I met with them in June and learned a bit more about how they approached this massive data set (N=18,000+).
First, a bit of background. I became acquainted with Ha Tu in 2003 when, unbeknownst to each other, we came out with what appeared to be different estimates of the e-patient population. After talking with her to better understand HSC’s data, I wrote:
Recent reports, one by the Center for Studying Health System Change and another by a team of Stanford researchers, have estimated a lower population of health seekers by limiting the scope of their survey questions. We deliberately keep the timeframe open, asking if respondents had “ever” searched for a health topic, because it has been our observation that once an Internet user has been successful in an online endeavor, she will return to it the next time she has a similar problem or question, no matter how much time has lapsed between the searches. We also do not limit respondents to thinking exclusively about their own health concerns since our past research shows that more than half of Internet health searches are conducted on behalf of someone else. Indeed, for most Americans, health is a family affair, not a solitary activity. Finally, for the first time, we asked questions of respondents about 16 specific health topics, rather than just asking them about general health searches, and this drove up the number of Internet users who say they have searched for health-related topics online.
(See p. 1 of Internet Health Resources for more data from this era.)
This time around, Ha and Genna invited me to preview their data so we could understand the differences and similarities of our findings before releasing them to the world. As I wrote to them afterwards, meeting with them was better than a cup of coffee to get me awake and excited about work that day. It was a wonderful collaborative moment and I very much appreciate their reaching out to me.
Their report finds that 56% of American adults sought information about a personal health concern:
Use of all information sources rose substantially, with the Internet leading the way: Internet information seeking doubled to 32 percent during the six-year period. Consumers across all categories of age, education, income, race/ethnicity and health status increased their information seeking significantly, but education level remained the key factor in explaining how likely people are to seek health information.
It turns out that my report E-patients With a Disability or Chronic Disease probably tracks most closely with HSC’s data on “personal health concerns” since that population is more likely than the general population to search on their own behalf instead of for other people and to talk with a medical professional about what they find. For example: “Fully 49% of e-patients with chronic conditions talked with a health professional about what they found online during their most recent search, compared with 30% of e-patients with no chronic conditions.”
Yet there was still a difference between our findings since Pew Internet combines all searches (the respondent’s own as well as her searches for others) whereas HSC chooses to keep them separate. Ha and Genna account for the “whole picture” of the e-patient/e-caregiver population in the following footnote: “In total, 65 percent of adults sought information either for themselves or other adults in 2007; 40 percent did so using the Internet.” It is an excellent solution, answering some observers’ questions before they are raised, yet forwarding what they believe is an important measure of health behavior.
I very much appreciate the work of the Center for Studying Health System Change and am thankful for the time the researchers took to explain their data. Research grows stronger through collaboration and good research is an element of participatory medicine.