[Statistical models with reference to their value for medical process quality assurance].

Since the revision of the KVG (Art. 58) (Health Insurance Law) in 1995, systematic scientific monitoring is laid down by statute in order to ensure quality (Health Insurance Regulations; KVV Art. 77). In addition, the statistics law of 1992 prescribes the BFS statistics (with ICD coding) (model 1). Since 1983 the "Arbeitsgemeinschaft Schweizerischer Frauenkliniken" (ASF) (The Swiss Working Group of Obstetrical and Gynecological Institutions) has been maintaining a common set of statistics which amongst other things also serves for quality assurance purposes (model 2). In 1995 a number of surgical hospitals joined together under the title "Arbeitsgemeinschaft für Qualitätssicherung in der Chirurgie" (AQC) (Swiss Surgical Quality Assurance Working Group) and now also maintain similar common statistics (model 3). In this paper the three above-mentioned models are described with regard to their suitability for process quality assurance. Whilst the BFS statistics are unsuitable for this purpose, the two other methods of data collection largely fulfil the requirements for process quality assurance by using statistical models. The largest deficiency in the ASF and AQC statistics is the lack of comprehensive geographical coverage which in contrast is provided by the BFS statistics thanks to statutory requirements. However, all three models are unsuitable for the areas of structure and outcome quality assurance. Therefore other solutions must be sought for these purposes.