Standardization of clinical decision making for the conduct of credible clinical research in complicated medical environments.

The likelihood that past experience will produce correct guides to current practice depends on the signal-to-noise ratio for the clinical problem of interest. If the signal-to-noise ratio is high, the decision will be sound and patient benefit likely to occur. If the signal-to-noise ratio is low, as is commonly the case with difficult clinical decisions, then personal experience and the best intentions will not assure sound clinical decisions. When the probability of benefit cannot be quantified, clinicians in complex settings are in danger of being misled by data and experience. Quantifiable probabilities established by group experiment or observation will be necessary for clinical decisions that can be expected to confer benefit on the patient. Explicit methods are necessary for interventions that can be replicated in experiments or in practice. Computerized protocols force the articulation of explicit clinical care methods and standardize clinical decision making. We have developed explicit, rule-based protocols, implemented them in our hospital, exported them to other hospitals, and successfully achieved a rigorous experimental environment in the clinical ICU. Exportation of such explicit methods may narrow the gap between efficacy (university hospital) and effectiveness (community hospital) research results.