At the center of the study is a proposed theoretical model the authors call EHBTAUT, which fuses three established pillars of behavioral science into a single architecture. The first is the eHealth context itself, encompassing how people orient themselves toward health information online. The second is the Health Belief Model, a classic framework that explains health behavior through perceptions of susceptibility, severity, benefits, and barriers. The third is the Unified Theory of Acceptance and Use of Technology 2, or UTAUT2, which captures how performance expectations, effort expectations, social influence, hedonic motivation, and habit shape whether people adopt a given technology. By weaving these together, the model allows the researchers to test, in a single integrated structure, how health beliefs and technology acceptance forces interact to produce actual behavioral intention and real utilization of eHealth services.

The empirical backbone of the work is substantial. The team collected two rounds of primary questionnaire data from young adults between the ages of 18 and 30, ultimately yielding 1,432 valid responses. Participants were assessed on a battery of constructs including health information orientation, online health behavior, eHealth literacy, social interaction needs, internet self-efficacy, and patterns of internet and social media use. The choice of age group was deliberate: young adults are the most digitally fluent generation, and if eHealth platforms are failing to capture them, the problem lies not in access to technology but in how these platforms are designed and communicated. Previous work by some of the same authors had already suggested that social media serves as a key gateway to health information for 18-to-30-year-old college students, making this cohort an ideal proving ground for testing a model of digital health engagement.

Methodologically, the study goes beyond the standard partial least squares structural equation modeling that dominates information systems research. The authors combined PLS-SEM with a probabilistic neural network, a hybrid approach the team labels PLS-PNN. Structural equation modeling identifies the strength and significance of the pathways connecting constructs, while the probabilistic neural network, a technique first formalized by Donald Specht in 1990, captures nonlinear classification patterns in the data that linear path models can miss. This combination reflects a growing movement in business research toward pairing interpretable statistical models with machine learning algorithms to boost predictive accuracy without sacrificing theoretical insight. To guard against fragile findings, the researchers supplemented their analysis with Monte Carlo simulation, repeatedly resampling and perturbing the data to confirm that the estimated relationships remained stable under uncertainty rather than emerging as artifacts of a single dataset.

The findings converge on a clear message: the desire for social interaction and the orientation toward health information are not peripheral factors but central engines of eHealth adoption. Health information-seeking motivations and internet-based information exchange, the study concludes, significantly shape healthcare information design. In practical terms, this means that a platform’s functionality must align with what its target audience actually expects and needs. A young adult who turns to Instagram or TikTok for health content does so partly for information and partly for the social texture surrounding it: comments, shares, peer validation, and community. An eHealth application that ignores this social dimension, offering only a sterile transactional interface, is fighting against the very motivations that bring people online in the first place. Conversely, platforms that weave in interaction opportunities can convert passive browsers into active users.

The study also gives weight to constructs that many commercial health apps undervalue. Internet self-efficacy, the confidence a person feels in their ability to navigate online environments, emerged as a meaningful contributor to acceptance, echoing earlier findings from online education research where interaction and self-efficacy predicted satisfaction. eHealth literacy, the capacity to find, appraise, and apply health information from electronic sources, similarly conditions whether digital health content translates into health action. These findings carry a pointed implication for health equity: populations with lower digital confidence or lower eHealth literacy may be systematically excluded from the benefits of digital health unless platforms are deliberately designed to lower those barriers, through simpler interfaces, clearer guidance, and built-in support.

What distinguishes this research from most survey-based studies is its final phase: an application-oriented demonstration. Rather than leaving the empirically supported mechanisms on paper, the team translated them into an iOS-based eHealth application and evaluated its practical relevance with 90 young adult participants. This design-to-validation loop is rare in information systems scholarship and gives the findings a concrete product dimension. It demonstrates that constructs measured in a questionnaire can be operationalized as design features, such as social interaction affordances and information architecture calibrated to user information needs, and that the resulting application resonates with the intended audience. For healthcare providers and technology firms, this offers a template for evidence-based product development in the digital health space.

The broader stakes are considerable. The eHealth market, spanning telemedicine, health information systems, mobile health, and e-pharmacy, has been projected to grow dramatically through 2030, and healthcare systems worldwide face chronic pressure on resources. The study’s framing of eHealth communication as a lever for the efficient utilization of healthcare resources highlights the practical payoff: when patients use digital tools effectively, they can make better-informed decisions, reduce unnecessary visits, and engage in preventive behavior, easing strain on overloaded systems. At the same time, the research implicitly acknowledges the darker currents of the online health ecosystem, including the infodemic of health misinformation that surged during the COVID-19 pandemic and the privacy concerns that shape acceptance of smart health technologies. Building platforms around verified information needs and genuine social interaction may be one of the most effective counters to misinformation, since trusted, well-designed channels can crowd out less reliable sources.

For the designers of the next generation of health applications, the study’s recipe is deceptively simple: start with the user’s information orientation and social needs, build confidence through ease of use and literacy support, and let health beliefs and technology acceptance theory guide the messaging. The researchers, whose work was approved by an Institutional Review Board and who report no competing financial interests, have made their data and materials available from the corresponding author upon request. As healthcare continues its migration to screens of every size, this research offers both a diagnostic of why adoption has lagged among the most connected generation and a validated blueprint for closing the gap between the digital health systems institutions build and the digital health behaviors people actually practice.

Subject of Research: Factors driving eHealth adoption and communication effectiveness among young adults

Article Title: Toward Effective eHealth Communication for Efficient Utilization of Healthcare Resources: A Social Interaction and Information Needs Perspective

Article References: Wang, Y., Prybutok, G., Gulzari, A., Peng, X., Prybutok, V., Lu, Y., & Cheng, F. (2026). Toward Effective eHealth Communication for Efficient Utilization of Healthcare Resources: A Social Interaction and Information Needs Perspective. Information Systems Frontiers. https://doi.org/10.1007/s10796-026-10819-y

Image Credits: AI Generated

DOI: 10.1007/s10796-026-10819-y

Keywords: eHealth, health communication, UTAUT2, Health Belief Model, eHealth literacy, social interaction, internet self-efficacy, PLS-SEM, probabilistic neural network, Monte Carlo simulation, healthcare platforms, young adults

The study, published in the Journal of New Approaches in Educational Research, employed a sequential explanatory mixed-methods design. In the quantitative phase, teachers from across Spain completed a 26-item questionnaire covering their professional profiles, school characteristics, available infrastructure, classroom technology use, perceptions of technology, and personalization practices. The researchers then fitted a structural equation model (SEM) to the data, using the lavaan package in R and following a rigorous three-step analytical approach: descriptive statistics and normality testing, confirmatory factor analysis (CFA) to establish construct validity, and multigroup invariance testing to determine whether the same statistical model could be applied to different groups. Model fit was excellent, with Tucker–Lewis and Comparative Fit Index values above 0.95 and error indices below 0.05.

That statistical rigor paid off when the multigroup analysis revealed something unexpected: while men and women could be validly compared within a single model, primary and secondary education could not. The constructs underlying technology use were simply not equivalent across the two stages. When the researchers fitted separate models for each stage, they found that in primary education, technology use was strongly determined by the availability of infrastructure in the school, with a standardized path coefficient of 0.600 — while teacher training and perception failed to reach statistical significance. In secondary education, by contrast, all three factors mattered: infrastructure, digital training, and teachers’ perceptions of what technology achieves all shaped classroom use. In other words, primary school teachers use technology mainly when it is put in front of them; secondary school teachers use it when it is available, when they know how to use it well, and when they believe it works.

The interviews illuminated why this divide exists. Primary teachers reported that the one-device-per-student model had been achieved in only a fraction of their schools, that equipment was often old and rarely renewed, and that internet connectivity was patchy. Their software use was limited to basic tasks — information searches, presentation programs, test creation tools. Sixty percent of the primary teachers interviewed expressed negative views of educational technology, 80 percent said it had not improved academic outcomes, and 70 percent reported that they did not actually personalize content, even though most believed they could. Secondary teachers, working with more demanding curricula and more autonomous students, showed the opposite pattern: better equipment, a wider variety of software including simulators and data analysis tools, and more tangible personalization practices, such as adapting materials for low-achieving students, gifted students, and students with cognitive or visual impairments.

Perception emerged as the study’s most powerful psychological variable. The researchers defined it as teachers’ expectations about whether technology improves student attention, engagement, motivation, and academic outcomes — or does nothing at all. Across both educational stages, perception drove the personalization of content, confirming the study’s hypothesis that what teachers believe about technology shapes whether they adapt materials to meet individual cognitive needs. This finding aligns with the UTAUT2 acceptance model, in which performance expectancy and social influence determine usage intention. Notably, however, perception did not directly drive teachers’ intention to pursue further digital training. Instead, training was linked to teacher profiles — age and years of experience — in both stages, with older and longer-serving teachers having accumulated more training opportunities over their careers.

The qualitative data exposed an uncomfortable truth about the training itself. Most teachers reported having completed fewer than ten formal digital courses and acquiring much of their competence through self-directed learning, which secondary teachers described as ‘incalculable’ in scope. More than half said training courses focused primarily on how to operate tools rather than on their didactic applications. The coordinators confirmed this, explaining that course evaluation was minimal: teachers submitted a classroom application project that was not rigorously graded, and certification amounted to a procedural pass or fail. One coordinator offered a memorable critique: teachers want ‘the flan recipe’ without caring what happens in the oven, and only those who understand what happens in the oven can create genuinely creative lessons. The system, she said, forces trainers to teach simple recipes.

The study also documented an emerging cultural headwind. Several primary teachers voiced resistance to digitalization, citing media coverage of technology’s negative effects on children and policy reversals abroad — such as Sweden’s reevaluation of digital efforts following disappointing PIRLS reading results. The researchers suggest these teachers may have been more influenced by news narratives than by actual institutional information, a mechanism consistent with research on how media shapes public opinion. In primary education, where concerns about child protection carry symbolic and emotional weight, this amplifying factor may reinforce conservative teaching practices and dampen both the willingness to integrate technology and the appetite for training.

There is also a structural peculiarity in the Spanish context: digital competence certification is often pursued not out of professional obligation but as a strategic asset, earning teachers points in public transfer competitions that determine where they can be posted. Two interviewed secondary teachers admitted they pursued ICT training primarily for exactly this reason. Combined with the finding that self-reported digital competence frequently exceeds assessed competence — a gap documented across multiple prior studies — this raises questions about whether certification systems are measuring anything meaningful at all. The authors argue for far more rigorous evaluation of training courses, including indicators of classroom transfer and observable improvement in teaching practice, along with systematic short-, medium-, and long-term follow-up of training impact.

The implications extend well beyond Spain. Governments across Europe and beyond have spent decades pursuing a supply-side strategy: provide the equipment, offer the courses, and assume integration will follow. This study shows why that assumption fails. In primary schools, infrastructure is the gatekeeper — but once passed, training and belief matter little, and usage remains shallow. In secondary schools, training and perception become decisive, but even there, half of the interviewed teachers did not believe personalized teaching was achievable with technology, citing class sizes and workload. If administrators want technology to move beyond replacing worksheets with screens, the authors conclude, they must act directly on teachers’ perceptions — demonstrating in concrete, everyday terms how these tools help solve real classroom problems — while redesigning training to be stage-specific, pedagogically deep, and genuinely evaluated. Each educational stage, the study insists, is its own ecosystem, and one-size-fits-all digitization strategies will keep producing islands of innovation rather than transformation.

The theoretical scaffolding of the study draws heavily on established technology-acceptance research. The UTAUT2 model, developed by Venkatesh and colleagues, holds that performance expectancy, effort expectancy, social influence, and facilitating conditions jointly determine whether a person adopts a technology. The Spanish findings map onto this framework in an instructive way: infrastructure operates as a facilitating condition, perception as performance expectancy, and training as a proxy for the self-efficacy that prior work by Hatlevik and others has identified as the starting point for overcoming everyday classroom obstacles. What the study adds is evidence that the relative weight of these predictors is not fixed but shifts across educational stages, a nuance that single-population acceptance studies have generally been unable to capture.

The distinction between instrumental and pedagogical technology use also connects to broader debates in the field. Frameworks such as TPACK have long argued that technical knowledge alone is insufficient, and that meaningful integration requires the intersection of technological, pedagogical, and content knowledge. The finding that most Spanish training courses emphasized tool operation over didactic application suggests that professional development has not yet internalized this principle, which may explain why self-reported digital competence, measured against the European DigCompEdu reference framework, consistently exceeds assessed competence in prior research.

The mixed-methods design itself deserves attention. Sequential explanatory designs of this kind allow qualitative interviews to probe the mechanisms behind statistically significant paths, and here the interviews revealed dynamics invisible to the questionnaire, such as the role of media narratives in shaping primary teachers’ skepticism and the strategic use of certification for posting competitions. This methodological layering strengthens the credibility of conclusions that might otherwise rest on correlational evidence alone.

Finally, the study’s regional sampling across two Spanish autonomous communities reflects the decentralized structure of Spanish education policy, where regional administrations design their own training offerings. This heterogeneity, while complicating generalization, mirrors conditions in other federal or quasi-federal systems such as Germany’s Länder, where neighboring research has documented similarly uneven integration trajectories. The authors’ call for stage-specific and perception-targeted strategies therefore carries weight for any jurisdiction pursuing supply-side digitization without accounting for the beliefs, profiles, and working conditions of the teachers expected to deliver it.

Subject of Research: Factors determining technology use, digital teacher training, and learning personalization in Spanish primary and secondary education

Article Title: Factors involving technology use, digital training, and teaching personalization: a national and regional mixed-methods study

Article References: Eduardo, C.-C., & María, N.-F. (2026). Factors involving technology use, digital training, and teaching personalization: a national and regional mixed-methods study. Journal of New Approaches in Educational Research, 15(1), Article 20. https://doi.org/10.1007/s44322-026-00067-y

Image Credits: AI Generated

DOI: 10.1007/s44322-026-00067-y

Keywords: educational technology, ICT integration, teacher training, digital competence, personalized learning, primary education, secondary education, structural equation modeling, UTAUT2, Spain, mixed methods, teacher perception