Surveillance for Certain Health Behaviors and Conditions Among States and Selected Local Areas — Behavioral Risk Factor Surveillance System, United States, 2013 and 2014

Problem Chronic diseases and conditions (e.g., heart diseases, stroke, arthritis, and diabetes) are the leading causes of morbidity and mortality in the United States. These conditions are costly to the U.S. economy, yet they are often preventable or controllable. Behavioral risk factors (e.g., excessive alcohol consumption, tobacco use, poor diet, frequent mental distress, and insufficient sleep) are linked to the leading causes of morbidity and mortality. Adopting positive health behaviors (e.g., staying physically active, quitting tobacco use, obtaining routine physical checkups, and checking blood pressure and cholesterol levels) can reduce morbidity and mortality from chronic diseases and conditions. Monitoring the health risk behaviors, chronic diseases and conditions, access to health care, and use of preventive health services at multilevel public health points (states, territories, and metropolitan and micropolitan statistical areas [MMSA]) can provide important information for development and evaluation of health intervention programs. Reporting Period 2013 and 2014. Description of the System The Behavioral Risk Factor Surveillance System (BRFSS) is an ongoing, state-based, random-digit–dialed telephone survey of noninstitutionalized adults aged ≥18 years residing in the United States. BRFSS collects data on health risk behaviors, chronic diseases and conditions, access to health care, and use of preventive health services and practices related to the leading causes of death and disability in the United States and participating territories. This is the first BRFSS report to include age-adjusted prevalence estimates. For 2013 and 2014, these age-adjusted prevalence estimates are presented for all 50 states, the District of Columbia, the Commonwealth of Puerto Rico, Guam, and selected MMSA. Results Age-adjusted prevalence estimates of health status indicators, health care access and preventive practices, health risk behaviors, chronic diseases and conditions, and cardiovascular conditions vary by state, territory, and MMSA. Each set of proportions presented refers to the range of age-adjusted prevalence estimates of selected BRFSS measures as reported by survey respondents. The following are estimates for 2013. Adults reporting frequent mental distress: 7.7%–15.2% in states and territories and 6.3%–19.4% in MMSA. Adults with inadequate sleep: 27.6%–49.2% in states and territories and 26.5%–44.4% in MMSA. Adults aged 18–64 years having health care coverage: 66.9%–92.4% in states and territories and 60.5%–97.6% in MMSA. Adults identifying as current cigarette smokers: 10.1%–28.8% in states and territories and 6.1%–33.6% in MMSA. Adults reporting binge drinking during the past month: 10.5%–25.2% in states and territories and 7.2%–25.3% in MMSA. Adults with obesity: 21.0%–35.2% in states and territories and 12.1%–37.1% in MMSA. Adults aged ≥45 years with some form of arthritis: 30.6%–51.0% in states and territories and 27.6%–52.4% in MMSA. Adults aged ≥45 years who have had coronary heart disease: 7.4%–17.5% in states and territories and 6.2%–20.9% in MMSA. Adults aged ≥45 years who have had a stroke: 3.1%–7.5% in states and territories and 2.3%–9.4% in MMSA. Adults with high blood pressure: 25.2%–40.1% in states and territories and 22.2%–42.2% in MMSA. Adults with high blood cholesterol: 28.8%–38.4% in states and territories and 26.3%–39.6% in MMSA. The following are estimates for 2014. Adults reporting frequent physical distress: 7.8%–16.0% in states and territories and 6.2%–18.5% in MMSA. Women aged 21–65 years who had a Papanicolaou test during the past 3 years: 67.7%–87.8% in states and territories and 68.0%–94.3% in MMSA. Adults aged 50–75 years who received colorectal cancer screening on the basis of the 2008 U.S. Preventive Services Task Force recommendation: 42.8%–76.7% in states and territories and 49.1%–79.6% in MMSA. Adults with inadequate sleep: 28.4%–48.6% in states and territories and 25.4%–45.3% in MMSA. Adults reporting binge drinking during the past month: 10.7%–25.1% in states and territories and 6.7%–26.3% in MMSA. Adults aged ≥45 years who have had coronary heart disease: 8.0%–17.1% in states and territories and 7.6%–19.2% in MMSA. Adults aged ≥45 years with some form of arthritis: 31.2%–54.7% in states and territories and 28.4%–54.7% in MMSA. Adults with obesity: 21.0%–35.9% in states and territories and 19.7%–42.5% in MMSA. Interpretation Prevalence of certain chronic diseases and conditions, health risk behaviors, and use of preventive health services varies among states, territories, and MMSA. The findings of this report highlight the need for continued monitoring of health status, health care access, health behaviors, and chronic diseases and conditions at state and local levels. Public Health Action State and local health departments and agencies can continue to use BRFSS data to identify populations at risk for certain unhealthy behaviors and chronic diseases and conditions. Data also can be used to design, monitor, and evaluate public health programs at state and local levels.


Introduction
Chronic diseases and conditions (e.g., cardiovascular disease, cancer, chronic lower respiratory diseases, and diabetes) are among the top 10 leading causes of death in the United States (1). Practicing healthy behaviors (e.g., quitting smoking, being more physically active, limiting alcohol intake, eating a nutritious diet, and maintaining a healthy weight) and using preventive health services (e.g., regular checks for high blood pressure and high blood cholesterol, screening for cancer on recommended schedules, and obtaining regular physical checkups) can reduce morbidity and premature mortality from chronic diseases and conditions (2).
BRFSS is an ongoing, state-based, random-digit-dialed cellular and landline telephone survey of noninstitutionalized adults aged ≥18 years in each U.S. state and participating territory. Since 1984, CDC has assisted state and territorial health departments in conducting the BRFSS survey each year. The survey is one of the main data sources that public health officials and practitioners use to track chronic diseases and conditions, health risk behaviors, use of preventive health services, and emerging health problems at state and local levels. The data are frequently used to set health goals as well as to monitor progress and success of public health programs and policy implementation at national, state, and local levels. BRFSS data collection is conducted by state health departments with assistance from CDC. The estimates in this report are calculated from BRFSS data sets, which are aggregates of the combined landline and cellular telephone data submitted during 2013 and 2014. Beginning in 2002, BRFSS data have been used to generate prevalence estimates from metropolitan and micropolitan statistical areas (MMSA) that meet the system's inclusion criteria. This report includes BRFSS findings related to selected chronic diseases and conditions, health risk behaviors, health care access, and use of preventive health services.

Methods
BRFSS is conducted in all 50 states, the District of Columbia, the Commonwealth of Puerto Rico, and Guam. BRFSS uses a multistage sampling design and random-digitdialing methods to select a representative sample from the noninstitutionalized adult population aged ≥18 years in each state and participating territory (3,4). Details on methodology, random sampling procedures, design (5), and reliability and validity of measures (6) used in BRFSS have been described in previous publications. Estimates are from 53 states and territories for both years, 145 MMSA for 2013, and 132 MMSA for 2014. A list of MMSA for each year is available on the BRFSS SMART website (https://www.cdc.gov/brfss/ smart/smart_data.htm). The 2013 and 2014 questionnaires and all related supporting documents are available at the BRFSS website (https://www.cdc.gov/brfss/). MMSA are defined by the Office of Management and Budget; respondents are assigned MMSA according to their county Federal Information Processing Standards code. MMSA were included in the data set if they met the selection criterion of ≥500 participants. Data are submitted monthly to CDC by the states or their designees. Data cleaning and weighting are conducted by CDC. Complete documentation of the BRFSS methodology is available at https://www.cdc.gov/brfss.

Questionnaire
The BRFSS questionnaire is designed to collect uniform, state-specific, self-reported data on a range of health behaviors and conditions (3,4). All questions undergo cognitive and field testing. The standard questionnaire consists of three parts: 1) core questions, 2) optional BRFSS modules, and 3) stateadded questions. The core consists of a set of demographic and standard health-related questions used by all participating states

Sampling
In 2013, BRFSS used a partially overlapping sample that, in addition to other eligibility requirements, screened out cellular telephone respondents who received more than 10% of all incoming calls on a landline telephone. In 2014, BRFSS adopted the use of a fully overlapping sample of landline and cellular telephone respondents aged ≥18 years. No minor children are included in the BRFSS sample. States designed samples using substate regions (e.g., public health districts or other jurisdictions) to ensure geographic representation within the sample. CDC assisted states with sample design and set minimum sample sizes for substate regions, split-sample versions of the questionnaire, and oversampling of populations that are hard to reach.

Data Weighting
BRFSS data were used to create direct estimates for each geographic area (i.e., state or MMSA). Data were weighted using a raking method. Raking (iterative proportional fitting) was applied using each demographic factor individually in an iterative process until demographic estimates matched control totals based on U.S. Census estimates for that year. Raking has improved the precision with which the BRFSS sample reflects the sociodemographic profile at the state level. Details of the BRFSS raking method are provided in the BRFSS weighting documents for 2013 and 2014 (7,8). The 2014 sampling overlap also prompted an adjustment to the BRFSS design weights. To account for overlap of the two samples, a composite factor was multiplied by the design weight for each dual user to create an adjustment that addressed and corrected for the respondent's probability of being selected in both frames. More information about the composite factor calculation is available at https://www.cdc.gov/brfss/annual_data/2014/pdf/ compare_2014.pdf.

Statistical Analysis
The analysis was conducted using statistical software, SAS-Callable SUDAAN release 11.0 (Research Triangle Institute, Cary, North Carolina), to account for the complex sampling design and calculate age-adjusted prevalence estimates, standard errors, and 95% confidence intervals. Sample sizes are unweighted in this report. Data with sample sizes <50 or having a relative standard error >30% were deemed unstable and less reliable and were suppressed in the tables, as noted by N/A (not available). Responses coded as do not know or refused were excluded from the analysis. Several chronic diseases and conditions (i.e., diabetes, arthritis, coronary heart disease, and stroke) were limited to participants aged ≥45 years (9). This is the first BRFSS report to include age-adjusted prevalence estimates. Age adjustment is a standard analytical technique used to compare estimates between populations with different age distributions (e.g., states) and over time. In this report, the estimates were age adjusted so that data could be compared across states, MMSA, and time, each having different age distributions. BRFSS age-adjusted estimates were standardized to the 2000 U.S. population using distribution No. 8, consistent with the current National Center for Health Statistics recommendations and practice (10).
Crude prevalence estimates for each individual state and MMSA are provided at https://www.cdc.gov/brfss/ brfssprevalence/index.html. Results should not be compared with those in previous BRFSS reports, which provided nonageadjusted estimates by state, territory, and MMSA. The ageadjusted prevalence estimates were calculated from results of BRFSS and might differ from those derived by other methods.

Results
In 2013, a total of 491,773 adults completed BRFSS interviews on landline and cellular telephones. Results were from 53 states and territories and 145 MMSA with sufficient sample sizes. A total of 360,079 respondents completed the interview by landline telephone (range: 1,461 in Guam to 27,763 in Florida;median: 5,668). A total of 131,694 respondents completed the interview by cellular telephone (range: 445 in Guam to 7,620 in Kansas; median: 2,291). In 2014, a total of 464,662 adults completed the interview on landline and cellular telephones. Results were from 53 states and territories and 132 MMSA with sufficient sample size. A total of 298,568 respondents completed the interview by landline telephone (range: 1,852 in Guam to 12,962 in Nebraska;median: 4,973). A total of 166,094 respondents completed the interview by cellular telephone (range: 686 in Guam to 9,952 in Nebraska; median: 2,868).
BRFSS uses the American Association of Public Opinion Research Response Rate 4 (defined as the number of complete and partial interviews divided by the number of contacted and eligible respondents) (11) as a method for calculating response. In 2013, landline response rates ranged from 28.0% in Alabama to 63.7% in Puerto Rico (median: 49.6%) and cellular response rates ranged from 19.1% in Washington to 62.6% in Alaska (median: 37.8%). The combined (landline and cellular) response rates ranged from 29.0% in Alabama to 60.3% in Puerto Rico (median: 46.4%). In 2014, landline response rates ranged from 26.7% in California to 61.6% in Kentucky (median: 48.7%) and cellular response rates ranged from 22.2% in California to 60.0% in Alaska (median: 40.5%). Overall, the combined response rates ranged from 25.1% in California to 60.1% in South Dakota (median: 47.0%). BRFSS Summary Data Quality Reports for 2013 (12) and 2014 (13) have detailed information on response, cooperation, and refusal rates.
Increasing use of cellular telephones (14) prompted BRFSS to move to an overlapping sample in 2014. This was a change from the 2013 screening process, which restricted the eligibility of cellular telephone respondents who also used landline telephones. Effects of the 2014 sample change include larger proportions of completed interviews among persons aged 18-44 years, men, and Hispanics. Moving to an overlapping sample increased the proportion of cellular telephone respondents eligible to participate in the survey (15).

Health Status Indicators Health Status
All respondents were asked if their general health was excellent, very good, good, fair, or poor. Respondents were then divided into two groups: those reporting their health was excellent, very good, or good and those reporting their health was fair or poor. In 2013, age-adjusted prevalence estimates for good or better health ranged from 66.4% in Puerto Rico to 88.7% in Vermont (median: 84.0%) ( Table 1). Among selected MMSA, estimated age-adjusted prevalence ranged from 66.7% in San Juan- Puerto Rico,to 91.2% in Sioux Falls,South Dakota (median: 83.8%) ( Table 2). In 2014, age-adjusted prevalence estimates of persons with good or better health ranged from 66.9% in Puerto Rico to 89.1% in Vermont (median: 84.1%) ( Table 3). Among selected MMSA, age-adjusted prevalence estimates ranged from 63.3% in Ponce,Puerto Rico,to 92.2% in Logan, Utah-Idaho (median: 83.7%) ( Table 4).

Frequent Mental Distress
All respondents were asked to determine how many days during the past 30 days their mental health status (e.g., stress, depression, and problems with emotions) was not good. The respondents were divided into two groups: those who reported frequent mental distress (≥14 mentally unhealthy days during the past 30 days) and those who reported no frequent mental distress (<14 mentally unhealthy days during the past 30 days). In 2013, age-adjusted prevalence estimates of frequent mental distress ranged from 7.7% in North Dakota to 15.2% in West Virginia (median: 11.3%) ( Table 5). Among selected MMSA, estimated age-adjusted prevalence ranged from 6.3% in Grand Forks, North and Minot,North Dakota,to 19.4% in Akron, Ohio (median: 10.9%) ( Table 6).

Frequent Physical Distress
Frequent physical distress included respondents who reported ≥14 days of poor physical health (e.g., physical illness or injury) during the past 30 days. In 2014, age-adjusted prevalence estimates of frequent physical distress ranged from 7.8% in North Dakota to 16.0% in Kentucky (median: 10.9%) ( Table 7). Among selected MMSA, age-adjusted prevalence estimates ranged from 6.2% in Logan, Utah-Idaho, to 18.5% in Aguadilla-Isabela, Puerto Rico (median: 11.1%) ( Table 8).

Binge Drinking
Respondents were considered to be binge drinkers if during the past 30 days a man had five or more drinks on one occasion and a woman had four or more drinks on one occasion. In 2013, age-adjusted prevalence estimates of binge drinking among both men and women ranged from 10.5% in Tennessee to 25.2% in North Dakota (median: 17.7%) (Table 35). Among selected MMSA, age-adjusted prevalence estimates ranged from 7.2% in Provo-Orem, Utah, to 25.3% in Buffalo-Cheektowaga-Niagara Falls, New York (median: 17.8%) (Table 36). In 2014, age-adjusted prevalence estimates ranged from 10.7% in West Virginia to 25.1% in North Dakota (median: 17.0%) (Table 37). Among selected MMSA, age-adjusted prevalence estimates ranged from 6.7% in Utah,to 26.3% in Fargo,.0%) (Table 38).

Chronic Diseases and Conditions Obesity
Obesity, calculated from self-reported height and weight, was defined as having a body mass index of ≥30 (weight [kg]/height [m 2 ]). In 2013, age-adjusted prevalence estimates of obesity ranged from 21.0% in Colorado to 35.2% in West Virginia (median: 28.2%) (Table 39). Among selected MMSA, age-adjusted prevalence estimates ranged from 12.1% in San Francisco-Redwood City-South San Francisco, California, to 37.1% in Huntington-Ashland, West Virginia-Kentucky-Ohio (median: 28.3%) (

Diabetes
Diabetes was defined as respondents aged ≥45 years who reported having ever been told by a doctor, nurse, or other health professional they have diabetes, excluding prediabetes or borderline diabetes and diabetes during pregnancy for women. In 2013, age-adjusted prevalence estimates of diabetes among adults aged ≥45 years ranged from 11.1% in Colorado to 27.5% in Guam (median: 16.2%) ( Table 43). Among selected MMSA, age-adjusted prevalence estimates ranged from 10.6% in Buffalo-Cheektowaga-Niagara Falls, New York, to 24.8% in Aguadilla-Isabela, Puerto Rico (median: 15.8%) ( Table 44).

Arthritis
Arthritis was defined as respondents aged ≥45 years who reported having ever been told by a doctor, nurse, or other health professional they have some form of arthritis, including rheumatoid arthritis, gout, lupus, or fibromyalgia. In 2013, age-adjusted prevalence estimates of arthritis among adults aged ≥45 years ranged from 30.6% in Hawaii to 51.0% in West Virginia (median: 39.4%) ( Table 45). Among selected MMSA, age-adjusted prevalence estimates ranged from 27.6% in Dallas-Plano-Irving, Texas, to 52.4% in Huntington-Ashland, West Virginia-Kentucky-Ohio (median: 39.4%) ( Table 46). In 2014, age-adjusted prevalence estimates ranged from 31.2% in Hawaii to 54.7% in West Virginia (median: 39.8%) ( Table 47). Among selected MMSA, age-adjusted prevalence ranged from 28.4% in College Station-Bryan, Texas, to 54.7% in Montgomery, Alabama (median: 40.1%) ( Table 48).
Health care coverage is associated with access to preventive health care, and lack of health insurance can often lead to adverse health outcomes (24). In 2010, one in four adults did not have health care coverage, and those who had a chronic illness and did not have health insurance were more likely to skip or delay medical care because of cost (25).
In the United States, cancer is a major public health problem and is the second-leading cause of death (26). In 2014, having a Pap test among women aged 21-65 years varied among states. This difference might be associated with lack of access to health care and lack of health insurance (27). Evidencebased public health approaches can improve cervical cancer screening among women in this age group. Colorectal cancer is the second leading cause of death from cancers that affect both men and women (28). Screening is key to finding precancerous polyps, and early detection makes colorectal cancer easier to treat. USPSTF recommends colorectal cancer screening for adults aged 50-75 years. The 2014 estimates suggest the need for continued population-level efforts to identify groups that are not receiving colorectal cancer screening.

Health Risk Behaviors
Staying physically active is an important part of improving health; it helps maintain a healthy weight, improves cardiorespiratory efficiency, strengthens muscles and bones, lowers stress, and can improve one's mental health and mood (29). Physical inactivity is a risk factor for chronic diseases and conditions (e.g., diabetes, heart disease, and arthritis) (30). BRFSS measured physical inactivity or lack of exercise as no leisure-time activity during the past 30 days. The varying prevalence of no leisure-time physical activity among states and MMSA indicates the need to implement strategies outlined in the CDC Guide to Strategies to Increase Physical Activity in the Community (31).
Good sleep is critical for good health and overall quality of life (32). The American Academy of Sleep Medicine and Sleep Research Society recommends that adults aged ≥18 years get at least 7 hours of sleep each night (33). Inadequate sleep (<7 hours) is associated with high blood pressure, asthma, arthritis, obesity, diabetes, coronary heart disease, stroke, depression, and other chronic diseases and conditions (34). The range of prevalence estimates for 2013 and 2014 shows greater efforts are needed to develop and implement interventions that address multiple health risk factors and conditions associated with insufficient sleep.
Tobacco use continues to be the single most preventable cause of morbidity and mortality worldwide; it is responsible for approximately 6 million deaths per year (35). Smoking causes various types of cancer, cardiovascular and pulmonary diseases, and reproductive and developmental disorders (36). Nicotine, found in tobacco products, is acutely toxic, and smoking has been linked to diseases of nearly every organ of the body (37). Implementing comprehensive tobacco control programs, which can include a combination of smoke-free laws, cigarette price increases, access to proven smoking cessation treatments and services, and direct media campaigns, can help reduce current smoking prevalence (38).
Binge drinking is the most common pattern of excessive alcohol use in the United States (39). It is a major risk factor for morbidity and mortality and other societal costs (40,41). Health-related risks extend beyond those stemming from alcohol abuse; binge drinking can lead to risky sexual activity, unintentional injuries, violence, fetal alcohol disorders, and suicide (42). In this report, estimated prevalence of binge drinking varies across the United States and might be associated with socioeconomic and demographic factors and alcoholrelated policies in states and MMSA (43).

Chronic Diseases and Conditions
Obesity is a national epidemic and a contributing factor in many health problems, including heart disease, stroke, diabetes, and certain types of cancer (44). It lowers quality of life and results in higher medical costs (45). A 2013 study on obesity reported that prevalence had increased during 1999-2002 and 2007-2010 among both men and women and substantial

Limitations
The findings of this report are subject to at least five limitations. First, because it is a household telephone survey BRFSS excludes information from persons in institutions, military installations, nursing homes, long-term care facilities, and correctional institutions. Second, the questionnaire is administered only to persons who speak English, Spanish, Mandarin, or Portuguese. Persons who do not speak these languages would not be able to participate in the survey.
Third, the BRFSS survey collects self-reported data that are subject to recall bias and social desirability effects. Fourth, because of small sample size or unreliable estimates, certain estimates could not be obtained for some MMSA. Finally, persons without a landline or cellular telephone are not able to participate.
Overall, BRFSS is a cost-effective, timely, and flexible surveillance system that provides state health departments and local communities with reliable estimates to monitor and track health status, health risk behaviors, chronic diseases and conditions, and access to preventive health care. Crude estimates obtained using BRFSS are comparable and consistent with other U.S. survey estimates (56). BRFSS questions have been shown to be valid and reliable (6).

Conclusion
Although chronic diseases and conditions are a challenge to the overall health of the U.S. population, prevalence of morbidity and mortality can be estimated and reduced by monitoring trends, promoting healthy behaviors, identifying emerging diseases, and building effective and sustainable public health community interventions. Results from this report reflect variations in health status, health care access, health behaviors, and chronic diseases and conditions at state and MMSA levels. Identifying areas with populations at risk can help public health officials address health needs and use limited resources more effectively. BRFSS results can be used to identify emerging health problems, support health-related legislative efforts, and develop and evaluate public health policies and programs at state and local levels. CDC will continue to work with states and territories to collect data, identify populations that are underserved and at risk, monitor chronic diseases and conditions and access to health care, and encourage the U.S. population to adopt healthy behaviors. 20 3-19.4 Abbreviations: CI = confidence interval; MMSA = metropolitan and micropolitan statistical areas; SE = standard error. * Age adjusted to the 2000 U.S. standard population. † Respondents were asked "How many days in the past month was your mental health (including stress, depression, and problems with emotions) not good?" § Metropolitan division.                                                                       2-20.9 Abbreviations: CI = confidence interval; MMSA = metropolitan and micropolitan statistical areas; SE = standard error. * Age adjusted to the 2000 U.S. standard population. † Including heart attack (also known as myocardial infarction) or angina. § Metropolitan division.