Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
28 Cards in this Set
- Front
- Back
if you are given a PROSPECTIVE study, you should narrow your options down to what two choices?
|
RCCT
Cohort (prospective obviously) |
|
if you see "pts were randomly assigned to a group" you know that shit is....
|
absolutely RCCT
|
|
if you see "we reviewed the medical records" or in the previous three years what should you be thinking
|
retrospective
|
|
if you see prevalence, what can you narrow it down to?
|
ecological and cross-sectional
|
|
if you see 'matched controls' what should you be thinking
|
case-control
|
|
if you are calculating an Odds ratio for case control what is your equation
|
A/C
----- C/D |
|
if you are calculating an Odds ratio for cohort what is your equation
|
A/B
____ C/D |
|
errors associated with cohort?
|
recall (for retrospective)
can only use common diseases |
|
cohort is used to study what?
|
rare risk
|
|
error associated with RCCT
|
external validity problem (hard to generalize to larger population)
|
|
why are RCCT most rigorous
|
minimze bias
Double-blinded |
|
errors with cross-sectional
|
late-look bias (this is the one to know)
recall |
|
cross-sectional outcome?
|
prevalence (one point in time)
|
|
why are cross-sectional studies better than ecological?
|
looking at individual level data vs population
|
|
case-control studies what?
|
rare diseases
|
|
Rare risk vs Rare disease
|
rare disease: case-control
rare risk: cohort |
|
if you have 2 groups, and you have the 95% CI, what do the values have to be in order to be statistically significant?
|
they 95 CI values CANNOT OVERLAP
if they overlap there is NO STATISTICAL SIGNIFICANCE |
|
if RR is less than one with respect to women exercising and women getting breast cancer, what does this mean?
|
That exercising decreases the chance of getting breast cancer
if the value were 1, there would be no relationship (they were equal) if it was greater than 1 that would mean exercise would actually make it worse |
|
what type of data contains the least amount of info?
the most? |
least: nominal (things like color)
most: ratio (temp kalvin or BP) |
|
describe the diff btw skewness and kurtosis
|
skewness: horizontal
kurtosis: vertical (will look all bumpy and shiz) |
|
does mean=mode=median if you have a kurtosis?
|
noooopeerrss
|
|
what are the 5 steps in EBM
|
Ask
Acquire Appraise Apply Asses |
|
disadvantages of meta-analysis?
|
publication bias
fact that most of what is published tends to focus on positive findings (it can be hard to publish negative) |
|
what type of research study is considered to be most scientifically rigorous?
|
meta-analysis
|
|
internal vs. external validity
|
internal: study actually studies what it says it is
external: generalize the info you got to the population |
|
statistical power is which box?
|
D
|
|
if you increase statistical power what does this do?
|
higher likely hood that you can reject your H0 and find significance
|
|
how do you increase statistical power?
|
increase sample size
|