Community health data in a wireless world: Inclusion of cell phone sample in a regional population health survey

Telephone surveys are a tested way to obtain health information from diverse populations. Historically, such surveys have relied on landline telephone samples, but the proliferation of cell phones has raised questions about the representativeness of landline telephone surveys. According to 2008 NHIS estimates, one in six U.S. homes (17.5%) has only cell phones. Some national and state health surveys have added cell phone only (CPO) samples, but there are few examples specific to local health surveys. This presentation provides a case study of the inclusion of CPO households in a local health survey, and examines distinctions between CPO and landline populations with a focus on disparities in access to care. We also examine overall differences between two sampling frames, RDD and Cellular. Using data from the PHMC 2008 Southeastern Pennsylvania Household Health Survey, the longest-running local random digit dial (RDD) telephone survey in the country, we examine CPO respondents (n=121) compared to landline respondents (n=4094) by demographic characteristics and indicators of access to care (e.g., regular source of care, insurance, cost barriers). Consistent with other research, findings show that CPO respondents are disproportionately young (53% ages 18-34), male (51%), low-income (23% below poverty level) and are more likely to be racial minorities (50% Black, 19% Latino, 6% Asian), compared to landline survey respondents. Preliminary findings indicate that CPO respondents experience reduced access to care even when controlling for age, gender, race, and income. We will discuss the research and policy implications of including CPO respondents in regional telephone health surveys.