This review follows the guidelines of the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) [11]. We were also inspired by the recommended framework for conducting systematic reviews in information-related fields by Okoli [12]. The bibliographic database search strategies were developed after consulting an academic librarian at the first author’s university.

First, we searched the following databases: Web of Science; Cochrane Library database; PubMed; MEDLINE; CINAHL Plus; APA PsycINFO; Library and Information Science Source; Library, Information Science and Technology Abstracts; Psychology and Behavioral Sciences Collection; Communication & Mass Media Complete; ABI/INFORM; and ACM Digital Library. These databases were chosen because they cover the academic disciplines (eg, medicine, medical informatics, communication, psychology, and information and library science) that are most likely to study older adults’ OHIS behaviors. Second, the search queries contained the following categories and keywords: people (older adults, elderly, aging, senior, seniors, older people, aged 60, aged 65), behavior (find, search, seek, access, retrieve), place (internet, online, web), object (information), and attribute (health, medicine, drug, nutrition, diet, wellness, illness). Specific queries were run in the topic, title, and abstract fields, depending on the database (see Multimedia Appendix 1). The initial search was performed in November 2020. Third, we captured additional articles using Google Scholar by tracking the citations and references in the articles found in the databases and in other relevant reviews. In addition, we supplemented relevant articles by searching Google Scholar directly. All the studies identified during the database searches were imported into the reference management software Zotero, and duplicates were removed.

We developed a series of inclusion and exclusion criteria to identify articles relating to older adults’ OHIS behaviors. The inclusion criteria were as follows: (1) The articles should pertain to health-related contexts, including areas such as health, mental health, diet, and nutrition. (2) The article should describe OHIS behaviors (eg, general OHIS, selection and use of health information sources, and adoption and use of health information). (3) The article should focus on older adults (Note that although the search strategies indicated 2 commonly accepted lower age boundaries, 60 and 65 years, to identify older adults, it did not exclude other ways to describe the population); studies that clearly mentioned the population of older adults or contained explicit, equivalent claims were eligible. (4) The research should be empirically based. (5) The articles should have been published in a peer-reviewed journal or in conference proceedings. (6) When we identified more than 1 paper published by the same author on the same topic, we selected only the most recent one. (7) The articles should be written in English.

Our exclusion criteria were as follows: (1) The articles did not pertain to a health-related context. (2) The articles were not about OHIS behaviors; for instance, some articles focused only on general ICT use or adoption behaviors, were more concerned with technology-related rather than information-related issues or addressed only older adults’ health literacy or eHealth literacy and did not investigate their OHIS. (3) The articles did not focus on older adults; we specifically excluded articles that treated age merely as a predictor or moderator in studying the OHIS of the general population, as it is evident that age influences people’s OHIS behaviors. (4) The articles were not based on empirical research; this criterion helped eliminate opinion pieces, brief communications, editorial commentaries, and reviews. (5) The articles were not peer-reviewed (eg, a self-archived manuscript). (6) The articles were not written as full papers (eg, abstracts, posters, or letters). (7) The articles were not written in English.

The procedure for screening articles was based on the eligibility criteria. The initial search used database search strategies and identified 8047 publications. After duplicates were removed, the data set consisted of 5949 publications for screening.

The screening involved 3 stages. In the first stage, all the 3 authors reviewed the titles and abstracts of a sample of 300 articles from the search results, and then discussed and refined the screening criteria. In the second stage, we selected another 300 articles randomly from the search results as a test set. The feasibility criteria were verified independently by 2 of the authors (SS and MZ). Intercoder agreement (κ=0.816) indicated satisfactory reliability. Discrepancies were discussed and resolved by involving the third author (YZ), and the eligibility criteria were further refined accordingly. In the third stage, author MZ screened the remaining articles based on the eligibility criteria using the titles and abstracts, and author SS validated the results. Discrepancies were resolved by involving author YZ. The whole screening procedure resulted in 279 articles for full-text analysis.

To read and code the full-length articles downloaded from the databases, we used the MAXQDA 2020 software, which is designed for analyzing computer-assisted qualitative and mixed methods data, texts, and multimedia data. During the full-text analysis, we excluded 211 articles by applying the eligibility criteria. The remaining 68 articles were retained, and 8 more eligible articles were identified through citation tracking with the assistance of Google Scholar. In total, 75 articles were selected for the systematic scoping review.

We used Excel (Microsoft Corporation) to extract and record the basic information of the articles in the sample, including the author(s), title, publication year, publication name, and publication type (eg, journal vs conference). We used thematic content analysis in an iterative manner to identify the evidence regarding our research questions [13]. Several lists of codes were generated during 2 rounds of full-text coding procedures. In the first round, all the authors participated in the open and selective coding processes until a coding schema emerged and converged. In the second round, MZ coded the full texts by applying the coding schema, and SS validated all the codes. The intercoder reliability of the thematic content analysis reached 85%. Discrepancies were solved by involving YZ in the discussion.

After screening, the final sample consisting of 75 articles was obtained, as shown in Figure 1. The articles were published between 1997 and 2020 (see Multimedia Appendix 2). Trend observations revealed that the number of publications in this subject area increased over time and that the OHIS of older adults began to receive considerable attention in the last 3 years (see Figure 2). The articles in the sample were mostly published after 2006 (n=69, 92%), which relates closely to the boom in social media. Of all the articles, 72 (96%) were published in journals, and the remaining 3 (4%) were published in conference proceedings. The articles originated from 17 countries (based on the first author’s affiliations), with the top 3 being the United States (n=44, 58.67%), Australia (n=5, 6.67%), and China (n=4, 5.33%). The top 4 journals publishing these articles include the Journal of Medical Internet Research (n=8, 10.67%), Educational Gerontology (n=4, 5.33%), Journal of Health Communication (n=3, 4%), and Library & Information Science Research (n=3, 4%), indicating the multidisciplinary nature of the sample.

The systematic scoping review first investigated how the included 75 articles defined the target population of older adults. The cutoff ages for defining older adults were determined. More than half of the articles used samples of older adults aged above 60 years. Furthermore, 16 articles (21.33%) defined older adults as those aged 65 years and above, and 23 (30.67%) had cutoff ages ranging from 60 to 64 years. In addition, we noted some papers that defined the older adult group more loosely. For example, the cutoff age in 17 articles (22.67%) ranged from 50 to 54 years, and 14 articles (18.67%) used samples with minimum ages ranging from 55 to 59 years. Moreover, 5 of the articles (6.67%) did not specify precise age distributions.

The research methods varied across the 75 studies. Regarding methodological approaches, we found that 45 studies (60%) used quantitative approaches, 22 (29.33%) employed qualitative approaches, and 8 (10.67%) were based on mixed methods designs, using a combination of quantitative and qualitative methods. As for specific methods, surveys (n=28, 37.33%) and interviews or focus groups (n=25, 33.33%) were the primary methods used, followed by secondary data analysis (n=6, 8%) and experiments (n=4, 5.33%). In terms of data sources, most of the studies were based on primary data (n=65, 86.67%) and a few on secondary data (n=10, 13.33 %). Concerning the types of data, we found 59 studies (78.67%) based on cross-sectional data and 16 (21.33%) based on longitudinal data.

Information types and information sources are 2 frequently reported aspects of information in OHIS studies [14]. For our analysis, we adapted the typologies of health information types from Kent et al [15] and Ramsey et al [16]. The results presented in Table 1 suggest that older adults often search the internet for information on specific diseases because they want to obtain a general idea of their condition before diagnosis or treatment so that they know what to expect and can be better prepared to face stressful situations [17]. The health problems mentioned in these 75 articles are mainly cancer (n=10, 13.33%), mental health problems (n=5, 6.67%), chronic conditions (n=4, 5.33%), and physical diseases (n=4, 5.33%). Aside from this disease information, the most frequently mentioned types of information are related to medication or treatment, nutrition or exercise, medical research, disease symptoms, and health promotion. Some articles mentioned that older adults also use the internet to seek support groups or interpersonal advice, health insurance information, health news, and health policy information. Of note is that more than half of the articles (n=40, 53.33%) used the umbrella term health information, without specifying any type of health information content. Furthermore, the types of content were not mutually exclusive. For example, a single article might mention more than 1 type of information (eg, older adults seeking information for cancer-related symptoms and medication).

Most of the articles in the sample (n=58, 77.33%) used the general internet to represent all the web-based sources of health information. Further, 26 articles (34.67%) described health websites as sources of internet-based health information for older adults; among these, the owners of the websites varied, consisting of educational, commercial, government, and nonprofit entities. Moreover, general search engines such as Google were the third most frequently mentioned sources in the studies (n=17, 22.67%), suggesting that older adults often use general search engines to start OHIS [18-20]. Further, 11 articles (14.67%) mentioned older adults’ use of social media (eg, Facebook, Twitter) and blogs in OHIS. Only 3 articles (4%) addressed older adults’ use of patient portals, and 2 articles (2.67%) were about older adults’ use of mobile internet services. Table 2 shows the health information sources mentioned in the studies.

Among the 75 articles, 35 (46.67%) treated OHIS as a variable or construct. These articles quantitatively measured OHIS with various scales or proxy variables. Among them, 27 (36%) regarded OHIS as a dependent variable and explored the antecedents of older adults’ OHIS. Further, 4 (5.33%) treated OHIS as an independent variable, and the remaining 4 (5.33%) treated OHIS as neither a dependent nor an independent variable but provided only descriptive analyses. Because the articles that employed quantitative approaches primarily concerned the antecedents of older adults’ OHIS, we summarize the antecedents in Table 3.

We summarize the main influencing factors that appeared in the investigations. The antecedents of older adults’ OHIS fall mainly into 2 categories, namely individual-related characteristics and source-related characteristics. Within the individual-related characteristics, 12 subcategories were observed, including demographics, anxiety, beliefs, attitudes, self-efficacy, personality, health status, medical history, health care service availability, source experience, health literacy and motivations. Among the source-related characteristics, credibility, usefulness, and ease of use were the 3 most frequently mentioned factors.

Rather than treating OHIS as a variable, 40 of the 75 articles (53.33%) treated OHIS as a process. Of these studies, 29 (38.67%) explored the barriers that older adults encounter during OHIS. The results suggest that older adults may experience many barriers preventing successful OHIS, as shown in Table 4. In the prior studies, we identified 3 main types of barriers (ie, individual, social, and ICT), 11 subtypes, and 38 specific issues.

Regarding individual barriers, some studies found that older adults’ OHIS could be hindered by age-related functional decline, including vision impairment, poor eye-hand coordination, physical challenges (eg, back pain), and illness. Moreover, some studies reported several aspects indicating low literacy among older adults that prevented effective OHIS, including limited English language skills, lack of basic health knowledge, limited digital literacy, undeveloped information literacy, and low health or eHealth literacy. Moreover, some studies found that older adults’ perceptions of low self-efficacy regarding computer use, reading, learning, and evaluation of health information reduced their willingness toward OHIS. Other findings revealed that negative attitudes toward internet use or general technology and privacy concerns about using technology decreased older adults’ intentions to search information on the internet. The results also revealed that beliefs regarding the external locus of the control of health care and fatalistic beliefs reduced older adults’ active OHIS.

As for social barriers, studies suggested that older adults may have some social stigma concerning OHIS when it comes to mental and sex-related health problems. Moreover, older adults often report a lack of social support in their OHIS, including informational, organizational (eg, health care services), instrumental (eg, instructions on computer use), intergenerational (eg, support from children), and peer support (eg, support from friends).

In terms of ICT use, analysis of the studies revealed that many older adults do not possess information technology devices, and they reported low accessibility to medical records. Moreover, the quality of general health information on the internet is problematic. Older adults are likely to encounter misinformation, conflicting information, and irrelevant information during their OHIS. Furthermore, they often confront information overload when reading health information due to overwhelming amounts of irrelevant information or pop-ups. Moreover, older adults’ OHIS may lead to some unpleasant and frustrating user experiences, such as unsatisfactory interactivity and navigability, unsuitable font sizes, dense text lacking visual elements, confusing layouts, and complicated site designs.

Given the abovementioned barriers, it is essential to provide older adults with additional support to facilitate their OHIS. We identified 11 studies (14.67%) among the 75 that used educational training programs to facilitate and intervene in older adults’ OHIS, as shown in Table 5. Among these, 10 of the 11 studies provided offline workshops, and 1 conducted an online workshop. The offline workshops were conducted in community settings (eg, public libraries, schools, or medical centers) and included face-to-face instruction. We identified only 1 study that used an internet-based tutorial to improve older adults’ ability to distinguish high-quality internet-based health forums from low-quality ones. Among the 11 articles, 9 described training programs with multiple sessions, each lasting 2 to 3 hours, and the duration of the programs varied from 1 to 4 months; the other 2 studies used 1-time training sessions.

Further, 4 of the 11 programs were guided by established theories, models, or concepts (eg, the self-efficacy theory and the health belief model). All the studies involved some form of evaluation, including postsession surveys or interviews, pre- versus postintervention comparisons, and experimental versus control group comparisons. In addition, 5 studies evaluated the effectiveness of the intervention outcomes from a longitudinal perspective over a period ranging from 1 month to 1 year to the competence of the program. Among all the studies, 9 statistically assessed the effects of the intervention. Measures varied across the studies; these included opinions from surveys on the internet, self-efficacy in seeking health information, and anxiety regarding computer use. All the articles reported some positive outcomes of the intervention programs.

This systematic scoping review provides an overview of OHIS behaviors among older adults, as shown in Figure 3. Overall, the findings of this paper reveal core elements of OHIS among older adults. First, the types and sources of health information that older adults search for were clearly presented. Then, a portion of the studies explored the main factors influencing older adults' OHIS behaviors, which can be categorized as individual-related and source-related characteristics. Then, we identified the barriers to OHIS behavior in older adults from existing literature, including individual barriers, social barriers, and ICT barriers. Finally, this paper provides an in-depth analysis of the interventions mentioned in some of the included papers to support OHIS behaviors among older adults. We believe that the framework of this paper can, to some extent, help researchers to better position their research objectives in future studies so that the objectives correspond to specific dimensions for in-depth empirical investigation.

Regarding the first research question, the results show that older adults sought various types of health information on the internet, including information about specific diseases, medication and treatment, nutrition and exercise, medical resources, disease symptoms, health promotion, support groups and interpersonal advice, health insurance information, and health news or policies. The information sources included health websites, general search engines, social media and blogs, patient portals, and mobile devices. The types of health information sought differed from those that interest young people. According to a recent systematic review [87], adolescents and youths (<24 years) search the internet for daily health-related issues, physical and psychological well-being, sexual health, social problems, and culturally sensitive topics. Compared to the adolescent and youth population, older adults tend to search more for disease-related health information topics.

As for the second research question, the results point to 2 main types of factors influencing older adults’ OHIS: individual-related characteristics and source-related characteristics. The individual-related characteristics include demographics, anxiety, beliefs, attitudes, self-efficacy, personality, health status, medical history, health care service availability, source experience, health literacy, and motivations. Among the source-related characteristics, credibility, usefulness, and trust were the 3 factors most frequently mentioned in the studies. We noted that the primary factors influencing older adults’ OHIS differ from those influencing young adults. A systematic review of studies investigating young adults’ (<24 years) OHIS [87] revealed that the most frequently mentioned influencing factors were gender, age, educational status, emotional characteristics, engagement in risky behaviors, and eHealth literacy.

The results for the third research question reveal that older adults might encounter 3 types of barriers during their OHIS, including individual barriers (eg, low literacy), social barriers (eg, social stigmas), and ICT-related barriers (eg, lack of ICT devices). These barriers may hinder effective OHIS behaviors of older adults. The results suggest some differences from the findings on young adults’ OHIS. For the adolescent and youth population (<24 years), the main barriers to OHIS include online privacy and concerns about information credibility [87]. Although some studies report low health literacy among adolescents [88], older adults seem to have more difficulties in this respect than adolescents [89,90].

As for the fourth research question, the review found that many intervention programs have been created to support older adults’ OHIS; they primarily use educational training workshops in offline and online formats. Most training programs contained multiple sessions, with each session lasting 2 to 3 hours; the duration of the programs varied from 1 to 4 months, and all the programs reported at least some positive effects in support of older adults’ OHIS.

Overall, this systematic scoping review identified the need for more in-depth research on older adults’ OHIS. As can be seen from the aforementioned evidence, a subset of studies have treated OHIS as a variable or construct and focused on exploring the factors influencing OHIS in older adults. Other studies treat OHIS as a process and investigate how the older adults search the internet for health information. However, given the complexity of the health conditions of older people and a projected future intensification of information overload, older adults will encounter more serious problems when searching for health information on the internet, such as how to select from among multimodal information sources, how to express health information needs, and how to evaluate health misinformation. Considering the growing population of older adults, the importance of internet-based information seeking for overall public health, and the lack of best practices, more research on this topic is needed. In this section, we propose several directions for future research based on gaps identified in the review.

This systematic scoping review has several limitations. The first one is in terms of search sources. Owing to the interdisciplinary nature of OHIS research in older adults, although we tried to search multiple databases using relevant keywords and consulted academic librarians to improve our search strategy, it was nevertheless inevitable that some literature would be missed, especially relevant research in unofficially published conference proceedings. The backward and forward strategy can be further used to expand the literature search sources in future [111]. Second, in terms of the literature type, this review mainly focuses on empirical studies, whereas some opinion papers, descriptive cases, and short communications on OHIS for older adults were excluded from our literature pool, and some complementary analyses of such nonresearch articles can be conducted in future. Finally, in terms of the analytical approach for searching literature, this study did not conduct a comparative chronological analysis of the literature in different periods, which to a certain extent could not fully reveal the impact of technological and sociocultural changes on older adults’ OHIS behavior. In future, the introduction of knowledge graphs can be considered to map the themes of the literature at different stages.

This review provides an overview of how older adults’ OHIS has been studied. It reveals that older adults search for various types of health information on the internet using different types of web-based sources and that their OHIS is jointly influenced by source-related and individual-related factors. Their difficulties in searching arise from individual, social, and ICT-related barriers. Some educational intervention programs that support older adults’ OHIS have been initiated in the form of web-based and offline workshops. Furthermore, the review reveals that the topic of older adults’ OHIS is understudied, although the number of studies is increasing. Nevertheless, more studies are needed to understand the problems associated with older adults’ interactions with health information and better support them in their decision-making when they are searching for medical and health information on the internet. Based on the findings of the review, the authors propose several objectives for future research.