Statistics and Epidemiology Flashcards

1
Q

What is standard deviation?

A

Square root of variance

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

What is variance?

A

Standard deviation squared

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

What is the definition of sensitivity?

A

The proportion of individuals with the disease that were correctly identified by the test

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

What is the definition of specificity?

A

The proportion of individuals who were confirmed not to have the disease who were correctly identified as normal by the test

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

What is a method of estimating distribution?

A

Histogram

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

What test is used to compare means between two separate groups of PARAMETRIC data?

A

Independant student t-test

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

What test is used to compare means between two separate groups of NON-PARAMETRIC data?

A

Mann-Whitney U

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

What test is used to compare results between two paired sample of PARAMETRIC data?

A

Paired student t-test

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

What test is used to compare results between two paired sample of NON-PARAMETRIC data?

A

Wilcoxon signed rank

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

What test is used to compare three or more measurements on one subject of PARAMETRIC data?

A

Repeated ANOVA

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

What test is used to compare three or more measurements on one subject of NON-PARAMETRIC data?

A

Friedman

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

What test is used to compare one variable between three or more separate variables of PARAMETRIC data?

A

One-way ANOVA

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

What test is used to compare one variable between three or more separate variables of NON-PARAMETRIC data?

A

Kruskal Wallis

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

What test is used to look for relationships between two continuous variables of PARAMETRIC data? (i.e. correlation)

A

Pearson coefficient
(Pearson = parametric)

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

What test is used to look for relationships between two continuous variables of NON-PARAMETRIC data? (i.e. correlation)

A

Spearman coefficient

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

What is level of evidence IV?

A

Panel of experts - weakest form of evidence

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

What is level of evidence III?

A

Evidence from case, correlation, and comparative studies

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

What is level of evidence IIb?

A

Evidence from at least one well designed experimental trial

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

What is level of evidence IIa?

A

Evidence from at least one well designed controlled trial which is not randomized

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

What is level of evidence Ib?

A

Evidence from at least one Randomized Controlled Trial

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

What is level of evidence Ia?

A

Evidence from Meta-analysis of Randomized Controlled Trials

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

How is SEM (standard error of the mean) calculated?

A

SEM = SD/square root of sample size

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

What is standard error?

A

Allows you to calculate how different the sample mean is likely to be from the population mean
It is a proxy for sampling error
Large SE - LESS likely to represent population
Small SE - MORE likely to represent a population

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

How do you calculate 95% confidence intervals using SEM?

A

(mean - 1.96SEM) to (mean + 1.96SEM)

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

What is the absolute risk?

A

= the actual risk

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

What is relative risk?

A

= the ratio of risk comparing the risk in exposed vs. controlled groups

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

How is sensitivity calculated?

A

A/(A+C)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
28
Q

How is specificity calculated?

A

D/(B+D)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
29
Q

How is negative predictive value (NPV) calculated?

A

D/(C+D)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
30
Q

What is the NPV?

A

If the result is negative what is the chance that the patient is ACTUALLY negative

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
31
Q

How is positive predictive value (PPV) calculated?

A

A/(A+B)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
32
Q

What is the PPV?

A

If the result is positive, what is the chance that the patient is ACTUALLY positive

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
33
Q

What is the negative likelihood ratio?

A

(1-sensitivity)/specifity

A multiplier for how much the risk of having the condition is decreased if the test is negative

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
34
Q

What is the positive likelihood ratio?

A

sensitivity/(1-specifity)

A multiplier for how much more the patient is likely to have the condition if the result is positive

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
35
Q

What is the most appropriate method to assess prognosis?

A

Cohort study

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
36
Q

What do parametric tests assume?

A

Assume a normal distribution of population data for the variable being tested and are used for testing variables within a population that are interval or ratio e.g. height, temperature or age

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
37
Q

What type of tests are used when the population distribution is not known, or variables are of a nominal or ordinal value e.g. eye colour or marital status?

A

Non-parametric

38
Q

What are the parametric tests?

A

Pearson; t-test; f-test; ANOVA; z-test

39
Q

What are the non-parametric test?

A

Spearman; Mann-Whitney; Chi2; Wilcoxen signed rank; Fisher exact probability; Kruskal Wallis; Friedman

40
Q

What are type I errors?

A

Incorrect rejection of a null hypothesis
False +ve
Detecting an effect that is not present

41
Q

What are the type II errors?

A

Incorrect acceptance of a null hypothesis
False -ve
Failing to detect an effect that is present

42
Q

What are ROC curves?

A

y = sensitivity
x = 1-specificty

43
Q

What is the most appropriate method to assess diagnostic tests?

A

Cross-sectional studies

44
Q

What is the most appropriate method to assess treatment/intervention?

A

RCT

45
Q

How should the power of a study be described?

A

The study’s ability to detect an effect of a specified size?

46
Q

What is the definition of incidence?

A

Number of new cases of a disease (usually per 100,000) that occur in a population previously free of it

47
Q

What is the definition of prevalence?

A

The frequency of existing disease at a given time

E.g. prevalence of ovarian cancer greatest in 60-69 age group, but incidence is greatest in the 80+ group

48
Q

What is a Kaplan-Meier plot used to graphically represent?

A

Probability of survival

49
Q

What are case-control studies useful for?

A

Rare diseases

50
Q

What are different types of categorical data?

A

Nominal/binary

Ordinal

51
Q

What are different types of quantitive data?

A

Continuous

Discrete

52
Q

What is mean sensitive to?

A

Outliers (extreme values)

53
Q

The higher the power of a study, the lower the probability of what?

A

The higher the power of a study, the lower the probability of type II errors

54
Q

High sensitivity is required for which sort of tests?

A

Screening tests, thus minimising missed cases

55
Q

High specificity is required for which sort of tests?

A

Diagnostic tests, thus minimising misdiagnoses

56
Q

What is the only thing the magnitude of the number needed to harm (NNH) depends on?

A

The absolute difference between treatment
groups in risk of the adverse event

57
Q

What does concealed allocation aim to do in an RCT?

A

Eliminate selection bias

58
Q

What are multifactoral RCTs?

A

When 2+ interventions are compared to the control

59
Q

What are crossover RCTs?

A

Where each participant is their own control (i.e. they have 2 different treatments at 2 different times)

60
Q

What is the principle of ‘intention to treat’ in RCTs?

A

The two groups of patients were compared on the basis of the treatment regimen assigned at the start of the trial

61
Q

What is the principle of ‘per protocol analysis’ in RCTs?

A

Only those fully compliant are included

62
Q

What is interim analysis in RCTs?

A

An independant body that monitors the treatment effects and early signs that it is working. Good for large/ethically dubious trials

63
Q

What are odds?

A

No. of time an event happens / no. of times an event DOESN’T happen

64
Q

What is an odds ration?

A

Odds of exposure to a RF in case group / Odds of exposure to a RF in control group

65
Q

What is absolute risk reduction (ARR)?

A

Difference in event rate in the intervention group and control group - i.e. the improvement in the intervention group minus the improvement in control group

66
Q

What is relative risk reduction?

A

ARR/placebo no improvement event rate

67
Q

What is NNT?

A

100/ARR

68
Q

What is relative risk?

A

Relative risk (RR) is the ratio of risk in an exposed group to a non exposed group

RR = Probability of an event when exposed/Probability of event in control group

69
Q

What is an absolute risk increase?

A

The absolute difference between treatment groups in the risk of adverse event

70
Q

What is the NNH?

A

100/ARI

71
Q

How do you calculate 95% CI for mean?

A

(Mean - 1.96xSEM) to (Mean +1.96xSEM)

72
Q

How do you calculate the upper CI limit?

A

Mean + (1.96 x SEM)

73
Q

How do you calculate the lower CI limit?

A

Mean - (1.96 x SEM)

74
Q

When would a risk be considered ‘very common’?

A

1 in 1 to 1 in 10

75
Q

When would a risk be considered ‘common’?

A

1 in 10 to 1 in 100

76
Q

When would a risk be considered ‘uncommon’?

A

1 in 100 to 1 in 1000

77
Q

When would a risk be considered ‘rare’?

A

1 in 1000 to 1 in 10000

78
Q

When would a risk be considered ‘very rare’?

A

Less than 1 in 10 000

79
Q

What is the ideal NNT?

A

1

80
Q

What is acquiescence bias?

A

‘Yes’ bias - when participants tend to agree with you/what you are saying

81
Q

What is reporting bias?

A

When researchers selectively report their findings/make omissions

82
Q

When is median used?

A

When date is skewed

83
Q

What does it mean when the CI includes zero?

A

> 5% chance (with 95% CIs) that there was no true change + the intervention is ineffective

84
Q

What is a p value?

A

The probability of any observed difference having happened by chance - i.e. lower the p value the more likely the effect is to be true

85
Q

What does concealed allocation aim to do?

A

To eliminate selection bias

86
Q

What does randomisation aim to do?

A

Prevent selection bias

87
Q

What does double-blinding aim to do?

A

To eliminate measurement bias

88
Q

What is exclusion bias?

A

When there are systematic difference in those that have withdrawn from the trial

89
Q

What is evidence is required to meet an FSRH ‘good practice point’?

A

Clinical experience of the guideline development group?

90
Q

Are ascertainment and sampling bias the same thing?

A

Yes

91
Q

Are measurement and detection bias the same thing?

A

Yes