Diagnostic Test Accuracy Flashcards
How do test detect disease
disease status in binary - there is a cut off score on the test you and either positive or negative
What does the score overlap between people with the disease and healthily people cause
false negatives as due to overlap people can have the disease but have a score lower than the cut off point and false positives
What is the issue with near perfect diagnostics test
they are expensive and time consuming eg biopsy (called references standards)
quick, cheap alternatives are much more commonly used
How are new diagnostic tests tested for accuracy
compared against the reference standard test
Sensitivity and specificity definitions
statistics that quantify the intrinsic ability of a diagnostic tests accuracy
How is sensitivity calculated?
proportion of those positive patients that correctly tested positive
How is specificity calculated?
proportion of those negative that correctly tested negative
How does disease severity effect sensitivity
the more severe/advanced the disease is the more likely it is to be picked up by the test
How do other unrelated symptoms in negative patients effect specificity
those with symptoms of the target disease that have been caused by something else will damage the test
What do positive and negative predictive value do?
quantify the likelihood that somebody has the disease based on their test result
- more useful and relevant in medical setting
How is positive predictive value calculated?
proportion of those with a positive test result that are truly positive
How is negative predictive value calculated?
proportion of those with a negative test result that are truly negative
How does disease prevalence effect PPV and NPV
greater prevalence = ^ PPV, lower NPV
lower prevlance = lower PPV, ^ NPV
How can known % prevalence be used
can be used alongside the number of participants to work backwards and fill in the table ( +, -, impaired, not impaired
