Small Area Estimation for Cancer Risk Factors and Screening Behaviors by Combining NHIS and BRFSS for years 2011 to 2016

National Cancer Institute, Frederick MD

Investigators

Abstract

Collaborating with NCHS, CDC, the University of Michigan, and the University of Pennsylvania, NCI has produced model-based small area estimates for several cancer risk factors and screening behaviors?including current and ever smoking prevalence, mammography screening, pap smear screening and colorectal (CRC) screening rates for all states, counties, and health service areas. The estimates were produced using data from the National Health Interview Survey (NHIS) from NCHS and the Behavioral Risk Factor Surveillance System (BRFSS) from CDC. This prior work utilized small area estimation methods for four data periods covering 1997 to 2010. The final estimates have been released on the NCI?s Small Area Estimates for Cancer-related Measures website (http://sae.cancer.gov) and have been included in NCI?s Surveillance, Epidemiology, and End Results Program (SEER) county attributes database (http://seer.cancer.gov/seerstat/variables/countryattribs). The most recent estimates from the prior collaborative research (2008-2010) have also been released on the State Cancer Profiles website (http://statecancerprofiles.cancer.gov). These previous small area estimates have become an important data resource for cancer researchers as well as health policy makers. We would like to continue this important effort to produce small area estimates using more updated data. For survey years 2011 and forward, BRFSS changed its design by including cell phone-only households using a dual frame design and implemented an improved weighting methodology. As a result, the previously developed small area models cannot be directly applied to data from 2011 forward. Also, access to the confidential state or county identifiers in the NHIS and BRFSS, is only available through a research data center (RDC) at the National Center for Health Statistics (NCHS) (due to its policy change). For this new research initiated in FY18, we have been collaborating with NCHS and BRFSS staff to develop a new methodology consisting of combining NHIS and BRFSS, implementing changes in the BRFSS design, and producing up-to-date small area estimates for the outcomes of interest. The purpose of this agreement is to provide funding to partially support the RDC cost and staff contributions of the DRM at NCHS to this ongoing project. One RDC staff member will set up the input files at the RDCs as needed. Two DRM staff members will provide programming support.