University based academic intensive care units serve as the driving force of advances in patient care in critical care. Each academic unit has multiple purposes.

Patient Care
First, they must provide outstanding patient care. They must also constantly seek to improve this care. The average pediatric intensive care unit admits less than 750 patients per annum. These patients may have greater than 100 different diseases. Thus only a small number of children with each disease and (generally with wide variability in their presentation) are available for research and training.

For a strictly controlled prospective randomized clinical trial (PRCT) each individual intensive care unit, to improve its care would have to study one question for many years. This is further confounded by the fact that data gathering can be time consuming and data analysis prohibitive.

The everyday urgent pressure of work in the new challenging economic environment makes it extremely unlikely that new knowledge or bedside application of basic research will be supported through clinical research if it depends upon individual intensive care units to provide it. Patient care will stagnate.

Teaching
The second role of academic intensive care units is education. Clearly, training intensivists to replace those in existence and to provide care for children who need it in underserved areas is a primary goal. To this end there are nationally recognized American Academy of Pediatrics certified fellowships in pediatric critical care which require three years of training.

Although there are guidelines for these fellowships, there is wide variability in their teaching. This variability occurs not only in the content of each curriculum, but also in the quality with which it is executed. In dozens of Centers throughout the United States, fellows are educated on the same subjects, such as statistics, research design, patient management, and specific care of various illnesses. Some units do this better than others. A vast amount of effort is repeated and redundant.

Minimal effort has been made to define a national curriculum. In the past we have relied on national conferences, academic journals and textbooks to educate new intensivists and re-educate practicing intensivists. This "old" technology is inadequate to the current environment. Children die, families suffer, society pays as we await this inefficient and slow transfer of knowledge-based information to all intensivists.

This educational paradigm is repeated in pediatric critical care units for nurses, medical students, pediatric residents, respiratory therapists, nurse practitioners, surgical residents and a host of other people who depend upon the critical care physicians and intensive care unit for education concerning the care of critically ill children. All of this effort is repeated in dozens of units throughout the country, with varying standards and with tremendous duplication of effort and energy and expense.

Research
The third purpose of intensive care units is research. Not only basic science research, but also clinically applicable outcomes based research. Basic science is needed to lay the foundations for the care of tomorrow. Clinical outcome research is necessary to apply todays knowledge to the care of todays children. The coordination of this research, down to the simplest questions such as: "Who is doing what?", "Are several programs doing the same thing?", and "Should they integrate?", is completely lacking.

Economic pressures, and the demands of critically ill children will further restrict basic and clinical scientific research. The effort in this direction must be optimized and shared to assure future improved outcomes for critically ill children. Improving the quality of care requires shared research.

Clinical research is central to the progress of clinical medicine. In a market environment providing the best care in the most efficient way is a national priority. Despite these facts, evidence based medicine in pediatric critical care has always been hampered by the absence of prospective randomized clinical trials (RCTs) which are the gold standard of study design. These RCTs are expensive and logistically complex because they rely on manual data acquisition and comparison.

Unless clinical equipoise is real within the study domain RCTs are fraught with ethical problems. The current situation is that the vast majority of clinical decisions are made by biased interpretation, albeit well intentioned, of clinicians acting on individual anecdotal experiences which are imprecisely remembered and from which, more often than not, wrong inferences can evolve. This form of medieval practice can only be dispelled by the provision of randomized prospective patient management protocols coordinated in multiple ICUs involving large numbers of patients.