We conducted a cross-sectional data analysis merging 3 cycles of HINTS: HINTS 5 Cycle 3 collected between January 2019 and May 2019 (n=5438), HINTS 5 Cycle 4 collected between February 2020 and June 2020 (n=3865), and HINTS 6 collected between March 2022 and November 2022 (n=6252), noting that data were not collected in 2021. We targeted respondents who identified themselves as nonprofessional, unpaid caregivers. HINTS is a nationally representative survey that has been administered by the National Cancer Institute since 2003. The survey targets noninstitutionalized adults in the United States and collects data on their need for, and access to, health-related information and health-related behaviors. The survey used a 2-stage stratified random sampling method. The sampling frame was derived from a database of addresses maintained by Marketing Systems Group. All nonvacant US residential addresses in the Marketing Systems Group database were eligible for sampling. In the first stage, a stratified random sample of residential addresses was drawn from the database. In the second stage, 1 adult was selected from each sampled household using the Next Birthday Method. Details on the sampling method and data collection are available in the methodology reports for each HINTS cycle [27-29]. For this study, we focused on respondents who identified themselves as nonprofessional, unpaid caregivers. This study was reported in accordance with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines (Checklist 1).

Descriptive statistics (means, SDs, frequencies, and percentages) were used to describe the respondents’ characteristics and digital health engagement behaviors in each HINTS cycle. Differences among the survey cycles were examined using Rao-Scott adjusted chi-square tests. Controlling for survey cycles, age, and sex, unadjusted logistic regression tests were conducted to examine the associations between digital health engagement behaviors and each independent variable. Independent variables with a P value<.10 in the unadjusted regression analyses were retained in the multivariable logistic regression models. The final models were selected using backward elimination with a removal criterion of P value<.10. We merged the datasets using IBM SPSS Statistics version 30 and performed all analyses using R version 4.6.0 (R Foundation for Statistical Computing). We applied the final sample weight and replicate weights to the analyses to obtain nationally representative estimates and SEs.

The Westat Institutional Review Board (IRB) approved HINTS 6 (project number 6632.03.51) and HINTS 5 Cycle 3 and HINTS 5 Cycle 4 (project number 6048.14). A “Not Human Subjects Research” determination was given to HINTS 5 from the National Institutes of Health Office of Human Subjects Research (Exempt number 13204) and HINTS 6 from the National Institutes of Health Office of IRB Operations (IRB ID: IRB002042).

Of the total respondents from the 3 cycles of HINTS (N=15,555), we selected those who fulfilled the following criteria: (1) answered “yes” to the question “are you currently caring for or making health care decisions for someone with a medical, behavioral, disability, or other condition?”; (2) answered “no” to the question “do you provide any of this care professionally as part of a job (eg, as a nurse or professional home health aide)?”; and (3) did not respond to the caregiving condition as “not sure/don’t know” only or “other” only. After excluding cases that did not meet the criteria, we included 1676 caregivers in the analysis (weighted N=91,486,345.9): 570 from HINTS 5 Cycle 3, 412 from HINTS 5 Cycle 4, and 694 from HINTS 6. The flow of the sample selection is shown in Figure 1.

Descriptive characteristics of the pooled sample (N=1676) are presented in Table 1. Approximately two-thirds of the sample were female participants (n=109, 59.8%), aged 50 years or older (n=1169, 61.5%), and non–Hispanic White (n=956, 64.9%). The majority were married (n=1083, 69.8%), had some college or higher level of education (n=1287, 76.8%), lived in urban areas (n=1483, 87.1%), and were covered by health insurance (n=1565, 92.8%). A similar percentage of the sample reported that they were caregiving for a child (n=523, 38.7%) and/or a parent (n=630, 37.8%). The most reported condition for caregiving was Alzheimer disease and other neurological and mental health issues (n=1055, 61.7%), followed by orthopedic and aging-related issues (n=709, 41.8%). Most of the sample had access to high-speed internet (n=1379, 84.4%) and owned a smartphone (n=1471, 91.4%) and/or a tablet (n=1066, 68.1%). The descriptive characteristics did not differ significantly between the survey cycles.

Table 2 presents the selected digital health engagement behaviors for each survey cycle, with significant differences found between the cycles. The percentage of caregivers accessing their own online medical records increased from around 50% (278/570) in HINTS 5 Cycle 3 (2019) and HINTS 5 Cycle 4 (2020) to 72.6% (487/694) in HINTS 6 (2022; P<.001). Compared to HINTS 5 Cycle 3 (2019; 168/570, 30.8%), the percentage of caregivers accessing the care recipient’s online medical records decreased in HINTS 5 Cycle 4 (2020; 106/412, 22.7%) but increased in HINTS 6 (2022; 307/694, 44.5%; P<.001). Regarding health-related use of social media, the percentages of caregivers sharing health information decreased in HINTS 5 Cycle 4 (2020; 74/412, 17.9%), then increased in HINTS 6 (2022; 263/694, 39.1%), compared to HINTS 5 Cycle 3 (2019; 110/570, 22.5%; P<.001). Interacting with others online (69/570, 16.6% in 2019; 55/412, 12.2% in 2020; and 176/694, 27.0% in 2022; P<.001) and watching health-related videos (238/570, 49.5% in 2019; 184/412, 45.6% in 2020; and 419/694, 60.9% in 2022; P=.005) showed similar trends.

This study examined trends in digital health engagement behaviors—specifically accessing online medical records and using social media for health-related purposes—among unpaid family caregivers in the United States using 3 cycles of nationally representative HINTS data collected before, during, and after the COVID-19 pandemic. Overall, caregivers’ engagement with digital health technologies increased substantially following the pandemic, particularly with respect to accessing online medical records for both themselves and their care recipients.

Consistent with trends observed in the general US population [30], in this study, family caregivers surveyed in the postpandemic cycle were significantly more likely to access online medical records for both themselves and their care recipients than those surveyed prior to the pandemic. This acceleration has been attributed to the rapid expansion of digital health technologies during the pandemic and to regulatory efforts—most notably the 21st Century Cures Act—that require health care organizations and health IT developers to provide patients with easier electronic access to their health information [31]. In addition, this finding is consistent with prior work showing that caregiving became a stronger predictor of online medical record use after COVID-19 than before the pandemic [32]. As caregiving responsibilities intensified during the pandemic—amid disruptions in in-person care and increased reliance on remote communication—caregivers may have increasingly relied on electronic health records to manage appointments, medications, test results, and care coordination. An important key contribution of this study is its longitudinal comparison of caregivers’ access to both their own and their care recipients’ online medical records across survey cycles spanning the pandemic. Prior studies have often focused on patient portal use among patients themselves or relied on cross-sectional designs [33-35]. By explicitly examining caregivers’ access to care recipients’ records, this study addresses a significant gap in the evidence, as caregivers frequently rely on online medical records to manage medications, monitor test results, and communicate with providers on behalf of care recipients [36-38].

In addition to survey cycles, several demographic and socioeconomic factors were associated with disparities in online medical record access among family caregivers. Higher educational attainment and health insurance coverage were associated with greater access to caregivers’ own records, consistent with previous studies among US adults [33,34,39]. These disparities may reflect differences in digital and health literacy, frequency of health system interactions, and familiarity with online patient portals [40,41]. Caregivers residing in urban areas were more likely to access care recipients’ online medical records than those living in rural areas, a pattern that mirrors broader rural-urban gaps in digital health engagement and infrastructure [35,42,43].

The caregiving context also shaped digital health engagement. Caring for a spouse or child, as well as for individuals with cancer or chronic conditions, was associated with a greater likelihood of accessing care recipients’ online medical records, but not caregivers’ own records. These findings suggest that caregivers’ use of online medical records is driven primarily by caregiving responsibilities and clinical complexity rather than personal health needs. Similar patterns have been observed in studies of caregivers managing serious or chronic illnesses, where electronic access to health information supports coordination, monitoring, and decision-making [36,44].

Access to high-speed internet and ownership of mobile devices were strong facilitators of online medical record use among caregivers in this study. These findings align with broader evidence that broadband access and mobile device availability are foundational to equitable digital health engagement [30,35,43,45]. Mobile-enabled access may be particularly important for caregivers, who often manage care across multiple settings and time constraints. Prior studies suggest that individuals who use mobile devices or health apps access health information more frequently and consistently than those who rely solely on websites [46]. Our results highlight the need for policies that focus on increasing high-speed internet access. Enhancing this digital infrastructure is increasingly critical in areas where digital health care needs and remote access are particularly important. In contrast to the steady increase in online medical record access, caregivers’ health-related use of social media followed a nonlinear pattern, declining during the early pandemic period and increasing substantially in the postpandemic cycle. Following the pandemic, caregivers were more likely to share health information, interact with others with similar health concerns, and watch health-related videos on social media. Social media platforms have become increasingly important sources of health information and peer support, particularly during and after the COVID-19 pandemic [47,48].

Disparities and cultural factors were also evident in health-related social media use. Older caregivers (aged ≥65 y) were less likely to interact with others or watch health-related videos on social media compared with younger caregivers (aged 18‐34 y). Access to high-speed internet and ownership of mobile devices were associated with greater engagement. These differences may influence the reach and effectiveness of social media–based health interventions, highlighting the need for tailored strategies that consider age and digital access [47,49]. In addition, non–Hispanic White caregivers were less likely to engage in certain health-related social media activities than non–Hispanic Black, Hispanic, and non–Hispanic Asian or other caregiver groups. National survey data show that racial and ethnic minority adults demonstrate equal or higher engagement with social media platforms compared with White adults [50]. Prior research also indicates that Black and Hispanic caregivers often rely more heavily on family-centered and kin-based support networks compared with non–Hispanic White caregivers [51,52]. Together, these patterns suggest that cultural differences are factors to consider, and social media may function as an extension of kin- and community-based information exchange among racial and ethnic minority caregivers. At the same time, the growing reliance on social media for health information raises concerns about health misinformation, underscoring the importance of digital health literacy and the promotion of credible information sources [53].

Notably, existing research on social media use among family caregivers has largely emphasized qualitative analyses of online interactions or associations with caregiver outcomes. Population-level studies examining trends and correlations of health-related social media use among caregivers remain limited. This study contributes to the literature by documenting temporal trends and identifying key sociodemographic and technological factors associated with caregivers’ health-related social media engagement.

This study has several limitations. First, the datasets lacked detailed information on caregiving intensity, duration, and caregiving tasks, which may influence digital health engagement. Second, harmonizing variables across survey cycles requires recategorization, which may have introduced measurement inconsistencies. Third, the repeated cross-sectional design does not allow the assessment of longitudinal changes within the same individuals. Finally, the analysis was limited to selected digital health engagement behaviors and did not capture the full range of caregivers’ digital health activities.

There was a clear increase in US family caregivers’ access to online medical records and health-related use of social media, following the COVID-19 pandemic. However, disparities by age, race/ethnicity, geography, education, insurance coverage, and digital access persist. These findings highlight the need for caregiver-centered, equity-focused digital health strategies and suggest that future technology-based interventions should account for the diverse characteristics and needs of family caregivers. The findings provide insight into the importance of digital health engagement among caregiver populations, with consideration of equity and cultural factors at the policy level, particularly for caregivers living in rural areas and/or those with limited access to broadband internet and mobile connectivity, which may affect digital health access. Future research should also examine caregivers’ access to and use of emerging digital tools such as AI-based conversational agents (eg, GPT-powered chat tools) to understand their potential roles in health information seeking, decision support, and emotional support, as well as the risk of misinformation associated with digital platforms, including algorithm-driven short-form video apps (eg, TikTok).