months: ß=1.7; p=0.13). No other covariates showed relevant effect moderation. Conclusions Given the post-hoc design and small proportion of patients with higher care levels in the study population, these results should be interpreted as hypothesis-generating. However, they suggest that structured aftercare programs may confer greater benefit in more dependent stroke survivors, a subpopulation underrepresented in aftercare trials. We therefore deem the targeted evaluation of such programs in this subpopulation in future trials warranted. Conflict of interest Bourry, Rafael: Nothing to disclose; Petter, Jonas: Nothing to disclose; Schrage, Theresa: Nothing to disclose; Molder, Janina: Nothing to disclose; Boskamp, Stefan: Nothing to disclose; Härter, Martin: Sourced funding with the Innovation Fund of the German Federal Joint Committee; Rosenkranz, Michael: Nothing to disclose; Schmidt, Holger: Nothing to disclose; Beisel, Andreas: Nothing to Disclose; Dams, Judith: Nothing to Disclose; König, Hans-Helmut: Nothing to disclose; Kriston, Levente: Sourced funding with the Innovation Fund of the German Federal Joint Committee; Thomalla, Götz: Sourced funding with the Innovation Fund of the German Federal Joint Committee
Bourry et al. (Fri,) studied this question.