Biostats 4- Cohort Studies Flashcards
This is a prospective (longitudinal) study design that compares the frequency of disease between 2 groups (exposed to risk and non-exposed) as it develops over TIME.
Cohort study
This is the risk of someone developing a disease over a time-period.
Absolute risk
Do you use rows or columns when determining AR?
rows
This is the risk that compares the risk between 2 different groups of people.
Relative Risk (risk ratio)
What is the eqn for RR?
RR = (AR of disease with independent factor)/(AR of disease withOUT independent factor)
What is the significance of a RR < 1.0?
There is a negative association between the disease and risk.
This is a retrospective study design that compares diseased and non-diseased groups on whether they were exposed toa hypothesized risk factor in the PAST.
Case-Control study
What is the eqn for Odds Ratio (OR)?
OR = (odds of exposure among those WITH disease)/(odds of exposure among those withOUT disease)
What is the significance of an OR > 1.0?
The disease is more likely to happen than not
This is the probability that an event will occur divided by the probability that the event will not occur.
Odd
For OR calculations, do u use columns or rows?
Columns
What is the eqn for the absolute risk ratio (ARR)?
ARR = CER - EER
CER = Control event rate ERR= Experimental event rate
What is the eqn for Relative risk reduction (RRR)?
RRR = ARR/CER
The inverse of the absolute risk reduction gives us what property?
The # needed to treat
if ARR = 20%, then 5 pts must be treated to save 1
This quantifies the harm caused by an intervention compated with the control.
needed to HARM
This is a longitudinal study design in which similar subjects are randomly allocated to 2 groups where the 2nd group is manipulated (treatment and control) and followed over time.
RCT
What is the double-blind method used in RCT’s?
where the tester doesnt know and the pt doesnt know whether theyre getting the treatment or placebo
What are the 2 tests u use when comparing a mean to a population?
CI or Z-test
What are the 2 tests u use when comapring group means?
T-test or ANOVA
What are the 2 tests u use when measuring association?
Correlation or Regression analysis
This is the test type to estimate an unknown population mean from sample data.
Confidence intervals.
What is the significance if the calculated value is WITHIN the 95% CI?
The value or result is NOT statistically significant
This is the half-width of a CI, and is a measure of the precision of an estimate for a given level of confidence.
Margin of Error (MOE)
What 2 things does a MOE reflect?
Mean and esimate of variability.
True or False: if the MOE > the % difference, you cannot estimate the results.
True
What is the eqn for SE?
SE = SD/(sample size)^-1/2
What is the eqn for CI?
CI = mean +/- CL(value)*SE
This is the tests that determins whethere a significant difference exists between a samples proportion (count) and the KNOWN population.
Z-test
