Federal aid to dependent variables: surveying the health services research landscape

The current status and future prospects of the Health Services Research and Development (HSR&D) Service are examined through a series of parables. The case of Goldilocks illustrates the fact that in health care research it is difficult to get things just right, and mischief may be caused, particularly by neglecting to carefully assess sample size. The HSR&D Service is now routinely returning proposals that do not pay sufficient attention to sample size estimates. Although models are appropriate tools for health services research, they must be well crafted and there must be evidence that prospective pilot studies have documented the clinical and methodological validity of proposed models. It is necessary to demonstrate that the techniques of health services research can have a practical, positive effect on the way medicine is practiced.