# Standardisation >[!Abstract] Standardisation is >- Indirect standardisation (SMR) >- Direct standardisation ```toc ``` Suppose we are comparing the COVID related mortality rates between groups with different age profiles? Like comparing Sweden and India. Standardisation helps us do that. Or, - Male vs Female mortality rate - Indigenous cultures vs General population - Country to country Direct comparison of these different groups will lead to different biases and therefore it is important to standardise the data. ## Indirect Standardisation Question: Do black Americans have a higher COVID-19 mortality rate than the general American population? - Study population: Black Americans - Standard population: General American population Step 1: Find mortality rates in standard population by age group - Age group (eg. 0-24) vs Corona deaths per 1000 people Step 2: Use study population age-distribution to find expected deaths - Use the percentage of black people (out of the general population) and calculate the expected number of deaths per 1000 people as though the two groups were identical Step 3: Compare total expected deaths to actual deaths - Calculate the standardised mortality rate (SMR) $SMR = \frac{\text{actual deaths}}{\text{expected deaths}} = 2.06$ - A value of SMR equal to 2.06 means that Black Americans have a mortality rate which is 2.06 times that of the general American population ![[indirect_standardisation_covid_black_america.png]] ## Direct Standardisation Question: Can we compare the COVID mortality rate in Kerala to the rest of India? - Study population: Kerala - Standard population: General Indian population Crude rate: - Kerala: 2.2 deaths per 1000 - India: 1.7 deaths per 1000 Looking at the crude rates, you might be tempted to conclude that Kerala did worse than the general population in India; but keep in mind that there might be a hidden issue whereby the age distribution of Kerala might be more heavily weighted towards the older ages. In order to compare these two populations, we first need to standardise the data in age distribution. What we need to do, in essence, is to reframe the question as follows: **What would the deaths have been if Kerala had the same age distribution as the rest of India?** Step 1: Get age-specific death rates for population of interest - Kerala Step 2: Use these rates to find the expected deaths in the standard population - India Step 3: Find the age-standardised mortality rate (SMR) ![[direct_standardisation_ex.png]]