Expanding the Reach and Impact of
Consumer e-Health Tools

June 2006

< Back to Chapter 2 (Part 1)

Chapter 2. Mapping Diversity to Understand Users' Requirements for e-Health Tools (Part 2)

User Characteristics That Influence e-Health Tool Use

Public health interventions typically rely on broad demographic categories to identify who is affected by an issue, risk factor, or disease. Those most affected become the targets for an intervention. These demographic categories—including race, ethnicity, gender, age, income and education levels, and disability status, among others—are the basis for much of the current debate on the nature and extent of health disparities (HHS 2000, 2005a).

One of the original purposes of the present study—a purpose that could not be wholly fulfilled because of a lack of existing research and publicly available data—was to identify and analyze factors in addition to demographics that affect the adoption of e-health tools by those population segments most affected by health disparities. As noted throughout the report, studies suggest that populations that experience health disparities are also likely to experience disparities in technology access and use. Beyond these broad observations, however, little information addresses factors related to users’ motivation, engagement, and understanding of e-health tools and their relevance to strategies to promote greater use. The IOM Committee on Communication for Behavior Change in the 21st Century found that “data that provide a much deeper and more sophisticated understanding of how specific beliefs and behaviors and health status covary across the U.S. population and of how health behavior is shaped by sociocultural processes are not available. . . .” (IOM, 2002, p. 15).

Demographic characteristics or functional skills, such as low literacy, novice computer skills, and limited English proficiency, are the main factors that have been used to characterize user groups to date. Gender, education, income, and age are strong determinants of interest and behavior in health information-seeking across media, according to a review of prevention communication and media use (Lieberman, Benet, Lloyd-Kolkin, et al., 2004). Regardless of ethnicity, well-educated, affluent women under age 65 are the most active health information consumers.

Studies suggest that race and ethnicity have some association with communication processes, perhaps because of the ways that race can act as a marker or proxy for cultural factors. The literature review conducted for this study (see Chapter 3) found that few studies explicitly assessed the significance of race, ethnicity, or culture on participants’ interaction with and response to technologies. A few studies did recruit participants on the basis of racial and ethnic characteristics, but they did not explore the significance of cultural influences.

Race and ethnicity are highly significant variables for health status, if only because of the impact of discrimination on health disparities. However, there is often more variation within traditional demographic categories than between them. Moreover, the IOM Committee on Communication for Behavior Change in the 21st Century cautions that the use of overly broad or rigid demographic characteristics can actually exacerbate inequities by reinforcing inaccurate assumptions and stereotypes. This Committee calls for a focus on “more meaningful ways of describing heterogeneity,” focused on cultural processes, life experience, sociocultural environment, economic contexts, community resources, and beliefs (IOM, 2002).

From a communication perspective, people attribute meaning and make sense of the messages, interactions, situations, and media around them; and they interact with and shape both the tools and the environments in which they live. Interactive media, including e-health tools, make these processes more obvious because they provide new opportunities to act as engaged users instead of passive receivers of information, “link(ing), think(ing) and interact(ing)” with information and other users (Cole, 2004). Individuals become involved in shaping an environment of highly personalized and private engagement with the Internet, Web sites, and interactive components.

Some researchers conceptualize the Internet as a “hybrid” medium with features of mass and interpersonal communication (Cassell, Jackson, and Cheuvront, 1998). Some of the many communication factors relevant to the analysis of e-health tools are patterns of media or technology use, values, beliefs, intentions, expectations, preferences, perspectives, capacities, and access to information and technology (Neuhauser and Kreps, 2003). The characteristics of technology are important in terms of its fit with, value for, and usability by different user groups (Badre, 2002; Nielsen, 1999; Norman, 2002).

The lack of research on psychosocial variables other than health information-seeking as well as the lack of multivariate analyses of demographic and communication factors are major gaps in the literature (Lieberman et al., 2004). A few studies have examined the motivations or level of interest of potential or actual users of e-health tools—typically health information Web sites, online communities, or provider-patient e-mails. It is easier to know who, in demographic terms, is or is not using computers and the Internet than it is to know how individuals think about what they do online and how the interaction reinforces or changes their attitudes, beliefs, values, and preferences.

Despite the paucity of research, however, some things are known about factors that influence health communication processes and audiences’ interactions with media. The most influential characteristics that have some evidence of their relevance are discussed briefly below.

Back to Top

Language Spoken

The relevance of language spoken to the use of e-health tools cannot be overstated. If individuals or groups use one language and the tool is based on a different language, users are very unlikely to make sense of the tool and the content. English-language materials dominate the Internet, which limits the utility of the content for those who read little or no English (The Children’s Partnership, 2000).

Approximately 19 percent of the population speaks a language other than English, according to 2004 Census Bureau data (U.S. Census Bureau, 2004). The majority of persons in this category are Spanish speakers (62 percent); Chinese is a distant second. Data from the Census and the U.S. Department of Education suggest that the majority of persons who speak a language other than English at home consider themselves able to function “very well” in English (Greenberg et al., 2001; U.S. Census Bureau, 2000). Overall, the Census Bureau reports that 92 percent of the population over the age of 5 years report that they do not have difficulty functioning in English (U.S. Census Bureau, 2000). Census data indicate that approximately 4 percent of the population is “linguistically isolated” (U.S. Census Bureau, 2000). Despite this picture of English-language functioning, these data do not speak to issues of language preferences of different groups, the significance of language as an element of culture, or the role of language in perceptions of health and illness.

“Linguistic appropriateness” may seem straightforward, but it is not. Fulfilling the proviso that communication should be in the primary language of the target audience is not simple for large and diverse population groups, given the number of versions of a given language. For example, Spanish speakers present an interesting example of the complexities of linguistic appropriateness. This population segment is both culturally and linguistically diverse, coming primarily from multiple countries in Latin America and the Caribbean and with distinct cultural origins related primarily to Africa, indigenous America, and Europe. Despite the cultural relevance of slang, dialect, and vocabulary, there is often an imperative to identify a “common” Spanish that will function cross-culturally (Schroeder, Trowbridge, and Price, 2002). One of the few general studies of factors relevant for Hispanic groups’ use of the Internet found that Hispanics encounter many barriers when trying to locate Spanish-language health information online (Schroeder et al., 2002).

At the same time, market research reports on Hispanics’ Internet use indicate that they are going online faster than any other segment and are finding content of interest in the categories of communication (e.g., instant messages), entertainment (particularly music), and product information (Hispanic Market Weekly, 2006). When they perceive the relevance of the content, Hispanics are willing to go online to “compare prices, see features, learn about benefits, and then decide on a brand or purchase,” according to the publisher of AOL Latino (cited in Hispanic Market Weekly, 2006).

Small-scale studies of the health information needs and preferences of Asian Americans, Native Hawaiians, Pacific Islanders, and Native Americans suggest that lack of content in the first languages of ethnic groups and inexperience with Internet resources are major barriers to greater use (Hsu, 2003a, 2003b). However, these factors have yet to be analyzed in terms of their contribution to overall lower rates of Internet usage and demand for e-health tools. For example, in a national survey of unpaid caregivers, only 5 percent reported that “finding non-English educational materials” was an unmet need (National Alliance for Caregiving and AARP, 2004).

In the scan of e-health tools conducted for this report (see Appendix 1), language and literacy emerged as two critical considerations in the design of successful tools. Even if developers did not report using any other methods to account for audience variations, they did mention creating understandable materials as design and content priorities. Designing for a stated reading grade level seemed to be the most popular strategy to make content more understandable. Providing content in Spanish was the most popular alternative to English.

Both these strategies have their own problems and raise a number of issues concerning the utility and comprehensibility of content. Even when content developers attempt translation, the quality of translations and the readability of materials can present problems. For example, translations can be of poor quality and reproduce problems, such as jargon and unfamiliar terms, that were features of the original text. Texts that meet a stated reading grade level can still make it difficult for users to understand the core meaning. Applying a health literacy approach that engages intended users in the development of the content from the beginning and focuses on assessing usability and understanding seems the most promising mechanism to address issues of language and literacy.

Back to Top

Socioeconomic Position

IOM proposes that the most important forms of diversity to pay attention to in health communication are those associated with “substantial disparities in health status and outcomes” that also represent differences in “health behavior and its antecedents” (IOM, 2002, p. 7). Individually and collectively, the components of socioeconomic position—including income, employment status, wealth, education, housing, and neighborhood environment—influence health, health behavior, and factors involved in health communication. IOM’s Promoting Health report discusses the relationships among these factors (2000). Communication theory from the 1970s proposed the existence of a “knowledge gap,” which represents the divide between higher socioeconomic persons who pay closer attention to and have greater access to information than lower socioeconomic persons (Tichenor, Donohue, and Olien, 1970). In the e-health arena, socioeconomic factors are major determinants of the elements of meaningful access, as discussed above.

Preliminary analysis of national data from the Health Information National Trends Survey, conducted by the National Cancer Institute (NCI), suggests that income and education levels, as well as gender and age, strongly influence the amount of attention people pay to health topics (Hesse, 2003). A study by Tu and Hargraves indicates that level of education is the most important predictor of health information-seeking; 55 percent of people with postgraduate education said they sought health information, compared with only 25 percent of those without a high school diploma (2003). Education level is also strongly associated with literacy skills, which are a component of health literacy. The relationship between education and literacy likely goes both ways: those who stay in school longer likely have stronger literacy skills, and those with stronger skills likely stay in school longer. This relationship indicates that there is much to learn about how both education and literacy affect people’s access to, interest in, and engagement with health information and the pathways for development of communication capacities.

Back to Top

Disabilities

An estimated 54 million Americans—20 percent of the population—have disabilities (HHS, 2000). Disability, by definition, involves the interaction of impairments and environmental barriers; removing or reducing a barrier can reduce a disability. The types of impairments can include visual, hearing, mobility, cognitive, and learning disabilities. Each type of impairment corresponds to a set of accommodations needed to reach a particular audience segment with effective e-health resources. Disabilities affect people of all ages, but the proportion of the population affected increases with age; therefore, because the U.S. population is aging, the proportion of Americans with disabilities is growing (HHS, 2005b). There are many crossovers between the topics discussed in this section and those on the characteristics and communication needs of older adults and family caregivers, described below. Although people with disabilities are not necessarily in poor health, they are at increased risk of secondary conditions and may have less access to health services and medical care. Health promotion to improve functioning and reduce the incidence of secondary conditions has been shown to be effective (HHS, 2000).

A report by the Pew Internet & American Life Project includes a “special analysis” on Americans with disabilities (Lenhart et al., 2003). The research shows that 38 percent of Americans with disabilities use the Internet, compared to 58 percent of the entire population. Users with disabilities are more likely than the general population to have access only at home (58 percent versus 44 percent, respectively) as well as more likely to look for medical information online (75 percent versus 59 percent, respectively). The Pew research also yielded insights into the reasons persons with disabilities give for not going online—some of which, such as misconceptions about the Internet, are amenable to solution (Lenhart et al., 2003).

For people with disabilities, digital divide issues apply not only to Internet access but also to a broad set of assistive and adaptive technologies that increase accessibility of all kinds. Some of these technologies, which have been likened to “electronic curb cuts,” enable access to the Internet and other digital resources for people with disabilities. Physical barriers to Internet use—or, alternatively, accommodations—can exist at many points, including the public access computing site, the computer terminal, the Web site, the Internet service provider, the browser, and the Web-based platform. Designing for persons with impairments was rare in the 40 e-health tools reviewed for this report (see Appendix 1). Only one makes specific accommodations for people with hearing or visual impairments.

Once physical access to computers and the Internet is achieved, the next set of issues relates to the design, content, and delivery of digital information resources. Paradoxically, although the Internet can reduce the isolation that can come with disability, it also presents its own barriers that must be overcome before it can be useful. The specific barrier, and thus the solution, varies with the impairment, and a detailed review of the often quite technical ways to achieve accessible Web design is beyond the scope of this brief overview. The creator of cascading style sheets, one such mechanism, points out that Web-based information involves the interaction of “content and presentation,” and these have to be addressed separately in order to successfully communicate with people with visual and hearing disabilities (Bartlett, 2002).

The types of accommodations in content and presentation for people with disabilities can be beneficial to other e-health audience segments as well, such as seniors and people with limited literacy or English proficiency. The accommodations include multimedia presentation, breaking text into small chunks, and allowing users to control font size and other visual attributes. Techniques such as these, together with general principles of user-centered design and usability testing (described below), can result in e-health resources that are beneficial to all people, including those with disabilities.

The problem of inadequate research to guide design and content decisions figures in this context as it does elsewhere. Apart from the few references noted above, the present study found no empirical research on health communication issues for people with disabilities. This finding was confirmed by staff members of the National Center on Birth Defects and Developmental Disabilities, Centers for Disease Control and Prevention, who conducted an unsuccessful literature search on health communication and disability in preparation for a health promotion campaign for women with disabilities (J. Thierry, personal communication, October 2004).

Developers can draw on a combination of laws, guidelines, and evaluation tools in achieving and measuring accessibility. Federal law on accessibility is in Section 508 of the 1973 Rehabilitation Act (revised based on the Americans With Disabilities Act), which requires that Federal agencies’ electronic and information technology be accessible to people with disabilities. An article in the Journal of Medical Internet Research reported on research that evaluated 108 Web sites for consumer health information according to disability accessibility guidelines; the researchers found that Government and educational sites are the most accessible, presumably at least partly because of Section 508 requirements for Government sites (www.section508.gov/). No site met all the criteria, however (Zeng and Parmento, 2004). Although the requirements only apply to Federal sites, some private Web developers choose to comply as well. (See Chiang and Starren, 2004, for another published evaluation of Web access for people with disabilities).

The World Wide Web Consortium (W3C) Web Accessibility Initiative has developed its own Web Content Accessibility Guidelines (WCAG) for determining Web page accessibility (www.w3.org/WAI/). The Web site of the International Center for Disability Resources on the Internet leads to a long chain of useful resources (www.icdri.org/prodserv.htm). The same is true of “Bobby,” a Windows-based tool that provides a free service to analyze Web pages for their accessibility to people with disabilities, to identify and repair barriers to accessibility, and to facilitate compliance with accessibility guidelines such as Section 508 and W3C’s WCAG (http://webxact.watchfire.com/). One expert reports that current Web accessibility guidelines do not address cognitive disabilities very well, as most of the focus to date has been on visual and sensory disabilities (R. Appleyard, personal communication, October 2004, citing Wehmeyer, 1998, 1999).

Back to Top

Age, Developmental, and Role Issues

As noted above, age is one of the most important factors affecting health status, information-seeking, media use, and Internet behaviors. Yet little attention has been paid to life course, roles (apart from parenting), and experiential variables that are often associated with age. Each phase of life has its own developmental perspective, obstacles and facilitating factors, and unique experiences that influence interests and capacities related to health communication. For example, unpaid caregiving by adults for adults is emerging as a critical policy issue as well as an experiential factor for millions of Americans. A survey by the National Alliance for Caregiving and AARP estimates that approximately 44 million adults provide unpaid care to other adults (National Alliance for Caregiving and AARP, 2004). The survey finds that “the typical caregiver is a 46-year-old woman who has at least some college experience and provides more than 20 hours of care each week to her mother.” Approximately one-third of caregivers rely on the Internet for information to help them cope with their caregiving (National Alliance for Caregiving and AARP, 2004).

Internet use is inversely associated with age. Only 22 percent of people older than age 65 have been online (Fox, 2004), compared with 96 percent of children and adolescents age 8 to 18 (Rideout, Roberts, and Foehr, 2005). The higher percentage of young people online is to a great extent due to school-based access, whereas home access remains a concern for the large segment of low-income children. Home-based access is also important for older adults, who are more likely to be out of the workforce or homebound. Partly because of young people’s greater exposure to technology, training, and technical assistance opportunities, they show greater comfort and facility with technology than older adults. (Indeed, some programs involve them as trainers, as seen in Chapter 5.) Older adults are more likely than persons in other age groups to have physical or cognitive impairments that further limit their ability to use computers and navigate the Internet (Morrell, Dailey, Feldman, et al., 2003; SPRY Foundation, n.d.).

However, both groups have shown considerable interest in health topics. Older adults use their Web access for health purposes more intensively than other age groups (Fox, 2004); and 68 percent of 15- to 24-year-olds and 50 percent of all 8- to 18-year-olds who have been online have used the Internet to get health information (Kaiser Family Foundation, 2001; Rideout et al., 2005).

One study is suggestive about the relationships among age, experience with both health and technology, and use of e-health tools. It examined participation and nonparticipation rates by primary care patients with type 2 diabetes in an Internet-based diabetes self-management support program (Feil, Glasgow, Boles, et al., 2000). The researchers found no significant differences in gender, insulin use, computer familiarity, or computer ownership. The significant differences between participants and nonparticipants were related to age and years since diagnosis; younger patients with more recent diagnoses were more likely to participate.

A relatively recent development of special relevance for older adults, including the significant percentage who are caregivers, is the growing use of disease management tools by healthcare organizations. Older adults have the largest incidence of costly chronic illnesses, and major institutions such as the Centers for Medicare & Medicaid Services (CMS) and the U.S. Department of Veterans Affairs (VA) are investing in the development of e-health tools to help patients manage their diseases. These programs provide training and sometimes the necessary equipment. If this trend continues, at least a small segment of older adults may be induced to become users of electronic communication and information for personal health management. In addition, the Web portal being developed for Medicare beneficiaries introduces them to an e-health tool that contains content of direct relevance.

Although the specifics vary considerably, both older and young age groups have style preferences, technology use characteristics, and health content interests that are often not served by standard e-health tool content, design, and architecture and that are best accommodated through targeted tools. The top priorities for meeting the needs of older and younger users include simplicity of design and content and the use of multimedia presentations. One example of applying good design practices and research-based knowledge of intended users is the Web site for older adults sponsored by the National Institutes of Health (www.nihseniorhealth.gov). The site is designed to accommodate limited literacy levels, cognitive and physical impairments, and different modes of learning (e.g., textual, visual, auditory). The Web site’s approach closely matches the general principles of good Web design for all users promulgated by the Federal Government (see www.usability.gov and www.firstgov.gov/webcontent). Many e-health tools designed for young people have behavior change and prevention purposes; here the challenge is to make them interesting and attractive.

Back to Top

Interest in Health Information

Health information-seeking attitudes and behaviors, as well as attitudes and behaviors toward health care and healthcare providers, have been identified as a useful basis for segmentation with respect to e-health communication. Researchers and expert observers classify people in terms of their degree of independence and initiative in relation to health care and health information-seeking. For example, research by the communication firm Porter Novelli found that the public can be segmented into five health information types, based on two broad sets of characteristics—degree of reliance on physicians for health information and level of activity in seeking out such information (cited in Lieberman et al., 2004).

• The Uninvolved (14 percent) are likely to describe their health as good or fair; value health less than others do; expend less energy on prevention; and exhibit low interest in health information.

• Doctor-Dependent Passives (20 percent) describe their health as excellent or very good; hold lower values for health and prevention; and express low interest in health information.

• Moderates (28 percent) are generally healthy adults; value good health and actively try to prevent disease; and value health information, but do not enjoy searching for it and may lack skills to do so.

• Doctor-Dependent Actives (20 percent) value health and prevention, but experience more health problems; and actively seek health information and are capable of finding it, but may have difficulty interpreting it.

• Independent Actives (19 percent) are in very good health; highly value health and prevention; place the highest importance on health information; and are very skilled at finding and understanding health information.

Long-time online health activist and analyst Dr. Tom Ferguson proposes a new vocabulary to capture the shift in individuals’ orientation to information and their health. Instead of “consumers” or “patients,” he sometimes speaks of “medical end users,” “e-patients,” and “prosumers,” the last term coined by Alvin Toffler in The Third Wave to capture the blurring of the distinction between service providers and recipients (Ferguson, n.d.). Similar to the Porter Novelli categories, Ferguson divides patients and consumers into three groups—passive patients, concerned consumers, and health-active prosumers—and he predicts an increasing shift into the third group. In addition to information, he stresses the importance of communication among consumers, such as in online and face-to-face support groups.

Dr. Judith Hibbard has developed a multifaceted typology to assess levels of “health activation” in patients and consumers (Hibbard, 2003).² Her work primarily concerns health behaviors, but it is highly relevant to health information-seeking and use. Hibbard’s “activation measure” assesses patients along two axes, one listing actions the individual can take related to personal health and the other listing the capacities to be assessed with respect to those actions (Table 3).

Back to Top

Table 3. Domains for Measuring Activation Measure

Back to Top

Hibbard states that consumers with higher activation are more likely to take such actions as read about possible complications when taking a new medication, seek out health information, visit a health Web site, and know about treatment guidelines for their condition. The relevance of her work for the present report is summarized in two questions she poses:

• What kinds of strategies will be most effective in increasing activation?

• How can we take advantage of knowing a patient’s activation level to tailor an intervention?

Attitudes About Privacy and the Protection of Personal Health Information

Since the initial framing of this project and drafting of the report, the issues of protecting personal privacy and ensuring the confidentiality of personal health information have moved to the top of the agenda in any discussion of consumer e-health tools, particularly personal health records. Numerous documents assert that there must be strong privacy protections for e-health tools that collect and store personal health information; the need for strong protections has been particularly noted in relation to personal health records (Markle Foundation, 2005; NCVHS, 2005a). Several national surveys have been conducted to gauge public understanding of privacy issues and the public’s expectations about privacy protections in an e-health environment (California HealthCare Foundation, 2005; Markle Foundation, 2005; Westin, 2005). The findings are consistent that a majority expect strong privacy protections, whether through policies, laws, or technologies.

The findings of two surveys suggest, however, that as in most other areas, segments of the public can be distinguished on the basis of their attitudes toward privacy, and likely by their privacy-protecting behaviors as well (California HealthCare Foundation, 2005; Westin, 2005). As with other factors discussed in this chapter, attitudes about health information privacy and e-health tools have not been well studied. It is possible to infer from user behavior in online communities, however, that participants do not perceive all disclosures of personal information as equal. Participants often post highly personal and identifiable information in online chats and blogs; yet a disclosure of the same or similar information as a result of a security breach of a digital record system would likely be treated as a privacy violation.

In numerous hearings on personal health records, the National Committee on Vital and Health Statistics consistently heard testimony that the key factor for consumers is their ability to control their own information and records and protect their privacy (NCVHS, 2005a, 2005b). In light of the preceding discussion on the diversity in information-seeking behaviors and activation toward health, the need for control and sensitivity to disclosures also should be treated as having a range of values rather than dichotomous values of either total or no control and sensitivity to disclosure of personal information.

Back to Top

Designing for Diverse User Groups

Given the number of factors that must be considered when designing tools to meet the needs of diverse users, it is clear that a focused effort by developers is required. Engaging persons with low income or education, different ethnic groups, and adults with limited literacy skills in health communication requires sophisticated audience segmentation techniques that involve intended users of the information in interactive roles (Freimuth and Mettger, 1990). Targeting (audience segmentation) and tailoring on communication factors are considered promising strategies for user-centric design in the electronic environment (IOM, 2002). Both are employed to engage users by personalizing and individualizing information based on demographic, behavioral, motivational, psychosocial, or physical characteristics (Brug, Oenema, and Campbell, 2003).

Targeting or audience segmentation is selecting groups of users based on common characteristics related to behavior, health status, or some other common factor. The process of targeting generally happens in the following sequence. First, a target audience or market is identified, related to a healthcare or public health need or a business opportunity. Then, the audience is analyzed, and if necessary, segmented, to optimize service and impact. In some cases, specialized products and services are developed for existing audience segments or new target audiences. Some tools integrate tailoring capabilities that make it possible to accommodate individual differences. This sometimes involves “cultural tailoring,” or tailoring to enhance the impact for individuals in targeted audience groups (IOM, 2002).

Tailoring is designed to simulate personal counseling in that the individual is surveyed and the responses are used to generate individualized information and feedback (Brug et al., 2003; IOM 2002).³ Tailored information has been shown to be more satisfying, read more deeply, seen as more personally relevant, and more often discussed with others (Brug et al., 2003). “First-generation” tailoring involves using a computer program to generate the individualized feedback that is presented to the user in a print-based format, such as a letter or newsletter. “Second-generation” tailoring takes advantage of the computer’s ability to immediately deliver tailored information and eliminates the lag time incurred while waiting for printed, tailored information to be presented (Oenema, Brug, and Lechner, et al., 2001).

Dr. Victor Strecher and Dr. Kevin Wildenhaus at the University of Michigan are leading practitioners of computer-based tailoring in health communication. They prefer tailoring over targeting to enhance the effectiveness of health communication messages. When asked to identify the intended user groups or populations served by the e-tools his lab develops, Dr. Strecher stated, “Targeted messages miss the important variation in behavioral predictors that are often found within demographic or even psychographic groups. Tailoring identifies these predictors at an individual level and addresses them.” He further stated, “Our most recent research suggests that deeply tailored materials seem to help the people who need them the most—those with low perceived capabilities in solving problems on their own. Tailoring may particularly help these individuals by providing a very individualized plan and by conveying information in a more vivid manner” (V. Strecher, personal communication, March 16, 2006). The National Cancer Institute (NCI) has funded Dr. Strecher’s lab to work on identifying the “active ingredients” that make computer-based tailoring successful.

Enhancing the usability of Web sites is another strategy to make e-health tools more fully accessible to all users (Koyani, Bailey, and Nall, 2003). In the Government context, the HHS Web team and NCI have played a leading role in developing and implementing a usability approach to improve the navigation of Web sites (http://usability.gov). Usability testing can be used on its own or as part of a broader approach known as user-centered design.

User-centered design is an iterative process that assesses tools throughout the design life cycle in terms of users’ preferences and performance. The process includes task and user analysis and participatory methods, such as focus groups and surveys, to determine the interests and capacities of prospective users. Later, usability testing determines how well users are able to use a given tool, with the goal of uncovering problems that can be fixed prior to launch. The Think Aloud protocol is a method in which users describe their thought processes as they make their way through a Web site. Other methods include contextual inquiry (observation and testing), interviews, journals, various forms of inspection, and performance measurement.

The major criteria are users’ success in finding information, including accuracy and speed; related criteria are likability, learning, and retention. For example, in one small study, adults with low literacy were able to learn Web navigation skills easily and use interactive features such as active graphics and pull-down menus when the instructions were simple, direct, and noticeable (Zarcadoolas et al., 2002).

In an effort to identify the types of user-centric strategies currently in use by e-health developers in the field, project staff interviewed 54 developers and other experts about 40 e-health tools designed wholly or partly for diverse users (see Appendix 1). Each of the tools proved to be distinctive in the way it combines functions and features to serve intended users. The analysis of this set of tools suggests the number of user variables that can be considered and the many ways developers think about enhancing relevance and engagement. These developers report that they often consider literacy levels relevant to the use of e-health tools, although the literature review in the next chapter indicates that few studies have systematically included persons with limited literacy skills, designed tools as health literacy interventions, or assessed health literacy as part of the evaluation of the tool.

The scan of 40 e-health tools indicates that developers employ a variety of strategies to enhance the connection between the tools and their intended users. The main strategy appears to be one of targeting or segmentation. The findings align with the observations made in the IOM report, Speaking of Health: Assessing Health Communication Strategies for Diverse Populations, about the adaptation of health communication for diverse audiences (2002):

• Some tools are developed for narrowly defined audiences (e.g., people over age 65 with chronic obstructive pulmonary disease; binge-drinking college students). Some developers have an array of such specialized tools or modules.

• Some tools are developed for a broad cross-section of users, but adapted to serve different audience segments (e.g., a Spanish-language version, a module for pregnant women, a chat room for caregivers). The broad cross-section may exist because the tool is available to all comers (e.g., through a public Internet site) or because it is distributed to a restricted but diverse constituency (e.g., employees of a distributor, health plan enrollees).

• Some tools are developed for a broad (and presumably heterogeneous) user group in a way that focuses on what all users have in common.

Often, tools are designed for large population segments based on public health priorities, such as kids with diabetes or adult smokers who are trying to quit. Several developers mentioned the economic impracticality of designing highly segmented or individually tailored tools. Many tools, such as public Web sites, serve anyone who finds the site on the Internet. Others may serve anyone in a more restricted but still heterogeneous group, such as members of a particular health plan or employees of a large organization. Targeting is often based on one or two dominant factors, such as shared health issues, gender, or age. Health condition, risk behavior, and age were the most popular factors for identifying intended users of e-health tools. Some developers stated that the most important characteristic in targeting was the shared health issue, such as people with cancer and their caregivers, rather than demographic factors. The implication is that shared health experience is the basis for coming together via technology. For the majority of the 40 tools, medical conditions (e.g., diabetes) or health-risk behaviors (e.g., smoking) define the audience.

In all, 19 of the 40 tools in the scan were described as having one or more special features for one or more diverse groups. Most consider multiple audience characteristics. The bases for audience segmentation among the tools (listed in order of frequency) are age, language, race/ethnicity, gender, income, geographic location, and disability or sensory impairment. The segments targeted by these tools include:

• Hispanics/Latinos

• Other non-English speakers

• African Americans

• Recent immigrants (e.g., Vietnamese, Caribbean)

• Women

• Teenagers

• Young children

• Elders

• People with low income

• Rural dwellers

• Inner-city dwellers

Added to these variations, several e-health tools have versions for intermediaries or adjunct users such as childcare providers, teachers, parents, school friends, and public health workers. The large group of healthcare tools (i.e., tools made available by healthcare providers or organizations for use by their consumers/patients) are also used by staff members of the healthcare organization, such as nurses, administrative staff, and personal physicians, and these are distinct user groups from the perspective of tool development and evaluation.

The interviews offer examples of developers who adapted a single basic program with multiple subprograms based on factors such as gender, age, or severity of disease. One company has 22 versions of its basic program. This finding suggests that the often-discussed potential of technology to create customized versions of generic interventions is starting to be realized in the marketplace through a variety of approaches.

Back to Top

Summary

This chapter identifies several concepts, factors, and strategies that can be used to design e-health tools for diverse users. The concepts of health literacy and meaningful access highlight the importance of ensuring physical access to information and technology and designing useful, understandable content. The IOM has already called for greater attention to communication factors in the design of health information, messages, and e-health tools. This chapter elaborates on many of the critical factors for user-centric design. If the vision of e-health tools for all is to be realized, these factors, along with others that have yet to be fully articulated, will require further research and integration into tool design and development. A scan of the current field of e-health tools indicates that developers are beginning to address issues of diversity, but do not yet have strategies and approaches that go much beyond traditional public health targeting based on demographic characteristics. Developers will need to engage consumers more fully in the research and design process and probe those factors that shape attitudes, beliefs, values, expectations, and experiences in relation to health and technology.

²This discussion is based on several of Dr. Hibbard’s articles and on her slides, “Measuring and Improving Patient Activation,” for a presentation to a September 2003 conference of the Center for Information Therapy. www.informationtherapy.org/conf_mat03/final_pres/Hibbard.pdf.

³For a book-length treatment on computer-based tailoring, see Kreuter, Farrell, Olevitch, et al., 2000.