This study was designed as an interventional RCT with a double-blind design, in which both participants and the principal investigator or statistical analyst were unaware of group assignments. The trial aimed to compare the effectiveness of an electronic nursing guidance platform on environmental hormones—specifically phthalates (eHealth intervention)—with that of conventional paper-based health education in improving women’s phthalate-related health literacy. The intervention period lasted 6 months, with assessments conducted at baseline (T0), 3 months after the intervention (T1), and 6 months after the intervention (T2). The study was reported in accordance with the CONSORT-EHEALTH (Consolidated Standards of Reporting Trials of Electronic and Mobile Health Applications and Online Telehealth) checklist (Checklist 1).

The required sample size was estimated using G*Power (version 3.0) for a repeated-measures design (F test, effect size f=0.25, 2-tailed α=.05, power=0.80, 3 measurement points). The minimum required sample size was 86 participants in total. Considering an estimated 30% attrition rate due to the COVID-19 pandemic, the target sample size was set at 120 participants (60 per group).

Participants were recruited from the outpatient departments of obstetrics and gynecology and family medicine at a regional teaching hospital in Taipei City, Taiwan. After obtaining institutional approval and ethics clearance, trained research assistants recruited participants on-site using convenience sampling. A total of 120 women were assessed for eligibility; 6 were excluded, and 114 eligible participants were randomized.

Inclusion criteria were as follows: (1) women aged 20 years or older, (2) ability to communicate in Mandarin or Taiwanese, and (3) willingness to participate with signed informed consent. Exclusion criteria included being male, being younger than 20 years of age, or having a diagnosed psychiatric or cognitive disorder. Eligible participants were randomly assigned to either the intervention group (n=58) or the control group (n=56).

A sequentially numbered, opaque, sealed envelope procedure was used for randomization and allocation concealment. The random allocation sequence and envelopes were prepared and safeguarded by an independent research assistant who was not involved in participant recruitment, outcome assessment, data management, or statistical analysis. After eligibility was confirmed and the baseline assessment was completed, envelopes were opened sequentially according to enrollment order.

Participants were informed only that they would receive health education and follow-up related to environmental hormones and phthalate exposure. They were not informed that the study included 2 different intervention conditions and were not told of their formal group allocation. Although participants were aware of the educational materials they personally received, they did not know whether those materials corresponded to the intervention or control condition.

Separate research assistants were assigned exclusively to the intervention and control groups and were responsible for intervention delivery. The principal investigator did not participate in sequence generation, envelope opening, participant allocation, or intervention delivery. Group identities were masked using coded labels in the analytic dataset, and the allocation codes were disclosed to the principal investigator and the statistical analyst only after the completion of the primary analyses.

Analyses were conducted in SPSS (version 26.0; IBM Corp). Continuous data were expressed as means (SD) and categorical data as frequencies (%). Baseline differences were assessed using 2-tailed t tests and χ² tests. Within-group changes (T0-T1-T2) were analyzed via paired 2-tailed t tests, and between-group differences via independent 2-tailed t tests. Longitudinal effects were examined using generalized estimating equations (GEE) including group, time, and group × time interaction, controlling for significant baseline covariates. Analysis of covariance (ANCOVA) was additionally performed to compare total and domain-specific health literacy scores between groups at 6 months after adjustment for the corresponding baseline scores. The homogeneity-of-regression-slopes assumption was examined before each ANCOVA. Adjusted mean differences and 95% CIs were reported. Participants were analyzed according to their original randomized group assignments. GEE analyses incorporated all available observations from the 3 assessment points without imputing missing outcome data. Participant attrition was reported at each time point. All statistical tests were 2-tailed, with a significance level of α=.05.

This study was approved by the Institutional Review Board of Cheng Hsin General Hospital, Taipei, Taiwan (CHGH-IRB number (864)110-10) and conducted in accordance with the Declaration of Helsinki. All participants provided written informed consent prior to enrollment and could withdraw at any time without affecting their medical rights. The data were anonymized and securely stored with access limited to authorized researchers. The trial was registered at ClinicalTrials.gov (NCT05024227).

This manuscript was prepared in accordance with the CONSORT-EHEALTH extension [42] to ensure comprehensive and transparent reporting of this web-based RCT.

A total of 114 women were enrolled and randomized in the intervention group (n=58) and the control group (n=56). After the 6-month follow-up, 1 participant from each group withdrew, resulting in 112 participants who completed the study (intervention group, n=57; control group, n=55; Figure 2).

At baseline, there were no significant differences between groups in most demographic or physiological characteristics, indicating good group homogeneity. The mean age of participants was 36.9 (SD 12.3) years (intervention group: 35.0, SD 10.7; control group: 38.8, SD 13.7; P=.098). The mean BMI was 23.3 (SD 4.5) in the intervention group and 24.7 (SD 5.3) in the control group (P=.12). Overall, 59.60% (n=68) of the participants were single, and 91.20% (n=104) had a college-level education or higher. The most frequently reported family medical histories were metabolic syndrome (n=77, 67.50%) and diabetes (n=57, 50.00%; Table 1).

However, there were significant differences between groups in the waist circumference (intervention group: 73.9, SD 9.3 cm; control group: 80.5, SD 15.7 cm; P=.038) and pregnancy experience (n=22, 37.90% vs n=36, 64.30%; P=.005; Table 2). These variables were therefore included as covariates in subsequent analyses to control for potential confounding effects.

Overall, both groups were comparable in baseline demographics, lifestyle behaviors, disease history, and reproductive characteristics, supporting the validity of the subsequent intervention outcome comparisons.

Phthalate-related health literacy was assessed at baseline (T0), 3 months (T1), and 6 months (T2). The intervention group’s mean total health literacy score increased from 68.10 (SD 9.76) at T0 to 76.29 (SD 7.77) at T1 and 84.34 (SD 10.76) at T2, whereas the control group’s score increased from 76.07 (SD 13.73) to 79.96 (SD 13.97) and 81.32 (SD 12.51), respectively. Although the intervention group had a lower baseline score, the difference between the groups decreased over time. At T2, the intervention group also had a significantly higher unadjusted score in the health promotion domain than the control group (10.35 vs 9.76; P=.04; Figure 3).

The adjusted GEE analysis showed a significant group × time interaction for total health literacy at T2. Compared with the control group, the intervention group demonstrated an additional increase of 9.93 points from T0 to T2 (95% CI 5.31‐14.56; Wald χ²₁=17.74; P<.001). Significant group × time interactions were also observed in the health care (β=1.52, 95% CI 0.65‐2.38; P=.001), disease prevention (β=1.32, 95% CI 0.57‐2.07; P=.001), and health promotion domains (β=1.12, 95% CI 0.43‐1.80; P=.001; Table 3).

ANCOVA results were consistent with the GEE findings. After adjustment for baseline scores, the intervention group had significantly higher total health literacy at T2 than the control group (F_1,109=11.43; P=.001; adjusted mean difference 7.15, 95% CI 2.96‐11.34). Significant adjusted between-group differences were also observed in the health care (adjusted mean difference 0.74, 95% CI 0.04‐1.43; P=.04), disease prevention (0.88, 95% CI 0.23‐1.52; P=.008), and health promotion domains (0.90, 95% CI 0.37‐1.44; P=.001; Table 4). Changes in total and domain-specific phthalate-related health literacy over time in the intervention and control groups are illustrated in Figure 4.

This randomized, double-blind controlled trial provides empirical evidence that an eHealth educational intervention can effectively enhance women’s environmental health literacy related to phthalate exposure, with significant improvements across all 3 literacy domains—health care, disease prevention, and health promotion.

A key finding was the positive association between user engagement and learning outcomes. Participants with higher platform engagement had higher total and domain-specific health literacy scores at T2. This finding suggests that user engagement may be an important correlate of learning reinforcement and health literacy outcomes. Frequent exposure to interactive educational content may facilitate habit formation, enhance cognitive retention, and strengthen motivation to adopt preventive behaviors—critical processes in the knowledge-to-action continuum.

These results suggest that digital health platforms, when designed according to theoretical models such as the Integrated Model of Health Literacy, can successfully operationalize the 4 competencies of accessing, understanding, appraising, and applying health information and were associated with measurable improvements in phthalate-related health literacy.

Our findings are consistent with the broader literature indicating that digital health interventions can improve health literacy by supporting users’ ability to access, understand, appraise, and apply health information. Previous studies have shown that web- and app-based educational programs incorporating multimedia content, interactivity, and repeated access can strengthen health literacy, self-management, motivation, and confidence across diverse populations [20,23-25,44].

However, the present study addresses a distinct and comparatively understudied area: health literacy related to environmental exposure. Phthalate exposure remains an important environmental health concern because of its associations with endocrine disruption, metabolic disorders, reproductive abnormalities, and hormone-sensitive cancers [5,45,46]. Although Taiwan strengthened regulatory controls following the 2011 plasticizer contamination incident, biomonitoring studies have continued to document phthalate exposure among Taiwanese adults and pregnant women [4,17,47-49].

Women may experience additional exposure through frequent contact with cosmetics, personal care products, and household materials. Therefore, an eHealth intervention specifically designed to strengthen women’s ability to identify, interpret, evaluate, and apply phthalate-related health information is both timely and relevant. The integration of structured educational content, behavioral guidance, multimedia materials, quizzes, and repeated digital engagement is consistent with established principles of technology-assisted health education. Nevertheless, because evidence from general digital health literacy interventions does not directly demonstrate effectiveness for environmental exposure reduction, the findings of the present study should be interpreted primarily in relation to prior interventions addressing phthalates and other endocrine-disrupting chemicals.

Previous interventions addressing phthalates and other endocrine-disrupting chemicals have used educational and behavioral strategies to improve exposure-related knowledge and reduce potentially harmful practices. In Taiwan, Chen et al [26] developed 7 behavioral strategies for girls with elevated phthalate exposure, including more frequent handwashing, avoiding plastic containers and plastic food coverings, refraining from microwaving food in plastic, and reducing the use of cosmetics and personal care products. Following the intervention, several exposure-reduction behaviors were associated with lower urinary concentrations of specific phthalate metabolites. These findings indicate that concrete and context-specific behavioral guidance may contribute to measurable reductions in exposure.

Digital approaches have further expanded the delivery of environmental exposure education. Kim et al [27] conducted an RCT of a web-based behavioral intervention among mothers with young children. The intervention included educational videos, an interactive game for identifying endocrine disruptors in the home, information on potential exposure sources, online resources, and a question-and-answer component. Compared with written information, the web-based program resulted in reductions in several urinary biomarkers, including phthalate metabolites, bisphenol A, and parabens after 1 month. This study supports the value of combining environmental health information with interactive tools and practical exposure-reduction guidance.

Internet-based environmental education has also been applied during pregnancy. In a nonrandomized controlled study, Kim et al [50] delivered an environmental prenatal education program through videoconferencing. The program improved perceived environmental severity, response efficacy, and overall environmental health perception, although significant changes in personal and community environmental behaviors were not observed. These findings suggest that improvements in awareness and risk perception may precede behavioral change and that sustained reinforcement may be required to translate knowledge into routine preventive practices.

More recently, Tomsho et al [28] developed a 1-hour virtual educational intervention for reproductive health professionals covering routes of phthalate exposure, associated reproductive health effects, and strategies for discussing exposure reduction with patients. The intervention improved phthalate-related environmental reproductive health literacy and substantially increased the reported frequency with which clinicians discussed phthalates with patients. This work illustrates how virtual education can strengthen not only individual knowledge but also the communication of environmental health information within clinical practice.

Collectively, these studies are consistent with our findings that technology-assisted environmental health education can improve the capacity to understand and respond to environmental exposure risks. However, our study extends the existing evidence in several ways. First, the intervention was specifically structured around the Integrated Model of Health Literacy, addressing the competencies of accessing, understanding, appraising, and applying phthalate-related information. Second, the platform combined structured nursing education, audiovisual materials, quizzes, reminders, and practical lifestyle guidance over a 6-month period rather than providing a single brief educational session. Third, the sustained improvement observed across the health care, disease prevention, and health promotion domains suggests that repeated digital engagement may support the maintenance of environmental health literacy beyond immediate postintervention effects.

Nevertheless, evidence across environmental exposure interventions remains heterogeneous. A scoping review by Yang et al [29] identified 26 interventions targeting synthetic phenols, glycol ethers, and phthalates, but only 6 relied solely on education; many studies instead removed or replaced exposure-related products. This distinction is important because educational interventions depend on participants’ ability and opportunity to translate information into behavioral change, whereas product-replacement interventions directly alter exposure sources. Our findings therefore contribute evidence that a scalable educational approach can produce sustained improvements in phthalate-related health literacy, although future studies should determine whether these improvements also lead to objectively measured reductions in exposure.

The positive association between platform engagement and literacy outcomes observed in this study suggests that user engagement may contribute to the effectiveness of digital health interventions. Repeated interactions with multimedia learning materials, quizzes, and tailored push notifications likely reinforced cognitive retention, strengthened motivation, and cultivated self-regulatory behaviors. These iterative feedback loops fostered reflective learning and enabled participants to internalize information, potentially supporting the application of phthalate-related information in daily decision-making. Such findings align with the knowledge-to-action continuum described in the Integrated Model of Health Literacy, wherein sustained exposure, self-efficacy, and reflection collectively facilitate durable behavioral change [18,50].

Nursing interventions, through structured education and health promotion strategies, have long been recognized as essential in addressing low health literacy and improving behavioral outcomes [51]. Within this framework, eHealth platforms represent an evolution of traditional nursing education, extending the reach of health professionals beyond physical clinical boundaries. By integrating behavioral feedback and evidence-based guidance, these platforms facilitate accessible, personalized, and continuous learning experiences. The Phthalates Free platform, for example, incorporated real-time feedback, tailored reminders, and interactive media that allowed participants to engage repeatedly and autonomously—features particularly conducive to adult learning and habit formation. No technical failures, data breaches, or operational disruptions occurred during the trial, further confirming the platform’s stability, usability, and data security safeguards.

From a clinical perspective, this model promotes patient empowerment by integrating information access with active engagement. Nurses and health educators can leverage such tools to reinforce counseling, enhance autonomy, and provide ongoing communication that strengthens preventive behaviors. Continuous reinforcement through digital channels—particularly those embedded within daily communication platforms such as LINE—bridges the gap between episodic clinical encounters and long-term self-management. This aligns with contemporary nursing practice emphasizing participatory care, patient engagement, and evidence-informed prevention [52].

At the public health and policy level, the findings underscore the potential of digital health interventions as scalable, cost-effective tools for promoting environmental health literacy. The Phthalates Free model can inform the development of national or regional digital campaigns targeting women and other vulnerable populations, particularly where traditional outreach channels are limited. Leveraging existing digital infrastructures—such as social media, mobile messaging systems, and eHealth portals—facilitates the broad dissemination of validated health information with minimal resource investment. Furthermore, embedding digital literacy competencies within health promotion policies directly supports several Sustainable Development Goals (SDGs), including SDG 3 (Good Health and Well-being), SDG 5 (Gender Equality), and SDG 12 (Responsible Consumption and Production).

Finally, behavioral insights derived from platform analytics—such as engagement frequency, completion rates, and interaction patterns—offer a powerful foundation for adaptive policy design. Tailoring public health messages based on real-time behavioral data can enhance adherence, minimize information fatigue, and improve message salience among target populations. Integrating behavioral analytics with population-level surveillance systems may further enable policymakers to identify knowledge gaps, optimize the timing and delivery of interventions, and design adaptive communication strategies that maximize public engagement. Collectively, these findings highlight how digitally enabled nursing and public health education can function as a sustainable bridge connecting evidence generation, individual empowerment, and population-level policy translation.

The findings of this RCT may be generalizable to other populations with comparable digital access and educational backgrounds. As the intervention was delivered through a widely used communication platform (LINE) and designed for mobile accessibility, it is applicable to general internet users and community-based health education settings. However, the generalizability to populations with limited internet access, lower digital literacy, or different cultural contexts may be restricted.

It should also be noted that several features implemented in the trial—such as automated push notifications, in-app reminders, and limited human assistance from research staff—were designed to enhance user engagement within the controlled setting. In routine applications without these supportive components, engagement levels and intervention effectiveness might be reduced. Future real-world deployment should therefore consider adaptive mechanisms (eg, AI-driven reminders or community health worker follow-ups) to maintain motivation and ensure sustainable outcomes across diverse populations.

This study’s strengths include its randomized, double-blind controlled design, which minimized bias and enhanced causal inference. The use of both GEE and ANCOVA models strengthened analytical robustness by accounting for within-subject correlations and baseline differences. Furthermore, the integration of engagement metrics provided valuable insights into digital behavior and learning mechanisms, allowing a nuanced understanding of how participation influences outcomes.

However, several limitations should be acknowledged. The study was conducted in a single hospital setting in northern Taiwan, which may limit generalizability to other populations or cultural contexts. Although the Phthalate Health Literacy Scale was specifically developed and validated among Taiwanese women, further psychometric evaluation is warranted before applying it to populations with different cultural, linguistic, educational, or demographic characteristics. Health literacy outcomes were based on self-reported measures, potentially introducing social desirability bias. Due to time and funding constraints, biological indicators (eg, urinary phthalate metabolites) were not collected, preventing direct validation of whether improved literacy led to reduced exposure. The 6-month follow-up period captures short-term effects but may not reflect long-term sustainability.

Finally, although a double-blind protocol and procedural separation were implemented, differences in the educational delivery format may have influenced participant engagement. The Hawthorne effect also cannot be entirely excluded, as awareness of study participation and follow-up may have affected participants’ responses.

Future research should aim to replicate these findings across multicenter, multiethnic, and cross-cultural settings to enhance external validity. Extending the intervention period to 12 months or longer would allow examination of the maintenance and decay patterns of health literacy and behavior change over time. Incorporating objective biomarkers, such as urinary or serum phthalate metabolite levels, would further establish the link between literacy improvement and biological outcomes.

As digital health technology evolves, future interventions should consider incorporating gamification, AI-driven personalization, and social network features to strengthen sustained engagement. Integration with wearable devices or mobile sensors could enable real-time exposure tracking and personalized feedback, connecting environmental data with health behavior analytics.

At the educational level, integrating digital health literacy modules into nursing curricula would equip health care professionals with the competencies needed to design, implement, and evaluate technology-assisted interventions. Cross-sector collaborations among health care, public health, engineering, and data science professionals are recommended to develop sustainable and adaptive eHealth ecosystems.

This RCT demonstrated that the Phthalates Free eHealth intervention improved women’s phthalate-related health literacy over 6 months across the health care, disease prevention, and health promotion domains. Greater platform engagement was associated with better literacy outcomes. The intervention may serve as a practical adjunct to nurse-led education in outpatient and community settings by providing accessible, continuous, and evidence-based guidance on reducing phthalate exposure. Further multicenter studies involving populations with more diverse educational and digital literacy backgrounds are needed to confirm its broader applicability.