📄 health informatics

Digital Health Adoption, eHealth Literacy, and Trust in AI Among Generation Z University Students in Sri Lanka: An Empirical Study

This study reveals that while Generation Z university students in Sri Lanka possess high eHealth literacy and increasingly rely on AI for health information, their adoption of digital health services remains constrained by significant privacy concerns and a persistent preference for traditional face-to-face consultations.

Original authors: Athukorala, S. C.

Published 2026-07-28✓ Author reviewed ⓘ

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Athukorala, S. C.

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

Imagine the world of healthcare as a giant, bustling library. For decades, the only way to get a book (or medical advice) was to walk up to a specific, quiet desk manned by a librarian (a doctor) and ask for help. But recently, the library has exploded with new technology: self-checkout kiosks, glowing tablets, and even a super-smart robot librarian that can whisper answers to your questions instantly. This is the realm of Digital Health, where apps, websites, and artificial intelligence (AI) try to help us manage our well-being without leaving our couches.

Two big ideas help scientists understand why people use (or ignore) these new tools. The first is eHealth Literacy, which is just a fancy way of asking: "Do you know how to find good health info online, and can you tell if it's real or fake?" The second is Technology Acceptance, which asks: "Does this tool seem useful to you, and is it easy to use without getting a headache?" We care about this because, while everyone says young people are "digital natives" (born with smartphones in their hands), we don't actually know if they trust these digital tools enough to replace a real doctor, especially in countries like Sri Lanka where the library is still building its new wings.

The Story of the Digital-First Generation in Sri Lanka

A researcher named Susith Athukorala decided to peek behind the curtain to see what's really happening with Gen Z university students in Sri Lanka. Think of this study as a giant, digital "show of hands" survey. The researcher asked 172 university students (mostly from private colleges and mostly studying medicine or health) a bunch of questions about how they handle their health.

The "Digital Native" Surprise
The students were surprisingly savvy. They scored high on "eHealth Literacy" (an average of 3.84 out of 5), meaning they generally feel confident finding and understanding health info online. In fact, 86.6% of them admitted to searching for health answers on the internet in the last six months.

But here is the plot twist that caught everyone's attention: AI is the new favorite detective.
While Google was still the most popular place to look (84.3%), the second most popular source wasn't YouTube or social media. It was Generative AI tools like ChatGPT and Gemini, used by 57.6% of the students! That's more than the 44.2% who used YouTube videos. It seems that when these students have a health question, they are turning to the AI robot librarian more often than the video channel. Interestingly, the medical students were even more likely to use these AI tools than students studying other subjects.

The "Trust Gap"
However, just because they use the tools doesn't mean they trust them completely.

The AI Hesitation: While they use AI to search, their trust in AI making actual medical recommendations is low (an average of 3.15 out of 5). They treat the AI like a quick fact-checker, not a final judge.
The Privacy Worry: Half of the students (50.0%) are worried about their personal data being stolen or misused.
The "Real Doctor" Preference: Despite all the cool tech, 76.7% of the students still said they would rather see a doctor face-to-face than use a digital platform for a consultation.

What Makes Them Click?
The study used some math (called regression analysis) to figure out what actually pushes a student to choose a digital app over a physical clinic. It turns out, it's not just about how "useful" they think the app is. The two biggest drivers are:

Ease of Use: If the app is clunky or hard to figure out, they won't use it.
Trust in AI: If they believe the AI's suggestions are safe, they are more likely to try the digital route.

Surprisingly, simply having high "eHealth Literacy" (knowing how to search) didn't directly make them prefer digital visits. They might know how to use the internet, but that doesn't automatically make them want to swap a real doctor for a screen.

The Roadblocks and the Roadmap
The study found that the biggest reasons students don't switch to digital health are:

They just prefer talking to a real human (76.7%).
They are scared of privacy leaks (50.0%).
They don't even know what local digital services exist (43.6%).

To fix this, the paper suggests that if Sri Lanka wants to get more students using digital health, the apps need to be incredibly simple to use, and the system needs to be crystal clear about how it protects private data. They also suggest a "hybrid" approach: using AI to get quick answers, but having a real doctor check the work before it counts as medical advice.

The Bottom Line
The paper concludes that Sri Lankan Gen Z students are ready and willing to use digital tools, and they are already using AI to learn about their health. But, they aren't ready to ditch the human doctor just yet. They want a system that is easy to use, keeps their secrets safe, and combines the speed of AI with the safety of a real medical professional. Until those bridges are built, the face-to-face consultation remains the king of the castle.

Technical Summary: Digital Health Adoption, eHealth Literacy, and Trust in AI Among Generation Z University Students in Sri Lanka

Problem Statement

Despite the rapid global integration of digital health technologies (mHealth, telemedicine, and AI), empirical data regarding their adoption among "digital native" Generation Z cohorts in lower-middle-income countries (LMICs) remains scarce. Specifically, there is a lack of understanding concerning how young adults in South Asian contexts, such as Sri Lanka, navigate the intersection of high digital literacy, emerging AI tools, and traditional healthcare preferences. While infrastructure has evolved (e.g., national teleconsultation channels), the systematic adoption of these tools for self-care and medical decision-making by university students is underexplored. Key gaps include the influence of AI trust on broader digital health acceptance and the specific barriers—such as data privacy and cultural preferences for face-to-face interaction—that hinder full adoption.

Methodology

The study employed a quantitative, cross-sectional survey design conducted between March 2025 and March 2026.

Participants: A convenience and snowball sampling method yielded 172 valid responses from Generation Z university undergraduates (aged 18–26) across Sri Lanka. The cohort was predominantly female (65.1%), enrolled in Medical/Health Sciences (73.8%), and attending private universities (94.8%).
Instrumentation: The survey utilized adapted, validated instruments:
- eHealth Literacy Scale (eHEALS): An 8-item scale measuring self-efficacy in locating, evaluating, and applying digital health information.
- Technology Acceptance Model (TAM): Assessed Perceived Usefulness (PU), Perceived Ease of Use (PEOU), Behavioral Intention (preference for digital vs. physical care), Trust in AI Recommendations, and Data Privacy Trust.
- Health Information National Trends Survey (HINTS): Captured online search behaviors, source trustworthiness, and specific concerns.
Statistical Analysis: Data were processed using Python (pandas, statsmodels). Reliability was confirmed via Cronbach's alpha (eHEALS $\alpha=0.863$ ; TAM $\alpha=0.800$ ). Analysis included descriptive statistics, Chi-Square ( $\chi^2$ ) tests for independence, Pearson correlations, and Ordinary Least Squares (OLS) multiple linear regression models to identify predictors of digital platform preference and perceived usefulness.

Key Results

1. High Literacy and Utility, Low Trust in AI:
Participants demonstrated high overall eHealth literacy (Mean = 3.84 ± 0.58) and strong endorsement of digital health utility (Mean = 3.99 ± 0.59). However, "Trust in AI Recommendations" scored the lowest among all TAM constructs (Mean = 3.15 ± 0.89).

2. Emergence of Generative AI as a Primary Information Source:
While Google/Web Search remained the most frequent source (84.3%), Generative AI tools (ChatGPT, Gemini, Copilot) emerged as the second most utilized channel (57.6%), surpassing YouTube (44.2%), medical websites (27.3%), and social media (26.2%). Medical students utilized AI tools significantly more than non-medical peers (67.2% vs. 37.8%; $\chi^2=8.70, p=.003$ ).

3. Predictors of Digital Platform Preference:
Multiple linear regression (Model 1, $R^2 = .352$ ) identified that preference for digital platforms over physical visits was significantly predicted by:

Perceived Ease of Use (PEOU): $\beta = 0.371, p = .001$
Trust in AI Recommendations: $\beta = 0.370, p < .001$
Notably, eHealth literacy scores and Perceived Usefulness were not direct independent predictors of the preference to replace physical visits. However, in Model 2, eHealth literacy was the strongest predictor of Perceived Usefulness ( $\beta = 0.625, p < .001$ ).

4. Barriers and Enablers:

Primary Barrier: A strong preference for face-to-face consultations (76.7%) and concerns regarding personal data security (50.0%).
Primary Enablers: Ease of use/accessibility (52.3%), professional endorsements (50.6%), and affordable/free services (45.9%).
Online Concerns: Misinformation/fake news (70.3%) and privacy risks (55.2%) were the top concerns.

Key Contributions

Empirical Evidence in LMIC Context: The study provides rare empirical data on Gen Z digital health behaviors in Sri Lanka, challenging the assumption that high digital literacy automatically translates to high adoption of digital clinical care.
AI Utilization Patterns: It documents a significant shift in health information seeking, where conversational AI has overtaken traditional video and social media platforms, particularly among health-science students.
Decoupling Literacy from Adoption: The findings reveal a nuanced relationship where high eHealth literacy drives the perception of usefulness but does not directly drive the preference to replace physical consultations. Instead, interface simplicity (PEOU) and specific trust in AI are the critical gatekeepers for adoption.
Privacy as a Structural Barrier: The study quantifies data privacy concerns as a primary impediment, equal in weight to cultural preferences for physical interaction.

Significance and Claims

The paper claims that while Generation Z students in Sri Lanka exhibit high digital readiness and substantial reliance on AI for information triage, this does not equate to a willingness to fully transition to digital-only care. The authors argue that the gap between high perceived utility and low adoption rates is bridged not by increasing literacy, but by addressing interface simplicity and trust in AI recommendations.

The study concludes that institutional deployment of digital health in Sri Lanka must move beyond standalone apps toward hybrid clinical workflows. These systems must integrate automated digital convenience with professional medical oversight to mitigate privacy fears and align with cultural expectations for face-to-face interaction. The authors emphasize that without transparent data governance (aligned with the Sri Lanka Personal Data Protection Act) and clinician-in-the-loop AI systems, high digital literacy alone will not overcome the barriers to widespread digital health adoption.

Note: The study acknowledges limitations including a sample skewed toward private university and medical students, the cross-sectional nature preventing causal inference, and reliance on self-reported data.