AFP Interview Flashcards

1
Q

Name the 6 levels of the hierarchy of evidence?

A

1) Meta-analysis/ Systematic Review (All relevant research regardless of design)
2) RCT (a type of cohort)
3) Cohort (Longitudinal observational - follows people with certain exposure over time)
4) Case control (Identifies two groups (with or w/o disease) and looks back for contributing factors
5) Cross sectional (snapshot of population, disease + exposure measured at same time)
6) Case series (collection of similar cases)

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

What is a cohort study?

A

Longitudinal observation study

Follows two groups with common exposure (smoker/ non-smoker) over time to determine risk of subsequent disease

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

What is a case-control study?

A

Retrospective observational

Identify 2 groups with different outcomes (disease vs. no disease) and looks back retrospectively for contributing factors

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

What is a cross sectional study?

A

Descriptive study
Takes a snapshot of a population measuring disease and exposure at the same time to make a link (but using individual data)

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

What is a case series study?

A

A collection of similar clinical cases derived from either one professional or a hospital etc

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

Name 5 pro’s of an RCT?

A
  • Studies effect
  • Controlled (balances the confounders)
  • Randomised
  • Powered for analysis
  • Blinding
  • Head-to-head comparison
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Name 5 con’s of an RCT?

A
• Expensive
• Time consuming
• Volunteer bias
• Not realistic clinical practice (e.g frequent
follow-ups)
• Underrepresent patients
• Can be ethically problematic
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Name 3 pro’s of a cohort study?

A

Safe (ethics)
Can study for risk factors
Realistic patients
Establishes timing

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

Name 3 con’s of a cohort study?

A

Can’t blind
Loss to follow up (attrition bias)
Needs large samples or long follow up for rare diseases

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

Name 3 pro’s of a case-control study?

A

Safe (ethics)
Quick and cheap
Small sample sizes
Good for rare outcomes/ diseases

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

Name 3 con’s of a case control study?

A

Recall bias
Selection bias
Difficult to select a control group

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

Name 2 pro’s of a cross-sectional study?

A

Quick and cheap

Ethically safe

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

Name 3 con’s of a cross-sectional study?

A

Association not causation
Recall bias
Unequal distribution of confounding variables

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

What is the difference between audit and research?

A

Audit - Are you doing what you should be doing?

Research - Looking for a better way of doing something

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

What is quality improvement?

A

A process by which we achieve better patient experience and outcomes

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

What is research governance?

A

The range of regulations, principles and standards of good practice which ensure high quality research

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

How is risk and risk ratio (same as relative risk) calculated?

A

AR = Absolute risk (No events / number people)

Risk ratio = (AR in treatment group/ AR in control group)

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

What does a risk ratio of 1 (or CI including 1) mean?

A

No significant difference between intervention and control

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

What does a risk ratio of 2.5 mean?

A

Events are more likely in the treatment group than the control group

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

What does a risk ratio of 0.4 mean?

A

Events are less likely in the treatment group than the control group

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

What is a type 1 error?

A

Rejection of null hypothesis when it was true (i.e. a false positive)

Think of a pregnant man (it’s obviously a false positive)

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

What is a type 2 error?

A

Failing to reject a false null hypothesis (false negative)

Think of someone who is pregnant but pregnancy test is negative < False negative (T2 error) is most dangerous because you are missing the disease

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

How is the credibility of a journal ranked?

A

Based on the number of times external work references an article in that journey/ number of articles

24
Q

What is the difference between a cohort study and an RCT? (2)

A
Cohort = Not randomised, looking at exposure (so better ethically)
RCT = Randomised, looking at intervention (may be more questionable)
25
Q

What observational study type is best for very rare outcomes?

A

Case-control

26
Q

What observational study type is best for very rare exposures?

A

Cohort

27
Q

Are case-control and cohort prospective or retrospective?

A
Cohort = Prospective (split groups according to exposure and look at outcomes) - note it CAN show causation 
Case-control = Retrospective (take cases and look back at exposures) - note it just shows association not causation
28
Q

What is a systematic review?

A

A rigorous protocol driven review of high quality literature on a specific area

29
Q

What is the difference between an ecological and cross-sectional study?

A
Cross-sectional = uses individual data
Ecological = uses population data
30
Q

What is CONSORT?

A

A validated tool used as a checklist to look at the design and reporting of an RCT?

31
Q

What is PRISMA?

A

A validated to used as a checklist to look at the design and reporting of meta-analysis and systematic review

32
Q

Name the benefits and drawbacks of SR/MA in terms of bias?

A

Reduces systematic bias (as more likely to find true result)

Increases publication bias (+ve result studies more likely to be published)

33
Q

What 4 biases should be commented on when appraising an abstract?

A

1) Selection bias
2) Performance bias
3) Observation bias
4) Attrition/ exclusion bias
5) Confounding factors

34
Q

What is selection bias and what are the two main ways it can arise?

A

1) How patients are chosen (must represent population)

2) If patient’s are randomised need to make sure intervention and control group are the same in terms of confounders

35
Q

What is performance bias?

A

Patients know which group they are in (no blinding = placebo effect)
OR
Healthcare professional knows the group and acts differently (not blinded)

36
Q

What is observation bias?

A

When you don’t measure control and intervention group using the same method
OR
Outcome of interest not defined well enough (so outcome treated a bit different in each group)
OR
Recall bias (in retrospective studies)

37
Q

What is attrition bias?

A

When patients drop out or don’t complete the study and you only analyse the one who completed (this is called per protocol analysis)

The way to get round this is Intention To Treat analysis (which includes those who dropped out)

38
Q

How do you reduce confounding factors?

A

Look at matching with your randomisation

Restrict your inclusion to those who do or don’t also have certain confounders

39
Q

What are the three key steps to appraising and presenting a paper?

A

1) Summarise in PICO format
2) Review internal validity (how good is the study, look at the 4 types of bias)
3) What do the results show? (Expand on primary outcome and add in some bits about any secondary outcomes, is there any needless speculation)
4) External validity (what is the clinical relevance)

40
Q

What is the difference between internal and external validity?

A

Internal validity - how good is the study

External validity - how relevant is the study to the real world/ clinical etc

41
Q

When would you use chi-squared, linear regression or t-test, ANOVA?

A

Both variables continuous = Linear regression
Dependant continuous, independent categorical = t-test (ANOVA if more than two groups)
Both categorical = Chi-squared

42
Q

What is a crossover study?

A

Intervention and control groups stop halfway through the study

43
Q

What is a confidence interval (95%)?

A

The range of values between which you are 95% sure the true value within the population lies

44
Q

What is the difference between incidence and prevalence?

A
Prevalence = Number of cases within the population at any given time
Incidence = Number of people who have outcome over a specified time period (i.e. 1yr)
45
Q

What is absolute and relative risk?

A

Absolute risk - Is risk of developing an outcome over time (i.e. 1 in 8 women get breast cancer)
Relative risk = Compares absolute risks in two groups (smoking makes you 3x more likely to get breast cancer). “Compares outcome of interest in the circumstances of with or without an intervention or exposure”.

46
Q

When do you use relative risk and when do you use odds ratio?

A

Relative risk used in RCT or cohort (prospective)

Odds ratio used in case-control (retrospective)

47
Q

What is an odds ratio?

A

The likelihood of an exposure in patients with the outcome of interest
(Odds of exposure in cases/ odds of exposure in control)

48
Q

What is absolute risk reduction and how is it calculated?

A

The fixed reduction in risk

AR in intervention - AR in controls

49
Q

How is number needed to treat (NNT) calculated?

A

NNT = 1/ ARR

50
Q

What is the difference between sensitivity and specificity?

A

Sensitivity - True positive rate (ability to correctly identify those with disease)
Specificity - True negative rate (ability to identify those without the disease)

51
Q

How is PPV calculated?

A

Positive predictive value
(Probability that subjects who tested positive, actually have the disease)

True positives / all those who tested positive

52
Q

How is NPV calculated?

A

Negative predictive value
(Probability that subjects who tested negative, do not have the disease)

True negatives / all those who tested negative

53
Q

What is z-value?

A

The z test is used to determine whether two population means are different (if variance is known and sample size is large)

54
Q

What is heterogeneity?

A

How similar or different two groups are

Homogeneity is same, heterogeneity is very different

55
Q

When is hazard ratio used?

A

Usually in survival analysis to describe the rate of outcome

i.e. death occurs at twice the rate in the control group vs. the treatment group

56
Q

When are kaplan-meier curves used?

A

As part of a survival analysis