Lecture 4: Standardization

What must you always include when standardizing data?

"After adjusting for ____" (ex. age)

What do you need for direct standardization?

# in each age group for populations to be compared Observed # of deaths for each age group # in each age group for Standard population

How does changing the population affect the standardization?

Affects the magnitude of the rates

Procedure for direct standardization:

Calculate age specific rates for each population to be compared Multiply age specific rate times by the standard population for each age group Divide expected total by the standard population

What methods are there for standardization?

Direct and indirect

Specific measures?

Measures for population subgroups. Restricts the numerator and denominator to specific subgroups.

What is indirect standardization?

Method of rate adjustment when the age specific rates are not available. Compute a standardized morbidity/mortality ratio (SMR). Comparison of observed to expected.

What is direct standardization?

Method of rate adjustment. Uses the age distribution of an external reference population (the standard). - Apply age specific rates observed in two or more study populations to standard population of known age structure. - Standard population used (US 2000 census data)

Indirect adjustment procedure:

Multiply standard population age specific death rate by # in each age group of populations to be compared Calculate SMR=observed/expected

How to go from SMR to standardized rate?

Multiply the SMR by the crude mortality rate of the standard population.

Is an SMR a final step?

NO! you can't compare SMRs

Why standardize?

Primarily to compensate for differential age distributions among comparison populations.

What are the pros and cons of frequency (raw count)?

Pros: actual number of events, useful for determining need for services Cons: influenced by size of population

What are the pros and cons of specific rates?

Pros: calculated for similar groups, comparisons of different populations, identifies subgroups at risk Cons: a lot of work to compare more than two subgroups

What are the pros and cons of adjusted rates?

Pros: differences adjusted between populations, permits unbiased comparison cons: artificial rate, doesn't represent the risk of dying, magnitude dependent on standard population

What are the pros and cons of Crude rates?

Pros: summary rate, easily calculated, risk of dying in the population for a given period of time Cons: influenced by population characteristics, comparisons may be confounded by these characteristics

What to consider for standard population?

Should be close to the distribution being compared. Could: combine the study populations, use the overall state or national population.

Crude measures?

Summary measure for the total population

Adjusted measures?

Summary measures for total population statistically transformed to remove the effect of differences in population composition (ex. age). Allows for fair comparisons.

What do you need for indirect standardization?

The number in each age group for populations to be compared The total number of deaths in the populations to be compared The age specific death rates for the selected standard population