Clinical trials and the real world: selection bias and generalisability of trial results.

The clinical effectiveness movement underpins a major health professional culture change which is occurring worldwide. Clinical practice guidelines are an important component of this movement and seek to promote evidence-based care, and combine quality and efficiency with equity of access in a patient centred approach. However some areas of practice are more amenable to guidelines than others and may thus be advantaged in terms of resource allocation. While high grade evidence from randomised controlled trials forms the basis for many cardiovascular guidelines, selection bias may considerably limit the generalisability of such data. This is particularly exemplified by clinical trials in heart failure which are not necessarily relevant to the majority of older patients with heart failure in the community. Future research in heart failure should be oriented towards the community and realigned in the context of health care priorities for the more elderly who have yet to benefit from previous advances.