HadPop - Lecture 3

Question 1

What is a difference between an absolute and relative comparison?

Answer 1

Absolute comparison - a subtraction

Relative comparison - a ratio

Question 2

What are the two types of comparison?

Answer 2

• Absolute

- Relative

Question 3

What is another name for a absolute comparison?

Answer 3

A attributable risk

Question 4

Define an attributable risk

Answer 4

The excess risk of developing the disease in a exposed group compared to the unexposed group.

Question 5

What do relative comparisons tells us?

Answer 5

The strength of relationship between the exposure and disease

Question 6

How do you calculate IRR

Answer 6

Number of new cases of disease in a particular area at a set time / total at risk person-time of observation.

Question 7

How to calculate RR

Answer 7

cumulative incidence in exposed / cumulative incidence in unexposed.

Question 8

What is a confounding factor?

Answer 8

Something that is independently associated with both the outcome and exposure but NOT present on the casual pathway.

Question 9

Give an example of 2 general confounding factors?

Answer 9

Age

Gender

Question 10

How can we adjust for confounding?

Answer 10

Design stage:

• Randomisation
• Stratification
• Restriction

Analysis stage:

• Standardisation
• Stratification
• Multi-variant analysis.

Question 11

What is standardisation?

Answer 11

method of adjusting for age differences to make a valid comparison.

Question 12

What are the two methods of standardisation?

Answer 12

• direct standisation

- in direct standisation

Question 13

How do you calculate standardised mortality rate?

Answer 13

Total number of observed deaths in index population / total number of expected deaths in index population

Question 14

How do you calculate standardised mortality rate?

Answer 14

Total number of observed deaths in index population / total number of expected deaths in index population
*X 100

If the SMR = 100 - then risk in the index population and reference population is the same

If the SMR > 100 - then the risk in the index population is higher.

Question 15

Is SMR direct or indirect standardisation?

Answer 15

In-direct

Question 16

What is a key feature of sampling?

Answer 16

Must be a representative of the population

Question 17

What is a sampling frame?

Answer 17

A list of people from whom a sample was taken.

Question 18

What are the two methods of sampling?

Answer 18

1) Random sampling

2) Non random sampling

Question 19

Discuss random and non random sampling.

Answer 19

Random sampling = everyone in the sampling frame has a equal chance of being chosen, hence you obtain a representative sample.

Non-random sampling = this is easier and more convenient however, the sample may not be representative.

Question 20

Describe the random sampling methods.

Answer 20

1) Simple random sampling - everyone is a number a they are selected randomly
2) Stratified sampling - Population split into two groups. People then picked from the group based on simple random sampling methods
3) Cluster Sampling - Using natural clusters in the population e..g school
4) Systematic sampling - Picking the Nth candidate.

Question 21

What are the two types of errors and what are they?

Answer 21

Random error - A error which is due to chance ALONE

Systematic error - this is a error which is due to either sampling or measurement

Question 22

Name the two main sources of bias and there subcategories:

Answer 22

1) Selection Bias
- Sampling bias (non representative sample)
- Non response bias

2) Instrument bias
- Equipment bias
- interviewer bias (due to things like training)

Question 23

What is response rate?

Answer 23

the percentage of the selected sample that take part in the survey / study

Question 24

Define precision

Answer 24

How close the results obtained are to each other

Question 25

Define accuracy

Answer 25

How close are the results to the TRUE value.

Question 26

What two components test the performance of your study?

Answer 26

1) Validity - How well did the test measure what it was suppose to measure
2) Repeatability - do you get the same results when redoing the experiment later on?