Brainscape's Knowledge GenomeTM

Browse over 1 million classes created by top students, professors, publishers, and experts.


See full index

Tutorials > EBM Final > Flashcards

EBM Final Flashcards

1
Q

What is the null hypothesis?

A

There is no significant difference between the groups being compared

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What is the p-value?

A

The probability of obtaining the observed result (or one more extreme) if the null hypothesis were true

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What is a confidence interval?

A

Quantifies the uncertainty in measurement - range of values in which we can be 95% confident that the true value for the population lies.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What is Type I error? (alpha)

A

The probability of rejecting the null hypothesis when it is actually true. (False positive)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

What is Type II error? (beta)

A

The probability of accepting the null hypothesis when it is actually false. (False negative)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What is “intention to treat” analysis?

A

We analyze outcomes based on the initial randomization, regardless of whether patient received the treatment or crossed over

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

What is relative risk and when is it used?

A

The ratio of risks between an experimental and control group. Used in cohort and RCT studies.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

What is absolute risk reduction?

A

The absolute difference between the exposed (experimental) group and the unexposed (control) group. It is used to estimate number needed to treat.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

What is the relative risk reduction?

A

The proportion of baseline risk reduced by the intervention (absolute risk reduction/absolute risk in control population).

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

How do you calculate the number needed to treat?

A

1/absolute risk reduction

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

How do you calculate the number needed to harm?

A

1/absolute risk increase

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

What is the sensitivity?

A

The probability of a positive test in people with the disease

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

What is the specificity?

A

The probability of a negative result in people without the disease

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

What is the positive predictive value?

A

The proportion of people who test positive who truly have the disease

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

What is the negative predictive value?

A

The proportion of people who test negative who truly do not have the disease

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

What is the likelihood ratio of a positive test?

A

The probability of an abnormal result in the population with the disease divided by the probability of that abnormal result in the population without the disease. A positive likelihood ratio of 6 means that the test result is 6 times more likely to occur in a patient with the disease than a patient without the disease.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

How do you calculate the likelihood ratio of a positive test?

A

LR+ = sensitivity/(1-specificity)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

How do you calculate the likelihood ratio of a negative test?

A

The probability of a normal result in the population without the disease divided by the probability of that normal result in the population with the disease.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

How do you calculate the likelihood ratio of a negative test?

A

LR- = (1-sensitivity)/specificity

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q

What is censoring?

A

In RCTs, it means that the duration of follow up is not the same in all surviving subjects, either because some were lost to follow up or because the study ended. Survival methods (like Kaplan Meier curves, log-rank tests, Cox models, etc) take censoring into account when comparing two groups.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
21
Q

What is a Kaplan-Meier curve?

A

It’s a survival curve that begins at 100% and plots the cumulative probability of survival in each study arm over time.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
22
Q

What is a Cox proportional Hazards Model?

A

Analyzes time-to-event data, comparing hazard functions between arms. Hazard = risk per unit time. It applies the principle of censoring and allows adjustment for potential confounders

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
23
Q

What is a case-control study?

A

A study that examines associations between exposure and outcome by sampling subjects with the desired outcome and comparing them to those without the outcome. Most important concern is bias (particularly recall bias).

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
24
Q

What is a Phase I trial?

A

Phase I: Researchers test a new drug or treatment in a small group of people for the first time to evaluate its safety, determine a safe dosage range, and identify side effects.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
25
Q

What is a Phase II trial?

A

Phase II: The drug or treatment is given to a larger group of people to see if it is effective and to further evaluate its safety.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
26
Q

What is a Phase III trial?

A

Phase III: The drug or treatment is given to large groups of people to confirm its effectiveness, monitor side effects, compare it to commonly used treatments, and collect information that will allow the drug or treatment to be used safely.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
27
Q

What is a Phase IV trial?

A

Phase IV: Studies are done after the drug or treatment has been marketed to gather information on the drug’s effect in various populations and any side effects associated with long-term use.

28
Q

What is the acceptable amount of Type I error?

A

Typically 5%

29
Q

What is the acceptable amount of Type II error?

A

Typically 20%

30
Q

What is the required amount of power?

A

80%. Power is 1-(Type II error), which means 1-0.2=0.8

31
Q

What might cause a decrease in power?

A 
Decreased number of events in control group
Lots of loss to follow up
Lots of cross over
Poor adherence
Increase in variability
32
Q

What is internal validity?

A

Homogeneity created by exclusion criteria ensuring that all participants benefit from treatment in a predictable manner

33
Q

What is external validity?

A

Generalizability of the study - the more exclusion criteria (and the more internal validity), the less external validity a study has

34
Q

What is allocation concealment?

A

Blinded randomization - reduces confounding and bias

35
Q

Do you see confounding in RCTs?

A

Not very large ones - they prevent confounding through the random assignment of a large number of people to the different study groups
Also stratify by site in large studies

36
Q

How can you reduce confounding in small RCTs?

A

Stratify the randomization

37
Q

Why would you stop a trial?

A
  1. Proven benefit
  2. Probable harm
  3. Futility (no difference)
38
Q

What is bias? Can it be controlled for?

A

Bias is error within the structure of the study that affects who is in the study, how we measure exposure or outcome, or how we analyze the results. It CAN’T be controlled for!

39
Q

What is effect modification?

A

Variation in the magnitude of measure of effect across levels of a third variable, which is not independently associated with either the exposure or the outcome.Effect modification is not a bias but useful information.
Remember the smoking and OCPs example

40
Q

What is an interaction?

A

When a treatment is shown to work differently depending on a given patient characteristic (race, gender, etc.), we call
that an interaction.

41
Q

What is an incident case?

A

Incident cases are those that are newly diagnosed (ideally those in whom the disease has just developed) with the disease of interest. In other words, all new cases of the disease.

42
Q

What is a prevalent case?

A

Prevalent cases are those whose disease

developed or was diagnosed before they were identified for the study. In other words, all cases, new and old.

43
Q

When is the odds ratio used?

A

Case-control study

Examines the probability of the exposure given the outcome

44
Q

How do you fill out a 2x2 table?

A

GO LOOK NOW

45
Q

How do you calculate sensitivity?

A

true positive/(true positive + false negative)

46
Q

How do you calculate specificity?

A

true negative/(false positive + true negative)

47
Q

How do you calculate positive predictive value?

A

True positive/(true positive + false positive)

48
Q

How do you calculate negative predictive value?

A

True negative/(true negative + false negative)

49
Q

When should a diagnostic test be used?

A

If patient has intermediate pretest probability.
If extremely low pretest probability, then a diagnostic test will just require additional testing. If extremely high, then we have already made the diagnosis and shouldn’t waste the time or resources.

50
Q

When should screening be performed?

A

• An accurate diagnostic test is available
• The disease is relatively prevalent
• The disease carries substantial morbidity
and/or mortality if left untreated
• There is an efficacious treatment for it, and
early treatment is better than late treatment

51
Q

How do we assess interactions when looking at subgroup analyses?

A

If the confidence intervals are not superimposed at all, we can be
certain that there is a significant difference between the two subgroups in regards to that variable/result. That is, the P-value for
that INTERACTION will be < 0.05. If either point estimate overlaps with the CI of the other subgroup, we can be certain that isn’t a significant difference between the two
subgroups in regards to that variable/result. That is, the P-value for that INTERACTION will be > 0.05. If there is minimal overlap of the CI’s then we are not sure: CI’s are
usually conservatively estimated, and there might be a tiny overlap when the P-value for an INTERACTION is slightly < 0.05.

52
Q

What is the checklist for a negative study?

A

• Did they recruit the number of patients they
intended to?
• Was the observed outcome in the control
group similar to what investigators
expected?
• Were the attrition (losses to follow up) and
cross-overs as infrequent as expected?
• Was the intervention carried out as planned?
IF YOU ANSWER “YES” TO ALL QUESTIONS,
THEN IT IS PROBABLY A TRULY NEGATIVE
STUDY.

53
Q

How do you calculate relative risk?

A

event rate treatment/event rate control

The event rates in each group are calculated as number of events/number of patients randomized to the group

54
Q

What are the limitations of relative risk?

A

Does not account for differences in follow up or confounders

55
Q

What is the checklist for the meta analysis?

A

• Was the publication search comprehensive?
• Is there evidence of publication bias?
• Is the quality of the individual studies appropriate?
• Were the populations and interventions similar to my own?
• Was there heterogeneity among selected studies?
• What method was used to pool the results? Fixed vs. random effects
• What are the pooled estimates?
• Were those estimates consistent acrosts sensitivity
and sub-group analysis?
• What is the clinical significance of the results?

56
Q

What is publication bias?

A

Authors and editors don’t often publish negative studies, particularly when they’re small, which can result in an exaggerated estimate of association or effect of interest.

57
Q

How can you tell if there is publication bias?

A

truncated funnel plot

58
Q

What is a non-inferiority trial and why do we need it?

A

Placebo-controlled are no longer ethical because there is available treatment.
The new treatment is expected to have similar efficacy, as compared to available
treatment/s and to have advantages in at
least one of the following: safety, convenience, tolerability, acceptability, or
cost.

59
Q

What is the difference in the null hypothesis in a non-inferiority trial?

A

• A new treatment is considered as noninferior to the available, older, treatment if it shows efficacy that is better, equal, or
“not significantly worse” than that of the old treatment.
• We accept some harm as being “not significantly worse”: how much harm we accept as part of noninferiority is the crucial issue.

60
Q

What is selection bias?

A

Do we choose the people in exposed/unexposed differently and in a way that could also be associated with the outcome?
Loss to follow up (attrition): If people are less likely to stay in the study based on their exposure status, it can affect the results

61
Q

What is prescription bias?

A

Prescription bias (aka confounding by indication or channeling bias): For studies involving medications, the decision to give a certain medication can be affected by other factors that also affect the outcome

62
Q

What is recall bias?

A

Need for interviews to collect data on past exposures

Those with disease are more likely to recall exposures than those without disease

63
Q

What is immortal time bias?

A

Cohort study: people are studied when they have an exposure, have to survive until they get the exposure, “protected” from dying/getting outcome before the exposure.
Those without exposure might die/get outcome very early

64
Q

What is a cross-sectional study?

A

Aka “prevalence study”
Define a population, gather data on presence/absence of exposure and disease at that time for each individual.
Not an attempt to collect data on exposures that existed PRIOR TO the development of disease

65
Q

How do we test for heterogeneity in a meta analysis?

A

I2 values less than 25, 25-50%, and 50-75% suggest low, medium, and substantial heterogeneity, respectively.
I2 greater than 75% suggests excessive heterogeneity — do not meta-analyze!

Tutorials (2 decks)

Brainscape helps you reach your goals faster, through stronger study habits.
© 2025 Bold Learning Solutions. Terms and Conditions