10 Understanding and Appraising Evidence

Question 1

Q: What is prevalence? What do prevalent cases include?

Answer 1

A: Number of cases of a disease within a defined population measured at a specific point in time

Prevalent cases include both new (incident) and existing cases

Question 2

Q: How do you think the prevalence of HIV was measured in the two tribes?

Answer 2

A:  Blood or saliva testing (e.g. ELISA)

 Self-report?

Question 3

Q: What are the problems with determining prevalence? (4)

Answer 3

A: When assessing ‘point prevalence’, need to consider whose data was and was not available on the ‘census day’; i.e. participation bias
 What is the relevant denominator population?
 Self-reporting errors (of HIV and circumcision)
 Sensitivity (positive/true positives) and specificity (negatives/true negatives) of the HIV test.

Question 4

Q: What is incidence?

Answer 4

A: Overall incidence = new cases in cohort/total number in cohort

Question 5

Q: What is relative risk?

Answer 5

A: The relative risk is the incidence in the exposed group divided by the incidence in the non-exposed group

Question 6

Q: What are the possible explanations for a study’s findings? (4)

Answer 6

A: biologically plausible
Chance…especially in small studies
Bias
Confounding

Question 7

Q: How can bias be minimised?

Answer 7

A: careful study design and aggregating data from different types of study

Question 8

Q: Should we circumcise men for HIV prevention?

Answer 8

A: Why?
Circumcision could dramatically reduce HIV infection rates in men and women, and prevent the epidemic spiraling further out of control until a vaccine, or other preventative strategy, is identified

Why not?
Circumcised men are not invulnerable, they are still able to become infected; circumcision should therefore be used as part of a wider strategy (e.g. encouraging condom use etc.) to reduce infection rates