Lecture 8: Screening & Diagnostic Tests 1 Flashcards
What is a test?
-Critical exam, observation, or evaluation
-Device to process designed to detect or quantify a sign, substance, or process in an individual or group
Used to classify
-Healthy vs sick
-Affected vs not affected
-Milk vs moderate vs severe
-Affected vs don’t know vs not affected
What is the difference between screening and diagnostic tests?
Screening (happens at Pathological process detectable but before symptoms)
-Focused on populations
-“healthy”
-Early detection of a pathological process or impaired productivity
-Sub-clincial disease
Diagnostic testing ( Happens in clinical disease evident)
-Focused on individuals
-“sick”
-Confirm, classify, guide treatment or aid in prognosis of clinical disease
What is a gold standard test?
-A test that is absolutely accurate but very few exist
-Diagnoses all individuals with disease and misdiagnoses none
-Allows us to determine whether a disease is truly present or not
What is the difference between True prevalence and apparent prevalence?
True Prevalence
-Actual level of disease present in population
-We dont know this so unless out test is perfect we are estimating TP apparent prevalence
Apparent prevalence
-What the prevalence “appears to be” if you use this particular test (going to be overestimated bc adding the false positives)
What are false pos and false negs?
False positives
-Cross-reacting with a similar disease agent
-Previous vaccination/exposure
-lab/test error
False negative
-Timing (daily variation in what the test is measuring)
-Test inhibitors, immunosuppression resulting in no production of antibodies
-Too early in disease course for response
-Lab/test error
How do we evaluate a test? What are the 2 questions we need to ask ourselves?
Assumptions to get us started:
-We have a gold standard test to determine which individuals are and are not diseased
-the outcome is dichotomous (ie diseased or not)
When considering a test, we need to ask 2 questions:
is it a good test?
-Tell us how well the test is able to pick up diseased and healthy animals
-ie if the individual is diseased, will the test identify them as so and vice versa)
-Sensitivity/Specificity
is it a useful test?
-Given that the test says individual is positive, what is probability that they actually have the disease (and vice versa)
Predictive values
What does sensitivity mean?
Sensitivity
-The proportion of truly diseased individuals (according to the gold standard) that test positive
-the ability of a test to identify truly diseased/infected
-If sensitivity is poor/low, then the test will miss finding diseased animals (increase false negative)
- 1-sensitivity= % of false negatives you expect
-If sensitivity is high, there will be few false negatives
What does specificity mean?
Specificity
-The proportion of truly healthy individuals that dont have the disease (according to the gold standard) that test negative
-if specificity is poor, then the test will call healthy individuals, diseased (increase false positive)
- 1-specificity= % of false positives
-If specificity is high, then there will be few false positives
What are some downfalls of sn and sp?
-They don’t tell us how useful the test might be when applied to individuals with unknown disease status
-Often want to know the probability that an individual has or doesn’t have the disease, depending on whether they test positive or negative–> predictive values
What are positive and negative predictive values?
Positive predictive value: (Predictive value of a positive test)
-Probability that given a test positive results, the individual actually has the disease
Negative predictive value: (Predictive value of a negative test)
-Probability that given a test negative result, the individual actually does not have the disease
What does the PPV mean?
-In an example 85% PPV means 85% of cows that tested positive using the ELISA actually have Johnes disease
-It also means that 15% of cows that test positive will actually be disease free
1-PPV= false positives
What does the NPV mean?
-In an example it means that 95% of cows that tested negative using the ELISA truly are disease free
-Only 5% of cows that tested negative will actually be diseased
What is the overall idea of the lecture?
-Sensitivity and specificity are qualities of a test that help describe how good it is
-Predictive values reflect the probability of the test results being correct and depend on: sn & sp & prevalence of disease in the population tested
-As prevalence decreases, PPV decreases, but NPV increases
-Neither sn nor sp are affected by prevalence
