Sensitivity And Specificity Flashcards
What is sensitivity?
The number of true positives correctly identified out of all the actual positives.
True positives/ true positive + false negative.
What is specificity?
The number of negatives correctly identified out of all the actual negatives.
True negative/ true negatives + false positives
What is the positive predictive value?
The proportion of positive results that are actually positive.
True positive/ true positive + false positive
What are the negative predictive value?
Proportion of negative results that are actually negative.
True negative/ true negative + false negative.
What is the null hypothesis?
The reasoning that there will be no difference when administering a new experimental drug.
What is a type 1 error?
False positive/ Rejection of the null hypothesis when it is true
What is a type 2 error?
False negative/ Accepting the null hypothesis when it’s false.
What is a lead time bias?
Survival seems longer because diagnosis occurs earlier, such as through screening.
What is a length time bias?
Overestimation of survival because Patients with slowly progressing diseases are more likely to be detected through screening than fast-progressing tumours, because fast-progressing tumours are detected through symptoms.
What is a case control study?
Selecting people for an outcome and asking them to recall for exposure. It is calculated using odds ratio.
It is cheap and easy to conduct, but recall bias is an issue
What is a cohort study?
Selecting a group of people based on an exposure and following them to see if they develop an outcome. It is calculated using relative risk.
It is expensive and risk of people dropping out of the study, and not ideal for rare disease. However it is good for risk factor analysis and determining disease time sequence.
What is a cross-sectional study?
Snapshot in time to provide information about a population and prevalence, such as questionnaires.
How to calculate odds ratio?
Odds of being exposed in the case/ odds of being exposed in control
How to calculate relative risk?
Risk in the exposed/ risk in the control.
What is ‘per protocol” analysis?
Analysing data from a Randomised controlled trial by removing any participant which did not comply with their treatment in either the experimental or control.