MEMORANDUM FROM THE SECRETARY
August 26, 1998
Heads of Operating Divisions
Heads of Staff Divisions
HHS Policy for Changing the Population Standard for Age Adjusting Death Rates
Agencies and programs of the Department use age-adjusted death rates for a variety of purposes ranging from routine data dissemination, research, public health surveillance, setting health goals, and monitoring health status. At present, a number of different population standards are used in the Department for producing age-adjusted death rates. The resulting multiplicity of rates creates confusion among data users, including the media; and it imposes an unnecessary burden on State and local health data users who must produce several data series in order to be consistent with Federal data which are often used as benchmarks. Accordingly, at the recommendation of the HHS Data Council, I am issuing the attached policy on adopting a uniform population standard within the Department for the dissemination of mortality data. That standard is the projected population for the year 2000 produced by the Bureau of the Census. The policy affirms the HHS effort to coordinate statistical activities in the Department for more effective communication with the public, and for reducing the statistical burden on State and local health agencies. This policy will produce mortality data that is uniform throughout the Department and data that is more consistent with the contemporary population structure than the 1940 population standard that it replaces.
In general, the policy described in the attached material requires the use of the new standard effective with deaths occurring in 1999, with the exception that researchers may use other standards as appropriate to their work. In such cases, the standard used should be clearly stated along with a rationale for departing from DHHS policy. The policy will go into effect on September 1, 1998. I am directing Heads of OPDIVs and STAFFDIVs to implement this policy within your organizations in accordance with normal agency data planning and analysis cycles. This policy covers all programs of the Department, including both health and human services. It reflects the outcome of deliberations in two workshops sponsored by the National Center for Health Statistics with broad participation from the National Academy of Sciences, the National Institutes of Health, the Centers for Disease Control and Prevention, academia and the States.
Please make every effort to assure the successful implementation of this policy. Questions about the policy should be directed to your agencys representative to the HHS Data Council (attached).
Donna E. Shalala
cc: Co-chairpersons, HHS Data Council
POLICY STATEMENT ON CHANGING THE POPULATION STANDARD USED FOR AGE ADJUSTING DEATH RATES IN DHHS PUBLICATIONS
SUMMARY: This document describes the Department of Health and Human Services (HHS) policy to change the population standard used for age adjusting death rates from the 1940 population to the year 2000 projected population. This policy covers all programs of the Department, including both health and human/social services. The policy reflects the outcome of deliberations in two workshops sponsored by the National Center for Health Statistics with broad participation from the National Academy of Sciences, the National Institutes of Health, the Centers for Disease Control and Prevention, academia, and the states. The need for the HHS-wide policy is in part caused by different standards used by different HHS agencies resulting in confusion among the media and the public, and creating extra work for States that attempt to provide data comparable to those of the Department. Further, the existing 1940 standard population is widely perceived as out of date. In addition, a more contemporary population standard produces indices of mortality that better reflect the structure of the current population. The policy does require the re-calculation of historic data using the new standard, and does require the recalibration of the Healthy People goals.
This policy reaffirms the HHS commitment to use statistical practices that are consistent from agency to agency, that speak clearly to the media, and that are sensitive to the statistical needs of State programs. Implementation of this policy will help reduce confusion among health data users including the media, and will reflect better statistical coordination within the Department. The policy will affect data on race differentials in mortality which are based on age-adjusted death rates. This policy will allow researchers to continue to use alternative standards when dictated by their research needs; but official data releases will use the new standard. Educational materials are being developed that describe the rationale for the change in population standards, and explain the impact on race differentials. The materials will cover topics such as the masking effect of a single ratio of age-adjusted rates upon important age-specific differences in mortality race ratios and note that to best understand race differentials in mortality, it is essential to augment analyses of age-adjusted death rates with analyses of age-specific rates.
I. INTRODUCTION AND POLICY RATIONALE
Since 1943, the National Center for Health Statistics and State health departments have used the 1940 population as a standard when age-adjusting death rates. However, DHHS agencies are currently using at least three different standards. Most agencies along with NCHS use the 1940 standard, but some agencies are using standards based on the 1970 and 1980 populations. In two departmental workshops, one in 1991 and one in 1997, there was consensus that DHHS agencies and States need to be consistent when publishing mortality statistics in order to minimize confusion and misunderstanding by data users and the media. Multiple standards also create burdens for the States, who attempt to make their data consistent with Federal practices. Thus, when DHHS releases official mortality statistics to the public, age-adjusted death rates should be based on a single, uniform standard.
In 1991, NCHS convened the first workshop on age adjustment to examine technical issues and problems related to the calculation and interpretation of age-adjusted death rates. Participants included representatives from NCHS, the Centers for Disease Control and Prevention, the National Institutes of Health, the National Academy of Sciences, State health departments, and academia. The first workshop recommended the continued use of the 1940 standard by NCHS and encouraged other federal and State government agencies to use this standard when publishing official mortality statistics. In addition, the first workshop recommended that NCHS study issues that might lead to the introduction of a new or additional standard by the year 2000.
A second workshop in 1997 was convened to examine policy issues related to age adjustment. In both workshops, there was agreement that when it is appropriate to standardize, the choice of population standard is arbitrary, and that trend comparisons and group differences tend to be similar regardless of the standard used. However, many perceive that the 1940 standard is grossly outdated. The second workshop considered this perception problem sufficient reason to replace the 1940 population with a standard that better reflects the current population. Thus, the second workshop recommended that the standard population be changed from the 1940 population to the 2000 population. Although there is no overriding statistical reason for choosing the year 2000 standard over other age distributions, the participants considered it appropriate to choose a year that represents the current population and with which data users can be comfortable.
The second workshop proposed that a new year 2000 population standard be implemented with data year 1999, i.e., beginning with publications reporting 1999 data. This necessitates the use of a projected 2000 population. The recommendations of the second workshop are listed below.
B. Rationale for an HHS Policy to Change the Population Standard for Age Adjusting Death Rates
The need for an HHS-wide policy for changing the population standard from 1940 to the year 2000 is based on the following:
- The multiplicity of standards currently used by HHS agencies reflects a lack of statistical coordination within the Department.
- The multiplicity of standards in different HHS programs is not based on a technical rationale.
- The change to a single standard will improve communication between the Department and data users including the media, which is confused by multiple standards.
- The change to a single standard will reduce burden for States, who attempt to produce data series that are consistent with those at the national level.
- The change to a more contemporary standard results in measures of mortality risk that better reflect the contemporary population structure.
The purposes of this policy, therefore, are (1) to ensure that age-adjusted death rates used in all HHS systems are comparable, (2) that age-adjusted death rates reflect a more contemporary population structure.
Agencies and Operating Divisions of HHS are expected to develop any mechanisms needed to implement this policy. When implemented, the policy will enable the Department to more effectively fulfill its mission to assure the health and well-being of the Nation and to ensure, on an ongoing basis, that the Department is sensitive to its need to effectively communicate with its data users including the media, and is sensitive to the burden it imposes on States who use national data sets as benchmarks.
II. POLICY ON CHANGING POPULATION STANDARDS IN AGE ADJUSTING DEATH RATES
The HHS-wide policy on changing the population standard for age adjusting death rates in the routine presentation of mortality data is as follows:
A. Single Standard: The population standard for age-adjusting death rates should be changed from the 1940 U.S. standard million population to the projected 2000 U.S. population available from the Census Bureau in the spring of 1998. A single standard should be used by all DHHS agencies for official presentation of data. For special analyses, alternative standards may be used as appropriate to the research, but the choice of standard should be clearly stated.
B. Implementation No Later Than for Deaths Occurring January 1, 1999: Agencies should implement the new population standard by data year 1999. The official implementation of the new 2000 standard will take place with data year 1999, i.e., beginning with publications reporting 1999 data. NCHS will begin publishing provisional 1999 data in June of 1999. Earlier implementation dates were considered. However, resource limitations particularly at the State level necessitate a somewhat later implementation date. Participants from all agencies were confident that implementation could take place effective with the data year 1999.
C. Earlier Implementation Procedure: Agencies should continue to use and publish their current standards until the official implementation data year (1999) when the new common standard will be adopted. To avoid confusion, agencies implementing the new standard prior to data year 1999 should simultaneously publish rates adjusted to both the old and new standards.
Some DHHS agencies have indicated a desire to implement the new standard prior to data year 1999. To maintain the current level of consistency, each agency will continue to publish age-adjusted death rates using their current official standard until data year 1999 when the new 2000 standard will be used exclusively for the publication of official mortality statistics. Rates based on the new standard may be published prior to data year 1999 provided that rates using the old standard are also published until data year 1999.
D. Use of New Standard in Communication. After the implementation date, agencies should use the new standard in all press releases and other communication with the public. Beginning with data year 1999, all age-adjusted death rates reported to the media or to the general public should be calculated using the new standard. These statistics will be considered official. Rates using the old or other standards may be disseminated for research purposes as appropriate.
E. Name of New Standard: The new standard is called the year 2000 population standard.
F. Age Groups: Agencies should continue to use the current eleven age groups (less than 1 year, 1 to 4 years, 5 to 14 years, 15 to 24 years, 25 to 34 years, 35 to 44 years, 45 to 54 years, 55 to 64 years, 65 to 74 years, 75 to 84 years, and 85 years and older) for calculating age-adjusted rates using the new standard.
G. Implementation Plan: NCHS will convene an implementation committee that will be responsible for developing a time table and strategies for implementation and for commissioning papers to publicize the change in standard.
The implementation committee will be responsible for addressing the practical matters related to implementation of the new age-adjustment standard. This will include the development of a time table and strategies for implementation and the commission of at least two papers--one technical and one non- technical--drawing on publications developed by the National Heart, Lung, and Blood Institute (NHLBI). The committee will also need to garner support from such offices and organizations as the Secretary, the DHHS Data Council, all DHHS agencies, the CDC Excellence in Science Committee, the Committee on National Statistics of the National Academy of Science, the National Cancer Policy Board, the National Committee on Vital and Health Statistics, and the Office of Management and Budget.
H. Publicizing Policy Change: NCHS will publicize the new standard in NCHS publications, the Morbidity and Mortality Weekly Report, Public Health Reports, and appropriate professional newsletters. Scholarly papers could also be published in appropriate professional and technical journals.
I. Evaluation of Population Standard: NCHS will convene a workgroup to evaluate the age-adjustment standard at least every ten years.
J. Data Collection and Reporting Activities Covered by this Policy: This policy applies to the following types of data collection systems:
- Statistical data collections (e.g. vital statistics, disease registries, and other research and survey data that present mortality data as outcome measures.)
- Research, evaluation, and other study projects (e.g. intramural research).
III. EXEMPTIONS FROM THE POLICY
Other standards may be used for research purposes. When doing so, a rationale for departing from DHHS policy should be given.
Anderson R, Rosenberg HM, eds. Proceedings of the Second Workshop on Age Adjustment. National Center for Health Statistics. Vital and Health Statistics (in preparation).
Curtin LR, Klein RJ. Direct standardization (age-adjusted death rates). National Center for Health Statistics. Healthy People 2000: Statistical Notes. 6. 1995.
Feinleib M, Zarate AO, eds. Reconsidering age adjustment procedures: Workshop proceedings. National Center for Health Statistics. Vital and Health Statistics 4(29). 1992.
Curtin LR, Maurer J, Rosenberg HM. On the selection of alternative standards for the age-adjusted death rate: Proceedings of the social statistics sections, American Statistical Association, pp 218-23. 1980.
National Center for Health Statistics. Vital Statistics of the United States, 1993, vol II, mortality, part A, section 7. Washington: Public Health Service.
Wolfenden HH. On the methods of comparing the mortalities of two or more communities, and the standardization of death rates. Journal of the Royal Statistical Society 86:399-411. 1923.