Topics
A to Z
Data &
 Statistics
Forms &
Publications
News &
Advisories
Licensing &
Certification
Rules &
Regulations
Public Health
Directory
Print this Article   Bookmark and Share
Mortality Patterns by Race and Ethnicity

Methodology:

Data were compiled from death certificates filed with the Oregon Health Division's Center for Health Statistics. Because the populations for minority groups are small and more youthful than the general Oregon population, an insufficient number of deaths occur in a single year to allow meaningful analysis. Therefore, data for a seven-year period (1989-1995) were aggregated to provide more reliable rates. The U.S. Census Bureau was the source of the population data. Because Indians, and perhaps other non-White races, are sometimes classified as non-Hispanic White on death certificates, the measures of mortality reported herein may understate the actual mortality experience of these Oregonians.

In these tables, the terms non-Hispanic White, African American, Indian, Asian and Pacific Islanders, and Hispanic were used. (The term "Indian" includes only American Indians.) These terms were chosen from several alternative choices. We recognize that within each group different individuals may have different preferences, but decided to use a single consistent term to describe each group to avoid nomenclatural confusion.

Several statistics are used in this report to assess a group's mortality patterns. One measure, the age-specific death rate, is the number of deaths in a certain age group divided by the population of that age group (times 100,000) while the crude death rate is the number of deaths per population regardless of age. (Rates based on a small number of events may be unreliable.)

However, the age-adjusted death rate is a more sophisticated index of mortality when comparing groups because, as the name implies, the rate derived is not biased by very young or very old populations. The age-adjusted death rate permits the comparison of populations with disparate age structures as if the populations had similar age distributions. The rate is computed by stratifying the populations into subsets by age, calculating an age-specific death rate for each group, then deriving a composite death rate by weighting each age category in proportion to its occurrence in a standard population. The estimated 1992 non-Hispanic White age distribution is the standard used here. For example, only 2.5 percent of the resident African Americans were 75 or older, compared to 6.5 percent of non-Hispanic Whites. Consequently, the crude death rate was lower for AFRICAN AMER. (698.6 per 100,000 population) than for non-Hispanic Whites (937.5). However, when compensation is made for the influence of the young African American population, the age-adjusted death rate is almost two times higher (1,240.2) than the crude rate.

Age-adjusted rates are meaningful only in comparison to other rates standardized in the same way and to the same population. Remaining differences among compared populations must be explained by differences other than age. Race-specific age-adjusted death rates are calculated only for causes which met the minimum criteria of 50 deaths. When fewer deaths occurred, age-specific rates may be used. All rates are per 100,000 population per year.

The comparative mortality figure is the ratio of the age-adjusted death rate for a minority group to the non-Hispanic White death rate. A comparative mortality figure of less than one indicates the group under comparison has a lower death rate than non-Hispanic Whites, while a figure of greater than one is higher. For example, the Indian comparative mortality figure for heart disease was 0.787 indicating that their age-adjusted death rate for this cause was only 78.7 percent of that for non-Hispanic Whites (i.e., 21.3 percent lower).

A measure of premature mortality is the years of potential life lost. This statistic emphasizes mortality occurring in younger age groups and assumes that each individual has 65 "productive" years, so that a death at age 21, for example, results in 44 (65 minus 21) years of life lost. The number of years are then summed for all individuals. Causes may be compared and ranked within each racial/ethnic category. For example, unintentional injuries are the leading cause of years of potential life lost for all groups except Asians and Pacific Islanders.

The years of potential life lost index is used to compare the extent of premature death among minorities to that among non-Hispanic Whites. It adjusts for the number of individuals in each racial/ethnic group as well as for their age distribution. The index value for a group is the ratio of observed years of life lost to the expected number of years based on the value for non-Hispanic Whites. For example, the African American years of potential life lost index of 6.870 for homicide means that proportionately almost seven times more years of potential life were lost by AFRICAN AMER. than by non-Hispanic Whites. Figures may be compared to Whites in the same manner as the comparative mortality figure.

Statistically significant differences are indicated by ".01",".02", or ".05", where appropriate, in the Tables. For example, the years of potential life lost index for deaths due to unintentional injuries for AFRICAN AMER. is 1.238 with the probability less than or equal to 0.05; this means that the likelihood of a value this different (from non-Hispanic Whites) occurring by chance alone is no more than one in twenty. Similarly, a significance level of 0.01 indicates that a value as different from the norm as the one under consideration would occur only one or fewer times out of 100 due to chance alone. Age-adjusted rates are calculated by the direct method and years of potential life lost by the indirect method.

In summary, crude death rates, age-specific death rates and the number of years of potential life lost are actual counts while age-adjusted rates, comparative mortality figures, and years of potential life lost indices are statistically created values that enable comparison of different populations regardless of the age-distribution of the populations. These measures of mortality are calculated for the leading causes of death.

For a detailed discussion of the above measures, see: National Center for Health Statistics. Mortality Statistical Notes for Planners, Number 3. Health Resources Administration. 1977. Washington, D.C.
Return to the top of the page.