Skip to content
Scintria
ToolsGrant writerFunding
HomeLiteratureClinical trialsEvidence LabProtocol packData analysisLearnProjectsLibraryReferencesApplications
Start free
ScintriaStart free

Learn / Epidemiology / Foundations: counting disease

Comparing populations: crude and standardised rates

14 minute read · 6-question quiz

Download this lesson as a colour Word or PDF study guide with the figures and answers. Included in paid plans; reading and quizzes stay free.Sign in

What you will learn

  • Explain why crude rates can mislead when age structures differ

  • Carry out direct age standardisation

  • Distinguish direct and indirect standardisation and say when each is used

Death and disease rise steeply with age. Compare Town A, with many older residents, and Town B, with a young population. Even if the two towns have identical health and health services, Town A will have a higher crude death rate simply because it has more old people. Age here is a confounder of the comparison between places.

Standardisation removes the effect of the differing age structure so that populations can be compared fairly.

Direct standardisation, step by step

  1. 1

    1. Get age-specific rates

    For each population, calculate the rate within each age group: deaths ÷ population in that group.

  2. 2

    2. Choose a standard population

    A reference age structure, for example the WHO World Standard Population, or simply the combined population of the two places.

  3. 3

    3. Apply the rates to the standard

    Multiply each age-specific rate by the standard population's share in that age group.

  4. 4

    4. Add up

    The sum is the age-standardised rate: what the death rate would be if each population had the standard age structure.

Worked example: two towns

Deaths per year and population by age:

AgeTown A popA deathsA rate /1,000Town B popB deathsB rate /1,000Standard share
0 to 142,000105.04,000246.030%
15 to 445,000255.05,000306.040%
45 to 643,0006020.02,0003618.020%
65 and over2,00012060.05003570.010%
Total12,00021511,500125

Crude rates: Town A = 215 / 12,000 = 17.9 per 1,000; Town B = 125 / 11,500 = 10.9 per 1,000. Town A looks 65% worse.

Standardised rates (weights 0.30, 0.40, 0.20, 0.10):

  • Town A: 5.0 × 0.30 + 5.0 × 0.40 + 20 × 0.20 + 60 × 0.10 = 1.5 + 2.0 + 4.0 + 6.0 = 13.5 per 1,000
  • Town B: 6.0 × 0.30 + 6.0 × 0.40 + 18 × 0.20 + 70 × 0.10 = 1.8 + 2.4 + 3.6 + 7.0 = 14.8 per 1,000

After adjustment Town B has the higher death rate. The crude comparison was reversed. Town A's crude rate was high only because 42% of its people are over 45, compared with 22% in Town B.

05101520Deaths per 1,00017.9A crude10.9B crude13.5A standardised14.8B standardised
The ranking reverses once age structure is taken into account.

Simpson's paradox in disguise

Here a pooled (crude) comparison points one way and the stratified (age-specific) comparison points the other. It is the same phenomenon as confounding in chapter 5 of the biostatistics course. Always look at the age-specific rates before trusting a crude comparison.

Indirect standardisation and the SMR

Direct standardisation needs age-specific rates in each study population, which are unstable when numbers are small (a small hospital, a rare cancer). Indirect standardisation instead applies the standard population's age-specific rates to the study population's age structure, to find the number of deaths expected:

SMR = (observed deaths)/(expected deaths)

An SMR of 1.0 means the death rate equals that of the standard population; 1.3 means 30% more deaths than expected. It is common in occupational epidemiology (miners, health workers) and for comparing hospitals.

DirectIndirect
NeedsAge-specific rates in each study populationAge structure and total deaths in the study population
GivesA standardised rateAn SMR
Best whenNumbers are largeNumbers are small
ComparingAny number of populationsEach against the standard (not directly with each other)

Standardisation hides as well as helps

A single standardised figure can conceal that a disease affects only the young. Report age-specific rates alongside it. Also, the result depends on the standard chosen: always say which was used. And standardise for whatever is the important confounder: age, sex, but rarely anything else.

In practice

Global health reports (WHO, Global Burden of Disease) use age-standardised rates to compare countries. If a table gives only crude rates, be careful interpreting differences between countries with very different demographics, such as Uganda (median age about 16) and Japan (about 49).

Summary

  • Crude rates are affected by age structure; standardisation removes this confounding.

  • Direct standardisation applies each population's age-specific rates to one standard population.

  • Indirect standardisation yields the SMR = observed ÷ expected deaths.

  • A crude comparison can be reversed after standardisation.

  • Report age-specific rates and name the standard used.

Check your understanding

Question 1
1

Town X has a higher crude mortality than Town Y but similar age-specific rates. The likely explanation is:

Question 2
2

In direct standardisation we multiply age-specific rates by:

Question 3
3

An SMR of 1.25 means:

Question 4
4

Indirect standardisation is preferred when:

Question 5
5

Standardisation is mainly a way of controlling for:

Question 6
6

Why should you also give age-specific rates?

Draft study guide. Spotted an error or something unclear? Tell us at hello@scintria.com. This is educational material, not clinical or regulatory advice.

Prevalence, incidence and ratesPerson, place and time