A substantial number of Americans have physical or other conditions that reduce their ability to travel. Such conditions hamper their ability to see, operate a vehicle, gain access to public transportation, or walk to a desired destination. Many of these individuals, therefore, rely upon family members, friends, or other modes of transportation for their travel needs [¹].

Rural residents have a slightly higher rate of disabling conditions than urban residents, particularly in the South [²]. These rural residents are especially vulnerable in regard to travel restrictions. The reduced availability of services, and relatively greater distance between services and housing centers, and the reduced availability of public transportation exacerbate these resident’s travel difficulties [³]. These barriers in available transportation can lead to reduced utilization of services [¹].

The evolution of the Internet as a resource, especially for health care information and services, may be an ameliorant for those with travel difficulties. Many patients rely on the Internet for gathering information about their conditions and treatment options, and for communication with their providers. Patients also use the Internet to garner social support, using the interface as a coping mechanism [^4-6]. The Internet also can play an important role in the education and recruitment of patients for specific services or programs [⁷,⁸].

Internet access is influenced by available telecommunication infrastructure and the affordability of Internet services [⁹]. The high cost of providing services across the more widely dispersed rural population is one barrier to the development of infrastructure in rural areas [¹⁰]. As a result, rural areas lag behind in the infrastructure required for optimal Internet use (such as broadband or other high-speed service), and rural residents have lower reported use of the Internet than urban residents [¹¹]. Since home availability of the Internet remains low in rural communities, and usage at work was also lower [¹²], rural residents were more likely than those in urban or suburban areas to use a source other than work or home for accessing the Internet [¹³].

Sociodemographic characteristics are also significantly associated with Internet use. African Americans and Hispanics were less likely than whites to report Internet access, and Hispanics were less likely than whites to report using the Internet for health-related issues [¹⁴]. Other socioeconomic characteristics, such as higher educational levels, younger age, and greater household income, were found to be associated with any prior use of the Internet among surgery patients [⁷,¹³]. A Pew Internet surveys found that Internet users who were female, were older, had a higher education and income, were white, were not employed full time, were married, and had a child under 18 living at home were more likely to report using the Internet to search for health information [¹⁵].

The digital divide between urban and rural populations has important implications for the health of rural residents, particularly those who are limited in their ability to travel. These individuals, as well as rural populations, generally have reduced access to primary care, coupled with greater travel distances to care [¹⁶,¹⁷]. They could benefit from Internet access, as Internet availability could facilitate research into health conditions, as well as providing additional links to services. The purpose of this study, therefore, was to examine Internet use among people with limited ability to travel. We also examined disparities in Internet use by race/ethnicity and rural residence, particularly among those with medical conditions.

In 2001, about two-thirds of Americans reported having accessed the Internet within the past 6 months (Table 1). Rural residents were less likely than their urban peers to report accessing the Internet (59.7 versus 69.4%, P < .001). Only about a third of persons who reported a medical condition that impaired their driving (32.6%) reported accessing the Internet, compared to 70.3% among those without a medically limiting condition (P < .001). A marked difference was also present across race/ethnicity. Nationally, 72.8% of white respondents, versus 65.7% of persons of “other” race, 51.5% of African Americans, and 38.0% of Hispanics, reported accessing the Internet (P < .001). Less than a third of rural African American or Hispanic respondents reported accessing the Internet compared to 64.5% of rural whites (P < .001, data not in table).

As might be expected, the likelihood of accessing the Internet increased linearly with education and income, and decreased with age (P < .001). Occupational differences may reflect job requirements; individuals in manufacturing and related industries were markedly less likely to report accessing the Internet than were those in other occupations (P < .001).

Among persons who did report accessing the Internet, the majority used it daily (54.2%; Table 2). Among persons with medical conditions, more than two-thirds (68.3%) reported accessing the Internet only from home, versus 38.7% of other individuals (P < .001). Rural residents were less likely to report daily use (47.0% vs 55.7%), and more likely to report use only once per month (13.3% vs 9.4%), than their urban peers (P < .001). Frequency of Internet use differed by race/ethnicity as well (P < .001): African Americans and Hispanics were less likely to report almost daily Internet use, and were more likely to report use only once per month. African Americans were less likely to have access at home (34.1%) than either Whites (40.8%) or Hispanics (40.3%), but were more likely to report use at work (9.9%, P < .001).

Adjusted odds for accessing the Internet and factors associated with intensity and location of use among persons who reported Internet access are presented in Table 3. With all personal and ecological characteristics held equal, rural residents were no less likely than urban residents to report accessing the Internet (odds ratio [OR] 0.89, 95% CI 0.76-1.04), and did not differ with regard to frequency or location of use. Among persons with a medical condition that limited travel, the odds of accessing the Internet were lower, even controlling for age and life cycle stage (OR 0.66, 95% CI 0.59-0.74). Medically impaired persons who did access the Internet were most likely to use it at home (OR 1.70, 95% CI 1.43-2.03).

The digital divide between the races in 2001 was extensive (Table 3). All minorities were less likely than whites to report any Internet access (OR 0.38, 95% CI 0.33-0.43 for African American; OR 0.20, 95% CI 0.17-0.24 for Hispanic; OR 0.51, 95% CI 0.45-0.58 for other). For African Americans, the odds of any use, of frequent versus infrequent use, and of use at home versus at other locations were all lower than for whites. Hispanics were similarly less likely to report any use and to report frequent use, although they did not differ in location of use from white respondents.

Other characteristics influenced accessing the Internet and type of use in a manner paralleling the findings shown in Table 1. In adjusted analysis, the odds of reporting any Internet access increased as education or income increased, and decreased as age increased. Women were less likely to report any Internet use and frequent use, with women who did use the Internet being more likely to access it at home than in other locations. Among persons using the Internet, lower income and education were associated with use at home versus other locations.

The present analysis sought to investigate differences in Internet access and use among persons affected by medical conditions, among rural residents, and across racial/ethnic groups. Persons with a medical condition that limits their availability to travel were deemed to have a particular need for Internet access, to allow them to obtain information and social support [^4-8]. We found, however, that Internet use and frequency were lower among persons with a medical condition than those without; persons with a medical condition were more likely to access the Internet only from home. Lower odds for any Internet use within the past 6 months and Internet use exclusively from home persisted in adjusted analysis. Other factors not captured by the present analysis, including personal preferences and/or the inability to use computers due to the person’s limitations, may account for this particular type of digital divide.

Further research is needed to explore barriers to Internet use among persons whose travel is limited by medical conditions. Such research must take into consideration that Internet access alone does not always translate into its use for health information and support. Previous research suggests that the proportion of patients with Internet access who use the Internet for health information ranges from 89% among bariatric surgery patients to less than 50% among primary or tertiary care settings [⁴,^19-22]. Thus, efforts should continue not only to improve Internet access among persons with medical conditions, but also to encourage their use of health-related information resources.

The unadjusted findings of the present study supported previous evidence of a geographic digital divide, as rural residents were less likely to use the Internet than their urban counterparts. Adjusted analysis, however, suggested that the characteristics of rural populations, rather than lower technology penetration in rural areas [¹¹], accounts for the differences. With demographic and ecological conditions held constant, rural residents did not differ from their urban peers. In particular, the job-related factors included in the model (job density and occupational type) may explain the rural-urban differences found in the unadjusted analysis. Rural residents were more likely to be in low job density areas and to work in nonwhite-collar occupations, both of which were associated with a reduced likelihood of Internet use [¹²,¹³]. This is further supported by rural residents’ report of higher Internet use at home only (44.7%) than urban residents (38.8%).

Our study also confirmed previous work suggesting lower Internet use among African American and Hispanic populations [¹⁴]. In both unadjusted and adjusted analyses, all minority groups were less likely than whites to report Internet access within the past 6 months. Disparities in frequent use and use at home persisted among African American and Hispanic respondents even after statistical adjustment for income, education, occupation, and other demographic characteristics. Further research is needed to determine whether these disparities, measured in 2001-2002, persist 8 years later. Should this be the case, additional research will need to explore whether minority populations perceive Internet access to be of lesser utility than do white populations, or experience other cultural barriers to use.

Our study has several limitations. First, the NHTS was not designed for health research; thus, using it to define medically limited individuals may lead to overestimation of those who may be clinically disabled. In addition, all data are based on respondent self-report, which may bias findings in an unknown direction. On the other hand, the NHTS was the only source for information on both travel limitations and Internet use from a random sample of the US population. A second limitation is that the NHTS defines rural differently from many traditional geographic analyses; however, the use of deciles closely mirrors alternative measures, providing a suitable proxy for rurality. The age of the data (2001-2002) may reduce the generalizability of the findings given the rate of technological advancement; future analyses will use newer data as it comes available. Finally, this survey did not inquire about what types of information the user was seeking while accessing the Internet. It would be helpful to know, for example, whether those who have a medical condition that limits travel are seeking health information on the Internet at a rate that differs from those who are not limited.

Despite the limitations, the findings of the present analysis remain important and relevant: the digital divide persists for several vulnerable populations. While it is posited that Internet access can make health expertise broadly available, persons with medical conditions that limit travel, who might benefit from such access, were less likely to use the Internet than their peers. African American and Hispanics also were affected by the digital divide. For rural residents, multivariate analysis suggests that personal characteristics, rather than geography, limit Internet access and use.

None declared

Dr. Wang performed the analysis, wrote the results, and materially contributed to the literature review and the discussion. Dr. Bennett contributed to the literature review and discussion sections, aided in editing of the manuscript, and prepared the manuscript for submission. Dr. Probst aided in the analysis and wrote the discussion.

This research was supported in part by Grant No. 6 U1C RH 00045-01 from the Federal Office of Rural Health Policy, Health Resources and Services Administration, and a grant from China Medical University, Taiwan (Grant No. CMU98-N2-18). We thank Dr. Andrew Johnson at the Department of Health Services Management, University of Kentucky, for his contributions to this study.