Interest in outcomes is universal. To patients, good outcomes represent their highest hopes for therapy; to health care professionals, good outcomes are the desired end-point of a complex web of care. More recently, politicians and health care managers too have shifted their emphasis away from health service activity and towards what is termed 'health gain'. The rise of the outcomes movement appears irresistible. However, the difficulties in interpreting outcomes data will not go away. Outcomes measured using routine data are subject to numerous biases and many practical difficulties. Despite recent statistical, methodological and technological advances, comparisons of outcomes at best provide us with weak evidence of either the effectiveness or the quality of health care. And sometimes they may frankly mislead. The apparent intuitiveness of outcomes monitoring has broad public appeal. But enthusiasm for outcomes needs to be tempered with a clear understanding of their limitations.