Innovation should not be likened to progress and cost, true innovation should be clearly distinguished from false, novelty for novelty's sake should be avoided, and medicoeconomic evaluations should be interpreted with caution as they are often too precocious and overfavourable. It seems preferable (taking into account examples of statins, anti-Cox 2 NSAIDs, antiTNFs, recombinants...) that therapeutic innovation, as concerns both the prescriber and the evaluator, should give rise to critical reflection rather than to blind enthusiasm and scientistic belief. Delayed use, focusing on real or potential dangers, can be as deleterious for our patients' health as precipitate or generalized use while only considering increased efficacy. Innovation is extremely desirable, particularly when it seems to correspond to a health requirement of prior concern. However, it should be evaluated with a cool head, without prejudice, and even with a hint of scepticism to act as a counterbalance. It should be used with discernment, without improper extrapolation to patients for whom it is not necessary. Careful assessment of the sanitary and economic consequences of therapeutic innovation--which constitute a source of expenditure or economy--should be a permanent concern of health authorities. They must be evaluated using reliable and partly independent data from firms and paying agencies.