BIOSTATS 4.6 - DIRECT AND INDIRECT ADJUSTMENTS

Learning Objectives

• The lecture compares different mortality measures and how to calculate them.
• It defines, calculates, and interprets direct and indirect age-adjusted mortality rates.
• The lecture also explores how to analyze morbidity and mortality measures in epidemiology reports.

Mortality Rate

• Mortality rate expresses the number of deaths in a population during a specific time.
• Mortality rate is calculated as the number of deaths during a period divided by the total population at risk during that period, multiplied by 10^n.

Crude Mortality Rate

• Crude mortality, also know as all-cause mortality rate measures all deaths in a population during a specific period of time.
• The crude mortality rate is calculated as the number of all deaths in a period divided by the total population at risk during that period, multiplied by 10^n.
• Globally, crude mortality rates varied in 2006, ranging from 0-4 deaths per 1,000 people per year to more than 11 deaths per 1,000 people per year.

Case Fatality Rate

• Case-fatality rate (CFR) represents the fraction of people who die from a disease, expressed as a percentage.
• Case-fatality rate equals the number of deaths due to a disease during a period of time divided by the number of persons with the disease, as a percentage.

Mortality Rate Comparisons

• In 1997, the death rate in Florida was greater than the death rate in Maryland.
• In 1997, there were 41,600 deaths in Maryland out of a population of 5,100,000.
• In 1997, there were 154,500 deaths in Florida out of a population of 14,670,000.
• In 1997, Maryland had 816 deaths per 100,000 while Florida had 1,053 deaths per 100,000.
• Crude rates in two populations may differ due to a lurking variable.

Age Distribution

• Mortality is highest in the elderly.
• Florida's population is older than Maryland's population.
• Differences in mortality rates might be skewed because of the age distribution between Florida and Maryland.
• Age distribution should be considered when comparing the mortality rates of Florida and Maryland.

Adjustment (Standardization)

• A factor can be statistically adjusted to eliminate the possibility that the observed differences in mortality rates between populations could be the result of differences in this factor.
• Adjustment can be performed for factors such as race/ethnicity, age, gender, and income.
• Adjustments are only necessary if the populations differ by the factor to be adjusted.

Types of Adjustment

• Direct adjustment uses the structure of a standard population.
• Indirect adjustment uses the specific rates of an unexposed population.

Direct Adjustment

• A standard population eliminates any affects of differences in a factor between two or more populations being compared.
• The factor-specific rates for each population being compared are applied to the standard population.
• Direct adjustment eliminates the possibility that the factor's distribution results in observed differences in a population.

Direct Adjustment - Standard Population

• Several methods exist for designating a standard population:
  • Use one of the study populations
  • Combine the study populations
  • Choose an external standard

Direct Adjustment Calculations

• Step 1: Create a standard population.
• Step 2: Calculate age-specific mortality rates for comparison populations.
• Step 3: Calculate expected deaths for each age group and add up the total deaths.
• Step 4: Calculate the age-adjusted mortality rates.

Direct Adjustment Example

• When considering combined state populations in 1997:
  • Maryland had 1,081 deaths per 100,000.
  • Florida had 973 deaths per 100,000.
• When considering the US standard million:
  • Maryland had 901 deaths per 100,000.
  • Florida had 820 deaths per 100,000.

Direct Adjustment Cautions

• The direct adjusted rate does not reflect the risk of the "real" population and rate depends on standard population.
• Direct adjustment is appropriate only when comparing groups or examining trends across time periods.

Indirect Adjustment: Standard Mortality Ratio (SMR)

• Calculated using the Standardized Mortality Ratio (SMR).
• The ratio of total observed deaths to total expected deaths, if study population had mortality experience of "parent" population.
• Used to study mortality in occupationally exposed populations.
• SMR equals Observed Number of Deaths divided by Expected Number of Deaths.
  • SMR = 1.0 when the observed number of deaths in the study population is statistically the same as deaths of parental population.
  • SMR > 1.0 when the observed number of deaths is greater than the expected level of deaths in parental population.
  • SMR < 1.0 when the observed number of deaths is statistically less than expected number from parental population.

Indirect Adjustment Cautions

• SMR's cannot be compared since the population distribution is of the exposed group, not the unexposed group.
• Two SMRs cannot be compared because they are standardized to different exposed groups, use different population distributions, and are therefore not comparable, even if they use specific rates from the unexposed group.