Primary Care's Denominator Problem.

It is usually impossible to determine the size of a primary care practice for use as a denominator in calculating rates of morbid and other events within a practice. Although this denominator problem has been a major hurdle for primary care researchers, its nature and significance are not broadly appreciated. This article describes the need for a denominator, defines the term 'practice population', outlines the need for the practice population as a valid denominator, and identifies the areas in which the practice population is the denominator of choice.