Are Medical Doctors Scientists? Causal Inference Based on Observational Data

Abstract

We present arguments for a view of what a clinician-scientist might be like, in the setting of everyday clinical practice. Clinical observations of good quality and completeness may, under suitable context or research framework, be viewed as evidence for claims of causation. Historical examples of clinicopathological observations in which causal claims were made and later confirmed, are briefly mentioned. We present in some detail a more recent causal inference framework in statistics, which is increasing accepted in practice, as well as a brief introduction to causal diagrams. Finally, we briefly present a view of a possible use of causal inference in the near future.