Selection, Verification and Validation of Qualitative Tests for Medical Laboratories

Course Lesson

1. Introduction to qualitative tests quality control

• In this lesson, we introduce evidence-based medicine into laboratory logic.
• Briefly discusses the importance of regulating in vitro diagnostic (IVD) medical devices for confidence in laboratory results.

Objectives:

• Review the evidence-based laboratory medicine
• Recognize standardization and good practices importance
• Understand "Emergency Use Only" validation principles
• Recognize in vitro diagnostic (IVD) medical devices regulation

2. Principles and guidelines

• In this lesson, you will learn the principles behind QC.
• We will review some of the major milestones in laboratory quality control over time.
• Briefly discusses the importance of the dynamics of quality cycles and associated practices.
• Introduces harmonized vocabularies to the medical laboratory.
• Finally, discuss some of the most important QC guides.

Objectives:

• Interpretation difference between GMP and GLP
• Learn the PDCA and TQM principles
• Demystifying the differences between selection, validation, and verification
• Introduction to the guidelines

3. ISO compliance

• In this lesson, you will learn which ISO 15189 and ISO 9001 requirements are applicable in the QC of qualitative tests.
• Primary introduction to ISO 15189 global guideline focused on selection, validation, and verification of qualitative med lab tests.
• Interpretation of ISO 9001 in fulfilling technical requirements in qualitative tests.
• Some myths associated with ISO standards and their application in the medical laboratory will be discussed.

Objectives:

• Identify ISO requirements for laboratory tests
• Demystifying some ISO pseudo-assumptions
• Learn which harmonized QC practices are consistent with meeting these requirements
• Know how to relate fit for purpose of results and technical requirements

4. Causes of uncertainty in qualitative tests

• In this lesson, you will learn the principal sources of error that can cause untrue binary results
• The impact of the analytical error on the cutoff trueness is discussed, as well as the effect of the analytical error on the accuracy of the classification of binary results
• The importance of the “gray zone” and the associated trinary classification to minimize the impact of analytical error in the results is debated

Objectives:

• Learn what are the most significant causes of uncertainty
• Interpretation of analytical uncertainty components
• Know how the analytical sensitivity and analytical specificity in NAAT
• Identify biological uncertainty components

5. Sampling principles

• In this lesson, you will learn the pro and cons of sampling
• Introduction to the epidemiological prevalence
• Techniques for collecting statistically and clinically representative samples

Objectives:

• Recognize the importance of the representativeness of the samples (fit for purpose)
• Know which are the best samples of individuals with and without a particular condition
• Know the importance of the sampling dimensions for the confidence interval estimates
• Recognize the critical role of samples for the robustness and reliability of estimates

6. Performance of binary classification tests

• In this lesson, we move to the discussion and application of models for calculating the accuracy of the condition, such as diagnostic accuracy
• We will focus on clinical sensitivity and clinical specificity
• However, we will also discuss the physician's perspective through predictive values
• The importance of the confidence interval will also be discussed. We explore misevaluations implications

Objectives:

• Learn how to calculate and interpret clinical sensitivity and clinical specificity
• Understand the limitations of estimates with poorly representative samples of the population
• Identify the importance of computing the confidence interval
• Understand how to assess the uncertainty of binary results

7. Agreement of binary classification tests

• In this lesson, we concentrate on the computation of binary results agreement
• The determination should only occur when it is not possible to calculate the condition's accuracy
• The misinterpretation can lead to weakly sustained decisions. For example, when referring to "clinical sensitivity" in real cases of agreement of positive results

Objectives:

• Know how to calculate and interpret concordances
• Recognize the limitations of this approach
• Recognize the importance of performance to comparative testing for a lower risk of misestimation
• Know how to interpret the confidence interval

8. Condition accuracy by analyzing numerical data

• In this lesson, you learn the importance of delta value assessment to differentiate mainly tests with identical clinical accuracy
• The delta value is associated with different levels of misclassification risk of binary results

Objectives:

• Learn when delta evaluation is important for binary results
• Know how to interpret the positive delta
• Know how to interpret the negative delta
• Assess the risk of false results in a series of qualitative tests

9. Seronegative window period

• In this lesson, we introduce the seroconversion window period using a binary and trinary results logic
• Recognize the seroconversion period as a primary source of biological bias

Objectives:

• Identify the pros and limitations of this approach
• Evaluate the seroconversion period with binary results
• Assess the seroconversion period with trinary results
• Characterize laboratory methods generations based on the window period

10. Limit of detection in nucleic acid amplification tests

• In this lesson, we introduce the evaluation of the limit of detection in nucleic acid amplification tests (NAAT).
• The statistical models are based on logit regression, probit regression, and hit rate.

Objectives:

• Learn the methodology for estimating the limit of detection in NAAT
• Introduction to logit regression, probit regression, and hi-rate
• Know how to determine the limit of detection through probit regression
• Know how to evaluate the detection limit