Comparing two identically protocolized, multicentre, randomized controlled trials on caregiver-mediated exercises poststroke: Any differences across countries?

Background: The evidence for rehabilitation interventions poststroke lack sufficient robustness. However, variation in treatment effects across countries have been given little attention. 

Objective: To compare two identically protocolized trials conducted in different western countries in order to identify factors that may have caused variation in secondary trial outcomes. 

Methods: Comparative study based on individual patient data (N = 129) from two randomized controlled trials, conducted in hospitals and rehabilitation facilities in the Netherlands (N = 66) and Australia (N = 63). Patients with stroke and their caregivers were randomly allocated to an 8-week caregiver-mediated exercises intervention (N = 63; 31 Australian and 32 Dutch) or to a control group (N = 66; 32 Australian and 34 Dutch). Patient characteristics, compliance, usual care and process measures were compared across countries. We examined if study setting significantly moderated the trial outcomes: Hospital Anxiety and Depression Scale, Fatigue Severity Scale and General Self-Efficacy Scale, measured at 8- and 12 weeks follow-up. In addition, we explored if factors that were significantly different across countries caused variation in these trial outcomes. 

Results: Most patients suffered an ischemic stroke, were in the subacute phase and participated with their partner. Dutch patients were younger (P = 0.005) and had a lower functional status (P = 0.001). Australian patients were recruited earlier poststroke (P<0.001), spent less time in exercise therapy (P<0.001) and had a shorter length of stay (P<0.001). The level of contamination was higher (P = 0.040) among Dutch controls. No effect modification was observed and trial outcomes did not change after controlling for cross-country differences.

Conclusions: The present study highlighted important clinical differences across countries whilst using an identical study protocol. The observed differences could result in a different potential for recovery and variation in treatment effects across trials. We argue that we can proceed faster to evaluating interventions within international pragmatic trials.