Independent quality control and its importance: Are you dependent?
The ISO 15189 standard recommends the use of independent quality controls. The IVDR also sets out new standards in relation to independent control material. Reagent, instrument, calibrators and quality control material from the same manufacturer represent a closed system, so that in Germany only the Proficiency Test stands as an independent assessment. Is this really enough? Based on a real laboratory situation and three examples from the literature, this article points to the urgent need to use independent control material.
Schlüsselwörter: Quality assurance, RiliBÄK, reagent lot, lot changes
Imagine the following scenario in the laboratory: the telephone rings and you pick up the phone. The clinician or submitter tells you that he has the impression that the sodium levels have been reading higher for about a week. What steps will you take now to handle the complaint?
Typically, quality control data is used to assess the performance of your analytical system. A look at the data shows you that the controls used do not show any abnormalities - your device system is within the limits allowed by the RiliBÄK.
You can look at the sodium levels for all patients in the last week. For example, a daily mean or better median value of patient results can be calculated and, in fact, find that the values are higher overall than in the previous month.
You have received a new reagent lot and have used it on your analyzer system.
Of course you measured the new reagent lot with a new calibration followed by quality control (two controls, one in the clinical decision area, the other in a higher concentration range). The controls showed no deviation from the stated target values and were within the permissible limits of the RiliBÄK.
After the production of a new reagent lot, the quality is tested by the manufacturer. One checks whether the performance of the new reagent lot meets the requirements and, if so, the final approved lot is made available for sale. Subsequently, the device system, the new reagent lot, the associated calibrator and the quality control material are used together. This means that a corresponding target value must be determined for the control material. This target value then depends on the device system, the new reagent and the calibrator lot. This in turn means that a change from reagent lot to reagent lot cannot be detected. There may be significant differences between individual reagent lots. This is precisely where the control material, which has not been optimized for the single reagent lot, is required as an independent review of the analytical process.
The controls were matched to the reagent lot and thus unable to detect deviations in patient samples.
In the future, independent control materials will be used in the laboratory.
The current RiliBÄK does not provide for independent control . In contrast, ISO 15189  recommends using an independent controlmaterial. Importantly, in the new IVDR , the reagent manufacturer is required to report and must, if requested, provide the user with details of the individual reagent lots (see the “IVDR Information” box).
The monitoring of the performance of the individual lots can only be carried out via an independent control material or by calculating the mean value of the patient results of the respective analyte. However, the calculation of patient mean values depends on many variables and therefore is not always the optimal solution.
Proficiency test samples are also independent control materials, but measurements are not done daily, they are typically at long intervals, as described in the RiliBÄK. This is far from enough considering the case as described at the beginning of this article.
Here the solution is the use of independent control material. Ideally, the control material should be ready for use, i.e. liquid, so that preparation errors can be avoided.
From the literature
Miller et al  report that an examination of 1483 reagent lot changes on average showed deviations in more than 40.9% (14.3-83.3%) of cases, affecting both quality and response the patient results had an influence.
Another publication  described the lot changes in a study that ran for more than 7 years on a collective of supposedly healthy people and diabetics. Again, however, after 7 years at the end of the study, it was concluded that the discovered drift effects were due to lot change and that wrong clinical conclusions had been drawn.