[General practice in a modern health care system--consequences for medical records, education and practice (2)]

Exemplified by studies on the predictive value of standardised history taking of angina pectoris, part 2 of this contribution describes the different significance of medical diagnostics in dependency from different levels of the health care system (general practice, cardiologic drop-in clinic, university hospital). Because of different prevalences (pre-test probability) and in spite of exactly the same symptoms the probability of actual coronary heart disease is highly different. The specific post-test probability can be explained and calculated by Bayes' theorem. It becomes evident why general practitioners have to be cautious in adopting recommendations, if those recommendations are based on studies which were done with patients in specialised health care settings. As a consequence from the reflections and empirical findings presented, it becomes better comprehensible what general practice is not and which perspectives for research, education and practice are resulting.