8: No. 1, January 2011
Yi Huang, MD; Peggy A. Hannon, PhD, MPH; Barbara Williams, PhD; Jeffrey R. Harris, MD, MPH, MBA
Suggested citation for this article: Huang Y, Hannon PA, Williams B, Harris JR. Workers’ health risk behaviors by state, demographic characteristics, and health insurance status. Prev Chronic Dis 2011;8(1).
http://www.cdc.gov/pcd/issues/2011/jan/10_0017.htm:A12. Accessed [date].
Employers often lack data about their workers’ health risk behaviors. We
analyzed state-level prevalence data among workers for 4 common health risk behaviors: obesity, physical inactivity, smoking, and missed influenza vaccination
(among workers older than 50 years).
We analyzed 2007 and 2008 Behavioral Risk Factor Surveillance System data, restricting the sample to employed respondents aged 18
to 64 years. We stratified health risk behavior prevalence by annual household income, educational attainment, health insurance status, and race/ethnicity.
For all 4 health risk behaviors, we found significant differences across states and significant
disparities related to social determinants of health — income, education, and
race/ethnicity. Among uninsured workers, prevalence of smoking was high and
influenza vaccinations were lacking.
In this national survey study, we found that workers’ health risk behaviors vary substantially by state and by workers’ socioeconomic status, insurance status, and race/ethnicity. Employers and workplace health promotion practitioners can use the prevalence tables presented in this
article to inform their workplace health promotion programs.
Back to top
Health risk behaviors are common among workers, are strongly related to chronic illness and death, increase health care costs, and reduce productivity (1). One key to a successful workplace health promotion program is to measure workers’ baseline health needs and use the data to inform the program (2,3). However, most employers do not have access to data about their workers’ health behaviors. Many midsized and small employers
the resources to conduct health risk appraisals (HRAs). In addition, employer-run
HRAs often have low response rates and overrepresent healthy workers (4).
Readily available data about risk behaviors could help employers plan and evaluate their workplace health promotion programs. Obesity, physical inactivity, and tobacco use are 3 of the most common lifestyle health risk behaviors in the United States (5,6) and cause approximately one-third of all deaths (7). Influenza vaccination is also of interest to employers because influenza leads to lost productivity and can trigger severe pulmonary and cardiovascular diseases. Vaccination reduces
the incidence of influenza and can save employers money in a short time frame (1 year or less) (8).
The objective of this study was to provide employers and other workplace health promotion practitioners with state-specific data for these 4 health risk behaviors (obesity, physical inactivity, smoking, and no influenza vaccination
[among workers older than 50 years]) among workers. We stratified the behaviors by insurance status and social determinants of health: annual household income, educational attainment, and race/ethnicity. To meet this objective, we show the prevalence of each health risk behavior by state and
workers’ characteristics, using data from the 2007 and 2008 Behavioral Risk Factor Surveillance System (BRFSS), the most recent data available.
Back to top
We conducted a cross-sectional study by using BRFSS data collected in 2007 and 2008. With assistance from the Centers for Disease Control and Prevention (CDC), state health departments conduct BRFSS surveys among US resident civilian, noninstitutionalized adults aged 18 years or older in all 50 states, the District of Columbia, and US territories (9).
Using a multistage cluster design, BRFSS selects state-specific probability samples of households to produce a nationally representative sample (5). After calling a selected home
telephone number, the interviewer randomly chooses 1 adult in that household to complete the
telephone interview. BRFSS data are weighted by race/ethnicity, age, and sex distributions found in each state, along with the respondent’s probability of selection.
The median cooperation rate, or the proportion of all respondents interviewed from all eligible units in which a respondent was selected and contacted, was 72.1% in 2007 and 75.0% in 2008 (10,11). Our study population included employed adults aged 18 to 64 years in 50 states and the District of Columbia. We considered adults employed if they were employed for wages or self-employed. We excluded adults older than 64 years because Medicare is available for most of this group.
The BRFSS questionnaire has 3 parts: core questions, optional modules, and state-added questions. All states must ask core questions every year or every other year. States may also choose optional modules or add their own questions to meet their specific data needs. Both English- and Spanish-language versions of the survey are provided to each state.
In this article, all data are from the core questions used in every state. The health risk behaviors are lifestyle behaviors (obesity, physical inactivity, and smoking) and no influenza vaccination in the past year. Obesity is defined as having a body mass index
of at least 30 kg/m2 (12). Physical inactivity is defined as not meeting the CDC physical activity guideline of at least 5 days
per week for 30 minutes per day of moderate-intensity activity or at least 3 days
per week for 20 minutes
a day of vigorous-intensity activity (13,14). Tobacco use is defined as ever having smoked at least 100 cigarettes and currently smoking every day or some days. Workers aged 50
to 64 years who reported no influenza vaccination in the past 12 months (either by injection or nasal spray) were defined as not vaccinated. We restricted the influenza vaccination analysis to workers older than age 50 because CDC’s Advisory Committee on Immunization Practices recommends influenza vaccination for
We analyzed workers’ socioeconomic status (SES), race/ethnicity, health insurance status, and health risk behaviors. The SES measures are annual household income and educational attainment as reported in the BRFSS data. We used 2007 BRFSS data for the physical inactivity measure because these questions were not included in the 2008 survey. We used 2008 data for the rest of the measures.
We calculated national and state rates for workers stratified by 1) annual household income (<$35,000, $35,000-$74,999, >$75,000),
2) educational attainment (high school
graduate or less, some college, college graduate), 3) health insurance (any, none), and
4) race/ethnicity (African American, American Indian/Alaska Native, Asian/Hawaiian/Pacific Islander, Hispanic, and white). We identified the national prevalence of each health risk behavior
among workers, the range across states,
and the range across states for characteristics associated with the highest risk behavior prevalence nationally.
Our analysis took into account the survey design and weighted sampling probabilities of the data source and was performed by using Stata version 10.0 (StataCorp LP, College Station, Texas). All the statistical tests were 2-sided
and significance was set at P < .05. We calculated 95% confidence intervals
(CIs) for all prevalence rates (versions of the tables with CIs are available from the corresponding author on request). Because of the very small numbers of respondents in some categories,
we restricted the prevalence estimates to the categories in which there were 50 or more respondents.
Back to top
There were 430,912 respondents in the 2007 BRFSS, and 414,509 respondents in the 2008 BRFSS. When we restricted our data sample to employed respondents aged 18
to 64 years, 48.3% of the 2007 sample (physical inactivity) and 47.5% of the 2008 sample (obesity, smoking, and influenza vaccination) remained. For each of the analyses described below, we excluded respondents who were missing data for the health
risk behavior under study; therefore, the number of subjects varies slightly across the analyses.
We further excluded respondents who were missing data for SES, insurance status, or race/ethnicity from all analyses stratified by these characteristics (8.3% in 2007 and 8.0% in 2008 were missing 1 or more of these variables). Thus, of the respondents who met our employment and age criteria, we were able to include more than 85% in our analyses (range: 87.0% for physical activity to 91.8% for smoking).
In 2008, 27.0% of employed adults in the United States were obese (Table 1); obesity rates were lowest in Colorado (19.5%) and
were highest in West Virginia (34.6%). Nationally, the highest obesity rates were reported by those with annual household incomes
less than $35,000 (30.2%), those who did not graduate from college (30.5%), and African Americans (37.3%). Obesity rates among workers with these characteristics varied significantly across states, from 21.8% (95% CI, 18.3%-25.2%) in
Colorado to 39.2% (95% CI, 35.0%-43.4%) in Mississippi for low-income workers; from 23.5% (95% CI, 21.0%-26.1%) in Massachusetts to 39.1% (95% CI, 33.1%-45.1%) in Tennessee among workers with a high school
education or less; and from 17.9% (95% CI, 6.5%-29.4%) in Nevada to 49.9% (95% CI, 33.3%-66.4%) in Nebraska for African American workers.
In 2007, 49.2% of employed adults did not meet physical activity recommendations (Table 2); physical inactivity rates were lowest in Alaska (37.2%) and highest in Louisiana (58.4%). Nationally, the highest physical inactivity rates were reported by workers with household incomes less than $35,000 (54.3%), high school education or less (52.5%), and Asians/Hawaiians/Pacific Islanders (63.1%). Physical inactivity rates for workers with these characteristics varied significantly across states, from 42.5%
(95% CI, 37.8%-47.2%) in Montana to 68.7% (95% CI, 63.0%-74.3%) in Tennessee for low-income workers; from 36.1% (95% CI, 29.4%-42.8%) in Alaska to 61.0% (95% CI, 57.0%-65.1%) in Louisiana for workers with a high school education or less; and from 40.1% (95% CI, 22.1%-58.1%) in Pennsylvania to 70.2% (95% CI, 63.3%-77.1%) in California for Asian/Hawaiian/Pacific Islander workers.
In 2008, 19.2% of employed adults reported that they currently smoke cigarettes (Table 3); smoking rates were lowest in Utah (9.8%) and highest in Indiana (27.6%). Nationally, the highest smoking rates were reported by workers with household incomes less than $35,000 (28.9%), high school education or less (29.3%), no health insurance (32.5%), and American Indians/Alaska Natives (27.8%). Among workers with these characteristics, smoking rates varied significantly across states, from 15.3% (95%
CI, 11.1%-19.5%) in Utah to 45.6% (95% CI, 38.4%-52.8%) in Indiana for low-income workers; from 17.6% (95% CI, 14.2%-21.0%) in Utah to 41.1% (95% CI, 35.7%-46.5%) in Indiana for workers with high school education or less; from 13.8% (95% CI, 9.1%-18.5%) in Utah to 54.9% (95% CI, 45.9%-63.9%) in Indiana for uninsured workers; and from 10.9% (95% CI, 2.3%-19.5%) in Arizona to 53.1% (95% CI, 32.6%-73.5%) in North Dakota for American Indian/Alaska Native workers.
No influenza vaccination
In 2008, 59.3% of workers aged 50 to 64 years reported no influenza vaccination (Table 4); the lowest
rate was in South Dakota (47.1%) and the highest was in Nevada (71.4%). Nationally, workers most likely to report no influenza vaccination had household income less than $35,000 (68.6%), high school
education or less (66.3%), no health insurance (77.1%), and were Hispanic (67.1%). Among workers with these characteristics, rates of no influenza vaccination varied significantly across states, from 49.0% in
Virginia (95% CI, 36.3%-61.7%) to
83.3% (95% CI, 77.1%-89.4%) in Nevada for low-income workers; from 51.6% (95% CI, 46.6%-56.6%) in South Dakota to 82.0% (95% CI, 75.5%-88.5%) in Nevada for workers with a high school education or less; from 59.5% (95% CI, 47.6%-71.4%) in Iowa to 90.2% (95% CI, 83.3%-97.1%) in Indiana for uninsured workers; and from 50.9% (95% CI, 34.7%-67.0%) in Hawaii to 84.3% (95% CI, 75.0%-93.6%) in Nevada for Hispanic workers.
Back to top
The most effective workplace health promotion efforts are tailored to the risk behaviors and needs of the workers (2,3). However, for many employers, data describing their workers are unavailable or unrepresentative of their workforce (4,16). To address this need, we used BRFSS data, a very large, recent data set of employed adults in the United States, and calculated prevalence for 4 common health risk behaviors stratified by state and by the worker characteristics that employers
routinely collect to describe their workforce.
In this national sample of employed adults aged 18 to 64 years, we found significant disparities related to SES and race/ethnicity for all 4 health risk behaviors and significant disparities by insurance status for smoking and influenza vaccination. We also found significant variations in health risk behaviors within and across states. Our findings both replicate and extend our prior study of employed workers’ health risk behaviors, which found significant disparities by SES and
race/ethnicity among insured workers (6). The findings make state-level data for workers available for the first time, include uninsured workers, and show that disparities are worse for the uninsured for influenza vaccination and tobacco use than for obesity and physical inactivity.
Our study and prevalence tables have several limitations. First, BRFSS includes only people who have
home telephones and speak either English or Spanish. Second, all of the health risk behaviors are self-reported. These 2 limitations suggest that our results may underreport the prevalence of workers’ health
risk behaviors. Third, in many states, fewer than 50 members of
some racial/ethnic groups were included in the sample, and we were not able to present health risk behavior rates in these
cases. In other states, we were able to present health risk behavior rates for every racial/ethnic group, but some of the confidence intervals are wide because of small numbers in these groups. Fourth, our study was cross-sectional; our findings show associations between characteristics and health
risk behaviors but not causation.
An important limitation of our study is that the prevalence tables are at the state rather than the local level. As such, they cannot provide employers with as accurate a view of their workers’ health risk behaviors as they could achieve by surveying their workers. For many employers, acquiring health behavior data from their own workers is often not feasible. Finally, our findings do not address the time and financial challenges employers face in implementing workplace health
promotion programs. However, our findings can serve employers by 1) providing data on the health risks of workers in their state with similar characteristics to
those of their own workforce (comparable to the intent of county health-ranking
systems that motivate policy makers to take action to improve health risks in
their counties ) and 2) serving as a planning tool for an individual employer’s health promotion efforts.
To our knowledge, this is the first time that state-level BRFSS tables summarizing health risk behaviors of the US employed population have been made available. We found significant differences in workers’ health behaviors across states and within states, depending on their SES, insurance status, and race/ethnicity. Employers, workplace health promotion professionals, insurers, and vendors can use these tables to inform workplace health promotion planning when data for a given
employer’s workers are not available.
Back to top
Research supporting the information in this article was sponsored by the University of Washington Health Promotion Research Center, a CDC Prevention Research Center (HPRC
cooperative agreement no. U48/DP000050-03). Additional funding support came from CDC and the National Cancer Institute through the Cancer Prevention and Control Research Network, a network within the CDC Prevention Research Centers
no. 1-U48-DP-000050), and the CDC Office of Public Health Research through its
Centers of Excellence in Health Marketing and Health Communication program (grant
Back to top
Corresponding Author: Peggy A. Hannon, PhD, MPH, University of Washington, 1107 NE 45th St, Ste 200, Seattle, WA 98105. Telephone: 206-616-7859. E-mail:
Author Affiliations: Yi Huang, Barbara Williams, Jeffrey R. Harris, University of Washington,
Back to top
- Loeppke R, Taitel M, Richling D, Parry T, Kessler RC, Hymel P, et al.
Health and productivity as a business strategy. J Occup Environ Med 2007;49(7):712-21.
- Goetzel RZ, Ozminkowski RJ.
The health and cost benefits of work site health-promotion programs. Annu Rev Public Health 2008;29:303-23.
- Sparling PB. Worksite health promotion: principles, resources, and challenges. Prev Chronic Dis 2010;7(1).
Accessed February 4, 2010.
- Wang PS, Beck AL, McKenas DK, Meneades LM, Pronk NP, Saylor JS, et al.
Effects of efforts to increase response rates on a workplace chronic condition screening survey. Med Care 2002;40(9):752-60.
- BRFSS: Behavioral Risk Factor Surveillance System: Survey data overview 2008.
Centers for Disease Control and Prevention.
Accessed January 10, 2010.
- Hughes MC, Hannon PA, Harris JR, Patrick DL.
Socioeconomic disparities in health behaviors of insured workers. Am J Health Promot 2010;24(5):315-23.
- Mokdad AH, Marks JS, Stroup DF, Gerberding JL.
Actual causes of death in the United States, 2000.
[Errata appear in JAMA 2005;293(3):293-4 and in JAMA 2005;293(3):298.] JAMA 2004;10;291(10):1238-45.
- Maciosek MV, Coffield AB, Edwards NM, Flottemesch TJ, Goodman MJ, Solberg LI.
Priorities among effective clinical preventive services: results of a systematic review and analysis. Am J Prev Med 2006;31(1):52-61.
- BRFSS: Behavioral Risk Factor Surveillance System: frequently asked questions 2009. Centers for Disease Control and Prevention. http://www.cdc.gov/brfss/faqs.htm#1. Accessed September 9, 2010.
- BRFSS: Behavioral Risk Factor Surveillance System: summary data quality report 2007. Centers for Disease Control and Prevention. ftp://ftp.cdc.gov/pub/Data/Brfss/2007SummaryDataQualityReport.pdf. Accessed September 9, 2010.
- BRFSS: Behavioral Risk Factor Surveillance System: summary data quality report 2008. Centers for Disease Control and Prevention. ftp://ftp.cdc.gov/pub/Data/Brfss/2008_Summary_Data_Quality_Report.pdf. Accessed September 9, 2010.
- Defining overweight and obesity. Centers for Disease Control and
Prevention. 2009. http://www.cdc.gov/obesity/defining.html. Accessed August 25, 2009.
- Haskell WL, Lee IM, Pate RR, Powell KE, Blair SN, Franklin BA, et al.
Physical activity and public health: udpated recommendation for adults from the American College of Sports Medicine and the American Heart Association. Med Sci Sports Exerc 2007;39(8):1423-34.
- Pate RR, Pratt M, Blair SN, Haskell WL, Macera CA, Bouchard C, 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(5):402-7.
- Centers for Disease Control and Prevention.
Recommended adult immunization schedule — United States, 2009. MMWR 2009;57(53):Q1-4.
- McLellan RK, Mackenzie TA, Tilton PA, Dietrich AJ, Comi RJ, Feng YY.
Impact of workplace sociocultural attributes on participation in health assessments. J Occup Environ Med 2009;51(7):797-803.
- University of Wisconsin Population Health Institute. Wisconsin County health
rankings. 2010. http://uwphi.pophealth.wisc.edu/pha/wchr.htm. Accessed March 28, 2010.
Back to top