Going Wireless: Initial findings from the inclusion of cell phones in the Household Health Survey

Telephone surveys are a tested way to obtain crucial health information from diverse populations.  Historically, such surveys have relied solely on landline telephones to collect data, but the proliferation of cell phone usage—and particularly of households that are completely wireless—raises important questions about the coverage for surveys that rely entirely on landline telephones.

In response to our changing landscape of communication technologies, in 2008 the Community Health Data Base incorporated a pilot study of cell phone interviews into its biennial Southeastern Pennsylvania Household Health Survey of 10,000 households in the region.  Focusing on Philadelphia alone, the 2008 survey included interviews by cell phone with 300 adults ages 18 and older in Philadelphia.

While the inclusion of cell phone-only households in survey research has been growing rapidly in recent years, the impact of cell phone-only users on health research is not well understood, and there are few examples specific to local surveys. For this reason the Community Health Data Base is very pleased to share with you a few initial findings from the 2008 Household Health Survey regarding similarities and differences in social and demographic factors between cell phone only respondents and others.

WIRELESS-ONLY HOUSEHOLDS: NATIONAL & LOCAL PICTURE

According to current national estimates from the National Health Interview Survey, one out of every five households in the U.S. (20.2%) is reachable only by wireless phone or cell phone only (CPO). In addition, 18.4% of all adults—nearly 41 million adults, live in households with only wireless telephones. In addition, certain sociodemographic groups are more likely to live in cell phone only households, which means that they those demographic groups may be more likely to be underrepresented in traditional phone surveys.  Most notably, adults living in or near poverty, young adults, and Latino adults are all more likely to live in cell phone only households than their counterparts.

In Philadelphia, the proportion of cell phone only households may be slightly lower than the national average.  According to recent models developed by Abt SRBI, the survey research firm that has fielded the Household Health Survey since 1996,  it is estimated that in Philadelphia about 14% of Philadelphia is cell phone only, 36% is landline only, and 50% have dual service.

CELL PHONE INTERVIEWS IN THE HOUSEHOLD HEALTH SURVEY

Inclusion of cell phone interviews presents a number of distinct challenges above and beyond the usual challenges of telephone survey research.  Among these: cell phone respondents must be compensated for the “minutes” used while they are completing the interview; interviewers must ascertain that the respondent is in a safe environment to be answering interview questions (e.g., not driving a vehicle); and all calls must be hand-dialed rather than computer-dialed, due to specific federal regulations.  Due to these challenges, including cell phones in survey research is a very costly—albeit increasingly necessary—undertaking. In spite of the challenges, the 2008 cell phone pilot survey was highly successful, and the target for number of pilot interviews (n=300) was met ahead of schedule.

Out of the 300 cell phone-based respondents, about 40% of them (n=121) were cell phone only (CPO) households. About 36% self-described as primarily but not entirely using a cell phone (“Cell Phone Mainly” users), about 2% reported using their cellular and landline telephones about equally, and 16% of cell phone respondents reported that they primarily used their landline (“Landline Mainly” users).

DEMOGRAPHIC CHARACTERISTICS OF CPO USERS IN PHILADELPHIA

The socio-demographic characteristics of the HHS cell phone sample are similar to those that have been found nationally.

Age
Cell phone only respondents are significantly more likely to be young than all other groups of phone users. However a gradient is seen: approximately 72% of CPO users are ages 18-39 compared to 53% of Cell Mainly users, 35% of those who use both phones equally, and 24% and 25%, respectively, of those who use mainly or only use landlines.

Race/Ethnicity
A higher proportion of CPO respondents are Latino than any other phone user group (20%, compared to 8.2% for Cell Mainly, 14% for equal use, 6.2% for Landline Mainly, and 11.0% for Landline Only users).  A slightly higher proportion of CPO respondents are Black or African American (48%) than other phone user groups (highest being Landline only at 43%).

Gender
CPO users were significantly more likely to be male than other phone user types. About 59% of cell phone only respondents were men; all other phone user groups were at least ten percentage points lower than that in their representation within their group.

Poverty Status
Notably, Landline Only users included the highest proportion of adults who lived below federal poverty level (30%). Cell Phone Only users were a close second (25.6% living below poverty level), followed by those who use landline and cell equally (21%), Landline Mainly users (12.4%) and Cell Mainly users (10.2%).

Uninsured
Cell Phone Only respondents were over two times more likely to be uninsured than those in all other groups. Three in ten CPO respondents had no private or public insurance (30%), compared to 14% of Landline Only, 12% of Cell Mainly, 9% of equal usage, and 7% of Landline Mainly phone users. This is likely accounted for by both the younger age and lower socioeconomic status (as seen above) of CPO respondents.

SUMMARY

The Community Health Data Base’s 2008 Household Health Survey presents a novel opportunity to examine the characteristics of cell phone survey respondents in the context of a local health survey.

As anticipated, we found important demographic differences between cell phone only respondents and others—as well, at times, as important differences between landline only respondents and others. The cell phone only users were more likely be male, to be Latino, and to be uninsured.  Cell phone only respondents were also more likely to live below poverty than most other groups (aside from landline only respondents, who were the most likely to be poor). The cell phone only respondents represent groups that are harder-to-reach through traditional telephone survey methodology, and thus by including cell phone sampling, the Household Health Survey is better able to include the voices of these underrepresented populations.

For more information about the Community Health Data Base or cell phone sampling, please contact Rose Malinowski Weingartner at rosemw@phmc.org.