We conclude that discovery of a variety of new principles makes possible the development of fast analyzers which are readily interfaced with and controlled by computers. However it is not yet possible to predict what configurations will prove most reliable and useful. For example, the reaction volumes may be varied greatly, and there appears to be no technical reason why devices having a total reaction volume of as little as 1 microliter and working at very high rotational speeds cannot be built. High-speed electronic computers have rendered secondary to present systems considerations of the cost or time required per addition, subtraction, or multiplication. The problem considered here is how to reduce the cost of individual analyses to the point where the cost of each is similarly inconsequential. The objective is not only to increase speed and lower cost, but also to solve one of the most difficult problems in the clinical laboratory, namely, mismatching of sample and result. As previously noted, if each analysis can be done in duplicate or triplicate at random, then there is a very high probability that, if the results agree, they are correct. Statistical analysis of the results will also provide a constant check on precision. In the lactate dehydrogenase assays presented, seven experimental assays with seven control were done in little more time than that required for assay manually. With larger rotors, the number may be easily increased. Even at this very early stage of development, when many concepts remain to be fully explored, the number of many kinds of tests done may be greatly increased, and the data in final form may be obtained quite rapidly. It is not too early to raise the question of the use to which these and similar devices will be put. There can be no question of the advantage of increased speed and decreasing cost of clinical tests involved in diagnosis and patient care. Mass screening for occult disease has been repeatedly proposed as a desirable objective, and one which required only the solution of technical problems. In screening, the presumption is (i) that the natural course of a disease is sufficiently well understood so that benefit is derived from its discovery before symptoms appear and (ii) that resources are available which will make it possible to effectively treat the patient. In the case of infectious diseases, such as tuberculosis, these requirements have been met.
