Stats Uworld + Notes Flashcards
Meta-analysis is what kind of study?
T-test?
Pulls from multiple sources
Compares means (interval data) between 2 groups
1 SD means what?
2?
3?
68%
95%
99.7%
What is the SNNOUT/SPPIN mnemonic?
Highly sensitive test when Negative, rules OUT the disease
If you have a PAP smear that is negative, you can r/o cervical cancer
Highly specificity test when Positive, rules IN the disease
If you have a + Colposcopy, you probably have cervical cancer
In what situations can the OR generally approximate the RR?
Rare diseases
LR(+) = ?
Always think what?
Sensitivity / (1 - specificity)
Sensitivity / specificity
Then manipulate the 1
If you have the 2 curves graph and then you move the 2 curves to make them smaller, what happens to specificity and sensitivity?
Why?
Both increase
Area of FN/FP is getting smaller when you smash the graph down
High sensitivity test used for screening in diseases with what?
Specificity used for what?
Low prevalence
After a + screening test
Odds ratio means what?
1 # over 1 #
Cross-sectional study is what kind of study design?
Examines what?
Observational
Associations and prevalence
ARR = ?
Event rate of control (in %) - Event rate of treatment (in %)
Case control mnemonic?
CaDORF
Based on disease
Odds ratio
Risk factor frequency
What is the relationship of mean, median, mode in + skew?
Mean > median > mode
What is a cause of lead-time bias?
Occurs in what context?
Detecting the disease earlier
Develop a new test and says it is better
Odds ratio equation?
(ad) / (bc)
Loss to follow-up creates what kind of bias?
Selection bias known as attrition bias
What kind of study utilizes 2 or more different interventions w/2 different variables studied independently?
Factorial
What is the formula for Attributable risk % (ARP)?
Question will ask you what?
ARP = (RR - 1) / RR
What % of blah blah blah
Variance = ?
SD squared
On the graph with the curves, if they increase the cutoff value what happens to FP and FN?
FP decreases
FN increases
- skew, what is the mean, median, mode like?
Mean < median < mode
What is tertiary prevention?
Quarternary?
Treat disease support people w/it
Not repeating tests
What is statistical power =?
1 - beta
Relative risk equation?
a/(a+b) / c/(c+d)
Checking BP at a health fair is what kind of prevention?
Describe it
Secondary
Screening for disease early to reduce impact
Cohort mnemonic?
CoExRI
Based on exposure
Risk Ratio
Incidence
Study says no association but in reality there is one, what kind of error?
Beta error (type 2)
Incidence = ?
of new cases / # of people at risk
May have to subtract from denominator the people that have it
Test shows 2x2 table and asks what kind of test is this?
Chi-square
If a study is looking into diseases with a slowly progressive, BENIGN case, what type of bias are they worried about?
Length time bias
CI formula = ?
mean +/- Z(SE)
Z = 1.96 for 95% CI SE = SD x (sq RF of n)
On the graph with the curves, if they lower the cutoff value what happens to FP and FN?
FN decreases
FP increases
Median of 20, 22, 23, 26 is what?
22.5
Take the average of 2 middle numbers
Confounding bias means what?
How to control for it?
Some outside influence is related to the exposure and the outcome
Matching
NNT = ?
1 / ARR
How does increased prevalence effect PPV?
NPV?
Inc
Dec
NNH = ?
1 / AR
A study that says there is an association but in reality there is not, what type of error is this?
Alpha error (type 1)
Cross-sectional study compares what?
Prevalence
CroP
ANOVA compares what?
Compares means of 3 or more groups
A-NOV-A —> 3 syllables
What is observer bias?
Investigator’s decision is adversely affected by knowledge of the exposure status