Clinical epidemiology diagnostic and prognostic

Question 1

Define validity and reliability:

Answer 1

Validity:

• the extent to which a test measures the true value of the variable we are interested in
• measured by comparing the test to the best available test-“gold standard”

Reliability:

• ## the extent to which a test will produce the same result if it is repeated

Question 2

Differentiate between diagnostic and screening tests:

Answer 2

Diagnostic Test: used to determine the presence or absence of a disease when the subject shows signs or symptoms of a disease

Screening test - identifies asymptomatic individuals who may have the disease

Question 3

Differentiate between sensitivity and specificity:

Answer 3

Sensitivity

• the ability of a test to correctly identify patients with a disease (true positives)
• Sensitivity= a/a+c

Specificity

• the ability of a test to correctly identify people without the disease (true
  negatives)
• Specificity= d/b+d

Question 4

What are positive and negative predictive values ?

Answer 4

Positive predictive values:

• the proportion of patients who test positive actually have the disease
• Positive predictive value = a/a+b

Negative predictive values:

• the proportion of patients who test negative that are actually free of the disease
• negative predictive value = d/ c+d

Question 5

Understand and interpret what a ROC curve is:

Answer 5

A graphical representation of a test’s discriminatory ability to distinguish between disease and non-disease cases

Plots the sensitivity (y-axis) against 1-specificity (x-axis)

Area under the curve (AUC) =the probability of the test to correctly classify the patient

Perfect Test-Sens = 100% and Specificity = 100% has an area under the curve of 1
ROC AUC > 0.9 = very accurate
ROC AUC: 0.7-0.9 = moderately accurate
ROC AUC: 0.5-0.7 = poor accuracy
ROC AUC:0.5 = performs no better than chance

Interpretation:

• Steeper curve = better test performance (higher sensitivity at lower false positive rates
• Diagonal line (AUC = 0.5) = worthless test (no diagnostic value)
• Closer the curve to the top left corner, the better the test

Question 6

How do you choose a cut off for what is considered test positive ?

Answer 6

A cut-off value is the threshold at which a test result is considered “positive” for disease
The choice of cut-off affects the balance between sensitivity and specificity

Youden’s Index (J): J = Sensitivity + specificity - 1
The point on the ROC curve where (sensitivity + specificity) is maximized

Question 7

How to minimise misclassification costs ?

Answer 7

If false negatives are dangerous, choose a cut-off that maximizes sensitivity (e.g., cancer screening).

If false positives cause unnecessary interventions, choose a cut-off that maximizes specificity (e.g., confirmatory genetic testing)

Question 8

What are prognostic models and how are they being used in research and clinical practice ?

Answer 8

Prognostic models predict future clinical outcomes based on baseline characteristics and biomarkers

Rarely use a single variable, multiple predictors therefore multivariable approach must be used and a multi stage process

Used in risk stratification, clinical decision-making, and resource allocation

Framingham Cardiovascular Risk score (determining indication for anti-
hypertensive and drugs for lowering cholesterol)

• Nottingham prognostic Index (estimate long term risk of cancer reoccurrence or
  death in breast cancer patients)
• Simplified acute physiology score (SAPS) predict hospital mortality in critically ill
  patients)