Intro To Epidemiological Invest

Question 1

Epidemiology

Answer 1

• study of distribution + determinants of health related events
• application of study to control

Question 2

Studies for distribution

Answer 2

Surveillance

Description

Question 3

Studies for Determinants

Answer 3

Analytical

Question 4

Statistics definition

Answer 4

• Collection, presentation, description & analysis of data
• same principles irrespective of field application

Question 5

Probability vs Statisics

Answer 5

Statistics - scale up (via sample)

Probability - scale down

Question 6

Sampling error

Answer 6

Variation in the samples (give / take 2%…)

Question 7

Measures of Frequency

Answer 7

How common

• Prevalence (existing cases)
• Incidence Rate (measuring new cases)
• Prevalence (P= I x D)

Question 8

Proportion

Answer 8

Proportion of people who already have it (per 1,000…)

• point prevalence (one time)
• period prevalence (across time period)
• no units (people/people)
• cross sectional survey

Question 9

Incidence rate

Answer 9

Number of new cases (not people)

• period at risk = period being counted
• person-years

New events
————— = events per person per time
Person x time

Question 10

Incidence rate: Person-Years

Answer 10

Sum total time everyone in study

3 people followed up for 2 years +
7 people followed up for 0.1 years
= (3 x 2) + (7 x 0.1)
= 6.7 p-y

Question 11

Prevelence

Answer 11

Prevalence = incidence x disease duration

Question 12

Measures of Random Variation

Answer 12

-

Question 13

Dealing with Random Variation

Answer 13

• need distinguish btw observed vs true

- confidence interval

Question 14

Tendency vs Observation

Answer 14

Tendency = true or underlying
Observed = best estimate of true

Question 15

Hypotheses

Answer 15

About true value

• statement: underlying tendency has particular quantitative value
• if probability sml –> hypothesis incompatible
• probability = p-value
• reject hypothesis not observation

Question 16

When to reject hypothesis

Answer 16

P-value < 0.05

• reject hypothesis
• observations statistically significant
• P-value >0.05 doesn’t prove hypothesis just doesn’t disprove it (i.e. Not guilty)

Question 17

Limitations of hypothesis testing

Answer 17

Not always helpful to reject hypothesis

• sample size may restrict
• statistically significant doesn’t equal clinically relevant

Question 18

Null Hypothesis

Answer 18

• two things are equal (no effect / difference)
• Usually test against it
• p < 0.05 doesn’t prove null hypothesis false (strong suggest)

Question 19

Odds Ratio

Answer 19

• ‘times as likely’
• 2 IVs comparison
• snapshot (one period)

Odds being exposed if case vs odds exposed if control
Odds being case if exposed vs odds being case if unexposed

• no cause and effect (only association)

Question 20

Incidence Rate Ratio

Answer 20

• risk of outcome over time
• can estab causeation

IRR = IR(expo) / IR(unexpo)