Oiling the wheels of intensive care to reduce "machine friction": the best way to improve outcomes?

INTRODUCTION: A number of costly trials of interventions for the critically ill have had results that are initially positive, and then subsequent trials are less positive. This has led to uncertainty and a feeling that our scientific approach may need reevaluation. What are we missing when performing these trials? Are there simple, less costly ways of improving outcomes for patients? DISCUSSION: Many of the large interventional trials have included patients on the basis of syndromic enrollment criteria. This inevitably leads to a heterogeneous profile of patients, precipitating conditions, and chronic health. The realistic effect of an intervention may vary depending on the individual circumstances of a particular patient, contributing to trial "noise." The assumption that trial size and randomization will deal with this may not hold true and in some circumstances may be biologically questionable. Perhaps we should consider alternatives? In common with other areas of medicine, it is well-known that intensive care outcomes are variable, reflecting differing performance. Until recently, little attention has been focused on a detailed understanding of variation in performance and delivery of care and how such knowledge might be used to improve patient outcomes. A number of recent efforts that have demonstrated positive improvements have been based around understanding and reducing individual and organizational underperformance. This human and organizational factors-based research does not conflict with more pharmacologic-based research and may be complimentary.
CONCLUSION: By properly understanding the way we do things and the optimal local balance of protocolized and individualized care, we may deliver greater improvements to the outcome of intensive care unit patients than many of the more expensive, pharmacologic, and technological attempts of recent years.