OCR Text |
Show Age Adjustment Abstract. Brian Paoli, Bureau of Surveillance and Analysis, Utah State Department of Health. Introduction Since 1943 the National Center for Health Statistics (NCHS) has recommended use of the 1940 U.S. population as the national standard for age adjustment of health measures. But in recent years this standard has come to be perceived as outdated because today's U.S. population includes a much higher proportion of older persons (1). Because of that perception, the 1940 standard has not been applied consistently, with some agencies using 1970 or 1980 as a standard. This inconsistent application of standards has led to confusion among data users and has created an additional burden for State and local agencies when trying to coordinate their statistics with those reported by Federal agencies. Beginning with the data year 1999, NCHS has implemented the year 2000 as a new standard for age adjustment that will be used by all agencies within the U.S. Department of Health and Human Services (DHHS) when reporting on mortality statistics. The year 2000 was chosen because the age distribution is similar to the current population and will remain so for at least 10 years. Weights for the new standard are based on 1998 U.S Census Bureau population projections for the year 2000. A projection was used since the figures for the upcoming 2000 census would not be available until 2002 or 2003 and would delay implementing the new standard. Rates Rates (e.g. incidence rates, prevalence rates, mortality rates) are fundamental tools of health data analyses. Rates are calculated by dividing the number of events such as births, deaths, illness episodes, et cetera by the population at risk at some point in time or over some time period. Rates of many events vary substantially among different age groups so that comparisons between populations with different age structures can be misleading. For example, the crude death rate (i.e. the number of deaths per 100,000 population) is a widely used method of reporting mortality. However, the crude death rate is influenced by the age distribution of the population being studied. Age standardization or "age adjustment" is a method that enables one to make meaningful comparisons of death or other rates between or among groups by eliminating the confounding effects that arise from differences in age distribution. For example, an area that has an older population is likely to have higher crude (not age-adjusted) mortality rates for cancer, even when the rates for specific age groups is identical across areas. Age adjustment requires the use of a "standard population". The age distribution of the standard population (i.e. the percentage of persons in each age group in that population) provides a basis for weights that are applied to data from the populations one wishes to compare. This weighting process essentially asks, "What if each population had the same age distribution as the standard population? Would differences in mortality be found?" Weights are chosen mostly for practical reasons since the only statistical criteria is that the weights should not be considered unusual relative the to the study population. What this Means Age-adjusted death rates will be affected by the change in population standards from 1940 to 2000. Because of the greater proportion of older persons in 2000, the 2000 standard places a greater weight on older age categories (Table 1). Greater weights in these categories will translate into higher age-adjusted death rates when evaluating illnesses that increase in frequency with age, such as cardiovascular or pulmonary disease. The impact from the change in standard is not as noticeable for conditions such as suicide or homicide where mortality risk is more uniform across age groups (Table 2). 125 |