Tutorial 1 - Tools of the trade

Question 1

what is a sample?

Answer 1

• a relatively small number of observations (or patients) from which we try to describe the whole population that the sample was taken - usually random

Question 2

what is typically calculated from a sample?

Answer 2

• mean - confidence intervals to describe the range within which we think the population mean lies

Question 3

what is a null hypothesis?

Answer 3

the hypothesis that there is no significant difference between study populations, any differences being due to chance

Question 4

when is a null hypothesis rejected?

Answer 4

when p < 0.05

Question 5

what does a confidence interval of 95% show?

Answer 5

that if the sampling was repeated 100 times, the result observed would fall within the CI range in 95/100 samples

Question 6

what are 3 examples of measures of association?

Answer 6

• relative risk
• attributable risk
• odds ratio

Question 7

what are odds?

Answer 7

another way of expressing probability - the odds of exposure is the number of people who have been exposed/the number of people who have not been exposed

Question 8

what is the mathematical relationship between odds and probability?

Answer 8

odds = probability/(1-probability)

Question 9

what is relative risk?

Answer 9

a measure of association between an exposure and disease

Question 10

what does a relative risk value of 1 show?

Answer 10

the incidence of disease in the exposed and unexposed is the same so the data shows no association between the exposure and the disease

Question 11

what does a relative risk value >1 show?

Answer 11

positive association or an increased risk among those that are exposed

Question 12

what does a relative risk value <1 show?

Answer 12

inverse association or decreased risk among those that are exposed (exposure is protective)

Question 13

what is attributable risk?

Answer 13

a measure of exposure effect that indicates on an absolute scale how much greater the frequency of disease in the exposed group is compared with the unexposed, assuming the relationship between exposure and disease is causal

Question 14

how do you calculate relative risk?

Answer 14

incidence in the exposed group/the incidence in the unexposed group

Question 15

how do you calculate attributable risk?

Answer 15

(incidence in the exposed - incidence in the unexposed)/100

Question 16

draw a grid that can be used to calculate risk

Answer 16

Question 17

what is attributable risk especially useful for evaluating?

Answer 17

the impact of introduction/removal of risk factors

Question 18

what is odds ratio?

Answer 18

the odds ratio (of exposure) is the ratio between 2 odds

Question 19

how do you calculate odds ratio?

Answer 19

odds of exposure in the diseased (cases)/odds od exposure in the disease-free (controls)

Question 20

what does an odds ratio of 1 show?

Answer 20

exposure is no more likely in the cases than controls (therefore exposure has no effect on case/control status)

Question 21

what does an odds ratio >1 show?

Answer 21

exposure is more likely in the case group (so exposure may increase the risk of disease)

Question 22

what does an odds ratio <1 show?

Answer 22

exposure is less likely in the case group (therefore exposure may have a protective effect)

Question 23

what’s the difference between relative risk and odds ratio?

Answer 23

• relative risk can be calculated from cohort studies as the incidence of disease in the exposed and unexposed is known
• odds of exposure (and therefore odds ratio) can be calculated from case-control studies where the subjects are selected on the basis of their disease status rather than exposure

Question 24

list 3 methods for dealing with confounding factors

Answer 24

• stratification
• standardisation
• regression

Question 25

how is stratification a method for dealing with confounding factors?

Answer 25

controls the effect of confounders at the analysis stage as risks are calculated separately for each category of confounding variable (e.g. age, sex etc)

Question 26

how is standardisation a method for dealing with confounding factors?

Answer 26

controls the effect of confounders at the analysis stage as it is used to produce a Standardised Mortality Ratio

Question 27

how is regression a method for dealing with confounding factors?

Answer 27

controls the effect of confouders at the analysis stage as statistical modelling is used to control for one/many confounding variables