Integrating a Multimode Design Into a National Random-Digit–Dialed Telephone Survey || Preventing Chronic Disease: November 2011: 10_0230

 

Integrating a Multimode Design Into a National Random-Digit–Dialed Telephone Survey

Shaohua Sean Hu, MD, DrPH; Carol Pierannunzi, PhD; Lina Balluz, ScD

Suggested citation for this article: Hu SS, Pierannunzi C, Balluz L. Integrating a multimode design into a national random-digit–dialed telephone survey. Prev Chronic Dis 2011;8(6):A145. http://www.cdc.gov/pcd/issues/2011/nov/10_0230.htm. Accessed [date].

PEER REVIEWED

Abstract

The Behavioral Risk Factor Surveillance System (BRFSS) was originally conducted by using a landline telephone survey mode of data collection. To meet challenges of random-digit–dial (RDD) surveys and to ensure data quality and validity, BRFSS is integrating multiple modes of data collection to enhance validity. The survey of adults who use only cellular telephones is now conducted in parallel with ongoing, monthly landline telephone BRFSS data collection, and a mail follow-up survey is being implemented to increase response rates and to assess nonresponse bias. A pilot study in which respondents’ physical measurements are taken is being conducted to assess the feasibility of collecting these data for a subsample of adults in 2 states. Physical measures would allow for the adjustment of key self-reported risk factor and health condition estimates and improve the accuracy and usefulness of BRFSS data. This article provides an overview of these new modes of data collection.

Introduction

The Behavioral Risk Factor Surveillance System (BRFSS) is a state-based random-digit–dial (RDD) telephone surveillance system that was established in 1984 by the Centers for Disease Control and Prevention (CDC) and state health departments. Information regarding health risk behaviors, clinical preventive health practices, and health care access, primarily related to chronic disease and injury, is obtained from a representative sample of adults in each state. For most states, BRFSS is the only source for this type of information. Data are collected monthly in all 50 states, the District of Columbia, Puerto Rico, the Virgin Islands, and Guam. Approximately 400,000 adult interviews are completed each year, making BRFSS the largest health telephone survey conducted in the world (1). Not only is BRFSS a unique source of risk behavior data for states, but it is also useful in measuring progress toward Healthy People 2020 objectives for the states and the nation (2).
For more than 30 years, RDD landline telephone surveys have been the workhorse of the survey research industry. During the past decade, however, participation in most RDD telephone surveys has declined because of changes in personal communication technologies, growth of call-screening technologies, and heightened privacy concerns resulting from increased number of telemarketing calls (3,4). Additionally, coverage provided by landline RDD survey samples has increasingly been questioned. RDD landline frames exclude households that do not have a telephone of any type (approximately 2% in 2009) (5). The increased use of cellular telephones has exacerbated this problem; 24.5% of households were reported to be cellular telephone–only (ie, households with no landline telephone) during the second half of 2009 (5-9).
As an RDD landline telephone survey, BRFSS has several specific challenges. First, households with only cellular telephone coverage or that lack landline telephones were not included; therefore, BRFSS may have been excluding people of low socioeconomic status. Second, the survey response rates have been declining over the past several years, making it increasingly difficult to collect survey data by using RDD landline telephone methods. Third, data are self-reported and are subject to recall bias.
RDD surveys are typically conducted using only the telephone survey mode of data collection. However, to meet challenges of increasing nonresponse and noncoverage rates because of households that use only cellular telephones and to facilitate validation of key BRFSS interview questions, beginning in 2009, BRFSS gradually integrated a multimode design into its RDD landline telephone survey. A survey of adults who use only cellular telephones was conducted in parallel with the ongoing landline-based health survey to reduce noncoverage and related bias in key estimates. A mail follow-up survey to nonrespondents of the RDD landline telephone survey was implemented in 6 states in 2010 to raise response rates and to assess nonresponse bias. A physical measurement study was piloted in 2 states in 2010 to adjust for recall bias with self-reported data. We describe changes that have been made as well as changes that are being implemented in BRFSS operations and comment on the effect these changes have had on coverage, response rates, and related survey bias.

full-text:
Preventing Chronic Disease: November 2011: 10_0230