A review was conducted of health focus area reports prepared by the six work groups of the HHS Initiative to Eliminate Racial and Ethnic Disparities in Health. The focus areas are (1) infant mortality; (2) cancer screening and management; (3) cardiovascular disease; (4) prevention of diabetes complications; (5) HIV/AIDS; and (6) child and adult immunizations.
The purpose of the review was to summarize existing data gaps and make recommendations for eliminating those gaps. A list of the reported gaps and the recommendations from each of the reports appears below. Where possible, similar topics are grouped together. The summary is divided into two sections: (1) cross-cutting gaps and recommendations; and (2) report-specific recommendations. In addition, several data gap sections which are not mentioned in the reports are listed.
For reference purposes, the new Federal standards for racial and ethnic data specify that data be reported for a minimum of five categories of race--American Indian or Alaska Native, Asian, black or African American, Native Hawaiian or Other Pacific Islander and white--and two categories of ethnicity--Hispanic or Latino and non-Hispanic or Latino.
Gaps and Recommendations
The following gaps and recommendations cross cut across all the reports.
The Department should encourage and sponsor the analysis of existing data by outside researchers. All types of data--vital statistics, hospitalization, medical care, immunization, cancer screening, risk factor data, and incidence and prevalence--should be included.
1. Considerable progress has been made in the publication of vital statistics data for the OMB minimum categories for data racial and ethnic groups. For example, Health U.S. contains population and selected mortality data for those groups. In addition, the website http://raceandhealth.hhs.gov/ contains data specifically targeted to the six focus areas. Because of the central role of vital statistics in public health, and because the availability and visibility of the current data can increase, this area deserves special attention. Although mentioning the existence of a gap, the six focus area reports did not mention the specific recommendations outlined here.
2. Examine the possibility of expanding the CDC WONDER system, especially the mortality data, to include data for: (1) the OMB minimum categories for race and ethnicity; (2) American Samoa, Guam, Puerto Rico, the U.S. Pacific territories, and the Virgin Islands; and (3) life expectancy.
3. Examine the possibility of increasing the visibility and availability of the data from the HHS Initiative to Eliminate Racial and Ethnic Disparities in Health website by: (1) increasing the number of agency websites that it can be hotlinked from, such as the CDC, NCHS, and Healthy People 2000 websites; and (2) increasing the usefulness of data on the HHS Initiative to Eliminate Racial and Ethnic Disparities in Health website by periodically publishing it as a government document and in the public health literature. For example, an apparent decline in childhood immunization rates between 1990 and 1995 would become more apparent if the data were published in a government document or as a journal article.
4. Health U.S. should consider including estimates of life expectancy, by sex, for the OMB standard/minimum groups.
- Gap: There are gaps in the analysis of currently available health data.
- Gap: There are gaps in the publication of vital statistics data.
- Gap: All reports mentioned the need for periodically updated data for ethnic subgroups, i.e., subgroups of the OMB specified racial/ethnic groups, at the national, State and local levels. From the existing data, it is clear that the larger racial and ethnic groupings can mask significant health disparities among racial and ethnic subgroups. The data gaps included all types of data, i.e. vital statistics, hospitalization, medical care, immunization, cancer screening, risk factor data, incidence and prevalence for each of the six focus areas (infant mortality, cancer, cardiovascular diseases, diabetes, HIV/AIDS, child and adult immunizations).
1. The OMB directives specify a minimum level for presenting data by race and ethnicity. The guidelines also encourage the presentation of data for racial and ethnic subgroups when and where possible, as long as the results can be collapsed back into the minimum categories. Where data currently exist for key subgroups, publication should be encouraged and, if possible, made available as public use data sets. For example, data for racial and ethnic subgroups for the six focus areas are contained on the Race and Health website, and cancer incidence and mortality rates from the National Cancer Institute's SEER have been published for both the OMB standard/minimum groups and for racial and ethnic subgroups. Where possible, collection of data for racial and ethnic subgroups should be encouraged.
2. Evaluate the possibility of collecting risk factor data for Hispanic subgroups, AA/PIs and their subgroups, and American Indians or Alaska Natives using the BRFSS, the Youth Risk Factor Survey and even a future census.
3. Encourage the development of population-specific examination surveys targeted to small ethnic subgroups or medically underserved populations which cannot be included in the NHANES through oversampling.
- Gap: Morbidity data for Asian and Pacific Islanders, Hispanics, American Indians and Alaska Natives is lacking.
This gap was mentioned in the Cardiovascular and HIV reports, but it may apply to the other areas as well. To fill this gap, it was recommended that:
1. HCFA collect hospitalization data on Hispanics and Hispanic subgroups.
2. NCHS expand the Hospital Discharge Survey to include data for those groups.
- Gap: The Cardiovascular and HIV Reports mention a need for including ethnic subgroups during the survey planning and data collection efforts.
Two recommendations for alleviating this gap were suggested in an unpublished report titled, Race Initiative Report on Cardiovascular Disease. They are:
1. Encourage representation on the National Committee on Vital and Health Statistics from the National Medical Association (NMA), COSSMHO, National Hispanic Medical Association (NHMA), Association of Asian/Pacific Community Health organizations, Indian Health Service (IHS), and an historically black medical school.
2. Solicit data concerns in the planning stages of the NCHS surveys NHANES and NHIS from the NMA, IHS, COSSMHO, NHMA, Association of Asian/Pacific Community Health organizations, and an historically black medical school.
- Gap: No data exists on the effects on rates of changes in the age adjustment standard and ICD-10.
1. Detailed studies have been conducted on the effects of the year 2000 standard population for African Americans and whites, but the effects on mortality rates on other racial/ethnic groups should be studied.
2. Evaluate the impact of ICD-10 coding for mortality and morbidity include analyses by racial/ethnic groups.
The following gaps and recommendations were identified in specific health focus area work group reports.
Cancer Screening and Management
- Gap: Yearly updates on screening rates are needed.
Explore the feasibility of moving basic questions on mammograms and pap smear screening from NHIS supplements to the core questionnaire.
- Gap: NCI's SEER has been very effective in documenting current levels and trends in site-specific cancer incidence and mortality rates; those data are invaluable. The problem arises from the fact that the rates for some of the minimum racial and ethnic groups are often unstable, so that measuring recent trends is difficult.
Although SEER has done an excellent job of reporting on racial and ethnic groups given current resources, NCI should explore the feasibility of expanding SEER and/or forming alliances with the State-based cancer registries in order to produce stable rates for the minimum racial and ethnic categories and to allow the system to produce rates for ethnic subgroups.
- Gap: Data are lacking on cardiovascular risk factors, including current risk factor data for Hispanic subgroups, AA/PI subgroups, and for American Indians or Alaska Natives.
The Cardiovascular Work Group placed special emphasis on this topic and proposed several possibilities for improvement.
1. Encourage evaluation of the possibility of collecting risk factor data for Hispanic subgroups, Asian/Pacific Islanders and their subgroups, and American Indian/Alaska Natives using the NHIS, BRFSS, the Youth Risk Factor Survey, and a future census.
2. Encourage the development of population-specific examination surveys targeted to small ethnic subgroups or medically underserved populations which cannot be included in NHANES through oversampling.
- Gap: There is a need for nationally representative data on the incidence of heart disease and stroke.
NHLBI has collected incidence data from different sources such as the Atherosclerosis Risk in Communities (ARIC) Study and the Minnesota Heart Survey. Although the data from those sources are relevant, the samples are too small to estimate current levels and incidence trends for the minimum standard racial/ethnic categories. Moreover, the trends for African Americans can be unstable. Therefore:
1. NIH, CDC, and/or their contractors should continue to analyze HCFA Medicare data to establish approximate national trends in incidence, case fatality, and one year survival rates for acute myocardial infarction (AMI), congestive heart failure, and acute stroke in each major racial/ethnic group.
2. Because HCFA data will not include incidence data on younger adults and older adults not receiving Medicare benefits, data will only be approximate. As a result, NIH, CDC, and HCFA should explore the feasibility of setting up a SEER-like registry for heart diseases and stroke.
- Gap: Local disease patterns may be masked by evaluating data at the national or State level.
Encourage geocoding of all health-related data to determine patterns of health and disease among minority populations. Patterns that may be masked when evaluating data at the State level include mortality, morbidity, risk factors, and health care utilization.
- Gap: Better and more accurate data on health care indicators of American Indians and Asian/Pacific Islanders are needed.
Work more closely with service providers, consumers, epidemiologists, and providers in native communities.
- Gap: Although information on mode of HIV exposure is not required to transmit a case report to CDC, collection of risk information is very important.
To allow the U.S. to target programs and resources most effectively, we must be able to monitor the epidemic. This means improving our ability to track early HIV infections before they progress to AIDS. As of December 1997, only 27 States sent data to CDC on all confidentially reported cases of HIV infection. Two additional States reported only cases of HIV infection among children under age 13, and one State reported cases only for children under the age of six. CDC and the State public health departments should consider the feasibility of reporting all HIV infections anonymously.
- Gap: There is a need for a standardized disease-based surveillance system for influenza, with a special emphasis on expanding networks of sentinel physicians and establishing HMO and occupational-based influenza surveillance.
CDC should be encouraged to determine the feasibility of setting up such a surveillance system.
- Gap: There is a need for rapid annual assessment of influenza vaccine effectiveness among vulnerable U.S. populations.
CDC should be encouraged to determine the feasibility of setting up such a surveillance system.
- Gap: There is a need for more timely data on vaccination levels.
Improve time of availability of NHIS- and BRFSS-reported vaccination levels to less than one year after data are collected.
- Gap: There is a need for more complete adult immunization data.
Assess the feasibility of using the NIS sample frame to survey adults.
Data on vaccination levels for American Indian/Alaska Natives and Asian/Pacific Islanders populations, age 65 years and older, are lacking.
Encourage a feasibility study on conducting special surveys of vaccination levels for persons ages 65 and over, or add vaccination questions on existing surveys of these populations.
- Gap: There are limited data on the success of programs targeting populations at risk of undercoverage.
Although some States have already implemented immunization registries, in the long term, a national system of registries is needed. The Data Working Group encourages a feasibility study on implementing a national system of State-based immunization registries. At minimum, the NHIS and the NIS should continue to collect information on vaccine coverage, along with information on race and ethnicity and socioeconomic status.
Recommendations Culled from Work Group Reports
Infant Mortality Work Group Report--Draft of December 16, 1998.
- To support research and effective interventions for racial and ethnic subgroups, national data collection efforts should include all key subgroups.
- Significant attention must be devoted to incorporating more useful and specific racial and ethnic subgroup classifications.
- Data collection for key subgroups must be updated, and a minimum set of categories for use in all Federal data collection activities should be constructed.
- In addition to race and ethnic subgroup, a mother's origin of birth and current residence may be correlated with birth outcomes. Research should seek to include indicators such as foreign vs. U.S. born, number of years of U.S. residency, and urban vs. rural residence.
Cancer Workgroup Report--Draft of December 16, 1998.
- Centers for Disease Control and Prevention (CDC), National Program of Cancer Registries (NPCR):
Improve the timeliness, quality, and completeness of cancer data in those 45 States, the District of Columbia, and three Territories that are funded by the Program.
Focus audits on the accuracy of classification of racial and ethnic data elements and the complete reporting of these data items.
Focus training on the collection of racial and ethnic data elements; the evaluation of the completeness and the accuracy of these data elements; and the importance of denominator data by race, ethnicity, sex, and age for State and county geographic areas.
Focus technical assistance on the generation and interpretation of cancer rates and percentages by race and ethnic group for reporting purposes.
- CDC, National Center for Health Statistics (NCHS):
Explore the feasibility of moving basic questions on mammograms and pap smear screening exams from NHIS supplements to the core questionnaire, so that they would be asked every year.
Cardiovascular Work Group Report--Draft of December 16, 1998.
- Current data needs for national data (NCHS/CDC/HCFA):
Vital statistics should be used to develop unbiased mortality rates for Hispanic subgroups, American Indian/Alaska Natives and for Asian/Pacific Islanders; and in conjunction with the Bureau of Census, accurate population data for smaller racial/ethnic groups.
Morbidity data should be used to develop hospitalization and prevalence rates for Hispanics, American Indian/Alaska Natives.
Needed data on cardiovascular risk factors include current risk factor data for Hispanic subgroups, Asian /Pacific Islander subgroups, and for American Indian/Alaska Natives.
Medical care databases should be exploited to yield current utilization data for Hispanics, Native America/Alaska Natives, and Asian/Pacific Islanders.
- Vital statistics may be better exploited for minority health by the following:
Encouraging representation on the National Committee on Vital and Health Statistics from the National Medical Association, COSSMHO, National Hispanic Medical Association, Association of Asian/Pacific Community Health organizations, Indian Health Services, and a historically black medical school.
Encouraging the solicitation of data concerns for planning NCHS surveys, including NHANES and NHIS, from the National Medical Association, Indian Health Service, COSSMHO, National Hispanic Medical Association, Association of Asian/Pacific Community Health organizations, and a historically black medical school.
Encouraging an assessment of the impact of race category changes for census 2000 and of ICD-10 on mortality rates.
- Morbidity among minorities should be monitored by the following:
Encouraging HCFA to collect data on Hispanics. Also, HCFA should collect information on race and ethnicityity, as described by OMB.
NIH, CDC, and/or their contractors should consider continuing to analyze HCFA Medicare data to establish approximate national incidence, case fatality, and one year survival trends for acute myocardial infarction, congestive heart failure, and acute stroke in each major racial and ethnic group.
NCHS should consider adding a Hispanic ethnicity category/methodology in data collection and improve completeness of race reporting in the National Hospital Discharge Survey.
- Data on risk factors in minorities may be enhanced by the following:
Considering a future census as a mechanism for selected risk factor data for smaller race and ethnic groups including American Indians.
Augmenting the BRFSS to obtain risk factor data on Hispanic subgroups, Asian/Pacific Islander subgroups, and American Indians/Alaska Natives.
Considering increased use of the Youth Risk Behavior Survey with oversampling of minority populations in order to obtain risk factor information on Hispanic, Asian/Pacific Islander, and American Indian/Alaska Native youth.
Considering implementation of cross-cutting recommendations on data issues from the COSSMHO National Hispanic Health Symposium and the HHS Hispanic Agenda for Action (http://www.cossmho.org/).
- Monitoring health care utilization by minorities should continue as follows:
Consider further analyses of HCFA and other national data sets by race and ethnicity.
Encourage HCFA to collect data on Hispanics.
Consider addition of Hispanic ethnicity to National Hospital Discharge Survey and other surveys.
Encouraging greater use of the BRFSS mechanism with oversampling of minority populations in order to calculate stable estimates of risk factors at both State and local levels.
Encouraging geocoding of all health-related data in order to determine patterns of health and disease among minority populations--mortality, morbidity, risk factor, health care utilization-- patterns which may be masked when evaluating data at the State level.
- Research data (NIH/CDC) may be enhanced by:
Encouraging funding of follow-up of adults examined in NHANES III for incidence of CVD (Cardiovascular Diseases [Major] ICD 9 Nos. 390-448).
- The Department should:
Complete implementation of unfinished items in Recommendation 7 (Improving and fully using available sources of data) of the 1985 HHS Secretary's Task Force on black and Minority Health.
Consider implementation of unfinished recommendations of the CVD Working Group of that Task Force (e.g. determining the incidence of sudden cardiac death in racial groups, and developing statistics for growing subgroups such as cardiac death in racial groups, and developing statistics for growing Asian subgroups such as Cambodians, Vietnamese, East Indians).
Diabetes Work Group Report--Draft of November 6, 1998.
- CDC should strengthen health surveillance data for racial and ethnic populations.
- Consistent data is needed for American Indians, Hispanic/Latinos; African Americans; and Asian/Pacific Islanders. The data should be accurate, timely, and complete.
- Existing surveillance systems should be expanded, and new data systems should be developed.
- An HHS Diabetes Data Coordinating Committee should be formed and charged with developing a plan for collecting, analyzing and disseminating timely public health surveillance data and information for all racial and ethnic populations. This committee will identify and collaborate with existing HHS committees working on data issues. The plan should be developed by March 1999, and systems in place by March 2000. The focus would be on improving data for monitoring trends in burden in the particular populations, for assessing diabetes care in health care systems, and for determining the extent of behavioral risk factors.
HIV/AIDS Work Group Report--Draft of September 30, 1998.
- There needs to be an increased awareness of HIV/AIDS disproportionate impact on communities of color. This should be measured by adding small numbers of psychometrically sound, relevant questions on national, representative surveys of the general population with over sampling of communities of color. (Measurement and analysis should be done before and after roll-out of new public information and education programs.)
- Agencies of the HHS and territorial, State and local health departments should modify current data collection methods and surveillance systems to disaggregate racial and ethnic communities by race, ethnicity, language, national origin, and tribal affiliation, report by oversampling where necessary, and conduct studies.
- Surrogate markers, such as sexually transmitted disease (STD) and unplanned pregnancy rates should be used to estimate HIV risk and to evaluate the effectiveness of risk reduction and prevention programs.
- NIH should actively recruit and retain members of racial and ethnic communities (in particular women, youth, and transgendered persons) for existing and new clinical treatment trials and vaccine trials. To facilitate this, NIH should develop a centralized data base containing demographic information and provide resources to address transportation, interpretation and translation needs. In particular, the use of community-based research sites will assist with this effort. In populations where there is under-representation, oversampling will be necessary. NIH should look for new and improved ways of addressing increased recruitment and retention needs among racial and ethnic groups.
- There is a need for better and more accurate data on health care indicators of American Indians and Alaska Natives and Pacific Islanders. This can be accomplished by closer work with service providers, consumers, epidemiologists, and providers in native communities.
- Although information on mode of HIV exposure is not required to transmit a case report to CDC, collection of risk information is very important.
- To allow the U.S. to target programs and resources most effectively, we must be able to monitor the epidemic. This means we need to improve our ability to track early HIV infections, before they progress to AIDS.
- As of December 1997, only 27 States send data to CDC on all confidentially reported cases of HIV infection. Two additional States report only cases of HIV infection among children under age 13, and one State reports cases only for children under the age of six.
- The CDC and Council of State and Territorial Epidemiologists have developed Suggested Guidelines for Developing an Epidemiologic Profile for HIV Prevention Community Planning. The Guidelines recommend that the epidemiologic profile in each prevention plan further characterize racial and ethnic groups such as the Asian/Pacific Islander population by origin (e.g., Chinese, Filipino, Japanese, Korean, Pacific Islander, etc.). However, such data is not readily available. Noting the need to collect valid information from smaller groups such as Asian/Pacific Islanders, the Guidelines also advise the collection of both quantitative and qualitative data. Qualitative data collection efforts include focus groups, key informant interviews, service provider surveys, community forums, and public hearings.
- Epidemiological research is needed on the incidence, prevalence, duration, and outcome of HIV/AIDS in various subgroups within the major racial ethnic categories presently reported. Data on the epidemiology of HIV disease within these subgroups are also needed. Epidemiologic data are needed not only to provide past prevalence risk profiles but also to project the estimates for new HIV infections. Differences for racial and ethnic groups in the rate, severity, and latency of HIV infection and the efficacy of various pharmaceutical treatments need to be examined.
Child and Adult Immunization Work Group Report--Draft of December 11, 1998.
- Systematic national surveillance of immunization rates in the target age group (Persons over 65 years of age) by race and ethnicity are needed to monitor progress toward the goal and evaluate the effectiveness of proposed interventions.
- Rigorous assessment of demonstration-scale programmatic interventions is needed before they are implemented nationally, as are criteria for selecting the most effective candidate interventions.
- In the long term, the Data Work Group of the Initiative to Eliminate Racial and Ethnic Disparities should create a Department-wide "warehouse" of results from immunization projects targeted at minority populations by community grantees or federal agencies. A centralized collection of data and systematic evaluation of effectiveness should help identify the approaches and types of intervention that are most successful in specific populations. Interventions that are consistently successful on a local scale should be funded on a national basis.
- Optimal disease surveillance is needed to detect changes in disease incidence and ensure that the impact of immunization programs on disease incidence can be evaluated.
- Ethnographic and other types of research on patient acceptability of immunization and reasons for non-immunization need to be conducted.
- Surveillance of immunization use must continuously assess progress in immunization coverage rates. These data are needed to ensure that immunization programs are achieving coverage and disease reduction goals within target populations, including patients who are traditionally underserved and who require special efforts to ensure that they are offered immunizations. As a result of this information, public health prevention strategies can be appropriately altered to improve their effectiveness.
Surveillance of Influenza and Pneumococcal Disease
- Expand and standardize disease-based surveillance for influenza both in the United States with a special emphasis on expanding networks of sentinel physicians and on establishing HMO and occupational-based influenza surveillance.
- Perform annual rapid assessments of influenza-vaccine effectiveness among vulnerable U.S. populations.
- Evaluate and modify current State activity assessment levels for influenza.
- Continue and enhance current active, population-based surveillance for invasive pneumococcal disease, including additional studies to identify causes of higher incidence rates among American Indians, blacks and possibly other racial and ethnic minority groups.
Research on Patient Acceptability of Vaccination
- Conduct ethnographic studies to identify factors influencing vaccination behavior among racial and ethnic minorities.
- Supplement vaccination coverage questions on surveys with additional questions related to doctor recommendations, reasons for vaccination and non-vaccination, and type of setting where vaccination is received.
- Conduct controlled studies of provider-based and patient-based interventions to improve vaccination levels in racial and ethnic minority populations.
Surveillance of Vaccination Use
- Improve time of availability of NHIS- and BRFSS-reported vaccination levels to less than one year after data are collected.
- Assess the feasibility of using the NIS sample frame to survey adults.
- Conduct special surveys of vaccination levels for Asian/Pacific Islander and American Indian/Alaska Native populations aged 65 years or more, or consider adding vaccination questions on existing surveys of these populations.
- Surveys of Hispanic, AA/PI and American Indian populations should collect information on national origin.
- Assess the validity of self-reported pneumococcal vaccination.
- Assess if managed care organizations using the HEDIS influenza vaccination measure can stratify vaccination levels by racial and ethnic group.
- Encourage development of a HEDIS measure for pneumococcal vaccination (CDC is currently working with a managed care organization to develop and evaluate such a measure).
- Consider development of mechanisms to monitor immunizations at workplaces and free-standing sites (e.g., grocery stores, pharmacies) during vaccine campaigns.
- Current efforts to assess vaccination coverage, including the National Immunization Survey, the National Health Interview Survey, and immunization assessment programs in the 50 States and 28 urban areas should continue.
- Collection of data on Hispanic ethnicity should be expanded to include specific data on ancestry (i.e. Cuban, Puerto Rican, Mexican, etc.) because this information would allow the design of more appropriate activities to increase coverage in different Hispanic communities. Similar efforts are needed for Asians and American Indian/Alaska Natives.
Efforts to collect information on race and ethnicity and socioeconomic status should be included in assessments of coverage at the local level.
- Studies of the effectiveness of interventions to increase coverage in specific minority communities should form the basis for expanding interventions beyond the pilot and demonstration scale.
- Additional progress in eliminating disparities in coverage can be expected from allocation of additional resources to existing programs and proven strategies (i.e., linkage of immunization efforts and the WIC program, targeted assessments of immunization coverage among private providers who serve poor infants and toddlers, partnerships exemplified by the TOTS program), especially if these efforts are buttressed by surveillance, immunization coverage by race and ethnicity, and socioeconomic status at the local level.
- In the longer term, implementation of a system of State-based immunization registries could be used to evaluate the success of programs that target populations at risk of undercoverage.
- Continue to collect information on vaccine coverage, representative of the U.S. population, and including race and ethnicity and socioeconomic status.
- Expand the Hispanic ethnicity question to assess ancestry information, since this information would allow the design of more appropriate activities to increase coverage.
- Increase efforts to assess coverage at the local level and include race and ethnicity and socioeconomic status with simultaneous efforts to protect confidentiality of participants.
- Conduct studies to evaluate interventions to increase vaccine coverage among specific racial/ethnic minorities at risk of undercoverage and implement activities that have demonstrated effectiveness."
- Although some States have already implemented immunization registries, in the long term, a national system of registries is needed. These systems will provide coverage information at the individual level as well as in aggregate summaries. On an individual level, each registry will serve to identify persons needing immunizations, generate reminders to the person, and provide information on contraindications and adverse events. On a population level, registries will provide summaries of the immunization status and identify populations in need of immunizations. Furthermore, data from registries could be used to evaluate the success of programs targeting populations at risk of undercoverage.
- Population based surveys will remain essential in the long term to evaluate representativeness and validity of registries.