Disease Occurrence and Effect Measures

Incidence proportion?

Probability (or risk) that a person will develop a disease over a specified time period

NEW cases of disease

• Numerator → # of people developing disease during the time period

  Denominator → # of people followed for the time period

What is problematic about incidence proportion (risk)?

It can be biased because it does not fully account for competing risks (death) or loss to follow-up, which can distort the true probability of developing disease

What is incidence rate?

The rate of disease occurrence using person-time.

• Numerator: new cases of disease

• Denominator: total time people are at risk (person-time)

How is person-time handled in incidence rate?

People contribute time while at risk:

Stop contributing after disease occurs (if disease occurs once)

Can re-enter time if disease is recurring

Stop contributing after death or loss to follow-up (no longer at risk)

Incidence proportion vs rate ?

Incidence proportion (risk): “WHO got it” → counts people

Incidence rate: “HOW FAST it happens” → counts person-time

Prevalence proportion?

The proportion of a population that has a disease at a specific point in time

• Numerator → # of people with the disease

• Denominator → total population size

What is risk difference / incidence proportion difference?

The difference in incidence proportion (risk) between exposed and unexposed groups

Measures how much extra risk is due to exposure

• Incidence proportion (exposed) - Incidence proportion (unexposed)

Incidence rate difference?

The difference in incidence rates (person-time rates) between exposed and unexposed groups

• Incidence rate (exposed) - incidence rate (unexposed)

How do you interpret a measure of association (difference)?

< 0: exposed group has lower risk/rate than unexposed

= 0: no difference between groups

> 0: exposed group has higher risk/rate than unexposed

Risk ratio (RR) or Relative risk?

Risk in exposed / Risk in unexposed

Compares the risk (incidence proportion) of disease between two groups

Incidence rate ratio (IRR)?

How much faster (or slower) disease occurs in the exposed group compared to the unexposed group

• Incidence rate (exposed) / incidence rate (unexposed)

Odds?

Risk = probability disease will develop → p

Odds = p / 1 - p

In case-control studies, you usually cannot calculate risk, so you use odds ratio

How are odds ratio (OR) and risk ratio (RR) different, and when are they similar?

OR and RR are not the same.

When disease probability is small → OR ≈ RR

When probability is large, odds and risk diverge → OR can overestimate RR

They are only similar when the outcome is rare