Ask the Expert: Dr. Stephen Blumberg

Beginning in 2003, NCHS’s National Health Interview Survey [NHIS] has collected information on cell phone ownership and use among its respondents. In the ensuing 10 years, cell phone usage has grown substantially and has become all but ubiquitous. A 2006 Health E-Stat revealed that, for the last six months of 2005:

• approximately 8.4 percent of households did not have a traditional landline phone, but did have at least one wireless phone
• approximately 7.8 percent of all adults lived in households with only wireless telephones
• approximately 7.7 percent of all children lived in households with only wireless telephones

Compare that to the last six months of 2012  [PDF – 375 KB]:

• approximately 38.2 percent of households did not have a landline telephone but did have at least one wireless phone
• approximately 36.5 percent of all adults lived in households with only wireless telephones
• approximately 45.0 percent of all children lived in households with only wireless telephones

As part of its NHIS Early Release Program, NCHS releases selected estimates of telephone coverage for the civilian noninstitutionalized U.S. population based on data from NHIS. Published twice yearly, Wireless Substitution reports deliver the most up-to-date estimates from the federal government concerning household telephones and wireless telephone usage.

In December 2013, NCHS published the latest wireless substitution data in two reports: Wireless Substitution: Early Release Estimates From the National Health Interview Survey, January-June 2013  [PDF – 363 KB], and Wireless Substitution: State-level Estimates from the National Health Interview Survey, 2012  [PDF – 400 KB]. We spoke with Dr. Stephen Blumberg, lead author, about wireless substitution and cell phone use.

 

Q: What are your favorite findings regarding wireless substitution?

A: What we used to say was – and I can’t claim credit for this phrase, it was a colleague of mine who originally came up with it – we described the wireless-only population as “the young and the restless.” They were young adults, they were renting their home, they were not married, and they didn’t have kids.  Now, more than half of wireless-only adults are age 35 or older, and adults with children are almost as likely as adults living alone to be wireless-only.

The other piece of this that has just been surprising to me… I’m not trained as an epidemiologist, and yet, in a way, I’ve been tracking an epidemic for the past 10 years. It’s not as serious as a health crisis, but there’s been steady growth, I’ve been looking at it on a regular basis, and I’ve been trying to identify who’s most “at risk,” if you will, of being wireless-only.

I sort of joke about that—using the language of “risk”—but one of the other findings that we’ve always seen is that those adults who are wireless-only also are more likely to engage in health risk behaviors. They’re more likely to binge drink. They’re more likely to smoke. They’re more likely to be uninsured. Other researchers have found that they’re less likely to wear seatbelts when driving.

And this remains even after controlling for age, sex, race, education, the usual controls. There’s something about being wireless-only that is appealing to the same group of people who are more likely to engage in health risk behaviors. A professor I was once talking to joked that therefore the wireless- only are a risk group and we should be targeting our interventions to them.

 

Q: There were children in 2003 who may have been in wireless-mostly or wireless-only homes. They’re now adults, 10 years later. Do you see this as part of the growth [in wireless-only], as these kids come out of their homes?

A: Certainly. We still talk about this as wireless substitution, that people are choosing to substitute a cell phone for a landline phone. But, in many cases, these are adults who have never had a landline phone as an adult. So they haven’t substituted anything in their lives, necessarily. We used to talk about “cutting the cord.” They never had a cord.

 

Q: What prompted including questions on wireless phones in the 2003 [National Health Interview] Survey?

A: We recognized that [the growing prevalence of wireless-only households] was going to have an impact on telephone surveys generally, and the CDC conducts several large telephone surveys, including the BRFSS [Behavioral Risk Factor Surveillance System], the National Immunization Survey, and SLAITS [State and Local Area Integrated Telephone Survey]. So it was a concern within the agency that we needed to know more about the impact this would have. It so happens that, being a door-to-door survey, the National Health Interview Survey talks to people who have cut the cord, as well as people who have no phones whatsoever. Therefore, it was an ideal vehicle for collecting this kind of information.

I am often asked if CDC is collecting this because of E-911, or if CDC is looking to these data because of an interest in being able to blast health emergency information via text message, or if CDC is concerned about health risks from using cell phones. Those weren’t the reasons we started collecting this information.

 

Q: Can you talk a little bit about the coverage bias, and how going mobile is, in many ways, a threat to getting good data.

A: It was originally a threat to getting good data. When we started this, only a few percent of adults were wireless-only. And, the concern was that that population would grow by a small amount, but a significant enough amount that the telephone surveys that were not dialing cell phones – was which just about every telephone survey – would start to see coverage problems.

The rate of growth in this wireless-only population was faster than anyone originally anticipated. It led to the survey research community doing a lot of research in a very short period of time regarding how best to call cell phones and conduct telephone interviews on cell phones. Nearly all major survey research organizations are now doing that, and they have identified best practices for doing that—they  have developed the operations, they have developed the weighting strategies that are necessary to combine the landline frame and the cell-phone frame—to the point that the wireless-only population isn’t as much of a threat to telephone surveys anymore.

Where it is still a threat, to some extent, [is that] some people are still less willing to participate in a survey on a cell phone than on a landline. So you can reach them, but they’re not as willing to participate. However, we’re seeing declines in participation rates in the landline population as well, so we don’t know that this is a device-specific effect.

 

Q: When you look at the state data, what accounts for the differences between states?

A: What we tend to find is that the strongest correlations to the state data that we have are living arrangements and income. Young adults who are living on their own, households where the oldest person living in the household is under 30 years of age…we see that states where that’s more prevalent are also more likely to be wireless-only. So, for instance, we see that in states with high wireless-only prevalence such as Utah or Idaho, there are lots of households headed by young adults. In contrast, in the Northeast, we still see a lot of households with older or elderly adults living in them, and they’re much less likely to be wireless-only. Then there’s also an income effect, so states in the South with lower incomes also tend to have higher wireless-only prevalence.

 

Q: Do you see, down the road, any way that mobile communications could open the door to different kinds of self-reporting?

A: Yes, we’re seeing that innovative research organizations are learning to use smartphones to deliver content and conduct surveys in ways that we’ve never considered before. You can do a survey on a smartphone and present pictures or video for that survey. You can also do real-time data collection with smartphones. So if you want to periodically sample snippets of time in a person’s life, where every so often you interrupt them, ask them what they’re doing, how they’re feeling, you can get much better experiential data through smartphone surveys. However, we’re not doing anything like that at NCHS yet.