[Statistical control of the quality of clinical laboratory analyses (a lecture)]

No measurement can be absolutely precise, but, by using the methods of the probability theory and by knowing the nature of errors, one can estimate a probable size. The paper considers the specific features of measurements during clinical laboratory analyses, as well as those of statistical methods for assessing their validity. The first part contains general provisions; the second one outlines numerical methods for assessing the quality of performed work. The procedures using both the data of analyses carried out in patients and those of special studies of a control biological material of known and unknown composition. Most numerical methods suggest that a laboratory has at its disposal a computer connected to an analyzer, but a few, quietly reliable assessments can be also manually made. The paper gives only final conclusions and recommendations and omits rather intricate mathematical proofs, which are for information available at the web site clilab.ru in the section "Introduction to statistical quality control".