Continuous-flow techniques for tests in clinical chemistry.

The title of this paper is, for the hospital biochemist, synonymous with AutoAnalyzer techniques, and it is with the development and evolution of the AutoAnalyzer that it must inevitably deal. The next paper gives an account of the intensive developments which are taking place in the application of discrete analysis systems. This is perhaps the more difficult task, since, while there are many examples of this type there is effectively only one continuous-flow analytical technique. To the industrial analytical chemist coming into the field of clinical chemistry 18 years ago, it seemed that the methods employed lacked the absolute character which obtain in the industrial field. The reasons implicit in the study of biological material soon became obvious and the work load overwhelming, so that any naive concept of bringing the cold purity of absolute analytical accuracy to bear in this new field was gladly abandoned. It is important, however, to remember this feeling, because there is a tendency to approach automated techniques in a highly critical fashion, examining the performance of this or that system against the yardstick of a sanctified manual system of clinical chemistry which never had any reality. The principles of the single-channel AutoAnalyzer are now widely known, since subsequent changes in design have not fundamentally altered the approach so brilliantly described by Skeggs (1957). It was his recognition of the non-absolute character of clinical chemical methods, to which attention has been drawn above, which is fundamental to the success of the AutoAnalyzer. It was not important or even desirable to produce a machine capable of estimating urea concentration with absolute accuracy. If the machine could be guaranteed to treat every sample in exactly the same way so that standards were treated in the same way as tests, then the approach usual in manual methods would be operative. Indeed since machines have no free will and do not tire, it would be reasonable to expect greater precision.