IUPAC/CITAC Guide: Classification, modeling and quantification of human errors in a chemical analytical laboratory (IUPAC Technical Report)

Abstract

Abstract:The classification, modeling, and quantification of human errors in routine chemical analysis are described. Classifications include commission errors (mistakes and violations) and omission errors (lapses and slips) in different scenarios at different steps of the chemical analysis. A Swiss cheese model is used to characterize error interaction with a laboratory quality system. The quantification of human errors in chemical analysis, based on expert judgments, i.e. on the expert(s) knowledge and experience, is applied. A Monte Carlo simulation of the expert judgments was used to determine the distributions of the error quantification scores (scores of likelihood and severity, and scores of effectiveness of a laboratory quality system against the errors). Residual risk of human error after the error reduction by the laboratory quality system and consequences of this risk for quality and measurement uncertainty of chemical analytical results are discussed. Examples are provided using expert judgments on human errors in pH measurement of groundwater, multiresidue analysis of pesticides in fruits and vegetables, and elemental analysis of geological samples by inductively coupled plasma mass spectrometry.

The prevention and detection of human error in ecological stable isotope analysis

“…We use “error” to refer to incorrect results that have some human behaviour contribution (Kuselman & Pennechi, ). This definition includes errors for which the human contribution dominates (e.g.…”

Validation of Test Methods, Human Errors and Measurement Uncertainty of Results

“…[10] In the lecture "Human error study as a part of method validation task", Dr. Kuselman proposed to use a method validation mapping possible human error scenarios according to the IUPAC/CITAC guide. [11] Results of such a study can be helpful in the correct formulation of the measurement uncertainty budget and the improvement of the standard operating procedure, as well as for training (how to avoid the errors) and for supervision. The map of the error scenarios, included in the validation report, may also be useful as a check list for prior assessment of an analyst before assigning the task, etc.…”

Correlation of test results and influence of a mass balance constraint on risks in conformity assessment of a substance or material

“…Validation of the SOP, training analysts and proficiency testing, supervision and quality control are elements of the quality system of an analytical laboratory that should prevent human error causing metrologically-related correlation, unless, as shown in Fig. 1, the Swiss-cheese model [9] lines up an error in each element. Therefore, correlations that have arisen in the routine measurement process should be in general negligible and test results for two or more components of the same item (sample) are expected to be metrologically independent.…”

“…For example, when a medication is tested routinely with a pharmacopeial HPLC procedure, correlations might be related to the resolution of the chromatography column used, not able to separate the analytes completely, and the column must be replaced by another one. In the IUPAC/CITAC Guide [9] such an event is rated as a skill-based mistake or omission error (lapse). However, test results of chemical compositions of different batches, lots or objects are inevitably correlated when their actual ("true") values are correlated.…”