|
| Parkland's Strategic Planning Office Director Sue Pickens, and staff member Brad Walsh |
Most public health professionals know the Behavioral Risk Factor Surveillance System (BRFSS), the annual phone survey meant to gauge the health behaviors and health status of Americans. The BRFSS survey effort is a triumph of public health information gathering, and an invaluable source of data for needs assessments and planning efforts. The survey helps us know, state by state and nationally, the prevalence rates of a dozen chronic diseases, as well as the effectiveness of flu vaccine efforts, the health insurance coverage rates and BMI statistics in all parts of the country.But the BRFSS survey is typically designed with a state-level sampling frame in mind, meaning that analyses at the regional, MSA or county level are problematic. Even if you're lucky enough to live in an MSA large enough that your BRFSS data are on the CDC Web site, the margins of error are usually large, and the data are often summaries of multiple years. Given the nature of the BRFSS question sets, many questions are not asked every year, making combining data from multiple years even more difficult.
But increasingly, BRFSS special studies are also being planned, paid for and conducted at the county and regional level. In Texas, special studies have been produced for Dallas County, Harris County (Houston) and Tarrant County (Fort Worth) in the past two years. County-level studies have recently been done in Pennsylvania and Michigan, and tribe-level studies have been done in North Dakota. In Columbus, Ohio, a coalition including the health department and a number of hospitals contracts for a county-level BRFSS survey every three years. In most of these cases the sampling frame is the county and sub-county level, although there is no reason a study couldn't cover multiple counties or cross state lines.
These studies were paid for by different groups. In Dallas County, a coalition of hospitals and community advocacy organizations raised the funds to do the special study, and provided the technical workers to handle much of the data analysis. In Tarrant County, funding came mostly from the county health department. The BRFSS office staff at the Texas Department of State Health Services, Center for Health Statistics, provided valuable advice and assistance along the way.
BRFSS isn't the only show in town. The U.S. Department of Health and Human Services' Steps to a Healthier U.S. program also has fostered several telephone survey campaigns. The New York state health department's Steps to a Healthier New York is conducting telephone surveys in four New York counties. Other Steps-funded telephone survey programs are in place for Philadelphia and the Minneapolis/St. Paul area. These surveys are similar in intent to the BRFSS, covering multiple chronic disease topics and behavioral risk factors.The robust BRFSS survey methodology, as well as the training programs and assistance available from the CDC and many state health departments, mean that there are fewer hurdles to conducting sub-state level BRFSS surveys than ever before. At the Texas Department of State Health Services the BRFSS staff list a range of options and prices, from adding a question or increasing the number of calls in a county on their annual statewide survey, to technical assistance with independent local studies. Their Web site (http://www.dshs.state.tx.us/chs/brfss/partners/partners.shtm) has more details. Your own state may offer similar assistance.
If you choose to carry out a special study, you may need to plan for data analysis, too. Your state BRFSS office or the survey company you hired can generate your results tables for you, for an additional fee. But remember that once the work is done, you may be called upon to examine the data in new ways later on down the road. If you have the staff to do it yourself, you can save costs and have more control over your output. Because these surveys usually require a sample plan with many demographic parameters, data analysis software with complex samples capability is required to generate accurate results tables. SUDAAN, SPSS and SAS are three software choices, although the complex samples procedures may require a separate module from the basic software package.
Data from the BRFSS survey are important for public health policy, hospital marketing and planning, health needs assessments, grant proposal writing, even police and emergency services planning. The types of data the BRFSS survey provides are available from few other places, and serve as an important addition to census, vital statistics and other data in informing public health policy. The numerous survey studies already being done, using both BRFSS and Steps methodologies, show that a regional, county level and sub-county level special study can add a needed dimension to public health policy work.
Brad Walsh is the strategic planning project director for the Office of Strategic Planning and Population Medicine of Parkland Health & Hospital System. Parkland is a county-owned hospital committed to serving the health needs of all Dallas County residents, regardless of ability to pay. Walsh's office conducts health needs assessment activities to help system administrators in planning and policy work, to use county funds more efficiently to carry out Parkland's charge. Parkland has taken its mission out into the community in a number of ways, building one of the first and largest outpatient clinic systems run by a county hospital, coordinating trauma services and helping build trauma capacity among hospitals across the region, assembling a fleet of vans providing complete primary care services for the homeless, supporting other non-profit clinics with supplementary services such as mobile mammography vans and specialty referrals, and partnering with school districts to build school-based clinics for both students and their families. Our innovative array of programs requires a rich basis in community health status and health needs data, and Strategic Planning and Population Medicine strives to fill that need.