Exam 3 Flashcards
List the types of biostatistical tests.
- T-test
- Paired T-test
- ANOVA
- Chi-square
(X) is the value to look at in order to determine if association exists.
X = p-value
(X) is the value to look at in order to determine the direction of an association.
X = regression coefficient (if it’s positive or negative)
(X)-values are used to determine (Y)-values, which tell us the significance of an association based on biostat test.
X = t Y = p
T/F: In multiple linear/logistic regression, values should be used only if significant association, based on p-value.
False - use even if not significant associatoin
Anytime you want to do a prediction, you should use (X) biostat tool.
X = regression (linear or logistic)
Anytime you want to do a prediction with dichotimous outcome, you should use (X) biostat tool.
X = logistic regression
You can use logistic regression if the predictor is (continuous/categorical).
Either!
Equation for maternal mortality ratio.
No. maternal deaths/No. live births
If you’re following a cohort over a period of time (to see if outcome develops, for example), which general type of measure will you be presenting?
A rate - any measure with time component is RATE
Loss to follow-up is a form of (X) bias.
X = selection
Phenomenon in which a
third variable, that is not associated with both exposure and outcome, modifies their relationship.
Effect modification
Phenomenon in which a
third variable, that is associated with both exposure and outcome, modifies their relationship.
Confounding
In order to examine relationships while controlling for confounders, we have to use
which statistical method?.
Regression
(X) is the probability of a negative result given the presence of disease.
X = false negative rate
(X) is the probability of negative test given absence of disease.
X = specificity
(X) is the probability of a positive test given presence of disease.
X = sensitivity
(X) is the probability of getting a positive test result given the absence of disease.
X = false positive rate
(X) is the probability of disease, given a positive test.
X = positive predictive value
T/F: Sensitivity and specificity are inherent characteristics of the test.
True
T/F: Predictive values (pos and neg) are inherent characteristics of the test.
False - depend on prevalence too
Absolute number of patients who would need to be treated to prevent one instance of the
bad outcome.
Number needed to treat
Absolute number of patients who would need to be treated for a bad outcome (ie. side effect) to occur.
Number needed to harm
When calculating (absolute/relative) risk reduction, you subtract (X) from 1.
Relative;
X = Risk ratio/relative risk/incidence ratio