Evaluation of a chemistry kit system: comparison of the performances of trained and untrained persons.

Abstract

The Dow Chemical Company Diagnostest kit system used by trained as well as inexperienced analysts was compared with standard chemical methods. In general, results with the kit method and reference methods were comparable. However, in some procedures, small but statistically significant differences between results obtained with the Dow methods and the reference methods were found. In precision the kit methods compared favorably with reference methods. Trained technical personnel averaged 1.4% errors in 360 analyses; untrained persons averaged 7.8% errors under the same circumstances. The proficiency of the untrained persons was quite variable. The number of errors made by individual untrained analysts varied from a low of 1.7% to a high of 16.7% errors. Therefore, it is not possible to make generalizations as to the performance of untrained persons using this kit system. Some did poorly, but half of the untrained persons did as well as a third of the experienced persons. Almost half of the errors made by untrained persons were due to errors in calculating the results from the data and could have been prevented by a simpler system for converting photometer readings to analytic results. When the calculation errors were excluded, the precision of the untrained persons, expressed as the average coefficient of variation of the analyses, compared favorably with the precision of the trained analysts and with the precision of the automated reference methods.