Step 1 Epi/biostats + Ethics

Front 1

Which varies with disease prevalence, PPV and NPV or sensitivity and specificity?

Back 1

PPV and NPV!

Front 2

Equatation for PPV?

Back 2

TP/(TP+FP)

Front 3

Equation for NPV?

Back 3

TN/(TN+FN)

Front 4

Equation for sensitivity?

Back 4

TP/(TP+FN)

Front 5

Equation for specificity?

Back 5

TN(TN+FP)

Front 6

What is absolute risk reduction? (ARR)

Back 6

Percent of disease in placebo - percent of disease in treatment group.

Front 7

How do you calculate NNT?

Back 7

1/ARR (ARR = percent of disease in placebo group - percent of disease in treatment ggroup)

Front 8

What is attributable risk?

Back 8

If 21% of smokers develop lung cancer and 1% of non smokers, 20% attributable risk to smoking.

Front 9

How do you calculate NNH?

Back 9

1/attributable risk (AR=% in group with risk factor who get disease - percent in group w/o risk factor)

Front 10

How do you calculate # of false negatives given the sensitivity and specificity of a study and the total number of positives and negatives? # of False positives?

Back 10

Sensitivity = TP/TP+FN. So, 1-sensitivity*total positives

Front 11

What is capitation? What does it incentivize?

Back 11

Physicians are paid by the number of enrollees they have, not by tests/procedures. Health maintenance and counseling are incentivized because these reduce costs in the long run.

Front 12

What is the Hawthorn effect?

Back 12

When a group being studied changes its behavior because they know they are being watched

Front 13

What is Berkson's bias?

Back 13

A type of selection bias from choosing hospitalized patients as a control group

Front 14

What is the pygmalion effect?

Back 14

When a researcher's belief in the efficacy of a treatment changes the outcome of that treatment. (aka, observer expectancy effect)

Front 15

What is recall bias?

Back 15

Subjects fail to accurately recall events in the past. (problem with retrospective studies)

Front 16

What is sampling bias?

Back 16

Subjects are not representative of the general population, therefore results are not generalizable. A type of selection bias

Front 17

What type of bias is loss to f/u?

Back 17

Selection bias!

Front 18

What is late-look bias?

Back 18

Individuals with severe disease are less likely to be uncovered in a survey bc they die first.

Front 19

What is procedure bias?

Back 19

Different treatment arms are treated differenty (ie, actually treatment arm gets more attention, and that affects outcome)

Front 20

Case control and cohort studies - which measure OR and which RR?

Back 20

Case contol - OR (odds of having had the exposure), cohort study - RR (RR of developing disease)

Front 21

Can case-control studies be prospective?

Back 21

No! Only observational of retrospective

Front 22

Matching (by age,m race, etc) in a study help control for what bias?

Back 22

Confounding (by age, race, etc). Matched things should be potential confounders.

Front 23

If there is an effect between a risk and a disease occuring, but then when stratified by some other factor that could effect both the risk and the disease, the significance goes away, what's going on?

Back 23

Confounding!

Front 24

If the p value of a study is just above 0.05, questions probably wants you to say the study was not significant because of...

Back 24

Sample size

Front 25

How do you calculate the CI?

Back 25

Mean +/- Z[SEM], where SEM = SD/square root n. Z = 2 for 95% CI, 2.5 for 99% CI

Front 26

How do you calculate risk?

Back 26

# of people who got disease/# of people who were at risk (DONT OVERTHINK IT)

Front 27

What are Type 1 and Type 2 error.

Back 27

Type 1 = upper right hand corner of box (alpha) = conlcuding there is a difference when there is not.
Type 2 = Lower left hand corner of the box, B, concluding there is no difference when there is!

Front 28

How do you calculate power? What is the definition?

Back 28

1-Beta (Beta = chance of concluding there is no difference when there is!) Power is the chance of seeing a difference when there is actually one!

Front 29

3 most common cancers in women by incidence and mortality?

Back 29

Incidence: breat>lung>colon. Mortality: lung>breast>colon

Front 30

Order mean, median and mode in terms of size for positive and negative skew:

Back 30

Positive skew: mean>median>mode. Negative skew: mode>median>mean.

Front 31

A t test looks at the difference between 2...?

Back 31

Means!

Front 32

An ANOVA looks at the difference between 3 of more...?

Back 32

means!

Front 33

What is a chi square test used for?

Back 33

To test for an association between two categorical variable (do things happen together by chance or by some kind of relation)

Front 34

Who participates in a phase I clinical trial? What is it looking at?

Back 34

Small # of healthy volunteers. Asseses safety, toxicty and pharmacokinetics

Front 35

Who participates in a phase II clinical trial? What is it looking at?

Back 35

Small # of ppl with disease. Asseses efficacy, dosing and adverse effects.

Front 36

Who participates in a phase III clinical trial? What is it looking at?

Back 36

Large # of ppl with disease assigned to an intervention and a control (best availble tx) arm. compares new treatment to standard of care.

Front 37

Who participates in a phase IV clinical trial? What is it looking at??

Back 37

Post marketing surveilance of pts after approval. Detecs rare or long term effects

Front 38

What is effect modification? Is it a bias?

Back 38

Effect modification is when the effect of the main exposure on the outcome measure is modified by another factor (effect of estrogen receptor on DVT modified by smoking status). NOT a bias, rather, a thing to be described

Front 39

As prevalence increases, what happens to NPV?

Back 39

It decreases! Can think of as prevalence of negatives in decreasing.

Front 40

Do NPV and PPV vary with pretest probability?

Back 40

Yes!

Front 41

What is alpha usually set at?

Back 41

0.05! 5% chance that correlation is actually due to chance

Front 42

What is the most preventative intervention ever in all people forever and always?

Back 42

Quitting smoking!

Front 43

What is the median of the data set 75, 75, 80,90,110,110?

Back 43

85! The median is the number that is located in the precise center of an ordered data set. If the data set has an een number of point, it is the number between the 2 central numbers.

Front 44

What are the ares under the cover within 1, 2, and 3 SDs of the mean?

Back 44

68, 95, 99.7 (1/2 of that on each end of the bell curve)

Front 45

How do you calculate the odds ratio? What is it in words?

Back 45

Think about case control study design. Odds that a patient who developed a disease was exposed to the risk factor vs patient who didn't develop the disease was exposed to it. a/c//b/d.

Front 46

When are a CI of 0 and 1 important?

Back 46

0 = when comparing the different between two means, 1 = when comparing two odds ratios

Front 47

Give an example of primary, secondary and tertiary prevention of cervical cancer?

Back 47

1 = HPV vaccine 2 = pap smear (early detection) 3 = chemo

Front 48

What is selection bias?

Back 48

The sample selected is not representative of the population

Front 49

What is lead time bias?

Back 49

Early detection confused with increased survival bc increased disease population

Front 50

What is a cross over study?

Back 50

All subjects receive intervention but at different times, either to increase power or bc unethical to deny anyone the intervention

Front 51

Types of Disease prevention:
1) Primary -
2) Secondary -
3) Tertiary -

Back 51

1) Primary - Prevent disease occurrence (e.g. vaccine)
2) Secondary - Early detection (screening)
3) Tertiary - Reduce disability from disease (i.e. treat but not cure)

Front 52

Relationship between sd and variance?

Back 52

The variance is the sd squared
(i.e. the sd is the square root of variance)

Front 53

Types of scales:
1) Nominal -
2) Ordinal -
3) Interval -

Back 53

1) Nominal - Bins without order
2) Ordinal - Bins with order
3) Interval - Continuous values

Front 54

Chi squared is used for what?

Back 54

Nominal data analysis

Front 55

Relative Risk equation

Back 55

a/(a+b) / c/(c+d)
The incidence of one divided by the incidence of the other

Front 56

Odds Ratio equation

Back 56

(a/b) / (c/d)
The ODDS of getting disease for exposed DIVIDED by the ODDS of getting disease for not exposed

Front 57

Attributable risk is?

Back 57

Incidence of A - Incidence of B
Tells us the Absolute increase in risk bc of A

Front 58

What is absolute risk reduction?

Back 58

Absolute value in percentiles that risk is reduced by intervention

Front 59

What is number needed to treat?

Back 59

1/absolute risk reduction

Front 60

What is number needed to harm?

Back 60

1/Attributable risk
(i.e. the absolute attributable risk)

Front 61

Significance of Quinlan case?

Back 61

Substituted Judgment Standard
The view most likely to represent a recently incompetent patient's best wishes is honored.

Front 62

Significance of Brother Fox (Eichner v Dillon)?

Back 62

Best Interest Standard
Never competent individuals cannot have substituted judgment so BEST INTEREST STANDARD is honored (everyone should come to same conclusion about what is best for patient)

Front 63

Significance of Infant Doe?

Back 63

Foregoing lifesaving surgery, parents withholding treatment if deemed in best interest of child (congenital anomalies)

Front 64

Significance of Roe v Wade?

Back 64

A) Abortion legalized
B) Patients can refuse care (even if it affects fetus)

Front 65

Significance of Tarasoff?

Back 65

Duty to warn others if patient is homicidal
1) Detain patient
2) Notify authorities
3) Warn victim

Front 66

Indications patient is not competent
(3)

Back 66

1) Suicidal
2) Grossly psychotic or dysfunctional
3) Communication not possible
*Competence is legal, capacitance is medical

Front 67

Surrogate decision maker hierarchy
(3)

Back 67

1) Advanced directive (oral or written)
2) Substituted judgment standard
3) Best interest standard

Front 68

Exceptions to informed consent
(4)

Back 68

1) Emergency
2) Waiver by patient
3) Patient is incompetent
4) Therapeutic privilege (pt unconscious, confused)

Front 69

Exceptions to child's incompetence until 18
(3+4)

Back 69

Emancipation
1) If older than 13 and living alone/self sufficient
2) Married
3) In military
*Pregnancy or parenthood NOT

Partial emancipation for 16 or older re:
1) Substance abuse tx
2) Prenatal care
3) STD tx
4) Birth control

Front 70

Parents withholding treatment

Back 70

1) If emergency --> treat
2) If critical to survival --> child becomes ward of state and treated
3) If not lifesaving --> honor parents' refusal

Front 71

Good samaritan law

Back 71

1) Not required to help
2) If help, protected if actions are w/in your competence level and w/in standard of care.
3) Must stay at scene until signout
4) No compensation

Front 72

Confidentiality exceptions:

Back 72

1) Duty to warn others of harm (HI or abuse)
2) Duty to protect patient from self (SI)
3) Infectious disease: must tell those at risk