Background: It has been recognised that trials of complex interventions are challenging. Systematic reviews of complex interventions tend to focus on the intervention, yet the usual care of both intervention and control groups may pose particular problems in these reviews. Trials in complex areas will have complex interventions which are compared to equally (or more) complex controls.

Objectives: To explore the potential risk of performance bias in a planned systematic review of unconfounded trials of systemic pharmacological interventions for the management of spasticity after stroke.

Methods: A recent search of the Cochrane Stroke Group's Specialized Trials Register identified 35 trials. Of the trials initially available we recorded the following information: 1) Type of comparison: a) other intervention, b) placebo, c) usual care. We assume that both intervention and placebo groups will have included some sort of usual care.
2) If rehabilitation and other interventions as part of usual care (which may have also had an impact on spasticity), were controlled during the trial.

Results: We found 35 trials (n = 1075) of systemic pharmacological interventions in stroke patients. Of these, 10 trials (n = 540) compared one treatment to another, three (n = 42) compared treatment versus treatment versus placebo, 20 (n = 462) compared drug versus placebo only, one (n = 24) compared drug to usual care, and one (n = 7) compared drug to control (unclear). Five trials (n = 96), all of which compared drug versus placebo, reported uncontrolled rehabilitation provided to both the treatment and control groups as part of usual care. In the remaining 30 studies the type of usual care was unclear.

Conclusions: The complexity of usual care may affect trial outcome. This aspect will be further considered in the planned systematic review.