Health information technologies (HITs), such as electronic health records (EHRs) and electronic medical records (EMRs), are pivotal tools in modernizing health care by enhancing efficiency and improving the quality of care [1-3]. Beyond these general applications, HIT has the potential to address the unique medical needs of the older population. Extensive research demonstrates that HIT use can significantly improve health outcomes for older adults [4], particularly through tools such as EMRs and electronic personal health records (ePHRs). For example, EMRs in hospital settings help health care professionals identify at-risk patients, such as those prone to falls, by analyzing factors such as age and medication use [5]. Similarly, ePHRs empower older adults by providing easier access to health information, fostering improved communication with health care providers, and encouraging active engagement in personal health care management [6].

Luo et al [7] highlighted that, among older adults, the use of ePHRs was positively associated with the ease of understanding health information. ePHRs enable patients to track their health history to better understand the progress or deterioration of their health conditions. Patients with access to their health information experience increased medical data transparency, which reduces medical errors, improves trust in care providers, and enhances patient satisfaction [8]. As such, in the United States, the adoption of ePHRs has steadily increased. This growth has been supported by policies such as the Health Information Technology for Economic and Clinical Health Act of 2009 [9], which incentivized health care providers to adopt EHR systems through financial rewards, indirectly creating an infrastructure supportive of ePHRs. The subsequent Meaningful Use Program set specific objectives for health care providers to demonstrate effective use of EHRs, including patient-access capabilities. Further support came from the 21st Century Cures Act of 2016 [10], which enhanced patient access to electronic health information and prohibited information blocking, thereby facilitating smoother integration and utilization of ePHRs.

While the benefits of HIT for older persons are well-documented, the adoption and sustained use of these technologies face notable barriers, such as lower levels of technology usage among older adults [11]. Recent data indicate that approximately 4 in 10 individuals aged 65 years and older utilize ePHRs, a rate significantly lower than that of younger populations [12]. Furthermore, a previous study reported that approximately 54% of US adults have been offered access to their online medical records, and among those offered, 57% accessed them at least once, translating to about 31% of the adult population accessing their online medical records [13]. However, specific statistics for individuals aged 65 years and older are not provided in this source. Given that older adults are generally less likely to use the internet, it is reasonable to infer that their rates of ePHR usage may be lower than those of younger populations [12].

Further, researchers have found that older age is negatively associated with HIT [14]. Reduced cognitive and motor abilities, common among older adults, often complicate interactions with modern digital applications [15,16]. van der Vaart et al [8] discussed how older age is associated with the nonuse of web-based portals. Czaja and Lee [17] argued that although older adults are willing to use technology, many report usability problems with existing systems, which may, in part, be due to the cognitive and perceptual demands placed on the user. Their findings show that self-efficacy is an important predictor of general use of technology and that people with lower self-efficacy are less likely to use technology. However, individual-level barriers such as confidence and usability are only part of the picture. A growing body of research suggests that technologies are often not designed with the specific needs of older adults in mind [18,19]. Studies have also shown that older patients may not be offered digital health tools for various reasons, including—but not limited to—age-related stereotypes held by health care professionals or concerns about overburdening older patients [20,21]. Together, these individual and environmental barriers create a complex landscape that can inhibit older adults’ use of health technologies.

Despite advancements in technology, significant gaps remain in understanding the factors that influence the regular use of ePHRs among older adults. Existing research has focused on general barriers and facilitators, but limited knowledge exists about the interplay between these factors and how they shape the frequency of ePHR usage. This knowledge gap calls for a deeper examination of the factors affecting ePHR use in older populations. Additionally, limited electronic health literacy remains a critical challenge that hinders ePHR use in older populations [22]. However, the potential benefits cannot be overlooked. ePHRs offer older adults opportunities to coordinate their health care by sharing critical health information with providers and other stakeholders, an approach particularly crucial for managing chronic conditions that require multifaceted treatment strategies [7].

There are still unanswered questions about how older adults use ePHRs, particularly regarding the specific factors that facilitate or hinder ePHR use and the extent to which these factors influence it. To address this gap, this study explores the following research questions:

To answer these questions, we build on established theoretical frameworks such as the Patient Technology Acceptance Model (PTAM) [23], which provides insights into technology adoption and use processes. We extend the PTAM by incorporating factors specifically relevant to older persons, such as usability challenges, limited education, and issue involvement, to develop a comprehensive understanding of the dynamics at play. This study aims to contribute to the literature by offering empirical evidence and actionable insights into how ePHRs can be effectively designed and implemented to support the unique health care needs of aging populations.

This study used publicly available cross-sectional data collected in 2019 by the National Cancer Institute (Health Information National Trends Survey [HINTS] 5 cycle 3). The cross-sectional analysis for this study was limited to a subset of the original dataset and included participants aged 65 years or older. The total number of individuals aged 65 years or older was 1961. Of these, 1234 had not used an ePHR at least once in the past 12 months and were therefore excluded from the analysis. Further, 195 responses were removed due to missing values for key variables. The final sample consisted of 532 respondents, representing 13,136,180 US adults aged 65 and older when survey weights were applied.

The HINTS 5 survey, conducted with the general population, underwent expedited review and received approval from the Westat Institutional Review Board on March 28, 2016 (project number 6048.14). This analysis used deidentified, publicly available data from HINTS, which did not constitute human research as defined by 45 CFR 46.102 and, therefore, did not require Institutional Review Board review.

In this study, the primary dependent variable, “frequency of ePHR use,” is operationalized using a single-item measure assessing how frequently patients accessed their online medical records within the past 12 months. The response options are categorized as follows: 1-2 times, 3-5 times, 6-9 times, and 10 or more times. Individuals who reported not accessing their online medical records during the past year were excluded from the analysis, as critical variables pertinent to these respondents were not captured.

Performance expectancy, representing the perceived usefulness of online medical records for health monitoring, is measured using a 5-point Likert scale, ranging from “5=very useful” to “1=don’t use.” Effort expectancy, reflecting the perceived effort necessary to understand health information in online medical records, is assessed using a 4-point Likert scale ranging from “4=very easy” to “1=very difficult.” Issue involvement captures an individual’s engagement with personal health management and is operationalized as the number of interactions a respondent has had with health care providers over the past year. Self-efficacy in accessing electronic health records is assessed through a binary-choice question indicating whether respondents have utilized any electronic method to access their medical records during the past 12 months.

Single-item measures are used for several constructs in this study. Such measures are considered acceptable when questions are clear, unambiguous, and not prone to multiple interpretations [52]. Additionally, single-item measures are prevalent and widely validated within information systems research, particularly in studies applying structural equation modeling (SEM) in health care contexts [43,44].

Gender, race, and income are incorporated as control variables, consistent with prior studies [43,44]. For comprehensive details on questionnaire items, scales, and specific variable operationalizations, please refer to Multimedia Appendix 1.

This study uses SEM to conduct a path analysis. Although SEM is predominantly used to model latent variables, it is also commonly applied to conduct a path analysis in mediation models. We examine 2 mediating relationships: first, self-efficacy mediates the relationship between age and the frequency of ePHR use; second, self-efficacy also mediates the relationship between education and the frequency of ePHR use. Accordingly, we use SEM to test the model, consistent with prior research [53,54]. We used SEM with weighted least squares mean and variance adjusted estimation to test the hypotheses. Weighted least squares mean and variance adjusted is well-suited for models with ordinal outcome variables [55,56]. The analysis was conducted in R (R Foundation) using the “lavaan.survey” package. We incorporated HINTS-supplied survey weights and used jackknife variance estimation techniques to account for the complex HINTS sampling design and to calculate nationally representative estimates [57].

Table 1 shows the descriptive statistics of the survey respondents. The survey included questions about the frequency of participants’ ePHR use, as well as questions related to the model variables, including performance expectancy, effort expectancy, issue involvement, self-efficacy, age, education, gender, race, and income.

Table 2 presents the correlations between all variables of interest. Correlation coefficients are important because high correlations among independent variables may indicate multicollinearity, which can introduce bias into the model results. Multicollinearity is not a concern in this analysis, as all correlations fall within the acceptable threshold of 0.6 [58], with the highest correlation being 0.37 between performance expectancy and effort expectancy. Table 2 also provides the means and SDs for the principal variables.

As data are self-reported and collected through a single survey, they may suffer from common method variance (CMV), which hampers the relationship between the variables [59]. We used the marker variable technique [60] to check if the data are suffering from CMV. Marker variable is a variable that is theoretically unrelated to 1 or more of the principal variables measured in the study and typically has a low correlation with the central variables.

The theoretically unrelated construct “morning person or night person” was used as a marker variable. The correlations between the marker variable morning person or night person and other principal variables were below the threshold of 0.1 (performance expectancy: −0.08; effort expectancy: −0.08; self-efficacy: 0.03) [60], except for issue involvement (0.17). The low correlation of the marker variable with the principal variables in the model indicates that CMV is not a problem.

As the National Cancer Institute administered both a paper-based questionnaire and an online questionnaire to survey participants, we first assessed whether the mode of survey introduced any biases. We regressed the dependent variable, “frequency of ePHR use,” on the mode of survey and found that the relationship between the 2 was nonsignificant (P=.48). Thus, the mode of survey administration did not introduce any bias in the main outcome variable, that is, “frequency of ePHR use.”

The SEM results are presented in Table 3. The overall fit statistics (χ²₁₇=13.230, P=.004; Comparative Fit Index=0.950, Tucker-Lewis Index=0.715, root-mean-square error of approximation=0.080, standardized root-mean-square residual=0.022) indicate a good model fit [61]. Table 4 presents the results of the mediation analysis for self-efficacy with age and education.

The results show that age is negatively associated with self-efficacy (β=−.9890, P=.03), suggesting that as age increases, self-efficacy decreases, supporting H1b. However, no significant (P=.85) relationship was found between self-efficacy and education (H2b) in our research context.

Further, the results show that age is positively associated with the frequency of ePHR use (β=1.4950, P=.03), suggesting that as age increases, ePHR use also increases. This is the opposite of what we hypothesized for H1a. This counterintuitive trend may reflect increased health care needs among older adults, leading to greater engagement with ePHRs for managing chronic conditions, medication schedules, and communication with physicians. No significant (P=.88) relationship was found between education and the frequency of ePHR use (H2a) in our research context. H3 proposed a positive relationship between issue involvement and the frequency of ePHR use. The path coefficient was positive and statistically significant (β=.2560, P<.001), suggesting that higher issue involvement leads to higher ePHR use.

Our analysis also revealed a significant positive relationship between performance expectancy and the frequency of ePHR use (β=.2470, P<.001), as well as between effort expectancy and the frequency of ePHR use (β=.1850, P=.04). These findings suggest that higher performance expectancy and higher effort expectancy both lead to increased ePHR use, supporting H4 and H5. We also found that self-efficacy is positively associated with the frequency of ePHR use (β=.4990, P<.001), supporting H6.

Further analysis confirms that self-efficacy partially mediates the relationship between age and the frequency of ePHR use among older adults, as evidenced by the statistical significance of both indirect (P=.03) and direct effects (P=.03). Specifically, self-efficacy accounts for 49.2% (−0.493/1.002) of the total effect, underscoring its substantial role as a mediator in this relationship. However, we did not find any statistically significant mediating effect of self-efficacy on the relationship between education and the frequency of ePHR use.

Our study aimed to update the PTAM by examining the impact of performance expectancy, effort expectancy, and self-efficacy on the frequency of ePHR Use. By using an integrated framework, this research provides actionable insights into the factors driving ePHR usage among older adults, incorporating variables that have not been widely explored in prior studies. This study makes both a significant theoretical and practical contribution.

This study contributes to the literature on ePHR use by building on and extending existing theoretical frameworks, including the Aging and Technology framework [28] and the PTAM [23]. Prior research has predominantly highlighted barriers to ePHR use among older adults, such as reduced self-efficacy, usability challenges, and low digital literacy [15,17]. While these studies provide an important foundation, our findings challenge and refine these perspectives by demonstrating a positive association between age and ePHR use, with self-efficacy serving as a partial mediator in this relationship.

Contrary to earlier work suggesting older age is a barrier to HIT use [8,14], this study reveals that age positively influences ePHR use when health needs increase, as chronic illnesses and frequent health care interactions necessitate greater reliance on digital tools. This finding aligns with studies emphasizing the contextual nature of technology use, where health-related motivations can offset age-related challenges [7]. By demonstrating that older adults use ePHRs more frequently despite lower self-efficacy, this study extends theoretical models by incorporating health-related drivers, such as issue involvement, into the broader narrative of technology adoption.

Self-efficacy, identified as a strong predictor and mediator, validates and expands prior work in the field. Studies have consistently highlighted self-efficacy as a key determinant of technology use [62]. Our findings reinforce this and reveal that self-efficacy mediates the impact of both age and education on ePHR use. This underscores the critical role of psychological factors, particularly in populations facing cognitive and physical challenges, and aligns with theories from the PTAM [23] and the UTAUT [29].

The study also affirms the relevance of performance expectancy and effort expectancy as critical drivers of ePHR usage, consistent with the PTAM and UTAUT frameworks [47,48]. However, by identifying issue involvement as a significant predictor, this study contributes a novel dimension to the literature. This finding supports prior arguments that perceived relevance and personal health involvement significantly enhance engagement with health technologies [43,45].

Overall, this research advances theoretical understanding by integrating psychological, motivational, and contextual factors into established frameworks for HIT use. It bridges gaps in the existing literature by demonstrating that age, when considered alongside mediating factors such as self-efficacy and health involvement, can positively influence ePHR use. These insights offer a nuanced understanding of older adults’ interactions with health technologies, offering a robust foundation for future research and practical interventions.

This study has several limitations that should be considered. First, the use of secondary data from the HINTS limited the analysis to variables available in the dataset. Important factors, such as detailed measures of digital literacy or prior technology experience, could not be examined. Further, this study included only individuals who reported using ePHRs at least once in the past 12 months, as key variables of interest were not captured for those who had not used ePHRs. Additionally, some variables in this study were measured using single-item scales, which may lack the robustness of multi-item measures. However, single-item measures are considered acceptable when the constructs are straightforward and unambiguous [52]. We acknowledge that single-item measures inherently do not allow for traditional assessments of internal consistency reliability, which presents a limitation in this study. This limitation necessitates cautious interpretation of the findings. Future research should aim to incorporate multi-item scales where feasible and conduct reliability analyses to ensure the stability and consistency of these measures.

Second, the reliance on self-reported data introduces the possibility of CMV, which could inflate relationships between variables [59]. To address this concern, the study applied the marker variable technique [53], which confirmed that CMV was not a significant issue. Third, the cross-sectional design of the study limits the ability to establish causal relationships among the variables. Longitudinal research is needed to explore how factors such as self-efficacy and performance expectancy influence the use of ePHRs over time. Fourth, although the use of survey weights enhances the representativeness of the sample for the United States, the relatively small sample size remains a limitation that future research with a more targeted approach could address. Additionally, we did not control for factors such as income, digital literacy, and prior technology experience, as these variables were not accessible in the dataset. The absence of these controls may limit the completeness of the study. Future research should incorporate these variables to provide a more comprehensive understanding of the frequency of HIT use among older adults. Finally, the findings are specific to the US context and may not fully generalize to other populations or health care systems. Despite these limitations, the study offers valuable insights and lays important groundwork for future research.

This study emphasizes the important role of ePHRs in empowering older adults to manage their health and maintain independence. By utilizing survey weights, the findings can be generalized to the broader US population, making them particularly relevant as the aging demographic continues to grow. Unlike previous research, this study reveals a positive relationship between age and the level of ePHR use, challenging the common assumption that older adults are reluctant to adopt health technologies.

Key factors driving ePHR usage are performance expectancy, effort expectancy, self-efficacy, and issue involvement, all of which offer actionable pathways for increasing the frequency of use among older adults. Raising awareness of the practical benefits and ease of use of ePHRs can encourage more frequent engagement. Additionally, addressing usability concerns and emphasizing the relevance of ePHRs to individual health needs can further promote their adoption. Frequent use of ePHRs empowers older adults to manage chronic conditions, access vital health information, and make informed decisions, ultimately enhancing their quality of life.

As health care systems strive to meet the challenges posed by an aging population, integrating ePHRs into routine care and ensuring equitable access can provide both social and economic benefits. This study lays the groundwork for targeted interventions aimed at bridging the digital divide and fostering a more inclusive, health-empowered society for older adults.