lecture 6

Question 1

Analytic studies generally test

Answer 1

hypotheses regarding causal effects.
Did exposure cause disease?
Did the treatment cause an improvement?

Question 2

Aim in a research study is to

Answer 2

Our aim in a research study is to obtain a reliable estimate of the impact of the intervention or exposure on the outcome of interest.

isolate the exposure/intervention to obtain a reliable estimate `

Question 3

What do you need in a hypothesis to make it specific enough?

Answer 3

Defined population
a quantity of the variable
a time frame

this gives us enough information for us to test it in a controlled environment

Question 4

Relative risk

Answer 4

ratio of the probability of an outcome in an exposed group to the probability of an outcome in an unexposed group.

measures the association between the exposure and the outcome

Question 5

RR = 1 means

Answer 5

no relationship - not protective/beneficial/harmful effect

Question 6

null value of RR is

Answer 6

1

Question 7

Sources of error in analytic studies - For the estimate of RR from the study sample

Answer 7

Study design and conduct

Question 8

Sources of error in analytic studies - for the true RR in the population

Answer 8

Inference ….

• confounding
• bias
• chance (random variation)

Question 9

Confounding

Answer 9

Confounding is a distortion of the association between exposure and outcome caused by the presence of a third factor (one or more additional factors that is not measured).

Question 10

Confounder

Answer 10

a variable which causes confounding/this distortion

also known as a lurking variable - can give illusion of an effect of something when there is not

Question 11

To be a confounder, a variable must be both

Answer 11

associated with the exposure (independent of outcome); and

associated with the outcome (independent of exposure);

Question 12

What does a good study do?

Answer 12

Isolate effect, control everything else so that the thing that is being measured can just be associated with the treatment/exposure

Question 13

What happens to the causal pathway off there is a confounder present?

Answer 13

causal pathway (which shows the association) disappears if effect is caused by the confounder

Question 14

When might we get confounding in an intervention study?

Answer 14

Poor selection of people

Question 15

Bias types in an analytic study

Answer 15

selection bias and information bias

Question 16

Systematic error in analytic studies =

Answer 16

bias specifically selection bias and information bias

Question 17

Selection bias

Answer 17

poor selection of participants

Systematic error arising from the way participants are selected for inclusion in the study.

In an analytic study, selection bias occurs if the selection processes cause a systematic difference between the groups of participants selected for the study.

Prospective analytic studies rarely obtain participants through random sampling from a population. The issue of representativeness must be considered, but for analytic studies we consider it a generalisability issue rather than bias - ensure that when you draw your conclusion you look at your hypothesis to see how much you can hypothesise

Question 18

Why are cohort studies more susceptible to confounders?

Answer 18

due to selection bias

Question 19

Information bias

Answer 19

Systematic error arising from the way study information is obtained, interpreted and recorded.

In an analytic study, information bias is a particular problem if there are systematic differences in the information obtained from groups under comparison in the study.

Information bias may be introduced by the:
Observer
Study individual (respondent)
Instruments used to collect the data (e.g. badly-designed questionnaire) Missing measurements (e.g. from loss to follow-up in a prospective study)

Question 20

Non response and its relationship to information and selection bias

Answer 20

non-response to a measurement is generally considered information bias, whereas non-response to participation in the study can be considered a selection bias.

Question 21

Example of information bias

Answer 21

Interviewers different and have different levels of interviewing (some hard card, some only ask once and move on etc) therefore combat by randomising which interviewer gets assigned to each participant so the enthusiastic interviewer is interviewing roughly the same amount of participants in each group

Question 22

Example of selection bias

Answer 22

selection method is biased for some reason, something causes a systematic difference in who is assigned to each group

Question 23

Classification by purpose of study

Answer 23

Descriptive (describe things) vs. analytic (testing hypotheses).

Question 24

Classification by form of the design

Answer 24

Experimental (researcher intervenes) vs. observational (researcher observes).

Question 25

3 most common analytic studies

Answer 25

RCT, cohort and case-control These classifications provide a useful framework for thinking about the strengths and weaknesses of different study designs, but they will not always work.

Question 26

Randomised control trial

Answer 26

Analytic, experimental, prospective.

Question 27

Cohort study

Answer 27

Analytic, observational, usually prospective.

Question 28

Case-control study

Answer 28

Analytic, observational, retrospective.