In 2007, the first commentary hosted by this Cochrane's Corner focussed on stroke units and discussed the principle that a valid combination of results from a series of unbiased primary studies can provide influential information that would not be otherwise available by individual studies [1]. We now present a case study of the role played by the stroke unit Cochrane review in the complicated process that led to changes in clinical practice and health policy. Our hypothesis is that the theoretical and pragmatic value of this systematic review goes beyond the mere increase in the number of patients considered in one study: robust results from a methodologically sound systematic review including trials performed in different setting and with different standards can increase the applicability of the conclusion.
Insight on the stroke unit Cochrane reviewThe Stroke Unit Trialists' Collaboration is an international, collaborative group of trialists who (since 1984) committed themselves to systematically reviewing the field. In fact, the role of stroke unit has been controversial for over 30 years [2]. A complete systematic review was published in 1995 and 1997 in the Cochrane Library and the BMJ, respectively [3,4]. This review can be consider as the benchmark of evidence availability on stroke unit. As any other Cochrane's review, its regular updating guarantees a cumulative assessment of information: the last update published in 2007 includes 31 trials, involving 6,936 participants, and showed that patients who receive this care are more likely to survive their stroke, return home and become independent in looking after themselves [5]. Stroke unit care reduces the odds of death or institutionalised care (OR 0.82; 95% CI 0.73-0.92; P = 0.0006) and death or dependency (OR 0.82; 95% CI 0.73-0.92; P = 0.001). This effect remains of moderate statistical significance for case fatality (median 1 year follow-up OR 0.86; 95% CI 0.76-0.98; P = 0.02). The sensitivity analysis based on those trials which used an unequivocally blinded assessment suggested that such bias has not seriously influenced the results. Finally, the review highlighted the particular problems related to the complex and heterogeneous nature of the interventions and its potential interaction with other aspects of care: the review was unable to disentangle precisely what confers to stroke units an advantage (i.e. early mobilisation, better diagnostic procedures, prevention of complications, etc.) in improving patient outcomes.