Three applications for quality management

Nós entendemos que a qualidade dos resultados é uma responsabilidade dos profissionais do laboratório clínico, nos seus diferentes papeis. The director has the task of establishing and leading his laboratory in the big picture, assuming this responsibility and translating it into the implementation of Quality Management. Deve cuidar do atendimento às normas legais (ANVISA Technical Regulation), do atendimento aos requisitos dos programas de acreditação e disponibilização de instrumentos, insumos e ferramentas adequados para o melhor desempenho do setor analítico. Professional analysts also have an important role to play

to gain in-depth knowledge of analytical systems, operational standards and statistical process control. Coordinated actions at these two main levels enable the creation and implementation of the best Quality Management (GDQ).

We all learned and recognized that Quality Management is a complex set of actions, which are based on planning, implementation, control and re-elaboration of processes for continuous quality gains. The scope we seek to address in this and other of our articles refers to the control of analytical processes, which is only part of GDQ. Our intention is to disseminate the need to carry out GDQ, which can initially be done through Internal Quality Control, in Statistical Process Control. Until then, a very widespread paradigm in our country is that if the laboratory carries out External Control through Proficiency Tests, it would guarantee the quality of its analyses. "This procedure, in addition to not being sufficient to add quality to laboratory processes, has the serious limitation of not carrying out daily verification of results. As evaluation results take a long time to return to program participants, process control procedures are no longer effective because, when corrective actions are implemented, many defective results may have been released." We understand

that participation in Proficiency Testing Programs is essential, but it is not enough.

Quando há 14 anos lançamos um software para o Controle Interno da Qualidade acreditávamos no seu valor como ferramenta de auxílio em uma dessas etapas, mas que não dispensava outras ações, mesmo anteriores, de planejamento e revisões para o IQC, bem como de participação do laboratório em ensaios de proficiência, ou seja, a External Quality Assessment.

We believe laboratories should have practical, effective tools to make quality control indicators tangible, bridging the gap between their daily work and the modern quality assessment methods available.

There are different approaches to understanding and planning actions for quality. Todas as definições da qualidade podem ser interpretadas, no campo de atuação do laboratório clínico, como sendo o estabelecimento de condições para que todos os testes realizados auxiliem os clínicos na prática da excelência da medicina⁹.

Modern quality management methods involve much more than statistical process control. Good laboratory practice, quality assurance, quality improvement and quality planning are also essential components of quality management^9,10.

In a general and very simplified overview of how we are going to take care here,

we can summarize this approach in the well-known cycle PDCA, acronym that comes from English Plan, Do, Check, Action, ou seja, Planejar, Executar, Controlar e Agir. This cycle is well known in society and by everyone who has ever worked with organizational development and implementation of quality systems. The beginning is always with planning.

The laboratory must then take care of at least these four management steps, which can be a good start. Adding other actions, for example, for the pre-analytical phase, will make the path to quality more accurate and customers will be better served.

These introductory words are important so that new students of the subject realize that here in this article we will cover just one stage of Quality Management, applying statistical methods there. In PDCA this step corresponds to Control (Check) which, as you can see in the diagram, is the third in the process. With this, we reaffirm that Internal Quality Control, although very important, is not the entirety of the actions that laboratory professionals must be concerned with to guarantee quality in their work.

Statistical Quality Control procedures are used in all branches of activity. In clinical laboratories, they are established to monitor the analytical performance of a system and alert professionals to the occurrence of problems that could compromise the usefulness of patient test results for medical purposes. In the PDCA cycle it corresponds to the third stage, or phase. This type of control is the "dream" of different organizational managers, who would like to have measurement systems and collection of indicators to assess their control status and quality. Clinical laboratories already have this facility, with IQC methods.

Work in clinical laboratories with analyzes that provide quantitative and other qualitative data. For those of the first type, statistics made a great contribution to the establishment of methodologies for controlling analysis processes. The use of appropriate control methods, implemented after planning, allows the Quality Manager to monitor their analytical systems in order to anticipate, preferably, a moment of deterioration. It is the idea that supports the continued performance of internal quality control, that is, carrying out Statistical Process Control (SPC) and detecting changes in stability that result in increased imprecision.

CEP is a control system applied in different sectors of society and applies very well to the clinical laboratory, when statistics are used to verify process variation. This runs in contrast to other components of the quality control plan, a broader and more fundamental issue, which involves preventive maintenance, instrument checking, personnel training, etc. CEP is only one part of the control strategy in a laboratory and should be understood as such.

Internal Quality Control is carried out through the repeated analysis of stable control materials, which are produced for this purpose, and the comparison of the results of these analyzes with parameters of acceptable variation under stable conditions. These comparisons are statistically based and provide elements for quality judgment. Our proposition is that laboratory professionals can have three approaches to analytical quality. They can thus make comparisons of the results of the analysis of control materials, in a standardized and practical way and with a broad scope for Brazilian clinical laboratories.

1. That the average values were below that indicated by the supplier of the control material, since the results fluctuated below the Xm line (green line), around 1 standard deviation.

2. That the oscillation was of small amplitude, indicating a CV smaller than that calculated by the manufacturer's values, which was good.

As the objective was to establish the laboratory's own mean and standard deviation parameters, these values were initially considered acceptable.

The most important aspect was the perception of the trend, of the analytical run from 9 to 15. It indicated a systematic error, which led the analyst to investigate the cause. He understood that the most likely cause would be deterioration of the reagent. The reagent lot was changed and the controls were retested. The results of run 16 and following were considered approved and the values returned to the levels before the deterioration, although still different from the Xm value offered by the manufacturer.

The interpretation of the control chart made a great contribution to detecting deterioration in the system, still in the preparation phase, that is, without yet applying the control rules. Measurements were made (analysis of control materials), application of statistics, plotting and analysis of the graph and the realization that there was a problem. From then on, the decision was made to change the reagent, perhaps considering the history of this analyte and reagent. The run was discarded and the controls were re-analyzed. He concluded that the analytical system recovered the desired balance.

In these actions we can see the last three phases of the PDCA cycle. The analyst carried out Quality Management, based on measurements of the controls and indicators that supported his decision making. It was also evident that in the

Internal Quality Control is true of Peter Drucker's maxim: what is not measured, cannot be managed. Here it was well measured and management provided decision and effective action.

The performance characteristics of an analytical system that are important for planning quality procedures are imprecision (random error) and bias. It is essential to critically analyze the performance of the analytical system to identify opportunities to improve measurement quality to obtain lower values for random error, reducing imprecision and bias, reducing systematic error⁹. The assessment of imprecision is part of the scope of Internal Quality Control and has been the subject of this work. Bias will not be addressed here, as it is beyond the scope.

The measure of imprecision can be given by the Coefficient of Variation (CV), calculated by the formula CV = SD / Xm

* 100. It is expressed as a percentage and provides a good estimate of the performance of the analytical method, regarding variability. The mean is a measure of central tendency and is related to accuracy, or the situation of systematic errors. The SD reflects the widening of the distribution of measured values and relates to imprecision, or random errors. The SD will be greater depending on the concentration of the analyte measured in the control material⁵. It is expected that the SD for level 1, with a mean value in the method's reference range, is smaller than the SD for level 2, with a generally high value, at the pathological level. As the CV reflects an index of the SD and mean, it is a better expression of imprecision for different concentrations and can be used for comparisons.

The CV can be used to compare the performance of one laboratory and another, in a specified method. The CV will be smaller, the smaller the variation in the control results, that is, there will be less imprecision⁵.

The worst case scenario for errors in the laboratory is given by the Total Error, which represents the combination of the systematic error (bias) and the random error (imprecision) estimated for a method and being in the same direction. Taking care only of imprecision, the lower the CV of the method, the better, because there is "left" more room for systematic error, before compromising the final quality of your result. Therefore, it is an interesting goal for the laboratory to achieve the use of a method that provides a reasonable CV, lower than the maximum desirable imprecision published in the literature².

The work of Ricós et al. It is very important for consultation by the laboratory, for reference of imprecision. It offers a very valuable table, with numerous tests and values for Maximum imprecision and Total Error, among others. Many laboratory professionals are using this data as a reference, especially for comparing their CV result with the Maximum imprecision value. The goal is to work to get the CV to be less than the published value for Maximum imprecision.