Background

In the context of the COVID-19 pandemic, the global public health environment has faced unprecedented challenges, significantly affecting healthcare-seeking behaviors, particularly among middle-aged and elderly individuals with critical diseases. Taiwan’s COVID-19 outbreak began on January 21, 2020, with measures such as border closures, mandatory mask-wearing, quarantines, and rapid testing in place until March 20, 2023, when restrictions were lifted. As the pandemic spread, heightened infection risks and stringent preventive measures imposed by healthcare institutions led to increased psychological stress and anxiety for patients seeking medical care. These factors influenced their behavioral intentions (BI) to seek healthcare services and their overall health outcomes.

The World Health Organization (WHO) projects that by 2050, the global elderly population will double from 700 million in 2019 to 1.5 billion, with the proportion of elderly individuals rising from 9.1 to 16.7% of the total population. Taiwan mirrors this demographic shift, facing severe challenges due to aging and declining birth rates. As of 2020, Taiwan’s elderly population reached 14.05%, officially categorizing it as an aged society, with projections indicating a super-aged society (20% elderly population) by 20251. The growing elderly demographic is associated with increased healthcare demands, particularly for patients with critical diseases, as defined by Taiwan’s Financial Supervisory Commission, including acute myocardial infarction, coronary artery bypass surgery, end-stage renal disease, post-stroke disability, cancer, severe paralysis, and major organ transplantation2. These conditions are leading causes of mortality in Taiwan, with chronic diseases such as diabetes, hypertension and hyperlipidemia contributing significantly to healthcare demands.

Given these trends, understanding healthcare-seeking behaviors during and beyond the COVID-19 pandemic is critical. Several theoretical frameworks have been commonly utilized in health behavior research, including the Health Belief Model (HBM), Social Cognitive Theory (SCT), Protection Motivation Theory (PMT), and the Theory of Planned Behavior (TPB). While the HBM primarily considers individuals’ perceptions of health risks and severity, SCT emphasizes self-efficacy and social observational learning. PMT focuses on the cognitive processes involved in assessing threats and coping strategies. However, the TPB uniquely integrates attitudes (individual evaluation of behavior outcomes), subjective norms (social influences), and perceived behavioral control (individual capability to perform behaviors), thus providing a comprehensive framework particularly suited to understanding multifaceted healthcare decision-making involving personal, social, and environmental factors.

Owing to the complexity and uncertainty created by the COVID-19 pandemic, traditional models might inadequately address critical contextual dynamics that influence patients’ decisions, such as institutional trust and heightened perceptions of risk. Therefore, this study explicitly extends the traditional TPB framework by incorporating two additional constructs: Perceived Risk (PR) and Trust (TR). The Perceived Risk construct captures individuals’ evaluations of potential adverse outcomes associated with healthcare visits, particularly relevant amid heightened pandemic-related concerns about infection risk, treatment uncertainty, and hospital safety protocols. Trust (TR) examines patients’ confidence and reliance in healthcare providers and institutions, acknowledging the intangible and service-oriented nature of healthcare, as highlighted in Zeithaml’s service evaluation spectrum3,4,5,6,7,8,9.

Study objectives and conceptual framework:

This study aims to investigate healthcare-seeking intentions among middle-aged and elderly individuals with critical diseases using an expanded Theory of Planned Behavior (TPB) model. While previous research has applied TPB in healthcare contexts, most studies have not explicitly incorporated institutional trust or perceived risk, particularly during public health crises. To address this gap, the current study expands the traditional TPB framework by integrating Perceived Risk (PR) and Trust (TR) to capture better the psychological and institutional influences on healthcare decision-making in the post-pandemic era10.

Specifically, the objectives of this study are:

  1. (1)

    To examine the direct effects of traditional TPB constructs (Attitude, Subjective Norm, Perceived Behavioral Control) and the added constructs (Perceived Risk and Trust) on behavioral intention.

  2. (2)

    To analyze the indirect (mediating) effects of secondary sub-constructs (e.g., personal factors, social norms, and risk domains) through primary TPB dimensions on behavioral intention.

  3. (3)

    To validate a hierarchical, second-order structural model where first-order latent variables influence second-order TPB constructs, affecting healthcare-seeking intention.

  4. (4)

    This framework allows for a nuanced understanding of both direct and mediated pathways, providing theoretical coherence and practical insights11,12,13.

Methods

Research model and hypotheses

The Theory of Planned Behavior (TPB), originally proposed by Ajzen (1991), asserts that behavioral intentions (BI) are primarily shaped by Attitude (AT), Subjective Norm (SN), and Perceived Behavioral Control (PBC). This foundational theory is widely applied in healthcare research because it comprehensively encompasses personal, social, and control-related determinants of behavior. Recent studies have shown the benefit of integrating TPB with other theoretical frameworks to better explain behaviors in uncertain and high-risk contexts. For example, Gumasing and Sobrevilla (2023) effectively combined TPB with Protection Motivation Theory (PMT) and ergonomic appraisal to understand protective behaviors during natural disasters. Their work underscores the value of extending TPB by incorporating context-specific constructs, thus enhancing its capacity to analyze complex decision-making processes14,15. However, given the unique complexities associated with healthcare-seeking behaviors during the COVID-19 pandemic, this study incorporates two additional constructs—Perceived Risk (PR) and Trust (TR)—to capture nuanced decision-making under conditions of uncertainty and risk. The following hypotheses were developed based on existing literature:

Attitude (AT)

Attitude, a key construct in TPB, refers to individuals’ positive or negative evaluations of performing specific behaviors. Empirical studies consistently highlight attitude as a primary predictor of healthcare-related behavioral intentions16. To more precisely address attitude formation within healthcare contexts, we delineate three attitude-related sub-dimensions:

  1. (1)

    Personal factors (ATP) These reflect individuals’ intrinsic beliefs, personal health values, and preferences. Paul et al. emphasize that alignment between personal values and health behaviors significantly enhances the likelihood of engaging in healthcare activities, thus supporting hypothesis H1a17.

  2. (2)

    Doctor factors (ATD) Representing patients’ perceptions of their physicians’ professional competence, communication skills, and empathy, these factors significantly shape patient attitudes and healthcare intentions18. Therefore, hypothesis H1b explicitly incorporates doctor-related factors to capture the relational dynamics between patients and healthcare providers.

  3. (3)

    Hospital factors (ATH) Shamsuddin and Yusoff argue that the quality of hospital facilities, equipment, and perceived effectiveness of services critically influence patient attitudes toward seeking medical care19. Thus, hypothesis H1c addresses the critical role institutional environments play in shaping healthcare-seeking intentions.

Subjective norm (SN)

Subjective norm refers to the perceived social pressure from influential reference groups regarding the performance of specific behaviors. According to Bélanger-Gravel and Amireault, and Tucker et al., individuals are significantly influenced by their immediate social networks (family, friends) as well as broader community or societal norms20,21. Thus:

  1. (1)

    Primary group norms (SNP) These encompass the expectations and pressures from close social circles, such as family and close friends. Prior studies consistently demonstrate these groups as key influencers of health behaviors21, supporting hypothesis H2a.

  2. (2)

    Secondary group norms (SNS) This dimension captures broader community influences, including peer groups, social networks, and organizational affiliations. Bélanger–Gravel and Amireault suggest that secondary groups often strongly influence elderly populations, thus supporting hypothesis H2b20.

Perceived behavioral control (PBC)

Perceived behavioral control is defined as the individual’s perception of the ease or difficulty associated with performing a behavior. Extant literature underscores the role of both internal capabilities and external resources in shaping intentions and behaviors14. Specifically:

  1. (1)

    Internal control factors (PBCI) These factors encompass personal abilities, knowledge, self-efficacy, and confidence in managing health behaviors. McEachan et al. argue that these internal factors are critical determinants of health-related intentions, directly supporting hypothesis H3a14.

  2. (2)

    External control factors (PBCE) External constraints or facilitators, such as the availability of healthcare resources, cost considerations, and social support, significantly influence behavioral control perceptions and, therefore, health-seeking intentions14. Thus, hypothesis H3b addresses external influences.

Perceived risk (PR)

Perceived risk addresses individuals’ assessment of potential negative consequences of healthcare actions, a particularly salient factor during health crises such as the COVID-19 pandemic13. We subdivide this construct into three distinct risk dimensions:

  1. (1)

    Financial risk (PRF) Concerns about affordability and potential economic burdens associated with seeking healthcare services have been empirically shown to influence health-seeking intentions negatively3, supporting hypothesis H4a.

  2. (2)

    Social-psychological risk (PRS) Fear, anxiety, and social stigma related to healthcare-seeking behaviors critically affect intentions13. Thus, hypothesis H4b addresses psychological barriers.

  3. (3)

    Ethical privacy risk (PRE) Issues concerning confidentiality and personal data security during medical care significantly impact healthcare behaviors (Ozawa & Sripad, 2013). Therefore, hypothesis H4c specifically targets privacy concerns.

Trust (TR)

Trust involves patients’ confidence in healthcare providers and institutions, serving as an essential factor in uncertain environments such as pandemics22. We distinguish between two forms of trust:

  1. (1)

    Trust in physicians (TRD) Patients’ trust in healthcare professionals’ competence, integrity, and ethical behavior significantly shapes healthcare-seeking behaviors, consistent with prior findings by Piette et al., supporting hypothesis H5a22.

  2. (2)

    Trust in hospitals (TRH) Institutional trust, encompassing organizational reliability, safety protocols, and healthcare quality, has a robust influence on healthcare-seeking intentions, especially in crisis scenarios23,24,25,26. Hence, hypothesis H5b explicitly includes institutional trust.

This hierarchical model enables us to examine mediation effects—i.e., how personal, institutional, and contextual factors exert their influence on intention indirectly through higher-order TPB constructs.

We had explicitly presented the mediation model in Fig. 1 to enhance theoretical alignment and conceptual clarity across the manuscript.

Fig. 1
figure 1

SEM of the Research Model.

Full size image

We employed mediation analysis based on a formal test of indirect effects of Zhao et al., which is appropriate for hierarchical structural equation modeling (SEM). This choice is theoretically grounded in Ajzen’s TPB, which posits that external influences (e.g., risk, trust, personal and social factors) operate through internal behavioral evaluations (i.e., Attitude, Norms, and Control) to form behavioral intentions27.

By incorporating this mediational logic into our research questions and conceptual model, we ensure consistency between the study’s objectives, statistical methods, and theoretical foundations.

The hypotheses are summarized below and illustrated in Fig. 1 for clarity.

Ethical considerations

This study was approved by the Ethics Review Committee of Antai Medical Care Cooperation, Antai-Tian-Sheng Memorial Hospital [approval number: 22-065-C]. All data were anonymized to protect participant confidentiality, in accordance with the Declaration of Helsinki. Informed consent was obtained from all subjects.

Sampling and data collection

A cross-sectional design with purposive sampling targeted middle-aged and elderly individuals (over 45 years old) diagnosed with critical diseases across Taiwan. Data were collected via structured questionnaires from July 2022 to March 2023. Of 550 distributed questionnaires, 526 were valid, yielding a response rate of 95.6%. Inclusion criteria focused on individuals diagnosed with at least one critical illness as defined by Taiwan’s healthcare guidelines. Although the data were collected between July 2022 and March 2023, during the later phase of the COVID-19 pandemic when Taiwan maintained specific public health protocols, the long-term influence of the pandemic on healthcare behaviors persisted. Therefore, the study captures healthcare-seeking intentions at a critical transitional moment, offering highly relevant insights in the post-pandemic era as healthcare systems adapt to sustained behavioral shifts, digital transformation, and lingering public concerns over safety and access.

Quality control measures

  1. (1)

    Pilot testing Conducted with 30 participants to refine questionnaire clarity.

  2. (2)

    Data verification Double data entry was implemented to ensure accuracy and reliability.

  3. (3)

    Anonymity assurance Participants’ identities were anonymized to reduce response bias.

Statistical analysis

Structural equation modeling (SEM) was performed using SmartPLS 3.3.7. Partial least squares (PLS) was chosen for its suitability in complex models and exploratory research28. The analysis included:

  1. (1)

    Model fit Assessment Evaluated using composite reliability, AVE, and Cronbach’s alpha.

  2. (2)

    Bootstrapping 5000 resamples were used to assess path coefficients’ significance28.

  3. (3)

    Indirect and total effects Besides direct effects, the study examined mediation effects through indirect and total effect analyses.

Results

The study included 526 participants, with 308 males (58.6%) and 218 females (41.4%). The prevalence of critical diseases among participants was as follows: hypertension (61.2%), hyperlipidemia (28.5%), heart disease (27.6%), cerebrovascular disease (9.5%), uremia (7.8%), diabetes (26.8%), chronic obstructive pulmonary disease (COPD) (7.8%), liver cirrhosis (7.2%), cancer (6.4%), and other conditions (11.6%). Additionally, 146 participants (27.8%) reported a prior medical history of critical diseases, including acute myocardial infarction (6.8%), coronary artery bypass surgery (2.1%), end-stage renal disease (8.0%), post-stroke disability (1.9%), severe cancer (6.5%), and major organ transplantation or hematopoietic stem cell transplantation (0.4%).

Reliability and validity of measurement model:

According to Hair et al., factor loadings above 0.5 indicate satisfactory item reliability28. In this study, the lowest Average Variance Extracted (AVE) was 0.526, and the highest was 0.893, indicating strong convergent validity. Cronbach’s alpha and composite reliability exceeded 0.7 for all dimensions, with values ranging from 0.776 to 0.960 and 0.847 to 0.961, respectively, confirming high internal consistency (Table 1).

Table 1 Results of reliabilities and AVE.
Full size table

Chin suggested that in the cross-factor loading matrix, the factor loadings of each item within its respective dimension should be higher than its loadings in other dimensions (see Table 2)29. According to the recommendations of Hair et al., the square root of the AVE of each variable should be greater than the correlation coefficients with other variables (see Table 3)28. The results show that in this study, the variables from different dimensions have low correlations, demonstrating discriminant validity.

Table 2 Cross-factor loading matrix.
Full size table
Table 3 Correlations among the constructs and the square root of the AVE.
Full size table

Hypothesis testing:

All 17 hypotheses, covering five primary dimensions and 12 secondary dimensions, were supported ( as Table 4).

  1. (1)

    H1 (Attitude) Personal factors (ATP), doctor factors (ATD), and hospital factors (ATH) significantly influenced attitude (AT), with ATH exerting the strongest influence, followed by ATD and ATP.

  2. (2)

    H2 (Subjective norm) Both primary group (SNP) and secondary group (SNS) significantly influenced subjective norm (SN), with SNS having a more substantial effect.

  3. (3)

    H3 (Perceived behavioral control) Internal (PBCI) and external control factors (PBCE) significantly influenced perceived behavioral control (PBC), with PBCI having a more substantial impact.

  4. (4)

    H4 (Perceived risk) Financial performance risk (PRF), social-psychological risk (PRS), and ethical privacy risk (PRE) significantly influenced perceived risk (PR). PRS had the highest influence, followed by PRF and PRE.

  5. (5)

    H5 (Trust): Trust in hospitals (TRH) and trust in physicians (TRD) significantly influenced trust (TR), with TRH having a more substantial effect.

Table 4 Hypothesis testing.
Full size table

Subgroup analysis: Clear differences in healthcare-seeking intentions emerged among patients categorized by specific chronic conditions. Specifically, significant variations were identified in the behavioral intentions of patients with hypertension (G2), hyperlipidemia (G3), and diabetes (G4) compared to the overall sample (G1). Statistical analysis indicated distinct differences in how TPB constructs influenced healthcare-seeking behaviors across these subgroups. These findings highlight the importance of developing condition-specific interventions tailored to address the unique behavioral determinants within each patient subgroup. (Fig. 2).

Fig. 2
figure 2

SEM of Research model assumptions established by cases of different diseases. Note: *p < 0.05; **p < 0.01; *** p < 0.001; n.s. = not significant. Group 1 (G1): All; Group 2(G2): Cases of Hypertension; Group 3(G3): Cases of Hyperlipidemia; Group 4(G4): Cases of DM.

Full size image

Mediating effect

To assess the presence and nature of mediation mechanisms within our proposed model, we conducted a bootstrapped indirect effect analysis (5000 resamples) and classified the mediating effects using the framework proposed by Zhao et al.27. As shown in Table 5, the majority of paths demonstrated complementary mediation, where both the indirect and direct effects are significant and point in the same direction. This pattern suggests that mediators, such as attitude, perceived risk, and trust, do not fully absorb the effect of their respective antecedents but rather enhance and explain part of the path to behavioral intention. For instance, both trust in hospitals and trust in doctors significantly mediated their respective antecedents’ effects on behavioral intention (β = 0.201 and 0.097, respectively), indicating the central role of trust in healthcare decision-making. Similarly, the pathways from personal, doctor, and hospital-related attitudes to behavioral intention, mediated by attitude, were all significant and consistent in direction, reinforcing the partial mediating role of cognitive evaluations. Only one pathPBCE → PBC → BI did not reach statistical significance (t = 1.915, p > 0.05), suggesting no mediation in this case. This indicates that external behavioral control factors, such as resource availability, may exert a direct rather than mediated influence on intention.

Table 5 Mediating effect.
Full size table

This analysis confirms that the integrated TPB framework, incorporating perceived risk and trust, demonstrates robust mediational dynamics, supporting both theoretical refinement and practical interpretation of behavioral pathways in healthcare-seeking contexts.

Discussion

This study offers significant theoretical and empirical contributions by enhancing the understanding of healthcare-seeking behaviors among middle-aged and elderly individuals with critical illnesses. Our expanded Theory of Planned Behavior (TPB) framework incorporates Perceived Risk (PR) and Trust (TR)30, providing richer insights into decision–making processes, particularly during global health crises such as the COVID-19 pandemic. First, our findings highlight that institutional trust significantly surpasses traditional TPB constructs, such as Attitude (AT) and Perceived Behavioral Control (PBC), in predicting healthcare-seeking intentions. This result supports Piette et al. identification of institutional trust as a critical determinant influencing patient decisions during uncertain situations22. Contrary to Gibson et al. balanced view between institutional and interpersonal trust23,24,25,26, our research distinctly demonstrates a more substantial influence of institutional trust, likely driven by concerns over infection control, hospital safety protocols, and institutional preparedness, as echoed by Han et al. during public health emergencies31,32,33,34,35,36. Practically, healthcare institutions should prioritize transparent communication about safety measures, infection control protocols, and emergency response capabilities to bolster patient trust and engagement.

Second, in examining Attitude (AT), hospital-related factors (ATH) emerged as the most influential, followed by physician-related (ATD) and personal factors (ATP). This aligns with Zeng et al. emphasis on the healthcare infrastructure’s impact on shaping patient attitudes16. The heightened importance of hospital infrastructure during the pandemic suggests that healthcare institutions should visibly invest in facility cleanliness, staff professionalism, and quality accreditations to positively influence patient perceptions and build trust.

Third, our findings regarding Subjective Norm (SN) indicate that secondary group norms (SNS), such as peer and community networks, exert a greater influence than primary group norms (SNP), like those of family members. This finding resonates with Bélanger-Gravel and Amireault’s research, reflecting Taiwan’s demographic shift towards aging populations, particularly in “empty-nest” households18. Thus, community-based interventions, including peer education programs, local health forums, and community support groups tailored to older adults, are crucial in shaping subjective norms and promoting healthcare utilization.

Fourth, consistent with McEachan et al. findings, internal control factors (PBCI)—specifically self-efficacy and health literacy—were more influential than external control factors (PBCE)14. Contrary to Shie et al. focus on healthcare accessibility34, Taiwan’s universal healthcare system likely minimizes external barriers. Therefore, interventions emphasizing patient education, chronic disease management training, and health coaching are crucial in enhancing self-efficacy, enabling older adults to navigate healthcare proactively.

Fifth, the influence of Perceived Risk (PR), notably Social-Psychological Risk (PRS), was more pronounced than financial or privacy concerns. This differs from Hsieh et al. emphasis on financial factors3 and aligns with studies by Dryhurst et al., Huang et al., and Han et al., which highlight psychological stresses such as fear and uncertainty during pandemics13,35,36. To address these psychological barriers, targeted interventions are recommended, including:

  1. (1)

    Transparent communication to clearly articulate infection risks and safety protocols.

  2. (2)

    Mental health support services are integrated within healthcare facilities.

  3. (3)

    Crisis-oriented patient education to reduce uncertainty around medical procedures and expected outcomes. These efforts can alleviate emotional barriers and encourage timely care-seeking among anxious or hesitant patients.

Sixth, our comparative analysis across chronic conditions offers additional insights. Patients with hypertension and hyperlipidemia showed stronger associations with healthcare-seeking intentions due to the severity of potential cardiovascular complications. This aligns with findings by Yoshimi et al. and Hsieh et al.3,7. Conversely, patients with hyperglycemia demonstrated weaker associations, possibly due to milder symptomatology, as reported by Chudasama et al.37. Tailored interventions, such as condition-specific education and targeted risk communication, are thus recommended, particularly for patients with less overt symptoms or those with progressive conditions.

Ultimately, our expanded TPB model remains relevant beyond the pandemic, offering a robust framework for understanding healthcare behaviors in an evolving landscape characterized by digital transformation and persistent risk perceptions. Trust, perceived risk, and self-efficacy will continue influencing patient decisions, particularly as telemedicine and AI-based healthcare solutions expand. Hence, integrating ethical transparency, digital health literacy, and system-wide innovations into healthcare strategies is crucial for maintaining patient engagement and trust in the long term22,30.

Limitations and future directions

Several limitations of this study should be acknowledged to enhance clarity and inform future research directions. First, the cross-sectional design restricts the ability to draw definitive causal inferences, and the reliance on self-reported data introduces potential biases, including recall and social desirability biases. Second, although moderation analysis was initially considered to investigate how demographic or contextual factors might influence the observed relationships, it was ultimately excluded due to the complexity and primary focus on mediation pathways.

Future research incorporating moderation analysis could provide valuable insights into factors that strengthen or weaken these relationships. Additionally, longitudinal studies are recommended to better understand how healthcare-seeking behaviors evolve, particularly in response to prolonged health crises. Furthermore, cross-cultural and international comparative studies would be beneficial to identify variations in healthcare behaviors influenced by diverse healthcare systems and cultural norms. Lastly, given the increasing adoption of digital healthcare platforms, future studies should investigate how digital interactions influence trust and perceived behavioral control within frameworks such as the Theory of Planned Behavior. Addressing these limitations will substantially contribute to the development of targeted interventions and policies, thereby facilitating equitable access to healthcare for vulnerable populations.

Conclusion

This study employed an expanded Theory of Planned Behavior (TPB) model to analyze healthcare-seeking intentions among middle-aged and elderly individuals with critical diseases, explicitly incorporating institutional trust and perceived risk alongside the traditional TPB constructs of attitude, subjective norms, and perceived behavioral control. The results reveal that institutional trust emerged as the most influential determinant, surpassing conventional TPB dimensions. This contributes meaningfully to behavioral health theory by underscoring that trust in healthcare systems—not just individual providers—plays a dominant role in shaping patient decisions, particularly during times of uncertainty and health crises.

In the post-pandemic era, healthcare systems face persistent challenges, including managing complications from deferred care, rebuilding institutional trust, addressing psychological hesitancy, and preparing for future public health emergencies. Our findings demonstrate that trust and perceived risk COVID-19 crisis. The expanded TPB framework presented here is thus well-suited to evaluating behavioral intentions under prolonged uncertainty and rapid digital transformation conditions. Moreover, our results underscore the increasing importance of institutional-level factors such as hospital safety, service transparency, and the usability of digital healthcare platforms.

Practical implications for stakeholders

To address these challenges and translate findings into actionable strategies, several concrete recommendations are proposed for healthcare institutions, government policymakers, and community organizations:

For Medical Institutions: Strengthen institutional trust by publicly communicating hospital safety protocols, infection control measures, and quality metrics (e.g., success rates, readmission statistics). Initiatives such as publishing safety dashboards or patient satisfaction testimonials on official websites can enhance institutional credibility. Clear labeling of infection control zones and professional staff behavior can further reinforce perceived reliability.

For Government and Public Health Agencies: Design large-scale, culturally sensitive public education campaigns that demystify hospital visits and treatment procedures, especially during health crises. These should target psychological risks such as fear, anxiety, and stigma, which were found to be highly influential deterrents to care-seeking.

For Policymakers: Integrate community-based health promotion strategies—especially for elderly populations in “empty-nest” situations—by building neighborhood health networks, peer support groups, and local screening programs that leverage social norms to promote early medical engagement.

For Society and Care Providers: Empower older adults with digital health literacy programs and chronic disease self-management workshops. These efforts build internal control (self-efficacy) and facilitate confident participation in in-person and telehealth consultations.

For Telehealth and AI Developers: Ensure platforms prioritize user-friendly designs and robust privacy standards to mitigate ethical risk perceptions and build trust in virtual care. Providing real-time human support, tutorials, and simple interface navigation are crucial to maintaining engagement.

These targeted strategies reflect the study’s core findings, offering tangible pathways to increase healthcare utilization and reduce disparities in care access for aging populations with critical conditions.

Theoretical and strategic contributions

Theoretically, this research enhances the TPB model by formally integrating trust and perceived risk as fundamental constructs, advancing its applicability to modern healthcare contexts. These insights extend the TPB’s relevance beyond traditional health behavior models by capturing systemic and emotional determinants critical in both crisis and recovery phases.

In conclusion, this study offers both conceptual innovation and practical value. By identifying institutional trust, perceived psychological risk, and self-efficacy as pivotal drivers of behavior, it equips healthcare institutions, policymakers, and practitioners with a clear roadmap for improving patient engagement, enhancing access equity, and building resilient, trust-based healthcare systems in the post-pandemic era and beyond.