CDC Health Survey System Meets the Cell Phone Challenge

Published: July 25, 2007

New technologies mean new ways of communicating, but it also creates unforeseen roadblocks. No one is more aware of this than those who maintain the Behavioral Risk Factor Surveillance System (BRFSS), CDC’s largest, on-going telephone health survey system, tracking health conditions and risk behaviors in the United States yearly since 1984.

Since its inception, BRFSS has used landline telephone sampling as a cost-efficient strategy for conducting health risk surveys of the US household population. However, in 2006, approximately 12 percent of US adults reported living in cell phone–only households. More than half of all adults living with unrelated roommates and one-in-four adults aged 18 to 34 years lived in cell phone–only households.

As a result, population coverage provided by landline telephones has been eroded to pre-1970s levels, raising concerns about the representativeness of landline telephone surveys. With the growing popularity of cell phones, this trend is expected to continue.

"Over the next few years you will see BRFSS moving away from reliance on only landline telephones as a means of reaching respondents and using instead a multimode approach," says head researcher Michael Link, PhD.

"Early on this will likely involve the addition of known cell phone numbers to our landline telephone sample," he adds. "By 2010, however, BRFSS could add a mail survey component as well. We have been conducting extensive pilot studies in both areas over the past few years in order to refine the proposed methodology of the future."

A recently completed pilot study assessed the feasibility and cost of conducting surveys with people sampled from known cell phone telephone exchanges. Three states (Georgia, New Mexico, and Pennsylvania) participated in the study.

The findings indicated that those in cell phone-only households were significantly different from those in the landline BRFSS survey in terms of both demographic characteristics and responses to key health conditions and risk behavior.

Based on the pilot study results, BRFSS is expanding the testing effort in 2008. Up to 25 states will be funded to conduct 200-300 cell phone interviews per state. BRFSS is also testing an alternative approach, sampling households through addresses rather than through telephone numbers and conducting interviews using a combination of mail and telephone surveys.

Whether surveys are conducted by landline or cellular, each presents its own set of challenges. "Surveys conducted by cell phone are more expensive for both the researchers and the respondent," says Link. "Because the sample frame of cell phone numbers is not currently screened for known business or out-of-service numbers, it requires a lot more interviewer time to screen these numbers, thereby increasing the costs.

"From the respondents' side, many are still charged for the minutes they use—although this varies significantly from plan to plan. As a result, most researchers calling cell phone numbers are offering a remuneration to help offset the real costs incurred by respondents. BRFSS is no exception. This will be the first time BRFSS has had to put such a plan into place, which raises some logistical challenges."

In the end, he says, the ultimate goal of the research efforts is to ensure the BRFSS meets the challenges of evolving technology and continues to produce data of the highest quality.