The purpose of this study was to examine the prevalence and predictors of participation in online health-related behaviors. Millions of people are now using the Internet on a regular basis, and much of this activity has been focused on health. Online eHealth tools have the potential to help people manage their health and health care, but little is known about the use of different kinds of eHealth tools across the population [1,2]. In 2003, an estimated 12.5 million online health-related searches were conducted globally each day [3]. However, aggregate figures like this do not tell us about what people are actually doing when they interact with online health-related content and tools. This study examined participation in three different online activities—health or medical information-seeking for oneself, use of online support groups for people with similar health and medical issues, and online purchase of medicine and vitamins—in order to better understand the potential of different health-related activities to influence people’s lives.

Understanding how people use the Internet for health-related activities can be critical to the design of relevant sites in order to attract and retain users [4]. One of the first steps in developing both traditional and computer-based health communication is an analysis of the target audience [5,6] to identify the audience’s preferences and needs and influence product acceptance and utilization [6]. Further, analysis of differential use of Internet applications by subpopulations can serve to identify potentially underserved groups and allow for the development of strategies to meet their needs [2].

Eysenbach [3] identified four broad applications of the Internet: (1) content (eg, information seeking), (2) community (eg, bulletin boards, chat rooms), (3) communication (eg, email), and (4) e-commerce (eg, buying or selling products and services on the Internet). The most commonly reported function for health has been content, or searching for health and medical information on the Internet [7]. Fewer Internet users have reported engaging in the other three functions for health-related purposes [7]. A recent study examined one health-related communication application—emailing a health care provider—and found a slow, but significant, increase in the prevalence of this activity from 2003 to 2005 (7% to 10% of Internet-using adults emailing a health care provider in the past year) [8]. The present study focused on health-related activities related to the other application areas of content, community, and e-commerce.

This study analyzed data from the National Cancer Institute’s Health Information National Trends Survey (HINTS) 2005 to answer the research questions. The HINTS survey collects nationally representative data about the US public’s cancer communication practices, information preferences, risk behaviors, attitudes, and cancer knowledge [20]. The HINTS survey has been administered twice, first in 2003 and then again in 2005.

Participants in the 2005 survey (n = 5586) were recruited via random digit dialing of all telephone exchanges throughout the United States and then randomly selected from among the adults in the household. Survey administration averaged 30 minutes per participant. Response rates were 34% at the household screening level and 61% at the sampled person interview level [21].

All analyses were conducted using STATA Version 9 [22] to properly calculate standard errors for the multistage sampling design. The present analyses were conducted on the subset of the sample who reported going online, based on their response to the item “Do you ever go online to access the Internet or World Wide Web, or to send and receive e-mail?” All results were weighted to be representative of the United States adult population. To correctly calculate standard errors for the subset of adult Internet users residing in the United States, the SUBPOP statement in STATA Version 9 was used.

Descriptive statistics were calculated for all variables. Logistic regression analyses were conducted to answer the research question of whether selected sociodemographic, health status, and Internet use variables predict the Internet behaviors of looking for health information, participating in an online support group, or purchasing medicines or vitamins online.

HINTS questionnaire items were used to measure the following demographic variables: age, gender, ethnicity and race, education, income, marital status, number of children under 18 years in the home, general health status, and cancer history. Data for each variable were grouped into categories consistent with other research using HINTS data [23-25]. Additional items were selected in order to measure factors related to Internet use: location of access, type of connection, and frequency of use. Outcome variables measured use of the Internet for personal health information, use of online communities, and the purchase of medications or vitamins.

The unknown rates, which included missing, don’t know, or refused responses, were less than 4% for all variables except income. For all variables except income, unknowns were removed from the denominators when calculating percents. In the regression analyses, all unknowns, except those for income, were excluded from all of the models run and the subsequent results that are presented. Because of the relatively large number of respondents with unknown income, “unknown” was included in the regression analysis as a separate category. The HINTS questionnaire can be viewed at the National Cancer Institute’s website [26].

This study examined the prevalence of participation in online health activities and whether participation in these activities is predicted by sociodemographic, health status, and Internet use factors. We found that participation in these activities varied and that different sets of variables predicted who was likely to engage in these different activities.

Almost 60% of the Internet users surveyed reported searching for health information for themselves in the previous year. Fox and Rainie [7] identified the search for health information as the most common online health-related activity, and the present study found it to be the most prevalent of the three studied behaviors. This finding has several critical implications for both information providers and seekers. First, health agencies can use findings such as these to justify making health information available online because many people are using this channel, often before talking to their clinicians. Although people may prefer to go to their clinicians first, they have reported that they actually go online first for health information [27].

However, just making information available is not sufficient. Health agencies need to take responsibility for understanding how best to meet the needs of online information seekers. The rate of Internet use and broadband adoption has continued to increase among those with less than a high school education [28], suggesting that the number of users with lower literacy levels will grow [29]. Although this study found that respondents with higher education were more likely to search for health information, people with low literacy have identified health information as one of the primary types of information for which they would search [30]. Given that much of the health information on the Internet has been written at too high a level for many population segments [1], the development of materials that can be used and understood by audiences with lower literacy levels will be critical.

Another implication of online health information seeking relates to whether consumers can assess the quality of online health information. Despite the existence of reputable health portals, most users begin their search from a search engine [7] and rarely go beyond the first page of returned results [31]. Consumers must possess the skills to sort through and critically evaluate online information if the Internet is to realize its full potential in helping people meet their information needs.

Health status has been associated with health-related Internet use in previous studies [3,10], but these studies have reported conflicting evidence on whether having a chronic condition predicted health information seeking on the Internet. This study found that neither health status nor having been diagnosed with cancer were associated with greater Internet use for health information seeking. However, on the HINTS survey, respondents who indicated that they had been diagnosed with cancer could currently be in remission and not actively coping with a current cancer diagnosis. Thus, the use of this variable as a proxy measure of health status may not truly distinguish healthier versus less healthy individuals.

Health information seeking was also predicted by Internet use variables. Although almost all (90%) users had Internet access from home, this study found that those with access primarily from work or from both home and work were more likely to search for health information for themselves. People were more likely to use the Internet to search for health information if they used the Internet for an hour or more on typical weekdays and, as in previous research [9], had faster connections.

Having Internet access in the home has been considered an important indicator of equitable access among population groups [32], but the current results suggest that work access may be a critical factor in using the Internet for health information seeking. Frequency of use on weekdays and not weekends was related to searching for health information, again suggesting use on typical workdays. Accessing the Internet from work may offer some users faster connections. Given that most contact with medical professionals occurs during traditional work hours, receiving a phone call at work with a new diagnosis or lab results may provide a cue to action for an Internet search for those with the means to immediately begin searching. Those with access from both home and work may simply have more opportunities to search for health information.

The use of online support groups for people with a similar health or medical condition occurred with the lowest prevalence, with only 3.8% of participants reporting participation in such a group. Online support group use appears to be an infrequent activity, and the low rate of participation in health-related support groups may coincide with the general decline in online chat room participation reported by Fallows [16]. However, this finding may also result from the survey methodology. The HINTS survey question asked only about participation in online support groups; however, people may use Internet forums, bulletin boards, online communities, mailing lists, chat rooms, wikis, and blogs as means of providing or receiving online support or to share information. Users may not identify experiences with these different varieties of online communication as “participation in an online support group,” so this study may be underestimating the prevalence of Internet use for social support. Future surveys should first ask generally about use of the Internet for health-related social support and follow up with questions asking about specific mechanisms (eg, message boards, blogs) for getting support.

The present study found that those reporting “fair” or “poor” health were more likely to use online support groups than those with “excellent” health. This suggests that those with greater health needs appear to be taking greater advantage of the Internet to help them cope with their conditions, even though online support groups usage was limited overall.

Those with lower incomes were also more likely to use online support groups, although previous studies found that those with higher incomes were more likely to search for health information [10] or purchase medications online [19]. Possibly, those with higher incomes have other means of support, while those with lower incomes are turning to the Internet for assistance. An implication of this finding may be that people with limited incomes are replacing medical care with online support groups, which may put them at risk if they do not get needed assessments and treatment. However, this finding could also be the result of statistical error based on the low rate of participation in this activity overall as well as the high number who did not report their income.

Respondents with Internet access only from their home were more likely to use online support groups than those with access at both home and work, suggesting that those who were “less connected” may have greater need for online support. Continued research on the impact of location of Internet access and use of online social support tools is needed to understand and validate this relationship.

Almost 13% of respondents reported purchasing medications or vitamins online, which is much higher than the 4% and 5% reported in previous studies [18,19] and the 9% in the 2003 HINTS survey [33]. Given the potential for negative outcomes if poor oversight and regulation of this activity occur, research on the impact of the online purchase of medications and vitamins could be expanded so that we better understand the nature of this behavior. In contrast to Fox’s [19] finding that income and years of Internet experience predicted online medication purchase, this study found only age and marital status to be significant predictors. This activity may have increased recently because of national discussions about prescription drugs and increased interest in complementary and alternative medicine. It may also reflect the increase in use of the Internet for e-commerce in general [34].

Across online health-related activities, this study examined the relative importance of specific sociodemographic variables. One of the variables most connected with these activities was gender. The current research found that women were more likely than men to engage in a search for health information for themselves. They were equally as likely as men to participate in online support groups and to purchase medications and vitamins. However, married participants were more likely to purchase medications and vitamins online, indicating that women may have additional influence on this behavior.

Traditionally, men have been less likely than women to engage in health information seeking and preventive health behaviors. Since men do use the Internet for so many other activities, perhaps the Internet provides an opportunity to provide non-traditional health education opportunities directed toward men. A comparative content analysis of existing websites for men’s and women’s health is needed to determine gaps in such resources. Continued formative research could also help determine the types of health-related sites or online formats that would better appeal to male audiences.

Despite implications from other Internet intervention studies [11-14], racial and ethnic minorities appeared to engage in health-related activities at rates similar as white users. These results may suggest that use of the World Wide Web is different than use of Internet-based systems on which these studies have been conducted. Even though African American and Hispanic populations have long shown lower rates of Internet adoption as compared to white users [32,35,36], the current research suggests that once online, they are as likely as white users to engage in health-related activities. Ongoing research is needed to understand the patterns of use by different subpopulations.