3: Age standarization

Question 1

Why do we need it?

If we dont use it, what can we use instead?

Answer 1

Why to we need it?
If we want to compare incidence risks in 2 areas we may have a problem:

Different age structures of populations in 2 areas leads to difference incidences cause disease is age-related

Age standarization allwos for fair comparison of frequencies of age-related diseases between groups with different age structures

• e.g. Clostridium dificle mostly affects old people

If we dont use it, what can we use instead?
age-specific risks for analysis

Question 2

Direct standarization
- What do we get?
- How do we get it?
- What do you need to consider?

Answer 2

What do we get?
Incidence or mortality rate corrected for age structure, same scale as crude incidence risk

-> Age standadized Rate (ASR)

How do we get it?
Apply age specific rates of your population to a standard population to cude overall risk

1. Calc age specific incidence/mortality risks
2. Multiple them by given weight (group size from standard population
3. Sum up age-adjusted number of cases, divide it by sum of weight

What do you need to consider?
- same standard population for every group
- standard population should ideally lie somewhere close to populations structures that you compare
- age of cases needs to be recorded

Question 3

Indirect standarization
- What do we get?
- How do we get it?
- Interpretation?
- What to consider?

Answer 3

What do we get?
1. A standardized incidence or mortality ratio (SIR or SMR), this is an estimate of risk ratio
- it compares total number of observed cases in your population with total number of cases that are expected bases on age-specific risks in reference group
- relative measure

2. A standardized incidence or mortality rate
   - SIR or SMR * incidence risk of reference group
   - absolute measure

How do we get it?
Apply the age specific risks of references group to age specififc population sizes of your population

1. Compute expected number of cases (/1000) in each age group for your population
2. Sum up the number of cases for each age group
3. Divide total observed by total expected number of cases to get SIR or SMR

Interpretation
The incidence in Area A is SIR-times as large as expected from the age-specific risks in Area B.

> 1 increased risk compared to references

What to consider?
- you only need total number of cases and age structure of your population, not the age for each case
- prefered when there are small numbers of cases in particular age groups