POPM 3240 Lecture Notes - Lecture 3: Unimodality, Net Present Value, Positive And Negative Predictive Values
Document Summary
Screening test: used as a population approach on apparently healthy individuals to detect subclinical disease; provides early indication of a pathological disease process/limited health/ productivity outcome. Diagnostic tests: used on sick individuals with clinical disease to con rm, classify, inform treatment or provide prognosis for a particular disease. Validity of a test is when average of all tests results (half-way point of the curve) is near the true value; graph b & pg. How accurately a test measures an individual"s true biological value or overall infection status/disease state. Gold standard: perfect validity; all diagnosis are correct (no misdiagnoses); very few exist in reality due to test imperfections & biological variability. Reliability of a test thought of as the variation in the test results after it has been repeated numerously or deviation from the average result (width of the curve) Graph c: high reliability is indicated by a narrow distribution of test results (rather than a wide range of possibilities)