Variation is widespread in all aspects of clinical
practice. Because only 10% to 20% of medical practice
is based on the type of scientifically sound knowledge
derived from randomized clinical trials, it is not sur-prising that there is so much variation in the way
physicians do just about everything. And with so
much variability in the processes of health care, vari-ability in outcomes and other quality measures is
inevitable. Some variability is unavoidable and even
desirable, but there is now sufficient evidence to show
that unnecessary variation in clinical practice con-tributes to poor quality and increased costs and that
real quality improvement in medical care processes
can best be achieved by eliminating inappropriate
variation in the execution of these processes