The Kaiser Permanente Medical Care Program in Northern California (KPNC) provides primary and specialty health care to a sociodemographically diverse membership that includes over 2.8 million adults who mostly reside in the San Francisco Bay Area, Sacramento area, Silicon Valley, and Central Valley. The KPNC adult membership is very similar to the insured population of Northern California with regard to sociodemographic and health characteristics [54]. For several years, the health plan has had a comprehensive website accessible to members and the general public, which provides information and advice about health conditions, medical procedures, medications and dietary supplements, and health and lifestyle risks and behavior change using online text, video, and podcasts, as well as online health behavior change programs available to members who register to use the patient portal.

Data for this study come from middle-aged and older adults who participated in the 2014 to 2015 cycle of the KPNC Member Health Survey (MHS). The MHS is a self-administered (mailed print questionnaire and online) survey that has been conducted with independent stratified random samples of English-speaking adults every 3 years since 1993. The survey covers sociodemographic and health-related characteristics, digital technology use, use of the patient portal and different types of health information resources during the previous 12 months, and preferred methods for obtaining information and advice about managing health conditions and making changes in health-related behaviors and lifestyle. Information about the survey is found in an earlier publication [55] and on the survey website [56]. The overall response rate for this age group in the 2014 to 2015 survey cycle was 49.3% (40.9% for those aged 45 to 65 years and 64.5% for those aged 66 to 85 years).

The sample used for these analyses was restricted to the 9005 respondents (4163 aged 45 to 65 years, 2656 aged 66 to 75 years, and 2186 aged 76 to 85 years) who were not missing data on internet use status and who indicated having at least 1 of the following chronic health conditions during the previous year: diabetes, prediabetes, high blood pressure, heart condition, high cholesterol, cancer, Parkinson’s disease, urinary incontinence, chronic obstructive pulmonary disease/chronic bronchitis, asthma, allergies, musculoskeletal pain, osteoarthritis, frequent migraines or other types of headache, chronic pain, frequent insomnia, depression, anxiety, frequent memory problems, or frequent problems with balance or walking. This subsample includes 81.1% of all respondents aged 45-85 years, with 93% of exclusions because of not meeting the chronic health condition criterion.

All analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC 2013) procedures for data from complex survey designs [57] and data weighted to the age, sex, and geographic composition of the KPNC adult membership in 2014. Proc Surveyfreq was used to produce weighted percentages with 95% confidence levels and Proc Surveylogistic was used to test whether differences between age groups and levels of educational attainment in access to digital devices and use of and interest in different HIA modalities were statistically significant after adjusting for race/ethnicity and sex. Analyses of patient-preferred HIA modalities were restricted to individuals who indicated interest in at least one modality in the HIA checklist. All differences between subgroups mentioned in the text are statistically significant at P<.05 or greater; if differences are not mentioned, they did not reach that threshold. Although we did not adjust for multiple comparisons, we have reported results of all statistical comparisons.

Use of MHS data for this study was approved by KPNC’s Institutional Review Board.

The sociodemographic characteristics of the middle-aged (45 to 65 years), younger senior (66 to 75 years), and older senior (76 to 85 years) study groups are shown in Table 1. Slightly over half of all 3 age groups are female. Compared with middle-aged adults, the 2 senior groups are significantly (P<.001) more likely to be non-Hispanic white, and by San Francisco Bay Area standards, more likely (P<.001) to be lower income (HHI≤US $35,000) and less likely to have an HHI>US $80,000, with older seniors being less financially well-off than the younger seniors. Older seniors are significantly (P<.001) more likely than the middle-aged and younger senior groups to have no college (a high school diploma or less education) and less likely to be college graduates (bachelor’s or postgraduate degree).

Table 2 provides data on access to digital technology by age group and level of education. Compared with middle-aged adults, after controlling for sex and race/ethnicity, both senior groups are significantly (P<.001) less likely to have easy access to a computer (desktop or laptop), to own a smartphone, and to be using the internet, email, text messaging, and health apps, and older seniors are significantly (P<.001) less likely than younger seniors to be engaging with these digital technologies.

Table 2 also shows that across all age groups, noncollege graduates are significantly (P<.001) less likely than college graduates to be using these technologies. In addition, the same age group differences are seen at each level of education, with the exception of being able to use apps on a smartphone, which was extremely low prevalence across all groups.

Approximately 62.5% (95% CI 61.2%-63.8%) of adults aged 45 to 75 years and 53.5% (95% CI 50.7%-56.4%) of adults aged 76 to 85 years had in the past 12 months used at least one of the HIA resources asked about in the survey. After controlling for age group, sex, and race/ethnicity, past-year HIA users with no college education were significantly (P<.001) less likely than college graduates to have used an internet-based HIA resource—odds ratio (OR) 0.60 (95% CI 0.48-0.74)—whereas those with some college did not significantly differ from college graduates.

Table 3 provides statistics on use of specific kinds of internet-based HIA resources in the past year by adults in these 3 age groups. Half of the adults aged 45 to 75 years and approximately one-third of those aged 76 to 85 years had obtained HIA from a website. However, no significant age group difference in accessing HIA from websites was observed among internet users. Approximately 10% of adults aged 45 to 75 years and 5% of adults aged 76 to 85 years had used a Web-based health education program, and the age group difference, although smaller, remained statistically significant among internet users. Health app use significantly declined with age (approximately 11% of those aged 45 to 65 years, 5% of those aged 66 to 75 years, and 3% of those aged 76 to 85 years, respectively), with prevalence of use among smartphone users approximately 3% points higher in each age group. Across all age groups, less than 3% of adults had listened to a podcast on the health plan’s website and less than 1% had participated in an online chat room on any website. Figure 1 shows that across all 3 age groups, adults with no college education were significantly (P<.001) less likely than college graduates to have obtained HIA from a website, and in the 2 older groups, those with some college were also significantly less likely than college graduates to have obtained Web-based HIA. However, among internet users, education-related differences in obtaining internet-based HIA were greatly diminished in all age groups (see Multimedia Appendix 1 for results of age group-specific multivariable logistic regression models of past year use of HIA from a website). Use of health plan Web-based health education programs and podcasts did not significantly differ by level of education, but among middle-aged and older adults, those with no college education were significantly (P<.01) less likely than college graduates to have used health apps (ages 45 to 65 years: 7.8%, 95% CI 5.7%-5.9% vs 12.4%, 95% CI 10.6%-14.2%; ages 66 to 75 years: 2.4%, 95% CI 1.1%-3.6% vs 6.3%, 95% CI 4.7%-8.0%).

Table 4 shows the percentages of all adults and internet users who indicated interest in obtaining HIA using specific internet- and noninternet-based modalities, restricted to the 86% of people who expressed interest in using at least one HIA modality asked about in the survey. Although 75% of all middle-aged and older adults expressed interest in using at least one of the internet-based HIA modalities, there were significant differences in interest among middle-aged, younger, and older seniors that persisted when restricted to internet users. Among all adults, after controlling for sex and race/ethnicity, 66 to 75 year olds and 76 to 85 year olds were less likely than 45 to 65 year olds to be interested in watching Web-based videos and webinars, using an interactive Web-based program, listening to podcasts, having a video visit with a patient educator, receiving HIA text messages, receiving emailed health newsletters, and using health apps. Adults aged 76 to 85 years were less likely than 45 to 65 year olds to be interested in obtaining HIA in messages sent through the patient portal, reading about health topics on a website, and receiving emailed health newsletters, but 66 to 75 year olds did not differ significantly from the middle-aged group. No age group difference was observed for print materials or counseling over the phone, but 76 to 85 year olds were less interested in counseling or classes that involved coming into the medical facility.

Figure 2 shows that across all age groups, adults with no college education were significantly (P<.001) less likely than college graduates to be interested in using at least one internet-based HIA modality, and among middle-aged and younger seniors, those with some college were also significantly less likely than college graduates to be interested in using any internet-based HIA modality. In Multimedia Appendix 2, we show that at all age levels, adults with no college education are significantly less likely than college graduates to prefer online HIA modalities.

As 97% of 45 to 65 year olds and 91% of 66 to 75 year olds interested in at least one HIA modality were internet users, prevalence of interest in using internet-based HIA modalities did not significantly differ between internet users and nonusers in those age groups. However, as only 73% of 76 to 85 year olds were internet users, interest in obtaining HIA from websites, patient portal messages, and emailed newsletters was significantly higher among online adults. Across all age groups, prevalence of interest in using health apps was significantly higher among smartphone users than all adults, but prevalence of interest in HIA text messages was not significantly higher among adults who currently use text messaging than all adults.

Figure 3 shows how the preference for obtaining textual HIA and health newsletters varies by age among those interested in textual HIA and health newsletters. Among those interested in textual information, preference for obtaining it only from print materials significantly increased with age (ages 66 to 75 vs 45 to 65: OR 1.44, 95% CI 1.22-1.69; ages 76 to 85 vs 45 to 65: OR 2.81; 95% CI 2.32-3.41 after adjusting for sex and race/ethnicity), whereas preference for obtaining information only from websites significantly declined with age (ages 66 to 75 vs 45 to 65: OR 0.77, 95% CI 0.66-0.89; ages 76 to 85 vs 45 to 65: OR 0.48; 95% CI 0.39-0.57). Similarly, among those who were interested in receiving health newsletters, preference for getting them only by mail significantly increased with age (ages 66 to 75 vs 45 to 65: OR 1.75, 95% CI 1.48-2.07; ages 76 to 85 vs 45 to 65: OR 4.05; 95% CI 3.31-4.95), whereas preference for getting them only by email declined with age (ages 66 to 75 vs 45 to 65: OR 0.60, 95% CI 0.51-0.70; ages 76 to 85 vs 45-65: OR 0.27; 95% CI 0.22-0.33). Compared with college graduates, after adjusting for age, sex, and race/ethnicity, adults with no college education or some college were significantly more likely to want print materials only (OR 2.11, 95% CI 1.72 to 2.58; OR 1.44, 95% CI 1.21-1.71, respectively) and mailed newsletters only (OR 2.83, 95% CI 2.30-3.48; OR 1.84, 95% CI 1.53-2.22, respectively), and significantly less likely to want emailed newsletters only (OR 0.41; 95% CI 0.33-0.51; OR 0.61, 95% CI 0.51-0.72, respectively); preference for internet-based materials only was not significantly associated with education.

Although we have previously reported on disparities in use of digital technologies and interest in using internet-based health information resources by seniors in this health plan population [23], in this study, which used data from a more recent survey, we extend the comparison to middle-aged adults and focus on adults who are managing chronic health conditions. In this study, we showed that there are significant age-group and educational disparities in access to or use of digital technologies used to access internet-based HIA, including computers, smartphones, email, text messaging, and apps. Specifically, younger (aged 66 to 75 years) and older (aged 76 to 85 years) seniors were significantly less likely than middle-aged adults (aged 45 to 65 years) to be using these digital technologies, and older seniors were significantly less likely than younger seniors to be doing so. Within each age group, we showed that use of these digital technologies was significantly lower among adults who have no formal education beyond high school or some college as compared with college graduates. Similar disparities by age group, and education within age group, in access to and use of digital technologies were also observed in the 2017 CPS-CIUS [6].

Across all age groups, only about half of adults had sought health information from Web-based sources during the past year or were interested in doing so in the future. Although younger seniors were less likely than middle-aged adults to use the internet alone or with help, we did not observe similar age group differences with regard to having used the internet to obtain health information in the past year or interest in using an internet-based modality in the future. Older seniors were less likely than the younger 2 groups to be using the internet and also less likely to have used or be interested in using internet-based health information resources. When we restricted our analyses to internet users, the difference between the older senior group and younger 2 groups in seeking health information from the internet in the past year substantially diminished, but the older group remained less interested in using an internet-based health information resource in the future. We found significant disparities by education in past year use and interest in future use of internet-based health information resources, although within age groups, the differences between college graduates and those with some college were much smaller than differences between college graduates and those with no college education, and the latter differences were still smaller among internet users.

With regard to interest in using specific health information and health education modalities, we found that interest in using Web-based health information resources (webpage information, Web-based videos, interactive patient education programs, webinars, podcasts/online audio programs, online chat rooms/communities, emailed newsletters, messages sent through the patient portal, text messages, and video visits with a patient educator) was substantially lower among older seniors than among the middle-aged and younger senior groups. For example, among those who were interested in textual health information and health newsletters, some of the differences (eg, interest in information from a website, emailed newsletters, and patient portal messages) were associated with not being an internet user or email user. This was not the case for most of the online modalities, where there was very little difference in percentages of all adults and online adults who were interested in using the modalities. Within all 3 age groups, interest in the online modalities was significantly lower among those with no college education than among college graduates.

Across all age groups, the percentages of adults who expressed interest in using health apps and podcasts in the future were about twice as high as the percentages of adults who had reported using these modalities in the previous year. Prevalence of interest in using health apps was also twice as high between younger and older seniors who owned a smartphone compared with all adults in those age groups, suggesting that as smartphone ownership increases in these older age groups, there is potential for greater uptake of health apps. Interest in listening to health podcasts was very low and was not substantially higher among smartphone owners than all adults for any age group.

The percentages of middle-aged, younger senior, and older senior adults in our study population who in 2014 used the internet to obtain health information were comparable with those observed in national samples for middle-aged and older adults [58] but substantially higher than estimates for all US adults in these age groups on the basis of the 2015 CPS-CIUS population (51.9% vs 39.2%, 51.1% vs 31.5%, and 38.1% vs 23.3%, respectively) [5]. This difference in internet-based health information seeking can be partially explained by differences in population demographics. Across all 3 age groups, compared with the US population, this study’s population had higher percentages of adults who had attended some college or were college graduates and lower percentages with lower household incomes (>US $35,000) [5]. As numerous studies have shown that use of the internet increases as educational attainment and HHI increase, it is not surprising that this study’s population had a higher proportion of internet users in all 3 age groups than the US population (approximately 95% vs 76%, 88% vs 64%, and 68% vs 42%, respectively) and thus had greater capability to search for health information online. When we restricted our comparison of these age groups to internet users, we found that the percentages in this study’s population who had sought health information from the internet were only slightly higher than among these same age groups in the US population (52.4% vs 51.1%, 55.5% vs 50.2%, and 53.0% vs 45.1%, respectively) [5].

Another potential reason for the higher prevalence of internet-based health information seeking in our health plan population is that this study’s population was restricted to adults who had at least one chronic health condition, and previous research has found that adults with chronic conditions are more likely to use patient portals and Web-based patient education resources [8,9,32,33,59,60]. The percentages of middle-aged and older adults who used the internet to obtain health information in the past year estimated from our survey and the 2015 CPS supplement are much lower than those reported in a 2012 Pew survey (71% of middle-aged and 58% of adults aged ≥65 years, who used the internet and 54% and 30%, respectively, of all adults in those age groups) [3].

The results of this study and other research suggest that when planning delivery of health information and patient education for adults with chronic health conditions, it is important to take into account the population’s age group composition and educational attainment to gauge the likely uptake of internet-based and mobile health (mHealth; mobile technology-based) resources. Although more middle-aged and older adults are using the internet now than in the past [61], they are still less likely than younger adults to be using the internet and using the internet for functions other than email [7,9,14]. Many noninternet users lack easy access to digital technology (internet-enabled devices, high speed internet connections) that could connect them to the internet [8,12], and many older adults with chronic health conditions have physical or cognitive impairments that make it difficult to use internet-based resources [11,16,62]. In addition, even those currently using the internet might lack internet-based health skills (ability to access and use DITs for health purposes), experience, comfort, and trust in accessing internet-based health information resources [4,8,63]. For various reasons, they might also just prefer to get health information through print materials or directly from a person rather than from an internet-based source [23,32,39-41]. As adults with lower levels of educational attainment are less likely to seek health information in general, let alone use the internet to do so [64], it is important to make sure that HIA remains easily available through modalities that noninternet using adults will be more likely to use.

Without encouragement and support from health care professionals, family, and friends, middle-aged and older adults with chronic conditions who are not currently using internet-based health resources and health apps are unlikely to make the transition to electronic health and Web 2.0 [37]. However, even with encouragement, these adults are likely going to need assistance in gaining access to Web-enabled computers and other digital devices that they can comfortably use to connect with, navigate, and read information on the internet, as well as use high-speed internet or Wi-Fi if they have their own devices. Although younger adults find smartphones and tablets work well for performing online functions, aging adults with poorer vision and less manual dexterity might need to use a desktop or laptop computer with a larger screen and manual keyboard. They will also likely need training and support in how to use these digital tools, navigate the internet, conduct Web searches, and download materials [4]. Most public libraries offer access to computers and printers, Wi-Fi for people who bring their own Web-enabled devices, and librarians or volunteers to assist those who need help with online tasks [65]. Many libraries and community centers also offer classes for adults in how to use different types of digital devices and interact with the internet [66].

Developers of Web-based health information resources and health apps must also test their products with a wide range of potential end-users to make sure that these programs and tools are both effective in what they aim to achieve and easy for older and less educated adults to use [37]. In addition, health care providers and patient educators should not assume that even patients who are using a patient portal or are college-educated will follow up on recommendations to access Web-based health resources. Some patients, who might be willing to use Web-based and mHealth patient education and self-management tools but lack the equipment to do so, might also need financial assistance to purchase digital technology or to be given access to loaner equipment.

This study has a number of strengths. First, the survey dataset enabled us to estimate the prevalence of use of multiple DITs and interest in using several different internet-based and mHealth modalities to obtain HIA in a population of insured patients with chronic health conditions. Second, because of the large sample size and sociodemographic diversity of the study cohort, we were able to show significant disparities in use of DITs and health information modality preferences across 3 age groups (middle-aged, younger seniors, and older seniors) and by education within age groups using directly observed weighted percentages, not just ORs from logistic regression models. Third, we were able to show how prevalence of previous use of and future interest in using different types of internet-based health information resources differed by age group and education among the segment of this patient population that was using the internet.

The survey was conducted with adults from 1 Northern California health plan membership that, while fairly representative of Northern California adults, is not representative of the US middle-aged and older adult population with regard to educational attainment, income, broadband internet access, and health care coverage. The health plan membership is better educated and has a lesser percentage of low-income adults than the general US adult population and primarily resides in urban and suburban communities with widespread access to home and workplace broadband internet and free Wi-Fi in commercial and community settings. Moreover, members of this health plan are encouraged by the health care staff to use the comprehensive health information and health education resources available on the health plan’s website. The confluence of these sociodemographic and internet-related factors might have increased the percentages of adults in all 3 age groups who used DITs and are interested in going online for health information. The survey did not include adults with a primary language other than English and with no health care coverage, and it did not include groups whose ability to access and preferences for using internet-based health information resources might differ from thus study’s population and thus limit generalizability to the entire US population. This study’s sample excluded adults who were missing data on internet use status. However, the percentages of respondents with missing data for this variable were so small (0.8%, 1.8%, and 3.7% of middle-aged, younger senior, and older senior adults, respectively, after weighting) that we do not believe this introduced much bias in the results. Finally, although we used logistic regression models to control for race/ethnicity and sex when we tested for age group and educational disparities in DIT use and health information modality preferences, we did not examine whether the same patterns of disparity were found across all race/ethnic groups. An earlier survey of seniors in this health plan membership found that within level of education, black and Latino seniors were less likely than non-Hispanic white and Asian seniors to be using the health plan’s patient portal [8,23]. Future research is needed to identify whether sociodemographic and sociocultural factors differentially influence use of DITs and preferences for using specific internet-based and mHealth information modalities among adults in different racial/ethnic groups. Such information would improve the evidence base for development and implementation of patient-centered resources at the population level to prevent chronic health conditions and improve CCM, health outcomes, and quality of life.

DITs and internet-based health information resources provide a relatively inexpensive and effective way for adults with chronic health conditions to access information that can help them learn about and manage their health. However, this study found significant digital divides by age and educational attainment among middle-aged and older adults in ownership of digital devices and preferences for using internet-based resources to obtain HIA. These digital divides could potentially limit access to valuable health information and chronic disease self-management resources among vulnerable adult populations. Bridging digital divides in use of internet-based health resources will require ongoing personal encouragement from clinical staff for patients to try these new resources, including talking up the advantages of using these as an adjunct to and not replacement of aspects of the way they currently receive health care and obtain and share health information that they value. Patients reluctant to engage with digital information resources might also need to be provided with print materials and personal (nonvirtual) learning opportunities to become comfortable using these resources. Health care providers and consumer health organizations should also user test their internet-based resources before implementation to make sure that older and less educated adults will be able to use them easily and effectively. Finally, as part of providing patient-centered care, it will be important for health care providers and other consumer health organizations to continue to make it easy for patients to obtain health information and patient education from print materials, one-on-one patient counseling, and other more traditional modalities.