Cell phone popularity a barrier for public health data collection: More Americans forgoing phone landlines
==========================================================================================================

* Kim Krisberg

The growing use of cell phones has reshaped day-to-day lives, from being able to reach almost anyone at any time to putting the Internet in a person’s pocket. And despite a few common gripes such as losing a signal or running out of minutes, cell phones have no doubt made life more convenient — for the most part.

There is at least one group of people for whom cell phones have become a bit of the proverbial thorn in the side: health researchers. The problem comes into play for researchers and surveyors who collect what is often essential health data via phone, relying on a randomized sampling of telephone numbers to bring back information and inform conclusions that users of such data can deem representative. Historically, this was not difficult — telephone numbers were readily available, corresponded with a person’s geographic location and usually belonged to an adult. But today, growing numbers of people have no landline at all, only a cell phone, and general assumptions about who will be at the other end of a phone call are quickly shifting. Still, these are secondary concerns. The primary concern is whether cell phone numbers are included in the survey design in the first place — a less-than-easy task, but one that is becoming more important in gathering truly representative data on people’s health and behaviors.

![Figure1](http://www.thenationshealth.org/http://www.thenationshealth.org/content/nathealth/39/7/1.2/F1.medium.gif)

[Figure1](http://www.thenationshealth.org/content/39/7/1.2/F1)

People wait to buy the new version of the iPhone in June in San Francisco. The popularity of cell phones is causing roadblocks for health researchers who collect data via phone calls.

Photo by Justin Sullivan, courtesy Getty Images

“The survey research industry is in a time of upheaval,” said Stephen Blumberg, PhD, senior scientist with the National Center for Health Statistics at the Centers for Disease Control and Prevention. “The survey modes and methods that worked reliably for the past decade or two need to be rethought. I fully expect that we will come out of this with new modes and new methods that continue to accurately reflect the behavior and attitudes of the public…Right now is a time of challenge, but a time of interesting experiments and new opportunities.”

Since 2003, the National Center for Health Statistics has been tracking the prevalence and characteristics of wireless-only — aka cell-phone only — users, putting out reports every six months on their numbers as well as their health-related behaviors. One such report, released in May, found that the “number of American homes with only wireless telephones continues to grow.” According to “Wireless Substitution: Early Release Estimates from the National Health Interview Survey, July-December 2008,” which Blumberg co-authored, more than 20 percent of surveyed homes had only wireless telephones during the second half of 2008, an increase of almost 3 percent from the first half of 2008 — a number that represented the largest six-month growth in wireless-only users since the center began collecting such data. In addition, more than 14 percent of American homes received all or nearly all phone calls on cell phones despite also having a landline. The biggest predictor of whether a household is wireless-only is whether the residence is owned or rented, with renters four times as likely to be wireless-only.

The demographics and health statuses of wireless-only users make their inclusion in health surveys important as well. According to the May report, men and those living in or near poverty were more likely to live in households with only wireless phones as are black and Hispanic adults. Binge drinking, smoking and being uninsured were more likely among wireless-only users, researchers found, but they were also more likely to be physically active, more likely to report being in good health and more likely to have been tested for HIV.

Blumberg noted that even when controlling for age, income and home ownership, such health characteristic differences still persist and while the data isn’t available to specifically explain why, Blumberg had some guesses. Perhaps, he said, “people who choose to be readily available to all of their friends, are those people who are more likely to drink and smoke with those friends…or perhaps, there are perceived risks associated with giving up a landline telephone and those people who are less likely to take risks are less likely to smoke or go uninsured.” Nevertheless, certain health surveys might end in biased results if wireless-only users are not included, he said.

“You should hope that your data collection contractors are cognizant of these issues,” Blumberg told *The Nation’s Health.* “But we are in a period when there will be question marks around some estimates and we need to be aware that some nonsampling errors may be growing.”

Including cell phone numbers in surveys, however, is not always easy, especially in terms of costs and in light of already-strained public health budgets. For example, automated dialers can be used to call landlines, but a law passed in the early 1990s requires cell phones be dialed by hand — a much more time consuming, and thus more expensive, method. Also, calling cell phones uses respondents’ minutes, so surveyors should ethically reimburse them, Blumberg said. In addition, calling cell phone numbers adds to the list of screening questions, such as whether the phone is a business or personal number. And because there is a higher chance the respondent is younger than 18 and thus not a valid respondent, many more calls are, simply put, a waste of the surveyor’s time. It does not cost more, however, to procure cell phone numbers than landline numbers. Safety and privacy concerns also pop up, Blumberg noted, such as whether a respondent is driving or in a location not suited for discussing personal health information.

“Overall, the cost of high quality population-based surveys is going up and there seems to be no way to avoid that increasing cost,” he said. “Certainly, we’ll see more lower-quality surveys…we may see fewer population-based surveys and more targeted surveys of known populations.”

Many researchers and organizations are taking up the challenge, however.

This year, for the first time, CDC’s state-based Behavioral Risk Factor Surveillance System will require all states to include cell phone interviews. Eighteen states conducted a pilot project in 2008 using cell phone numbers, according to Lina Balluz, ScD, MPH, chief for CDC’s Behavioral Surveillance Branch, and this year, states are being asked to call a minimum number of cell phone users, with requirements differing based on a state’s capacity. Balluz said the new requirement may reap better data on younger people, especially those ages 18 to 34, significant percentages of whom forgo the traditional landline.

“I think many telephone surveys are already moving in that direction,” she told *The Nation’s Health*.

![Figure2](http://www.thenationshealth.org/http://www.thenationshealth.org/content/nathealth/39/7/1.2/F2.medium.gif)

[Figure2](http://www.thenationshealth.org/content/39/7/1.2/F2)

Wireless-only users are not alone in their potential to muddy up survey results, however. The largest group of phone users are those with both a cell phone and landline, and because such people have a higher probability of being included in a study, “dealing with the overlap can be tricky,” said Michael Davern, PhD, research director at the State Health Access Data Assistance Center at the University of Minnesota. Davern, who co-authored May’s wireless estimates from the National Center for Health Statistics, noted that while address-based mail surveys can be a good alternative to telephone surveys — and the issues that wireless-only users present — speaking to people via phone still reaps the richest details.

“Where we are as a field is that we pretty much agree that something has to be done,” Davern said. “The telephone is such a robust mode for data collection that it won’t be going away.”

For more on cell phones and data collection, including CDC wireless estimates, visit [www.cdc.gov/nchs](http://www.cdc.gov/nchs) or [www.shadac.org](http://www.shadac.org).

*   Copyright The Nation’s Health, American Public Health Association