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;
49 Cards in this Set
- Front
- Back
begins with "cases" and
looks for prior exposure among cases and among controls. |
case-control study
measure: OR |
|
case-control study.
measure? |
OR.
begins with "cases" and looks for prior exposure among cases and among controls. |
|
cohort study, begins with?
|
the exposure.
given the exposure. exp(+)-------> how many cases? exp(-)--------> how many cases? measure: Relative risk |
|
cohort study, measure?
|
RR.
Risk of the exposed/risk of the unexposed. |
|
twin concordance study - measures what?
|
heritability
|
|
postmarketing surveillance?
|
phase IV
|
|
large nr. of patients, randomly assigned to the new drug or placebo.
|
Phase III.
Compares the new tx to the current standard of care. |
|
Compares the new tx to the current standard of care. Also called: "the definitive test"
|
Phase III.
large nr. of patients, randomly assigned to the new drug or placebo. |
|
assesses efficacy, optimal dosing, adverse effects; which phase study?
|
Phase II.
|
|
assesses safety, toxicity, pharmacokinetics. which phase study?
|
Phase I.
healthy volunteers. |
|
SNOUT?
|
sensitivity rules out.
If negative, it is a true negative. (if %100 sensitivity) sensitivity is used for screening in diseases with low prevalence |
|
SPIN?
|
Specificity rules in.
If positive, it is a true positive. (if %100 specificity) so, specificity is used as a confirmatory test after a positive test result |
|
given a negative test result, probability of being true negative is:
|
NPV
|
|
given a positive test result, probability of being true positive is:
|
PPV
|
|
used for screening in diseases with low prevalence
|
sensitivity
sensitivity=detection of the disease |
|
used as a confirmatory test after a positive test result
|
specificity
|
|
1-false negative rate
|
sensitivity
|
|
1-false positive rate
|
specificity
|
|
prevalence equals incidence for ... diseases
|
acute
(e.g. common cold) |
|
for incidence calculation, shall we include current patients?
|
No.
|
|
prevalence = incidence x duration
true? |
True.
prevalence = incidence x duration |
|
Pr. of getting the disease in the exposed / Pr. of getting the disease in the unexposed
|
RR
(measure of Cohort studies) |
|
odds of having the disease in the exposed / odds of having the disease in the unexposed
|
OR
(measure of case-control studies) |
|
one is proportion (probability), one is odds. OR and RR.
|
OR: odds (case-control)
RR: probability (cohort) |
|
if we do not make "dividing", instead if we calculate the difference between the risk of the exp. and unexp. - ANSWER TO: "how many more cases in one group?"
|
Attributable risk
(risk of the exposed - risk of the unexposed) |
|
the reduction of risk associated with a tx compared to a placebo
|
absolute risk reduction
|
|
1/absolute risk reduction
(absolute risk reduction: the reduction of risk associated with a tx compared to a placebo) |
NNT
|
|
NNT?
|
1/absolute risk reduction
(absolute risk reduction: the reduction of risk associated with a tx compared to a placebo) |
|
the consistency and reproducability of a test
|
reliability=precision
also, the absence of random variation in a test. so, random error > reduces precision. |
|
trueness of measurements
|
accuracy=validity
systematic error - reduces accuracy (validity) in a test. |
|
random error reduces . . . . . . . in a test
|
precision
(reliability) |
|
systematic error reduces . . . . . . . in a test
|
accuracy
(validity) accuracy (validity) = how often the test is right overall accuracy= (TP+TN) / TP+TN+FP+FN |
|
accuracy (validity) = how often the test is right overall
true? |
true.
accuracy= (TP+TN) / TP+TN+FP+FN |
|
subjects are not representative. which bias?
|
sampling bias
|
|
each subject has its own control (in time). which study?
|
cross-over study
each subject receive intervention, but at different times. |
|
randomization limits mainly this bias:
(also known as non-respondent bias) |
selection bias
also known as non-respondent bias. bias in loss to follow ups. |
|
non-random assignment. which bias?
|
selection bias
also known as non-respondent bias. bias in loss to follow ups. |
|
mean + 1 SD ve mean - 1 SD aralığı (bell curve) % kaç?
|
%68
|
|
mean + 2 SD ve mean - 2 SD aralığı (bell curve) % kaç?
|
%95
|
|
mistakenly reject the null.
|
type I (alpha) error.
Pr. (making a type I error) is p. convicting an innocent man. you sAw. p<0.05 dediydin ya, işte bu olasılık gerçekleşmiş! |
|
stating that there is not a difference when one exists.
|
type II error.
Pr.(making a type II error) = Beta. you were Blind. Power = 1 - beta |
|
Pr. of making a type I (alpha) error is . . . . .
|
p
(generally 1/20 is accepted) |
|
1 - beta is
Pr.(making a type II error) = Beta. |
Power.
Pr. of rejecting the null, when it is in fact false. Sample size increases power. There is power in numbers. |
|
how about statistical power in meta-analyses?
|
high.
but it is prone to selection bias (study selection) |
|
nr. of persons "exposed to risk" during a time period
|
denominator of "incidence"
|
|
total population "at risk" at a given point or period
|
denominator of "prevalence"
|
|
what happens to incidence and prevalence if; new effective tx is initiated?
|
Incidence: NO CHANGE
Prev: DECREASES (Remember the incidence-prevalence pot) |
|
what happens to incidence and prevalence if; new effective vaccine gains widespread use?
|
Incidence: DECREASES
Prev: DECREASES (Remember the incidence-prevalence pot) |
|
what happens to incidence and prevalence if; long-term survival rates for the disease increases?
|
Incidence: NO CHANGE
Prevalence: INCREASES (Remember the incidence-prevalence pot) |