Description of the included studies and pooled analyses.

*Ref is the reference number listed at the end of this paper.
†Intention to treat analysis unless otherwise indicated.
‡The primary author confirmed through e-mail correspondence that all residents had onsite 24-hour per day nursing care available.
σTime from allocation to treatment to end of follow-up period (allocation in January, 2004 and March, 2005 to end of follow-up in March, 2006).
λ A Bayesian random effects meta-analysis could not be performed for this data.
φCalculated.
∂Results reported in the primary paper.
£The treatment effect reported by the primary author is not an intention to treat analysis as the number randomized is higher than the number for which the statistical test was performed.
μThe baseline variables adjusted for in this model included: Mini-Mental Status Examination score, Barthel index score, physical restraints and delirium, sex and history of falls, and age.