Statistics

Question 1

Sensitivity - Definition & Calculation

Answer 1

Proportion of patients with a condition who have a positive test result.
TP/(TP + FN)

Question 2

Specificity - Definition & Calculation

Answer 2

Proportion of patients without a condition who have a negative test result.
TN / (TN + FP)

Question 3

Positive Predictive Value - Definition & Calculation

Answer 3

Chance that a patient actually has the condition given a positive test.
TP / (TP + FP)

Question 4

Negative Predictive Value - Definition & Calculation

Answer 4

Chance that a patient does not have the condition given a negative test
TN / (TN + FN)

Question 5

Likelihood Ratio (Positive) - Definition & Calculation

Answer 5

How much the odds of a disease increase when a test is positive.
Sensitivity/(1-Specificity)

Question 6

Likelihood Ratio (Negative) - Definition & Calculation

Answer 6

How much the odds of a disease decrease when a test is negative.
(1 - Sensitivity) / Specificity

Question 7

Numbers Needed to Treat (NNT) - Definition & Calculation

Answer 7

Measure indicative of how many patients would require an intervention to reduce the number of outcomes by one - 1 / Absolute Risk Reduction) - Expressed as a whole #

Question 8

Absolute Risk Reduction - Calculation

Answer 8

Control Event Rate - Experimental Event Rate

Question 9

Relative Risk (Risk Ratio) - Calculation

Answer 9

Experimental Event Rate / Control Event Rate

Question 10

Relative Risk Reduction / Increase (RRR/RRI) - Calculation

Answer 10

RRR = (EER - CER) / CER

Question 11

Odds / Odds Ratio - Definition and Calculation

Answer 11

Odds - A ratio of people who incur an outcome to the number who do not incur e.g. 80/20 (80 out of 100).

Odds Ratio - Ratio of the odds of a particular outcome with that of a control (Used in Case-control studies).

Question 12

Stages of Clinical Trials

Answer 12

Stage 0 - Exploratory Studies - Small Numbers of participants assessing pharmacokinetics/dynamics.
Stage 1 - Safety Assessment - Determines side effects prior to larger studies (Healthy Volunteers)
Stage 2 - Efficacy Assessment - Small # of patients w/ a disease.
- Stage IIA - Assesses optimal dosing.
- Stage IIB - Assesses Efficacy
Stage 3 - Effectiveness Assessment - thousands of patients in RCT trials comparing the novel treatment to a placebo.
Stage 4 - Postmarketing Surveillance - Monitors for long-term side effects.

Question 13

Randomised Control Trial (RCT)

Answer 13

Random Allocation to a control or intervention group. (Can have practical and ethical problems)

Question 14

Cohort Study

Answer 14

Observational and Prospective
Groups selected based on exposures and are followed up to determine relative risks of outcomes.

Question 15

Case-Control Study

Answer 15

Observational & Retrospective
Patients are selected based on outcomes and are surveyed to determine exposure (Prone to confounding.
Outcome measure is usually odds ratio

Question 16

What is Categorical Data

Answer 16

Data w/ Names & no logical order

Question 17

What is Ordinal Data

Answer 17

Data with a logical order but unequal gaps between values

Question 18

What is Continuous Data

Answer 18

Data on a scale with equal gaps between values.

Question 19

Features of a “Perfect” Normal (Gaussian) Distribution

Answer 19

Mean = Mode = Median
68.3 % of Values lie within 1 SD of Mean
95.4 % of Values lie within 2 SD of Mean
99.7 % of Values lie within 3 SD of Mean
95 % of Values lie within 1.96 SD of the Mean

Question 20

Characteristics of Skewed Gaussian Distribution

Answer 20

Positive Skew Mean> Median > Mode
Negative Skew Mean < Median < Mode

Question 21

What is the Null Hypothesis (H0)

Answer 21

Two treatments are equally effective (Significance tests assess how likely H0 is to be true)

Question 22

What is the Alternative Hypothesis (H1)

Answer 22

The opposite of H0 - there is a difference between the two treatments

Question 23

What is the P-Value

Answer 23

The probability of obtaining a result that is at least as extreme as the observed (Chance of making a Type I Error)

Question 24

What is a Type I Error

Answer 24

The null hypothesis is rejected when it is true

Question 25

What is a Type II Error

Answer 25

The null hypothesis is accepted when it is false.

Question 26

How is Statistical Power Calculated

Answer 26

Power = 1 - The chance of Type II Error

Question 27

How to Calculate Standard Error of Mean (SEM) & the 95% Confidence interval?

Answer 27

SEM = SD/root n (n = sample size) 95% CI = Mean +/- (1.96 * SEM)

Question 28

What is Confounding?

Answer 28

The phenomenon when an apparent association between A & B is actually due to an unmeasured variable C.

Question 29

What is Selection Bias?

Answer 29

Errors assigning individuals to groups where doing so may influence outcomes.

Question 30

What is Sampling Bias?

Answer 30

Subjects are not representative of the population. - Volunteer Bias - Certain groups are more likely to volunteer for a study. - Non-responder Bias - Certain groups are less likely to respond to surveillance.

Question 31

What is Recall Bias?

Answer 31

Differences in accuracy of recollections retrieved by study participants.

Question 32

What is Publication Bias?

Answer 32

Publications are more likely to be accepted if they have positive findings.

Question 33

What is Expectation Bias?

Answer 33

Observers may subconciously rport data in a way that favours the expected outcome.

Question 34

What is the Hawthorne Effect?

Answer 34

Described a group changing its behaviour due to the knowledge that it is being observed.

Question 35

What is Late-look Bias?

Answer 35

Gathering information at the wrong time

Question 36

What is Lead-time Bias?

Answer 36

Seen when a test identifies a condition earlier giving the impression that survival times have increased.

Question 37

What are Funnel Plots used for?

Answer 37

They are primarily used to detect publication bias in meta analyses. A symmetrical inverted funnel shape indicates that publication bias is unlikely.