Quality Assurance in clinical laboratory part 03.

Components of  validation Studies 

Validation studies encompass various components such as.

•     Reproducibility.
•     Method comparison.
•     Recovery.
•     Interference.
•     Carryover.

To ensure accuracy, these studies necessitate the use of matrix-matched samples. These samples should possess similar characteristics to the patient specimens (e.g., urine, serum, CSF) that will be analyzed on a particular instrument, including viscosity, turbidity, analytes, and color. For instance, when validating a urine assay, it is essential to employ a urine matrix.

 

Furthermore, validation studies involve the utilization of a blank or background matrix, which lacks the presence of analytes. This blank matrix serves the purpose of evaluating the background signal inherent in the analysis process. The background signal ought to fall below the assay's detection limit to ensure its reliability.

 

 

 

Reproducibility Study 

Reproducibility is conducted to assess the precision or random error associated with a specific method. In this study, a relatively large portion of control material containing the analyte of interest within the matrix is utilized. For instance, when validating a method to measure plasma glucose, a plasma glucose control is employed. Similarly, when validating a method to measure urine glucose, a urine glucose control is utilized.

 

During the reproducibility study, it is recommended to incorporate at least two levels of controls, including samples with analyte concentrations near the critical decision thresholds used in medical settings. For instance, cholesterol has a medical decision point at 200mg/dL, which determines whether treatment is required. Therefore, when validating a cholesterol assay, it is essential to include a control with an analyte concentration around 200mg/dL.

 Once the controls are prepared, a within-run and between-run precision study is conducted to evaluate the consistency and variation of the method's results.

 To conduct a within-run precision study, the analyte level is measured a minimum of 20 times on the same day. Conversely, a between-run precision study involves measuring the analyte level at least 20 times over multiple days. Ideally, the results obtained from both precision studies should be consistent. However, in reality, there will be variations within a certain range. This range signifies the random error (RE) associated with the method. Subsequently, the mean (x̄), standard deviation (SD), and coefficient of variation (%CV) are calculated. A higher degree of variation among individual measurements corresponds to higher values of SD and CV.

A method that exhibits excessive random error is deemed medically impractical. The SD and CV values are influenced by the concentration of the analyte in the control and the specific method employed. %CV tends to be higher when working with very low analyte concentrations. In most chemistry tests, CVs are generally below 5%, while immunological tests typically yield CVs ranging from 5% to 10%. Ultimately, the acceptability of the method relies primarily on the medical utility and intended application of the test, taking into account the %CV.

 

Source of Errors

During validation studies, where most factors are typically controlled and the timeframe is relatively short, it is important to consider additional factors that can contribute to random error.

 In the within-run precision study, the following factors related to the reaction process can contribute to random error:

 

• Pipetting of samples
• Mixing of the samples
• Timing
• Temperature variations (e.g., during incubation)
• Measurement techniques

In the between-run precision study, random error must be taken into account due to  long-term variations,  specifically day-to-day changes. Factors that can contribute to such variations include:

 Fluctuations in room temperature, both on a daily basis and across different seasons

Variations in technique between different technologists (particularly relevant for manual methods)

Considering these factors allows for a comprehensive assessment of random error in both within-run and between-run precision studies during the validation process    (Continue.... )