Surveillance of Health Status in Minority Communities --- Racial and Ethnic Approaches to Community Health Across the U.S. (REACH U.S.) Risk Factor Survey, United States, 2009

Youlian Liao, MD

David Bang, PhD

Shannon Cosgrove, MHA

Rick Dulin, BS

Zachery Harris, BS

Alexandria Stewart, BS

April Taylor, MPH

Shannon White, MPH

Graydon Yatabe, MPH

Leandris Liburd, PhD

Wayne Giles, MD

Division of Adult and Community Health, National Center for Chronic Disease Prevention and Health Promotion

Corresponding author: Youlian Liao, MD, Division of Adult and Community Health, National Center for Chronic Disease Prevention and Health Promotion, CDC, MS K-30, 4770 Buford Hwy, N.E., Atlanta, GA 30341. Telephone: 770-488-5299; Fax: 770-488-5974; E-mail: ycl1@cdc.gov.

Abstract

Problem: Substantial racial/ethnic health disparities exist in the United States. Although the populations of racial and ethnic minorities are growing at a rapid pace, large-scale community-based surveys and surveillance systems designed to monitor the health status of minority populations are limited. CDC conducts the Racial and Ethnic Approaches to Community Health across the U.S. (REACH U.S.) Risk Factor Survey annually in minority communities. The survey focuses on black, Hispanic, Asian (including Native Hawaiian and Other Pacific Islander), and American Indian (AI) populations.

Reporting Period Covered: 2009.

Description of System: An address-based sampling design was used in the survey in 28 communities located in 17 states (Arizona, California, Georgia, Hawaii, Illinois, Massachusetts, Michigan, New Mexico, New York, North Carolina, Ohio, Oklahoma, Pennsylvania, South Carolina, Virginia, West Virginia, and Washington). Self-reported data were collected through telephone, questionnaire mailing, and in-person interviews from an average of 900 residents aged ≥ 18 years in each community. Data from the community were compared with data derived from the Behavioral Risk Factor Surveillance System (BRFSS) for the metropolitan and micropolitan statistical area (MMSA), county, or state in which the community was located and also compared with national estimates.

Results: Reported education level and household income were markedly lower in black, Hispanic, and AI communities than that among the general population living in the comparison MMSA, county, or state. More residents in these minority populations did not have health-care coverage and did not see a doctor because of the cost.

Substantial variations were identified in the prevalence of health-related risk factors among minority populations and among communities within the same racial/ethnic population. In 2009, the median prevalence of obesity among Asian/Pacific Islander (A/PI) men and women was 10.3% (range: 4.8%--45.3%) and 6.7% (range: 4.5%--38.2%), respectively, whereas it was 46.2% (range: 39.4%--53.6%) and 45.5% (range: 35.1%--55.1%), respectively, among AI men and women. The median percentage of cigarette smoking among black (28.0% in men and 19.9% in women) and AI communities (36.1% in men and 36.0% in women) was much higher than the national median (19.6% in men and 16.8% in women) among the 50 states and the District of Columbia (DC). Among the four minority communities, blacks had the highest median percentage of persons who reported engaging in no leisure-time physical activity (28.5% in men and 31.6% in women). A much lower percentage of black women met physical activity recommendations in almost all communities compared with that in the corresponding MMSA, county, or state.

Substantial variations were identified in self-perceived health status and prevalence of selected chronic conditions among minority populations and among communities within the same racial/ethnic population. In 2009, the median percentage of men who reported fair or poor health was 15.8% (range: 8.3%--29.3%) among A/PI communities and 26.3% (range: 22.3%--30.8%) among AI communities. The median percentage of women who reported fair or poor health was 20.1% (range: 13.3%--37.2%) among A/PI communities, whereas it was 31.3% (range: 19.4%--44.2%) among Hispanic communities. AI and black communities had a high prevalence of self-reported hypertension, cardiovascular disease, and diabetes. For most communities, prevalence was much higher than that in the corresponding MMSA, county, or state in which the community was located. The median percentages of persons who knew the signs and symptoms of a heart attack and stroke were consistently lower in all four minority communities than the national median.

Variations were identified among racial/ethnic populations in the use of preventive services. Hispanics had the lowest percentages of persons who had their cholesterol checked, of those with high blood pressure who were taking antihypertensive medication, and of those with diabetes who had a glycosylated hemoglobin (HbA₁C) test in the past year. AIs had the lowest mammography screening rate within 2 years among women aged ≥40 years (median: 72.7%; range: 69.4%--76.2%). A/PIs had the lowest Pap smear screening rate within 3 years (median: 74.4%; range: 60.3%--80.8%). The median influenza vaccination rates in adults aged ≥65 years were much lower among black (57.3%) and Hispanic communities (63.3%) than the national median (70.1%) among the 50 states and DC. Pneumococcal vaccination rates also were lower in black (60.5%), Hispanic (58.5%), and A/PI (59.7%) communities than the national median (68.5%).

Interpretations: Data from the REACH U.S. Risk Factor Survey demonstrate that residents in most of the minority communities continue to have lower socioeconomic status, greater barriers to health-care access, and greater risks for and burden of disease compared with the general populations living in the same MMSA, county, or state. Substantial variations in prevalence of risk factors, chronic conditions, and use of preventive services among different minority populations and different communities within the same racial/ethnic population provide opportunities for public health intervention. These variations also indicate that different priorities are needed to eliminate health disparities for different communities.

Public Health Action: These community-level survey data are being used by CDC and community coalitions to implement, monitor, and evaluate intervention programs in each community. Continuous surveillance of health status in minority communities is necessary so that community-specific, culturally sensitive strategies that include system, environmental, and individual-level changes can be tailored to these communities.

Introduction

Substantial racial/ethnic health disparities have been identified in the United States (1). In 2006, one of every four U.S. residents identified themselves as being a racial or ethnic minority (2). The U.S. Census Bureau projects that by 2050, populations that have historically been called "minorities" will make up approximately 50% of the U.S. population (3). Achieving a healthy nation is impossible without healthy minority populations and without eliminating racial/ethnic health disparities. Eliminating health disparities was one of the goals of Healthy People 2010 (4). Individual health is closely linked to the health of the community* and environment in which persons live, work, and play (4). The health of a person is inseparable from the health of the larger community and the health of every community in every state/territory determines the overall health status of the nation. Healthy People 2010 called for community partnerships in building healthy communities (4).

CDC launched the Racial and Ethnic Approaches to Community Health (REACH) in 1999 (5). During 2000--2006, CDC funded the initial phase of the REACH project, REACH 2010, supporting 42 community coalitions in designing, implementing, and evaluating community-driven strategies to eliminate health disparities (5). The project supported the development and implementation of innovative approaches to working with racial and ethnic minority populations. REACH 2010 demonstrated that health disparities can be reduced and that the health status of populations traditionally most affected by health inequities can be improved (6). Building on the successes of the initial phase of the project, a new funding phase, REACH Across the U.S. (REACH U.S.), was launched in 2007. REACH U.S. endeavors to address the social determinants of health through policy, environmental, and system change and to disseminate effective strategies to more community partners (7). A total of 40 communities were selected competitively and funded by CDC on the basis of the cultural relevance of the proposed intervention, investigator expertise, community participatory strategies, and demonstrated success in prior training, translation, and dissemination activities. These 40 funded communities focused on one or more racial and ethnic population, including blacks, Hispanics, Asians (including Native Hawaiians and Other Pacific Islanders), and American Indians (AIs). The health focus areas include cardiovascular disease, diabetes mellitus, breast and cervical cancer, adult/older adult immunization, hepatitis B, asthma, and infant mortality.

In 2001, to monitor the health of racial/ethnic minority communities and as part of the REACH project evaluation, CDC began to conduct annual REACH 2010 Risk Factor Surveys. Baseline data for the REACH 2010 project have been published previously (8). This report presents data from the first survey year of the REACH U.S. project collected during May--November 2009.

Methods

Survey Communities

CDC contracted with the National Opinion Research Center (NORC) at the University of Chicago to conduct the REACH U.S. Risk Factor Survey in 28 of the 40 U.S. REACH communities.† The survey included adult residents aged ≥18 years. The 28 communities participating in the survey were located in 17 states.§

The survey areas and populations were consistent with the focus of the intervention programs. The areas included specific counties, census tracts, zip codes, neighborhood areas, or tribal areas (Table 1). The size of the survey area varied by community, ranging from a small neighborhood (e.g., within the boundaries of four specific streets in west Philadelphia) to a whole county (e.g., Pima County, Arizona) and from several census tracts or zip codes to an entire state (e.g., Oklahoma). Among the five communities in California, some geographic overlapping occurred. The survey was conducted independently in each community. The REACH project led by the Morehouse School of Medicine, the Southeastern U.S. Collaborative Center of Excellence for the Elimination of Disparities, involved several areas in south Georgia, South Carolina, and North Carolina. The survey for this project included only the intervention areas in Fulton County, Georgia. In 20 REACH communities, just one minority population was focused (black or Haitian American: eight; Hispanic: four; Asian or Native Hawaiian/Other Pacific Islander [Asian/Pacific Islander (A/PI)]: four; and AI: four); six communities focused on both blacks and Hispanics; and two communities focused on three minority populations (blacks, Hispanics, and A/PIs) (Table 2).

Address-Based Sampling Design

An address-based sampling method was used in the REACH U.S. Risk Factor Survey to reduce the potential coverage bias of traditional random-digit--dialing. When the previous REACH 2010 Risk Factor Survey was initiated in 2001, random-digit--dialing was the preferred mode of data collection, except in three communities in which in-person face-to-face interviews were performed because of low telephone coverage (8). Since then, use of cellular telephones has increased. In 2009, 25% of U.S. homes had only cellular phones (9). In addition, 15% received most or all of their calls on cellular phones even though they had a landline. Therefore, up to 40% of U.S. homes might not be reachable by traditional random-digit--dialing used in most surveys targeting landlines. Minority populations have embraced cellular phones at a higher rate than the majority, increasing the risk for coverage bias in REACH communities (9). The basis of the address-based sampling frame is the U.S. Postal Service delivery sequence file. This file contains nearly all addresses in the United States that receive mail. Geographic information systems technology was used to construct an address frame that matched the intervention geographies of the REACH program. After a sample of addresses was selected randomly, the addresses were matched to telephone numbers. The median matching rate was 60% (range: 40%--71%). Advance letters describing the survey were sent to sampled households with known telephone numbers. The survey was conducted by telephone for these addresses. Self-administered questionnaires were mailed to households without a phone match and to those who did not respond by telephone. Finally, an in-person follow-up to a subset of nonresponders was conducted.

To increase the efficiency of the survey, designers purposely selected those addresses that were more likely to be households of the survey race/ethnicity. These addresses were identified by aggregating data from multiple sources (e.g., residential directory listings, administrative data, and consumer transactions).

For telephone and in-person interviews, the household screening was conducted with any household member aged ≥18 years to ascertain the age and racial/ethnic eligibility of each adult household member. Up to two eligible adults were selected for further household member interviews. For the survey by mail, all household members were invited to complete the mailed survey. The goal was to survey 900 adult household members in each community.

Complete data were collected from 24,117 eligible household members. Of these, 19,177 (79%) were from telephone interviews, 4,022 (17%) from questionnaire mailings, and 918 (4%) from in-person interviews. Among household members who were contacted by telephone, 43% cooperated with the screening interview (range: 25%-- 65%). The completion rate of detailed household member interviews was 42% (range: 26%--58%) for eligible household members. In the questionnaire mailing, the mail return rate was 20% (range: 8%--25%). In the in-person survey, the screening completion rate was 75% (range: 48%--88%), and the completion rate of household member interviews was 72% (range: 50%--92%).

Questionnaires

A uniform questionnaire was used for all communities and was administered in English, Spanish, Vietnamese, Khmer, Haitian Creole, or Chinese (Mandarin or Cantonese). The questionnaire included questions regarding respondents' demographics, socioeconomic status (e.g., education and income), perceived health status, health-care access, self-reported height and weight, leisure-time physical activity, fruit and vegetable intake, cigarette smoking, awareness of hypertension, cardiovascular diseases, diabetes and diabetes care, and receipt of preventive services (e.g., cholesterol screening, mammography, Papanicolaou [Pap] smear test, and influenza and pneumococcal vaccination). The questions were identical to those used in the Behavioral Risk Factor Surveillance System (BRFSS) (10).

Comparison Populations

Health status and risk factor level in the REACH U.S. communities were compared with those in the general population from BRFSS. BRFSS is a state-based telephone survey of civilian residents aged ≥18 years (10). Typically, BRFSS reports state-level estimates. Since 2002, BRFSS has had sufficient samples to produce area-level estimates for selected metropolitan and micropolitan statistical areas (MMSAs), metropolitan divisions, and selected counties. For 13 REACH U.S. communities, data from each community were compared with data from BRFSS in the MMSA in which the community was located (Table 1). For eight communities, community data were compared with county data from BRFSS. For the remaining seven communities that could not be matched to a specific MMSA or county, state-specific BRFSS data were used for comparisons.

The sample size of respondents from the comparison populations was increased by combining MMSA- or county-level data from 2007--2009 survey years and state-level data from 2008--2009 survey years (Table 2). Certain BRFSS data were available only in the 2007 and 2009 survey years (e.g., meeting physical activity recommendation, fruit and vegetable intake, having cholesterol checked, awareness of hypertension, and the use of antihypertensive medication among those who reported having high blood pressure). Certain questions were included in BRFSS as the optional modules that were asked only in certain states (either all samples or subsamples) for selected years. Therefore, state-level data from the combined years of 2007--2009 were used whenever available to compare with community-level data for the following modules: signs and symptoms of heart attack and stroke (2007 and 2009), diabetes care (glycosylated hemoglobin [HbA₁C] test, foot, and eye exams) (2007--2009), and women's health (mammography and pap smear test) (2007--2009). Aggregated data across communities by race/ethnicity also were compared with the national estimates for all 50 states and the District of Columbia (DC) in the 2009 BRFSS (women's health data in 2008).

Data Analysis

The prevalence of risk factors, chronic conditions, and access to and use of preventive services was estimated by community, four racial/ethnic populations, and sex. In Boston, Massachusetts, the surveyed minority population, Haitian Americans, were reported under the "black" racial category. The term "Asian/Pacific Islander (A/PI)" included various Asian populations, Native Hawaiians, and Other Pacific Islanders. For example, the community in Waianae, Hawaii, focused on Native Hawaiians and Other Pacific Islanders. A community in New York City focused on Chinese, Koreans, and other Asians. Special Services for Groups in California focused on Cambodians, Filipinos, Laotians, Vietnamese, and Samoans. Orange County Asian and Pacific Islanders Community Alliance focused on Cambodians, Hmong, Thais, Vietnamese, Laotians, Native Hawaiians, Samoans, Tongans, Marshallese, and Guamanians/Chamorros, all of whom were grouped under A/PIs. So that sample sizes could be increased, data from men and women were combined in the analyses for the following variables: taking antihypertensive medication among those with reported hypertension, HbA₁C tests and foot and eye examinations among persons with diabetes, and vaccinations among persons aged ≥65 years.

In the calculation of prevalence, persons who replied "don't know" or who refused to answer the questions were excluded from the denominator. If a denominator was <30, the prevalence estimate was considered unstable and was not presented. SUDAAN was used in the analysis to account for the complex sampling design and to calculate the 95% confidence intervals for both the REACH U.S. and BRFSS data. For the REACH U.S. survey, each sample was weighted to reflect the probability of selection, the number of eligible members, and the number of selected members at the sampled address and was adjusted by age-gender population sizes of members of the surveyed minority population. For BRFSS, each sample was weighted to reflect the adult population for each MMSA, county, or state. Use of MMSA, county, or state BRFSS as the standard permits a percentage estimated from a community to be described as being higher than the standard if the percentage in the MMSA, county, or state is lower than the lower limit of the confidence interval of the percentage in the community (11). Conversely, the percentage estimated for the community can be described as being lower than the standard if the percentage in MMSA, county, or state is higher than the upper limit of the confidence interval. These comparisons were not used as an indicator of statistical significance in a formal statistical test.

Results

Social Demography and Access to Health Care

Education

Among men, the median percentage of adults who reported having less than a high school education ranged from 6.3% in A/PI communities to 29.5% in Hispanic communities (Table 3). Among women, the median percentage ranged from 9.3% in A/PI communities to 31.5% in Hispanic communities. Except for A/PI men, these medians were higher than the national median among the 50 states and the DC in the 2009 BRFSS (8.5% in men and 7.5% in women). With a few exceptions, a much higher percentage of black and Hispanic men and women in the surveyed communities reported having less than a high school education compared with that in the general populations from the same MMSA, county, or state in BRFSS. Among minority populations, Hispanics reported the lowest education level, whereas A/PIs had the highest education level. However, substantial community variations were identified within certain racial/ethnic populations. For example, <2% of A/PI men or women reported having less than a high school education in the surveyed community in Los Angeles and Orange Counties, California, compared with 21.3% of men and 27.1%, of women from the surveyed community in New York City.

Household Income

The median percentage of men who reported having annual household income of <$25,000 ranged from 27.2% in A/PI communities to 45.4% in Hispanic communities (Table 4). The median percentage among women ranged from 27.6% in A/PI communities to 50.5% in black communities. All these medians were higher than the national median percentage of household income of <$25,000 among the 50 states and DC in 2009 (21.3% in men and 26.4% in women).The median percentages of persons with incomes of <$25,000 in black and Hispanic communities were about twice the national level. A substantially higher percentage of men and women reported low income in all Hispanic and AI communities and in the vast majority of black communities compared with that in the comparison MMSA, county, or state populations. Substantial community variations were identified within the same racial/ethnic population. For example, 12.1% and 19.2% of A/PI men and women, respectively, had an income of <$25,000 in the surveyed community in Seattle and King County compared with 41.6% and 46.9%, respectively, in the community in New York City.

Health-Care Coverage

Respondents were asked if they had any kind of health-care coverage, including health insurance, prepaid plans such as HMOs, or government plans such as Medicare. The median percentage of men who reported having no health-care coverage ranged from 14.8% in A/PI communities to 29.2% in Hispanic communities (Table 5). The median percentage among women ranged from 15.2% in A/PI communities to 26.7% in AI communities. Except for A/PI men, these medians were higher than the national median among the 50 states and DC in 2009 (16.5% in men and 12.4% in women). Higher percentages of men and women without health-care coverage were reported in the majority of black, Hispanic, and AI communities compared with those in the corresponding MMSA, county, or state in which the community was located. Within the same racial/ethnic population, a two- to fourfold difference was identified in coverage rates across communities.

Cost as a Barrier to Obtaining Health Care

Respondents were asked whether at any time in the previous 12 months they had needed to see a doctor but could not do so because of cost. The median percentage of men who could not see a doctor because of cost ranged from 10.6% in A/PI communities to 21.8% in Hispanic communities (Table 6). The median percentage among women ranged from 14.0% in A/PI communities to 25.6% in Hispanic communities. Except for A/PIs, these medians were higher than the national median in 2009 (12.2% in men and 14.7% in women). A higher percentage of adults who had not seen a doctor because of the cost was reported consistently in Hispanic communities compared with the corresponding MMSA, county, or state in which the community was located. A greater variation in the one-on-one comparison between the individual community and BRFSS counterpart was observed in black, A/PI, and AI communities.

Chronic Disease Risk Factors

Obesity

Obesity is defined as body mass index ≥30 kg/m2 calculated from self-reported height and weight. The median percentage of obesity among men ranged from 10.3% in A/PI communities to 46.2% in AI communities (Table 7). The median percentage among women ranged from 6.7% in A/PI communities to 45.5% in AI communities. Except for A/PI communities, these median percentages were higher than the national median among the 50 states and DC in 2009 (28.6% in men and 26.0% in women). The prevalence of obesity was substantially higher among both men and women in AI communities and among women in black communities compared with that in the comparison MMSA, county, or state in which the community was located. Overall, approximately 45% of AI men and women and black women were obese in the surveyed communities, whereas slightly more than one fourth of adults were obese at the national level. Obesity was uncommon in most A/PI communities. However, obesity prevalence in Native Hawaiians/Other Pacific Islanders (45.3% in men and 38.2% in women) was much higher than in the general population in Honolulu (24.5% in men and 19.9% in women). Two- to threefold differences in obesity prevalence were reported across black communities.

Cigarette Smoking

Cigarette smokers were defined as those who had ever smoked ≥100 cigarettes and who currently smoke. The median percentage of cigarette smoking among men ranged from 13.8% in A/PI communities to 36.1% in AI communities (Table 8). The median percentage among women ranged from 3.7% in A/PI communities to 36.0% in AI communities. The median percentages among black and AI communities were higher than the national median among the 50 states and DC in 2009 (19.6% in men and 16.8% in women). Substantial community variations in prevalence of smoking within the same racial/ethnic population were identified. The difference could be up to fourfold across communities. For example, the prevalence of smoking was <5% in women in the majority of A/PI communities, compared with 21.3% in Waianae, Hawaii.

Leisure-Time Physical Activity

Respondents were asked if they had participated, other than as a part of their regular job, in any physical activities or exercises (e.g., running, calisthenics, golf, gardening, or walking for exercise) during the past month. The median percentage of men who reported having no leisure-time physical activity ranged from 23.6% in A/PI communities to 28.5% in black communities (Table 9). The median percentage among women ranged from 25.6% in A/PI communities to 31.6% in black communities. Except for women in A/PI communities, these median percentages were higher than the national median among the 50 states and DC in 2009 (21.5% in men and 25.6% in women). Substantial variations were identified within the same racial/ethnic population in one-on-one comparisons between individual communities and the corresponding MMSA, county, or state.

Met Physical Activity Recommendation

Respondents were asked to recall their overall frequency and duration of time spent in moderate activities (e.g., brisk walking, bicycling, vacuuming, or gardening) and vigorous activities (e.g., running, aerobics, or heavy yard work) in a typical week. Persons were defined as meeting recommended physical activity levels if they reported participating in either moderate physical activity ≥30 minutes/day, 5 days/week, or vigorous physical activity ≥20 minutes/day, 3 days/week (12). The median percentage of men who met physical activity recommendations ranged from 39.8% in A/PI communities to 47.9% in AI communities (Table 10). The median percentage among women ranged from 34.6% in black communities to 44.9% in AI communities. These median percentages were all lower than the national median percentage among the 50 states and DC in 2009 (52.4% in men and 47.8% in women). With very few exceptions, a much lower percentage of black women met physical activity recommendations in the surveyed communities compared with that in the corresponding MMSA, county, or state in which the community was located. The results were less consistent in one-on-one comparisons between individual communities and the corresponding BRFSS counterparts for Hispanics, A/PIs, and AIs.

Fruit and Vegetable Intake

Fruit and vegetable intake was calculated from six questions regarding the intake of fruit juices, fruit, green salad, potatoes, carrots, and other vegetables. A national education program has advocated eating at least five servings of fruits and vegetables daily (13). The median percentage of men who reported eating at least five fruits and vegetables daily ranged from 19.6% in Hispanic communities to 23.8% in black and A/PI communities (Table 11). The median percentage among women ranged from 23.5% in AI communities to 33.6% in A/PI communities. Overall the differences between these medians and the national median percentage (19.2% in men and 27.7% in women) were small. The percentage of women who reported eating at least five fruits and vegetables daily was much lower in most of the Hispanic communities than in the corresponding MMSA, county, or state. One-on-one comparisons between individual communities and the corresponding BRFSS counterparts were less consistent in other racial populations.

Health Status and Selected Chronic Conditions

Perceived Health Status

Respondents were asked to rate their own general health as either "excellent," "very good," "good," "fair," or "poor." The median percentages of men who reported fair or poor health ranged from 15.8% in A/PI communities to 26.3% in AI communities (Table 12). The median percentages among women ranged from 20.1% in A/PI communities to 31.3% in Hispanic communities. These medians were all higher than the national median percentage among the 50 states and DC in 2009 (13.4% in men and 15.3% in women). With few exceptions, a substantially higher percentage of men and women in Hispanic communities and women in black communities reported fair or poor health compared with that in the comparison MMSA, county, or state populations. The one-on-one comparisons between individual communities and the corresponding BRFSS counterparts were less consistent for A/PI and AI communities.

High Blood Pressure

The prevalence of high blood pressure was assessed by asking respondents, "Have you ever been told by a doctor, nurse, or other professional that you have high blood pressure?" Adults who reported prehypertension or borderline high blood pressure, and females who reported high blood pressure during pregnancy, were not considered as having hypertension. The median prevalence of high blood pressure among men ranged from 22.8% in Hispanic communities to 43.9% in AI communities (Table 13). The median prevalence among women ranged from 24.0% in A/PI communities to 46.2% in black communities. The medians among black and AI communities were much higher than the national median among the 50 states and DC in 2009 (29.8% in men and 27.8% in women). With very limited exceptions, the prevalence of high blood pressure was substantially higher in black and AI communities than in the corresponding MMSA, county, or state. The results of one-on-one comparisons between individual communities and the corresponding BRFSS counterparts were less consistent for Hispanic and A/PI communities. Substantial community variations in the prevalence of high blood pressure were identified within the same racial/ethnic population. For example, the prevalence of high blood pressure was 12.9% and 15.0% in Hispanic men and women, respectively, in the surveyed community in Seattle/King County, whereas it was 42.5% in Hispanic men in the surveyed community of Grant and Hidalgo Counties, New Mexico, and 36.5% in Hispanic women in the surveyed community of East Harlem, New York. The prevalence of high blood pressure was 23.7% and 16.8% in A/PI men and women, respectively in Seattle/King County, whereas the prevalence was twice as high in A/PI men in the surveyed community in Los Angeles and Orange Counties, California (46.7%) and in A/PI women in the surveyed community of Waianae, Hawaii (38.1%).

Cardiovascular Diseases

The percentage of cardiovascular diseases was assessed by asking respondents if they had ever been told by a doctor that they had any of the following conditions: heart attack or myocardial infarction, angina or coronary heart disease, or stroke. The median prevalence of cardiovascular diseases among men ranged from 6.6% in A/PI communities to 13.4% in AI communities (Table 14). The median prevalence among women ranged from 4.4% in A/PI communities to 12.3% in AI communities. The medians were higher among AI communities, and lower among Hispanic and A/PI communities than the national median among the 50 states and DC in 2009 (8.8% in men and 6.3% in women). A higher prevalence of cardiovascular diseases was apparent among women in the majority of black and AI communities compared with that in the corresponding MMSA, county, or state. One-on-one comparisons between the individual communities and the corresponding BRFSS counterparts were not consistent for Hispanic and A/PI communities.

Diabetes

The prevalence of diagnosed diabetes was assessed by asking respondents, "Have you ever been told by a doctor that you have diabetes?" The median prevalence of diabetes (excluding gestational, borderline, or prediabetes) among men ranged from 10.0% in Hispanic communities to 18.0% in AI communities (Table 15). The median prevalence among women ranged from 10.3% in A/PI communities to 18.4% in AI communities. All these medians were higher than the national median among the 50 states and DC in 2009 (8.8% in men and 8.2% in women). With a few exceptions, the prevalence of diabetes was substantially higher in both men and women in black and AI communities than in the comparison MMSA, county, or states. The prevalence was also much higher in women in Hispanic communities compared with their BRFSS counterparts. Substantial community variations (two- to fivefold difference) were identified in the prevalence of diabetes within the same racial/ethnic population.

Knowledge of Heart Attack Symptoms

A respondent was considered as having knowledge of heart attack signs and symptoms and the action to take if he or she correctly answered "yes" to all five questions on symptoms of heart attack, "no" to the incorrect symptom, and "call 911" when the responder thought someone was having a heart attack (14). The five symptoms of heart attack were 1) pain or discomfort in the jaw, neck, or back; 2) feeling weak, lightheaded, or faint; 3) chest pain or discomfort; 4) pain or discomfort in the arms or shoulder; and 5) shortness of breath. An incorrect symptom was sudden trouble seeing in one or both eyes. The median percentage of men who knew heart attack symptoms and action ranged from 3.2% in Hispanic communities to 8.7% in AI communities (Table 16). The median percentage among women ranged from 4.7% in Hispanic communities to 13.5% in AI communities. These median percentages were all lower than the median percentage of 33 states¶ and DC in the 2009 BRFSS (10.4% in men and 14.7% in women). The percentage of adults who knew heart attack symptoms and action was substantially lower in almost all black, Hispanic, and A/PI communities than it was in the comparison states or the median percentage of the 33 states and DC when state-specific data were not available. Within the same racial/ethnic populations, community variations in the percentage of persons who had knowledge of heart attack symptoms were substantial.

Knowledge of Stroke Symptoms

A respondent was considered as having knowledge of stroke signs and symptoms and the action to take if he or she correctly answered "yes" to all five questions on symptoms of stroke, "no" to the incorrect symptom, and "call 911" when the respondent thought that someone was having a stroke. The five symptoms of stroke were 1) sudden confusion or trouble speaking; 2) sudden numbness or weakness of the face, arm, or leg, especially on one side; 3) sudden trouble seeing in one or both eyes; 4) sudden trouble walking, dizziness, or loss of balance; and 5) severe headache without known cause. An incorrect symptom was sudden chest pain or discomfort.

The median percentage of men who knew stroke signs and symptoms and what action to take ranged from 7.9% in A/PI communities to 14.5% in AI communities (Table 17). The median percentage among women ranged from 7.0% in Hispanic communities to 15.4% in AI communities. These median percentages were all much lower than the median percentage among 33 states and DC in the 2009 BRFSS (19.4% in men and 21.0% in women). With very few exceptions, the percentage of adults who knew stroke symptoms and action was substantially lower in black, Hispanic, and A/PI communities than that in the corresponding states or the median of the 33 states and DC when state-specific data were not available. Within the same racial/ethnic population, substantial community variations were identified in the percentage of persons who had the knowledge of stroke symptoms.

Clinical Care and Preventive Services

Blood Cholesterol Checked

The median percentage of men who reported having their blood cholesterol checked during the preceding 5 years ranged from 60.5% in Hispanic communities to 73.6% in black communities (Table 18). The median percentage among women ranged from 67.4% in Hispanic communities to 80.5% in black communities. The median percentage was much lower than the national median percentage in both men (74.4%) and women (79.2%) among the 50 states and DC. A substantially lower percentage of men and women reported having cholesterol checked in the preceding 5 years in almost all Hispanic communities compared with that of the population in the comparison MMSA, county, or state in which the community was located. One-on-one comparisons between the community and corresponding MMSA, county, or state were less consistent in other minority populations.

Antihypertensive Medication Usage

The median percentage of adults with hypertension who reported taking antihypertensive medication ranged from 68.7% in Hispanic communities to 81.0% in black communities (Table 19). The median percentage among Hispanic communities was lower than the national median percentage (79.2%) among the 50 states and DC in 2009. The difference between the median percentage and the national median was small for the other three minority populations.

Preventive Care in Persons with Diabetes

Respondents who reported having diabetes were asked whether in the preceding 12 months, they had 1) an HbA₁C test, 2) their feet checked for any sores or irritations by a health professional, and 3) a dilated eye exam. The median percentages of respondents with diabetes who had none of these three exams were all <3% in the four minority populations. The median percentage of persons with all three exams ranged from 46.2% in Hispanic communities to 59.6% in A/PI communities.

HbA₁C. The median percentage of adults with diabetes who reported having had an HbA₁C test within the previous year ranged from 64.1% in Hispanic communities to 79.5% in AI communities (Table 20). The median percentage among black, Hispanic, and A/PI communities was lower than the national median percentage (79.7%) among 45 states (all states except Arkansas, Idaho, Mississippi, Oklahoma, and South Dakota) and DC in 2009. Substantially lower percentages of adults having an HbA₁C test within a year were reported in almost all Hispanic communities compared with those in the comparison state populations. One-on-one comparisons between individual communities and the corresponding states were less consistent in other minority populations.

Foot exam. The median percentage of adults with diabetes who reported having had their feet checked by a health professional within the previous year ranged from 68.7% in Hispanic communities to 82.0% in black communities (Table 21). The median percentage among Hispanic communities was lower, but higher among black and AI communities, than the national median percentage (73.4%) among 45 states and DC in 2009. Much higher percentages of adults who had their feet checked within the previous year were reported in the majority of black and AI communities compared with those in the corresponding states.

Dilated eye exam. The median percentage of adults with diabetes who reported having had a dilated eye exam within the previous year ranged from 71.3% in Hispanic communities to 78.3% in A/PI communities (Table 22). These median percentages were somewhat higher than the national median percentage (69.2%) among 45 states and DC in 2009. However, no substantive difference in the eye exam rate was identified in the majority of the one-on-one comparisons between the individual community and the corresponding state in which the community was located.

Women's Cancer Screening

Mammography

The median percentage of women aged ≥40 years who reported having had a mammogram during the previous 2 years ranged from 72.7% in AI communities to 80.7% in black communities (Table 23). These medians were lower than the national median percentage (83.2%) among the 50 states and DC in 2008 BRFSS. A lower mammography screening rate was reported in AI communities than in the corresponding state in which the community was located. The results of one-on-one comparisons between individual communities and their BRFSS counterparts were less consistent for the other minority populations.

Pap Smear Test

The median percentage of women with an intact uterine cervix who reported having had a Pap smear screening during the previous 3 years ranged from 74.4% in A/PI communities to 85.0% in black communities (Table 24). The median percentage among A/PI communities, but not among the other three minority populations, was lower than the national median percentage (82.9%) in 2008. A lower percentage of women received Pap smear screening in the majority of the A/PI communities compared with that in the corresponding state. The differences were mostly small in the one-on-one comparisons between individual communities and their BRFSS counterparts for the three other minority populations.

Immunization

Influenza Vaccination

The median percentage of adults aged ≥65 years who reported that they had an influenza vaccination in the previous year ranged from 57.3% in black communities to 79.5% in A/PI communities (Table 25). The median percentage was lower than the national median (70.1%) in 2009 among black and Hispanic communities, while higher among A/PI communities. With very few exceptions, black communities had a much lower rate of influenza vaccination than that in the corresponding MMSA, county, or state population. The differences were not significant in most of the one-on-one comparisons between the individual community and BRFSS counterpart for the three other minority populations.

Pneumococcal Vaccination

The median percentage of adults aged ≥65 years who reported that they had ever had a pneumococcal vaccination ranged from 58.5% in Hispanic communities to 78.7% in AI communities (Table 26). Except among AI communities, the median percentages among the other three minority communities were lower than the national median (68.5%) among the 50 states and DC in 2009. However, the vaccination rates in the majority of the communities and those in the corresponding MMSA, county, or state were not substantially different.

Discussion

Several population-based surveys have been conducted in the individual states and the nation, including the National Health Interview Survey, the National Health and Nutrition Examination Survey, and BRFSS. These surveys were designed to collect data in national or statewide probability samples to obtain national or state-level estimates. They were not designed specifically to monitor the health status of persons at the community level or to focus on minority communities. As a result, surveillance data for racial/ethnic minorities often is lacking. The previous REACH 2010 Risk Factor Survey (8) and the current REACH U.S. Risk Factor Survey provide valuable information at the community level on socioeconomic demographics, risk factors, chronic conditions, and the use of preventive services in the four minority populations surveyed.

Data from the REACH U.S. Risk Factor Survey indicate that for the majority of health and socioeconomic indicators, black, Hispanic, and AI communities do not fare as well as the general populations in their respective MMSA, county, or state, or in the United States as a whole. Socioeconomic status as measured by education level and household income was substantially lower among these communities. Minorities in these communities had less health insurance coverage, higher cost barriers to access health care, and worse self-rated general health. Variations were identified in the prevalence of risk factors and chronic condition burden among the four minority populations. Obesity, smoking, high blood pressure, cardiovascular diseases, and diabetes were the major health risk and chronic conditions in black and AI communities. Lack of physical activity was prominent in black communities, especially among women. Lack of knowledge of heart attack and stroke symptoms was prominent in black, Hispanic, and A/PI communities. The cholesterol screening rate and the percentage of persons who reported having high blood pressure and were taking antihypertensive medication was lowest in Hispanic communities. Underuse of selected preventive services was apparent, including underuse of mammography screening in AI communities, of Pap smear testing in A/PI communities, and of influenza vaccination in black and Hispanic communities.

The substantial variations identified among the four racial/ethnic minority populations in different risk factors and health conditions indicate that different priorities are needed to eliminate health disparities. In the REACH 2010 and REACH U.S. project, community-based coalitions were formed to address community-specific health issues. These coalitions were driven primarily by residents of the community at every stage of the program, including setting health priorities, planning, implementation and evaluation. A unique feature of REACH was that the project did not use a standardized intervention protocol but was sufficiently flexible to allow community choices on the basis of priorities, available resources, and local realities (6).

These survey data show an aggregation of socioeconomic demographic factors (e.g., education and income), risk factors (e.g., obesity, smoking, and physical inactivity) and chronic diseases (e.g., high blood pressure, cardiovascular diseases, and diabetes). This indicates that multifaceted and multisectoral strategies are needed to make effective changes. REACH U.S. grantees have identified numerous societal, policy, environmental, cultural, and individual-level factors that must be changed to eliminate racial and ethnic disparities. Thus, each grantee develops appropriate programs that address the complex root causes of racial and ethnic health disparities.

The data demonstrate a substantial heterogeneity across communities within the same broad racial/ethnic population. For example, obesity prevalence among black men was 13.0% in a community in Los Angeles and Orange Counties, California, whereas it was 41.6% in a community in southern West Virginia. The smoking prevalence among Hispanic men was 10.3% in a community in Santa Clara Valley, California, whereas it was 34.4% in a community in Southeast Chicago. Although smoking was rare among women in Asian communities, it was prevalent among Native Hawaii/Pacific Islander women in Waianae, Hawaii. Differences in percentages of persons receiving preventive services were also substantial across communities. The rate of mammography screening during the previous 2 years among women aged ≥40 years was only 66.3% in the Hispanic population surveyed in Seattle and King County, Washington compared with 93.5% among the Hispanic population surveyed in Lawrence, Massachusetts. Similarly, the rate of pneumococcal vaccination among blacks aged ≥65 years was 42.7% in Humboldt Park and West Town, Chicago, whereas the rate was 73.5% in the black population surveyed in southern West Virginia. These wide variations in prevalence of health-related behaviors, chronic conditions, and use of preventive services across communities likely are related to demographic, cultural, local laws, policies, and environmental influences among residents of these communities and the majority of these influences are more or less modifiable, depending on local circumstances. The salient variations and intrapopulation differences also indicate that opportunities and the possibility for change exist.

The baseline data of the previous REACH 2010 Risk Factor Survey were collected in 21 communities in 2001 (7). Among them, 11 communities also participated in the current REACH U.S. Risk Factor Survey with complete geographic match in seven and partial matches in four communities. Although a list-assisted random-digit--dialing design instead of address-based sampling, was used in the previous REACH 2010 survey, the survey questionnaires were very similar. Since the previous survey was conducted 8 years ago (8), many improvements have occurred in the health and health care indicators in the four minority populations. For example, among men, the median prevalence of smoking was 28.8% in Hispanic communities 8 years ago, whereas it was 17.6% in the current survey. Likewise, the median prevalence of smoking decreased from 30.5% to 13.8% in Asian men, consistent with the 5-year smoking trend in REACH 2010 Asian communities reported previously (15). These improvements are greater than the national decreasing trend in smoking (25.5% in 2001 and 19.6% in 2009). Although an increase occurred in the percentage of adults who met physical activity recommendations in the nation for men (from 49.6% to 52.4%) and women (from 42.9% to 47.8%), the increase was larger among REACH communities during the same time period. Among men, the median percentage increased from 36.2% to 43.8% in blacks, from 35.1% to 45.8% in Hispanics, from 24.1% to 39.8% in A/PIs, and from 42.9% to 47.9% in AIs. Among women, the corresponding median percentage increased from 25.6% to 34.6% in blacks, 33.7% to 44.7% in Hispanics, 17.3% to 40.5% in A/PIs, and 35.9% to 44.9% in AIs, respectively. Similar trends also were observed in fruit and vegetable intake (data not reported). Improvements in the measures of health care and preventive services were also evident in these communities. For example, fewer than half of Hispanics who reported having hypertension were taking medication for high blood pressure in the 2001 REACH 2010 survey; although still lower than the national level, this percentage increased to more than two thirds in the 2009 REACH U.S. survey. The vaccination rate for pneumonia increased from 50.5% to 60.5% in black communities, from 46.0% to 58.5% in Hispanic communities, from 37.5% to 59.7% in A/PI communities, and from 67.3% to 78.7% in AI communities. Racial and ethnic health disparities have long been recognized and actions have been called for to close the gap (16,17). Many improvements in health indicators have occurred during the past decade in minority communities (6). However, vital statistics data indicate that black-white mortality disparities widened during 1990--2005 in certain diseases (e.g., heart disease, breast cancer, and diabetes) at the national level, especially in such urban areas as Chicago (18).

Limitations

The findings in this report are subject to at least eight limitations. First, because minority populations are not homogeneous, substantial ethnic, cultural, and social diversity exists within any racial/ethnic minority population. For example, Haitian Americans were grouped under blacks. Hispanics comprised multiple diverse subpopulations (e.g., Mexicans, Puerto Ricans, Cubans, and Dominicans). A/PIs consisted of Native Hawaiian/Pacific Islanders, and various Asian populations (e.g., Chinese, Koreans, Vietnamese, Cambodians, and Filipinos). Second, because AIs were sampled in only four different communities, the data reported might not represent AIs from other communities in the United States. Third, although the survey sampled an average of 900 adult residents in each community, sample sizes in certain communities (e.g., Eastern Band of Cherokee Indians, North Carolina, and Intertribal Council of Michigan) were relatively small. Fourth, certain communities included multiple racial/ethnic populations, reducing the sample size for each minority population. Sample size was reduced further when prevalence estimates were limited to women, specific ages (e.g., age ≥65 years), or those with certain chronic conditions (e.g., diabetes). As a result of limited sample size, estimated confidence intervals were wide for certain health indicators in certain communities. Fifth, the response rates of the telephone and questionnaire mailing components of this survey were somewhat low. The survey aimed to collect representative samples from each racial/ethnic minority population surveyed. Respondents were weighted according to age and gender distribution of the minority population in the community and accounted for varying inclusion probabilities among them. Approximately 42% of the mail questionnaire respondents reported having only cellular phones or no phone service. When compared with national data for each racial/ethnic population, participants in the REACH Risk Factor Survey tended to have lower income and educational attainment. This most likely reflects the lower socioeconomic status of the REACH communities along with the higher cellular phone coverage that was obtained in the REACH survey compared with the BRFSS. Sixth, because estimates were based on "having been told" and self-reported data and subject to recall errors/bias or social desirability effects, the prevalence of certain chronic conditions and use of preventive services might be under- or overestimated. Seventh, the questions used in this study related to physical activity were appropriate to evaluate the percentage of persons who met an earlier physical activity recommendation (12) but not the most current 2008 guideline (19). Finally, the prevalence estimates for small area (i.e., MMSA and county) might not be the same as those reported here if other statistical methods (e.g., Bayesian methodology) are used.

Despite these limitations, the REACH U.S. survey has multiple strengths. Unlike previous national or state-based surveys, it is the largest community-based survey that focuses on multiple minority populations in the United States. The use of an address-based sampling design, which included multiple data collection modes (telephone, mailing, and in-person interviews), enabled the survey to reach those without telephones or with only cellular phones. The survey was conducted from a single center using a uniform methodology across all communities. The questions used in the survey were identical to those used in BRFSS, thus allowing data from the two surveys to be compared.

Conclusion

Despite measurable improvements in the overall health of the nation including minority populations, the REACH U.S. Risk Factor Survey demonstrates that health disparities remain widespread among members of racial and ethnic minority populations. The data from this survey provide important information for assessing, prioritizing, and planning intervention efforts in each community. These results underscore the need for community-based approaches that include policy, systems, environmental, and individual-level changes. They also underscore the need to tailor prevention strategies to the needs of specific communities to eliminate health disparities. Continuing data collection is necessary for evaluating the effectiveness of the interventions and for enhancing existing programs and disseminating the lessons learned.

References

1. CDC. CDC Health disparities and inequalities report---United States, 2011. MMWR 2011;60(Suppl).
2. CDC. Health, United States, 2006: With chartbook on trends in the health of Americans. 2006. Hyattsville, MD: US Department of Health and Human Services, CDC, National Center for Health Statistics; 2006. Available at http://www.cdc.gov/nchs/data/hus/hus06.pdf. Accessed March 30, 2011.
3. US Census Bureau, US interim projections by age, sex, race and Hispanic origin: 2000--2050. Available at http://www.census.gov/population/www/projections/usinterimproj. Accessed March 30, 2011.
4. US Department of Health and Human Services. Healthy people 2010. Washington, DC: US Department of Health and Human Services; 2000.
5. Giles WH, Tucker P, Brown L, et al. Racial and ethnic approaches to community health (REACH 2010): an overview. Ethn Dis 2004:14(Suppl):S5--8
6. CDC. REACHing across the divide: finding solutions to health disparities. Atlanta, GA: US Department of Health and Human Services, CDC; 2007. Available at http://www.cdc.gov/reach/pdf/health_disparities_101607.pdf. Accessed March 30, 2011.
7. CDC. At glance: racial and ethnic approaches to community health (REACH U.S.). Atlanta, GA: US Department of Health and Human Services, CDC; 2008. Available at http://www.cdc.gov/reach. Accessed March 30, 2011.
8. Liao Y, Tucker P, Okoro CA, et al. REACH 2010 surveillance of health status in minority communities---United States, 2001--2002. MMWR 2004;53(No. SS-6).
9. Blumberg SJ, Luke JV. Wireless substitution: early release of estimates from the National Health Interview Survey, July--December 2009. Available at http://www.cdc.gov/nchs/data/nhis/earlyrelease/wireless201005.pdf Accessed March 30, 2011.
10. CDC. Public health surveillance for behavioral risk factors in a changing environment. MMWR 2003;52(No. RR-9)
11. Rosner B. Fundamentals of biostatistics. 5th ed. Pacific Grove, California: Duxbury Thomson Learning; 2000.
12. Pate RR, Pratt M, Blair SN, et al. Physical activity and public health: a recommendation from the Centers for Disease Control and Prevention and the American College of Sports Medicine. JAMA 1995;273:402--7.
13. CDC. About the National Fruit and Vegetable Program. Atlanta, GA: US Department of Health and Human Services, CDC; 2011. Available at http://www.fruitsandveggiesmatter.gov/health_professionals/about.html. Accessed March 30, 2011.
14. Greenlund KJ, Keenan NL, Giles WH, et al. Public recognition of major signs and symptoms of heart attack: seventeen states and the US Virgin Islands, 2001. Am Heart J 2004;147:1010--6.
15. Liao Y, Tsoh JY, Chen R, et al. Decreases in smoking prevalence in Asian communities served by the Racial and Ethnic Approaches to Community Health (REACH) Project. Am J Public Health 2010;100:853--60.
16. US Department of Health and Human Service. Report of the Secretary's Task Force on Black and Minority Health. Bethesda, MD: US Department of Health and Human Services; 1985. Available at http://minorityhealth.hhs.gov/assets/pdf/checked/1/ANDERSON.pdf. Accessed March 30, 2011.
17. US Public Law 106-525. Minority Health and Health Disparities Research and Education Act of 2000. 106th Congress 2nd session, November 22, 2000.
18. Orsi JM, Margellos-Anast H, Whitman S. Black-white health disparities in the United States and Chicago: a 15-year progress analysis. Am J Public health 2010;100:349--56.
19. US Department of Health and Human Services. 2008 physical activity guidelines for Americans. Washington, DC: US Department of Health and Human Services; 2008. Available at http://www.health.gov/paguidelines. Accessed March 30, 2011.