14 - Cross Sectional and Cohort Studies

Question 1

What are some confounding variables in studies?

Answer 1

• Age
• Sex
• Ethnicity
• Deprivation
• Smoking
• Alcohol

Question 2

Where does the observed ratio lie in terms of the 95% confidence interaval?

Answer 2

ALWAYS WITHIN THE CONFIDENCE INTERVAL

Question 3

How can you conclude that a finding is statistically significant?

Answer 3

Question 4

How do you work out 95% confidence limits?

Answer 4

• 95% CI is the range within which one may be 95% certain that the true value of the quantity of interest really lies.
• If 95% CI includes the value corresponding to the null hypothesis p>0.05 if null hypothesis outside CI then p<0.05

Question 5

What are some issues with cross-sectional surveys?

Answer 5

• Sampling bias
• Responder bias
• Chance

Question 6

How can you compare two groups of individuals internally in a cohort study?

Answer 6

• Incident rate ratio

Question 7

How do you calculate 95% CI?

Answer 7

Worked example!

Question 8

What are the advantages of cohort studies over routinely available data?

Answer 8

Question 9

How does error factor relate to subcohort sizes in a study?

Answer 9

Small sub-cohort then larger error factor

Question 10

How can you compare two groups of individuals externally in a cohort study?

Answer 10

- SMR: standardised mortality ratio. represents increase or decrease in mortality of a study cohort with respect to the general population

• Work out error factor using expected in Lexis diagram so you can work of 95% CI by dividing and multiplying the SMR by the e.f

Question 11

What are the limitations of external comparison in cohort studies?

Answer 11

• Limited data available for reference population
• Often no incidence data
• Usually have to make do with mortality data
• Study and reference populations may not be comparable (selection bias e.g healthy worker effect in study)

Question 12

What are the advantages and disadvantages of cohort studies?

Answer 12

Question 13

What are some specific issues with cohort studies?

Answer 13

Question 14

Why may it appear that there is an increase in SIDS deaths from 76 to 88 but there actually is not and what can account for the falls from 88 to 92?

Answer 14

• WHO defined SIDS in the 70s so there is a transfer of diagnosis making SIDs to appear to increase
• Could be chance

Question 15

What does a high sensitivity and high specificity in a study correlate to?

Answer 15

Question 16

Using the following data, how well did the patients self administered HADS-D predict MDE?

Answer 16

Predict: want as few false negatives as possible so looking at sensitivity

Question 17

How would you answer a patients question who has come for a cervical smear test and is asking ‘what is the likelihood of me getting a high risk score and having cervical cancer’?

Answer 17

• Would work out PPV
• High Risk +ve / Total High Risk

Question 18

What table should you draw when faced with trying to work out PPV, NPV etc

Answer 18