Ppal99 Dissecting The Usmle 1,2,3 Biostatistics

Front 1

Mean

Back 1

average

Front 2

Median:

Back 2

middle value

Front 3

Mode:

Back 3

most frequent value

Front 4

Endemic:

Back 4

localized epidemic

Front 5

Pandemic:

Back 5

world-·wide epidemic

Front 6

which of the following central tendency measures is sensitive to outliers?

Back 6

mean

Front 7

which of the following central tendency measures is used in a skewed population?

Back 7

median

Front 8

Variability:
what are the different standard deviations and their percentage?

Back 8

1 STD: 68%
2 STD: 95%
3 STD: 99%

Front 9

Chi square

Back 9

compare percentages

Front 10

T -test

Back 10

compare 2 things

Front 11

ANOVA

Back 11

compare 3+ things

Front 12

Skewing:

Back 12

shift to right (mean> median > mode)

Front 13

Standard Error of Mean

Back 13

= S/ (square root) of N => precision of mean
S = standard deviation, N = sample size

Front 14

1 -> 2 Question Survey:
what will happen to sensitivity? NPV? specificity? and PPV?

Back 14

Decrease: Sn, NPV
Increase: Sp, PPV

Front 15

PPV
equation
effect on prevalence
example

Back 15

= TP / all positives
(increases w / prevalence)
Ex: ( +) ELISA -> Do you have HIV?

Front 16

NPV
equation
define

Back 16

= TN/ all negatives
(probability of not having a dz if have negative test)

Front 17

Sensitivity

Back 17

= a/ a+c = TP /all diseased (people that have dz)

Front 18

Specificity

Back 18

= d/ b +d = TN/ all non-diseased (people that don't have dz) "

Front 19

Incidence
equation

Back 19

= new cases/ total population

Front 20

Prevalence= equation
example

Back 20

= all diseased/ total population
(Ex: improved quality of care)

Front 21

Odds Ratio= equation
what does it mean in layman's terms
• OR <1 =>
• OR >1 =>

Back 21

Odds Ratio= (ad+ bc) cross product
(diseased are x times more likely to see risk factor)
• OR <1 => protective factor
• OR >1 => risk factor

Front 22

Correlation Coefficient:
• 0:
• +:
• -:

Back 22

• 0: nothing
• +:correlation
• -: negative correlation

Front 23

CI

Back 23

= 95% => 95% sure it lies within the interval (cannot include 1.0)

Front 24

RR= equation
explain in layman's terms

Back 24

= exposed/ unexposed
(risk of getting dz w/ known exposure)

Front 25

define NNT
p value < 0.5

Back 25

= number needed to treat to change 1 life
p value < 0.5 = random chance that you will be wrong 1 time out of 20

Front 26

Null hypothesis

Back 26

=> nothing's happening

Front 27

Power:
equation
explain in layman's terms

Back 27

1-β = probability of detecting a true intervention

Front 28

what is improved by ⇧size of study

Back 28

power

Front 29

Effect size

Back 29

= how different two groups are

Front 30

α type I error

Back 30

(FN), P value error "too optimistic''

Front 31

β: type II error

Back 31

(FP), Power error "too pessimistic"

Front 32

what is the cause of type β: type II error?

Back 32

small samples

Front 33

Prevention:
1°:
2°:
3°:

Back 33

Prevention:
1°: ⇩Incidence
2°: ⇩Prevalence
3°: Slows disease progression

Front 34

Accuracy:

Back 34

validity "truth"

Front 35

Precision:

Back 35

reliability "keep making the same mistake"

Front 36

Admission rate:

Back 36

hospitals have certain populations

Front 37

Confounding bias:
define
what are the 3 things that should be done to prevent this

Back 37

forgotten variable => use matching, restriction, randomization

Front 38

Lead time:

Back 38

time between diagnosis and treatment

Front 39

Hawthorne effect:

Back 39

the "watched" change their behavior

Front 40

Unacceptability:

Back 40

subjects lie

Front 41

Observer bias:

Back 41

the observers have knowledge about control and study samples

Front 42

Recall bias:

Back 42

inaccurate recall of past events

Front 43

Respondant:

Back 43

subjective diagnosis

Front 44

Sample distortion:

Back 44

sample does not represent population

Front 45

Selection

Back 45

who is in or out

Front 46

I) Clinical Trials: experimental
Phase I:
Phase II:
Phase III:
Phase IV:

Back 46

Phase I: Toxicity "burt pt?"
Phase II: Efficacy "help pt?"
Phase III: Comparison "any better?"
Phase IV: Post-marketing surveillance "can they screw it up?"

Front 47

Cohort study:
define
what happens in a cohort study?
what does it determine?

Back 47

Prospective
Observe a "cohort" (group of people with similar characteristics) to see how many develop a specific
disease after exposure to a risk factor.
• Determines incidence: new cases of disease

Front 48

Cohort study:
what risk it uses
occurs where
ex

Back 48

• Uses Relative Risk
• Occurs in community
• Ex: finds new cases of common disease

Front 49

Case control:
type of study
what happens in a case control
what does it use?
where does it occur?
bias
examples

Back 49

Retrospective
• Select subjects with a disease, compare to controls, and study the differences
• Uses Odds Ratio
• Occurs in hospital
• Has more selection bias
• Ex: finds risk of developing rare disease

Front 50

Cross sectional:
define
what does it determine
example

Back 50

snaphot in time
• Determines prevalence: total cases of diseased
• Ex: polls/ surveys

Front 51

Case report

Back 51

describe an unusual pt

Front 52

Case series report

Back 52

describe several unusual pts

Front 53

Consensus panel

Back 53

panel of experts provides a recommendation

Front 54

Clinical wisdom

Back 54

"I think .... "

Front 55

Meta-Analysis

Back 55

tries to combine data from many trials