Appraising Pronostic Research Flashcards
Which research design is MOST COMMONLY associated with prognostic studies?
Observational studies
In cohort studies, Subjects who are likely to develop a certain condition or outcome, are recruited _____ the condition has occurred. For what?
Before,
To see if they develop the condition
How are groups in cohort studies separated?
Based on exposure
Prospective cohort study
Gather cohort> Determine exposure status> Followup to determine outcome.
Retrospective Cohort
Identify Cohort> Determine exposure status> Determine outcome
Which study is conducted after an outcome has occurred?
Case-Control Study
When subjects are studied over time, it is considered
Longitudinal
When subject data collected at one point in time?
Cross-sectional
Factors other than exposure that could account for the results in a study are
Cofounders
The greater the number of factors in a study, ______
The greater the sample must be
Prognostic studies are of _____ not ______
Association
Causality
What are the most appropriate statistics for analysis of Prognostic Studies?>
Association statistics
Most prognostic studies are
Multi-factorial
A measure of the extent to which two variables are associated, a measure of how these variables change together
Correlation Statistics
mathematical statement regarding the extent of the association between the two variables (-1 to 1)
Correlation coefficient r
What happens with negative correlations?
As one variable increases, the other decreases
What happens with Positive correlation?
As one variable increases, so does the other
An “r” of 0 means
There is no correlation
0-.25 .26-.49 .5-.69 .7-.89 .9-1
Strengths
Little to none Low Moderate High Very high
expresses the percentage of variance that is shared by the two variables
Coefficient of determination r^2
R^2 of 0 means
Of 1 means
Y cannot be predicted from X
Y can be predicted from X w/o error
used to make prediction from 1+ variables to a variable of interest
Regression analysis
the process of determining a regression equation to predict Y values based on a linear relationship with X values
Simple linear regression
Y=a+bX
Y= outcome
A=y intercept
B+slope
Is r a predictor of accuracy of prediction from regression line?
NAWW
statistical technique to establish the predictive relationship between one DV (Y) and multiple IV (X), DV is continuous
Multiple Regression
multiple regression procedure with dichotomous DV; predicts odds associated with presence or absence DV (Y) based on IVs (X)
Logistic Regression
What can only be used in cohort studies?
Risk ratios
Risk ratio of 1 indicates
Unchanged risk compared to control
Risk ratio of >1 means
Increased risk
Risk ratio of <1 is
Decreased risk
Odds of an event occurring in one group relative to the odds that it will occur in another group
Odds ratio
Odds ratio can be used in which studies?
BOTH cohort and case-control
