The movement in the past decade toward patient-centered care has increasingly emphasized patient empowerment in health care. In particular, the Chronic Care Model, characterized by the interaction between an “informed and activated patient” and a “prepared and proactive practice team” has been highlighted as a fundamental model for optimum care [1,2]. Alongside the changing tide in health care delivery, there has been a revolution in information technology. With the development of new technology and Web 2.0 and 3.0 communication media, the field of eHealth has emerged and with it a plethora of new opportunities for individuals to access and exchange health information, manage their health through electronic platforms, and participate in “peer-to-peer health care” [3-5]. These online opportunities have been identified as a means to better enable patient empowerment and self-management of care [6].

The field of eHealth has enabled public health and medical practitioners to communicate with patients in both traditional and novel ways to address health concerns such as diabetes management [7], heart health [8], cancer prevention and health promotion activities [9-11], and smoking cessation [12,13]. Online strategies range from adaptations of more traditional communication methods, such as the delivery of tailored information and the creation of support networks [14], to more innovative developments, such as the implementation of smartphone applications for disease prevention and management [15-17].

These eHealth developments have elevated the importance of assessing the extent to which technology has empowered patients and improved health in general and among the most vulnerable populations in particular [18-21]. With noted disparities across racial and social groups in chronic health outcomes, such as cancer, obesity, and diabetes, it is essential that researchers thoughtfully examine any differences in the implementation, uptake, and impact of eHealth strategies across groups that bear a disproportionate burden of disease [20,22,23]. Current eHealth studies are limited in that many, such as those published by the Pew Internet & American Life Project, report national percentages without rigorous statistical control to determine what factors may be true drivers of any eHealth disparities. Other research has focused on issues of access to technology based on the original concept of the digital divide, which formulated 2 groups—those with access to the Internet and those without [5,24]. Gaps in access to the Internet have been persistent in that lower socioeconomic status, minority racial/ethnic groups, older age, and poorer health, among others, are associated with decreased access to the Internet [19,21,25-29]. These access patterns can lead to differential access to health information that might intensify health disparities [30]. Yet, this dichotomous oversimplification wrongly suggests that all “haves” use the Internet in a similar manner.

However, the limited number of published studies focused on use of the Internet for health point to noted communication inequalities or differences in use and engagement across important racial and social groups [18,22,31]. For example, those in the lowest income and education brackets are shown to be considerably less likely to seek out health information online compared to those in higher income and education brackets. Similarly, non-Hispanic blacks and Hispanics are significantly less likely to seek out health information online compared to their non-Hispanic white counterparts, yet these differences by race/ethnicity have begun to narrow overall. Recent unadjusted Pew Internet & American Life Project data highlight differences in topic-specific seeking behavior in that more non-Hispanic blacks and Hispanics report using the Internet to find information on how to lose weight and pregnancy compared to non-Hispanic whites, whereas a larger percentage of non-Hispanic whites report using the Internet to find information on a specific disease or problem [5,20,32,33]. These differences in eHealth use are of importance to public health practitioners and health care providers in that these communication behaviors could lead to significant health-related disparities [22].

Evidence in support of communication inequalities in health-related Internet use is building [20,34,35]; however, we lack a clear understanding of comprehensive differences by sociodemographic factors, such as race/ethnicity, socioeconomic status (SES), and sex in the utility of the Internet for eHealth tasks. Past studies have primarily examined online disparities in isolation, have not adequately adjusted for confounding factors that could drive use, or have used older datasets that do not reflect the ever-changing digital landscape [5,34,36-38].

This study aims to employ an up-to-date, comprehensive examination of eHealth use by sociodemographic factors to illustrate potential profiles of disparities across a number of communication domains. We hope this work will assist future health communication interventions and efforts that seek to use the Internet, email, and social media to reach and engage underserved populations.

Data for this study were drawn from the National Cancer Institute’s 2012 Health Information National Trends Survey (HINTS). HINTS is a nationally representative survey of the US noninstitutionalized adult population that collects data on the American public’s need for, access to, and use of health-related information [39]. HINTS is publicly available online [40]. Data used in this study are from HINTS 4 Cycle 1, collected from October 2011 to February 2012 (N=3959) through mailed questionnaire. The sample design was a 2-stage stratified sample with addresses selected from a comprehensive United States Postal Service national residential file, and individual respondents were selected per each household in the sample. The final response rate for HINTS 2012 was 36.7%. Further details on survey design and sampling strategies are published elsewhere [41].

To assess hypothesized differences in eHealth usage and engagement, we used 11 HINTS variables that were asked of those respondents who reported yes to ever going online to access the Internet or World Wide Web or to send and receive email (N=2358). The 11 eHealth tasks are presented in 3 domains relevant to health communication (health care, health information-seeking, and user-generated content/sharing). Items were grouped into domains in effort to illustrate trends across eHealth tasks and for purposes of informing future health communication-related interventions [42]. The categorization of items into domains was informed by both mass communication theory, such as uses and gratifications theory, as well as recent health care policies, specifically the Affordable Care Act and Healthy People 2020, in which there is interest to track progress in goal achievement [43-45]. For example, one of the goals outlined in Healthy People 2020 is aimed at improving access to comprehensive, quality health care services, research is emerging that correlates increased engagement with the Internet and access to health care services [46]. The eHealth items assessed in this study were “In the past 12 months, have you used the Internet to look for health or medical information for yourself?” (yes/no) and “In the past 12 months, have you used the Internet for any of the following reasons?” (yes/no) as listed in Figure 1.

For the purpose of this analysis, primary predictor variables included in each model represent sociodemographic characteristics: place of birth, race/ethnicity, home ownership, education, income, age, and sex. Hot-deck imputation was used to replace missing responses with imputed data for the race and ethnicity variables in HINTS 2012. With this approach, the resulting distribution preserved the distribution of values observed for respondents.

All models adjusted for occupational status, marital status, children, health information–seeking (ever sought health information from any source), regular access to a health care provider, insurance status, health status, personal cancer history, and family cancer history.

We used multivariable logistic regression to model the fitted odds that SES (education and income), race/ethnicity, age, and sex independently and differentially predicted eHealth usage in the population of online US adults. We used SAS-callable SUDAAN 10.0.1 to account for the complex sampling design used in HINTS and to incorporate jackknife replicate weights needed to compute accurate standard errors. All analyses were weighted to provide nationally representative estimates. We calculated weighted percentages, odds ratios (OR), and 95% confidence intervals (CI) utilizing complete case analyses with listwise deletion for each model (N=2358).

Being younger and female has consistently been predictive of increased use of eHealth [5,21,25,47]. Younger generations who have grown up with technology have been labeled “digital natives” and they are more comfortable using technology for everyday needs, including management of their health care needs. In comparison, older generations, labeled “digital immigrants,” have had to learn and acquire the necessary skills needed to navigate the Internet and are generally less comfortable using technology [48]. Females also tend to have increased eHealth utilization due in part to their higher engagement in both health care-related online activities and increased use of general social media, such as SNS [3-5,49]. This could be because of their role as the health care liaison for their family members.

Our analysis identifies specific proxies of SES that are more reliably associated with eHealth use. Education was more consistently predictive of eHealth use across the health care and information-seeking domains, whereas household income was somewhat less predictive across items and domains. Although both education and income have been used to describe SES, in considering technology use and health communication, our analysis suggests that education may be a more salient proxy as compared to income.

Furthermore, this distinction between common proxies for SES may offer insight into the fundamental drivers of online use for health and subsequently assist in the development of more effective interventions and programs. Divides among those less educated indicate that issues of health literacy and eHealth literacy may be important factors. Research is emerging that advocates matching eHealth technology to the eHealth literacy (defined as “the ability to seek, find, understand, and appraise health information from electronic sources and apply knowledge gained to addressing or solving a health problem”) of the intended user [50-52]. Future analyses of online communication-based interventions should better investigate and address issues pertaining to eHealth literacy in an effort to reduce communication inequalities across low-SES groups.

It is also important to examine the direction of the relationship between education and engagement in eHealth tasks. Lower levels of education were associated with increased use of social media for health, whereas higher levels of education were associated with increased use of the Internet to engage in health care-related activities and search for information—much like the patterns we see for static Web 1.0 engagement. Similar patterns are seen in the Pew Internet & American Life Project’s 2013 Health Online study in that higher percentages of non-Hispanic blacks (35%) and Hispanics (38%) reported using their mobile phones to access health information compared to non-Hispanic whites (27%) [5]. These differences in use across domains may indicate potential inequalities in not only the quality of information obtained, but also patients’ engagement with the health care system. Further empirical examination is warranted to better objectively assess quality of health information shared via social media compared with more basic Web 1.0 sites as well as quality and satisfaction of care between those that engage in eHealth care and those that do not.

On the other hand, our study also indicates the potential for eHealth technologies to aid in reducing communication inequalities and disparities in health. Our data found no racial divides among the most vulnerable groups in eHealth use once access was achieved; there was only increased use of downloading information to a mobile device and looking online for a health care provider among other race individuals. This finding is consistent with several prior studies [5,25,35] and points to the opportunity to better explore the association between eHealth utilization and health outcomes among Hispanic and non-Hispanic blacks more directly.

Although our research is an important addition to the literature, we note its limitations. First, the low survey response rate may increase sampling error in our estimates; however, overall sampling coverage was enhanced through the stratified design. Also, because this was a cross-sectional survey study, it is challenging to account for unmeasured confounding variables. Future studies should examine potential factors related to those sociodemographic variables that were predictive of our outcomes, such as literacy, to determine their role in eHealth use. In addition, our study does not attempt to examine factors that may moderate or perhaps mediate the main associations presented within this analysis. For example, past research has examined the role of trust in terms of patients’ trust of the Internet as well as their own physicians and/or the health care system [53,54]. Examination of these psychographic variables is an important addition in better understanding the complex predictors of eHealth use. Yet, with such persistent inequalities in health, we cannot underestimate the importance of maintaining a constant understanding of eHealth use by sociodemographic characteristics.

A more qualitative examination into the rationale offered by patients for their eHealth utilization would build upon this work and offer a more robust understanding of why certain groups do and do not use the Internet for health purposes. A recent study conducted by the Pew Internet & American Life Project reports that 36% of non-Internet users cited that they did not use the Internet because they did not think that it was relevant to them, whereas 32% reported difficultly in use [55]. Investigation as to whether or not these reasons are applicable to eHealth use is warranted.

This study illustrates that lower SES, older, and male online US adults were less likely to engage in a number of eHealth activities compared to their counterparts. In an effort to reduce existing disparities in health outcomes, clinical and public health communication strategies should be attuned to these differences in online use.

Our results have important implications for clinical care and public health communication efforts attempting to leverage Web 2.0 and 3.0 platforms. It is evident that a one size/platform would not fit all, as significant demographic and individual factors influence eHealth engagement. For example, campaigns and interventions targeting women or younger populations may see success in utilizing user-generated content and sharing sites. Whereas, in the current Chronic Care Model of medical care, offline materials should perhaps supplement online health information if practitioners would like to ensure equal access to information across educational strata, age, and sex.