The Health Belief Model (HBM) was initially developed by social psychologists at the U.S. Public Health Service in the 1950s, aimed at explaining the psychological motivations behind individuals’ participation in disease prevention and detection behaviors. Becker et al. (24) further elaborated that the HBM comprises four core components: the perceptions of susceptibility to and severity of a specific health threat, as well as the perceived benefits and barriers to taking preventive or curative actions. Additionally, Rosenstock (25) introduced the construct of “cues to action,” emphasizing that external or internal stimuli can trigger individuals’ awareness of potential adverse health consequences, thereby motivating them to act. Building on this, Jeong & Ham (26) noted that these cues could stem from internal perceptions (such as physical symptoms) or external factors (such as social interactions and media influence). Over time, research on the HBM has deepened, with the model being expanded to include more variables that influence individual health behaviors, such as self-efficacy, motivational factors, and personality traits, all playing significant roles in individuals’ health behavior decisions (27). The HBM has become one of the mainstream theoretical frameworks for explaining and predicting individual health behaviors, not only elucidating people’s participation in disease prevention and detection but also providing a theoretical basis for health behavior interventions (27).

HBM’s empirical research shows that the main feature of this model is that it can effectively predict and explain the behavior of individuals in different health fields, such as preventive health behaviors (28), promoting health behaviors in patients with chronic kidney disease (29), HIV/AIDS prevention behaviors (30), and the health beliefs and promotive behaviors of middle-aged women (31). Carpenter (32) provided some of the most compelling evidence for the HBM’s predictive power regarding health-related behaviors through a meta-analysis encompassing 18 studies with 2,702 participants.

Therefore, using the Health belief Model (HBM) to study the relationship between information overload and overuse of health services has several unique advantages. First, HBM is able to explain how individuals make health behavior decisions based on their perceptions of susceptibility and severity of disease, as well as cognitive benefits and barriers to preventive or therapeutic measures. Second, HBM can integrate the two important factors of self-efficacy and action prompting. This mechanism could help explain why users tend to adopt excessive health behaviors when faced with a constant barrage of information from mobile health apps. Through this theoretical lens, we can gain deeper insights into the psychological and social drivers behind the overutilization of medical services, providing a theoretical basis for the development of corresponding intervention measures.

Information overload refers to a scenario where the demand for information processing surpasses an individual’s actual capacity to process information. Time is a critical dimension in assessing the imbalance between information processing demands and capabilities, that is, whether the capacity to process information within a given time aligns with the volume of information that needs processing (33). Information overload is commonly encountered during information retrieval, analysis, and decision-making processes (34) and can lead to difficulties in synthesizing information, increased anxiety and stress, ultimately affecting the quality and efficiency of decision-making (35).

In the healthcare sector, the issue of information overload has garnered considerable attention. Information overload in healthcare poses significant challenges to patients’ health behaviors. Firstly, the rapid evolution of medical information may leave patients unable to keep up with the latest developments, leading to an information asymmetry between patients and healthcare providers. Secondly, patients may feel overwhelmed during communications with doctors, struggling to fully comprehend medical advice and diagnoses (36). For instance, information overload has been identified as a significant predictor of the lack of knowledge about oral anticoagulants, which can result in patients’ non-adherence to oral anticoagulant therapy plans (37). Furthermore, when faced with an abundance of information, some patients might attempt self-diagnosis and treatment, searching online for symptom or medication information, which could lead to misunderstandings and inappropriate treatment actions, increasing health risks. For example, information overload has been found to impact patients’ willingness to undergo regular health check-ups (38), and affect the intention to receive vaccinations during a pandemic (20). Lastly, healthcare information overload can lead to psychological issues such as anxiety and distress in patients (39). When confronted with contradictory or inaccurate information from the internet or media, patients’ concerns increase, and the uncertainty about diseases and treatments intensifies. Studies confirm that people exposed to online information are more prone to experience information overload, which can lead to cyberchondria and an increased perception of the severity of diseases (21).

In the healthcare sector, there is a potential link between information overload and the overutilization of medical services. Amidst the vast amount of information patients encounter, some may be inaccurate, inconsistent, or challenging to comprehend. When experiencing information overload, patients might lean toward seeking excessive medical interventions, such as unnecessary tests or treatments, possibly due to misunderstandings of medical information or an exaggerated focus on health risks. Furthermore, information overload can heighten patients’ expectations regarding diagnosis and treatment, leading to unnecessary demands or interventions with their healthcare providers, based on the expectations formed from the information they have encountered.

According to the Pew Research Center, health-related information is among the most searched-for topics on the internet (40). Online health information searches have replaced traditional face-to-face medical consultations, reducing the frequency of traditional medical visits. However, with the rapid growth of mobile health applications and health information, patients find it challenging to efficiently locate accurate information, leading to a health information processing burden that exceeds individual capacity and triggers information overload (39). While the swift development of mobile health brings value to public health and simplifies the connection between patients and healthcare providers, it also carries the potential risk of exposing patients to an overwhelming amount of unfiltered or complex information. Therefore, examining the relationship between information overload in mobile health and the overutilization of medical services is particularly crucial.

In this study, we employ the Health Belief Model (HBM) as the theoretical framework to explore patients’ health behaviors in the context of mobile health applications. The HBM posits that an individual’s health decision-making process is influenced by their perceived severity of and susceptibility to a disease, perceived effectiveness of treatment or preventative measures, perceived barriers to performing specific health behaviors, self-efficacy, as well as personality and motivational factors (27). Building on this, our research incorporates perceived severity, perceived susceptibility, perceived benefits of treatment, perceived barriers to treatment, self-efficacy, and intention to act into the model to predict patients’ overutilization of medical services when using mobile health applications.

Mobile health applications, being a primary source of modern medical information, provide two main functions: one is the search for medical knowledge, where the app provides relevant information; the other is online medical consultation, allowing users to interact with online doctors. This study aims to examine how information overload in mobile health affects the six key factors in the HBM to predict patients’ behavior regarding the overutilization of medical services. The research model for this study is illustrated in Figure 1.

Perceived severity is defined as an individual’s assessment of the seriousness of their health condition and its potential consequences, reflecting the individual’s perception of the severity of their current and future health status. In the digital health environment, users are exposed to a vast amount of health information, which can trigger information overload. This phenomenon can affect users’ perception of the severity of health threats. Previous research indicates a positive correlation between information overload and perceived severity (21). Particularly in mobile health applications, users are confronted with an abundance of health warnings and risk data that may exceed their information processing capabilities, increasing health-related anxiety. Prolonged exposure can lead to excessive vigilance, self-diagnosis, and even symptoms such as heightened anxiety, fatigue, and insomnia in users (41). Information overload in the mHealth environment may lead users to erroneously believe their health condition is extremely severe, a perceptional imbalance due to inadequate information processing capacity rather than a rational assessment based on objective medical evidence. Hence, this study proposes the following hypothesis:

In the digital health era, frequently searching for medical information on the internet can exacerbate illness anxiety, a phenomenon referred to as “cyberchondria,” which is closely associated with excessive searching of health information online (42). With the widespread adoption of mobile health applications and online health platforms, the amount of medical information users are exposed to via mobile devices has significantly increased. While this enhances user awareness of health issues, it can also potentially lead to information processing overload. Information overload may make it difficult for users to discern the accuracy and reliability of information, thereby increasing unnecessary concerns about illness. For instance, among men who have sex with men/bisexual men who are at high risk but not infected with HIV, continuous attention to the threat of HIV infection has led to depression and anxiety (43). Furthermore, due to personalized recommendation algorithms, users might be inundated with a large volume of similar information, isolating them from other sources of information and further exacerbating anxiety and panic. The personalized health information provided by mobile health apps could intensify users’ concerns about illnesses, especially when the apps offer detailed information on disease progression, real-time health metrics feedback, or personalized health predictions, making users feel more sensitive and vulnerable to health issues. Therefore, this study proposes the following hypothesis:

Information overload has a significant impact on individuals’ perceptions and decision-making across various domains. In the field of public administration, Cao et al. (35) demonstrated that information overload has a positive impact on individuals’ perception of policy benefits. Similarly, in the educational sector, Chen et al. (44) found that information overload could lead to inefficient learning and deplete students’ cognitive resources. In the realm of public health, Hoffmann & Del Mar (45) observed that most patients tend to overestimate the benefits of medical interventions due to unrealistic expectations while underestimating potential harms.

Particularly in the context of using mobile health apps, the influx of a large volume of medical information can lead to information overload during medical decision-making, especially when processing information about the benefits of treatment methods (46). Information overload can exacerbate cognitive biases towards the perceived advantages of treatment options, prompting users to selectively focus on information that supports the benefits of certain treatments, thus reinforcing their perception of these benefits. The overestimation of treatment benefits caused by information overload could lead to an increase in placebo effects, enhancing individuals’ confidence in the effectiveness of treatments (47), thereby potentially impacting health outcomes. Therefore, this study proposes the following hypothesis:

Information overload is considered one of the main factors hindering individuals from understanding their health conditions and taking appropriate medical actions (48). Information overload occurs when the volume of information exceeds an individual’s processing capacity, requiring users to discern between beneficial, superfluous, erroneous, and meaningless information (49). Users are constrained in evaluating the accuracy and reliability of information, which significantly impacts the quality of decision-making. Information overload can lead to reduced decision accuracy (50), cause information to be misinterpreted or over-interpreted, provoke unnecessary anxiety (51), and potentially mislead users into making incorrect self-diagnoses or inappropriate treatment choices, thereby enhancing perceived treatment barriers.

The research by Pchelina et al. (41) corroborates the negative impacts of information overload on individuals’ health and sleep, potentially leading to psychological states such as anxiety and depression. In the mobile health context, information overload influences users’ perception of treatment barriers. Faced with an abundance of information, users may experience cognitive confusion and uncertainty, hindering their ability to make rational health and treatment decisions. This uncertainty can lead to hesitancy in seeking medical help, becoming an obstacle to effective treatment and recovery. Treatment barriers may encompass aspects such as the cost, time, and accessibility of specific treatments. Information overload intensifies patients’ confusion and stress regarding the treatment process, increasing perceived barriers. When dealing with a large volume of complex information, especially information closely related to their health status, patients may feel overwhelmed, thereby heightening the sense of uncertainty around the treatment process, as well as psychological and practical barriers. Therefore, this study proposes the following hypothesis:

Self-efficacy refers to an individual’s belief in their capability to execute specific behaviors to achieve desired outcomes. In the digital health environment, accessing a large amount of medical information can impact patients’ self-efficacy. This change in self-efficacy might stem from patients’ perception that they have acquired more health information, enabling them to make more informed personal health decisions. For instance, Wiljer et al. found that breast cancer patients’ access to electronic health records was positively correlated with their perception of self-efficacy (52). This impact could be influenced by various factors, including the individual’s health literacy, the clarity and relevance of the information received, and the individual’s prior experience in health decision-making.

Positive experiences can enhance an individual’s sense of efficacy, while negative experiences might lead to its reduction. However, information overload can trigger heuristic thinking in users (53), affecting the efficacy of experiences and possibly leading to individuals overestimating their abilities (54). This overestimation can result in patients developing blind confidence, which might ultimately lead to decision-making errors. In the mobile health context, users are confronted with a vast amount of information and may not be able to process this information effectively, potentially misleading their evaluation of their own capacity to handle health information and make health decisions. Therefore, this study proposes the following hypothesis: Therefore, this study proposes the following hypothesis:

In the digital health environment, accessing medical information is a double-edged sword for patients. On one hand, it empowers patients to effectively search for online health information, potentially increasing action cues, thereby promoting the occurrence of health behaviors (48). On the other hand, information overload can influence users’ behavioral patterns through psychological mechanisms. For instance, research by Honora et al. (20) shows that information overload can reduce vaccination intentions by increasing skepticism about vaccines. Similarly, Laato et al. (21) found that information overload leads to cyberchondria, triggering panic-buying behaviors. This dichotomy highlights the complex interplay between information access and user outcomes in the realm of digital health.

The relationship between information overload and the effectiveness of action cues is complex and varies depending on situational factors. While information overload can sometimes lead to confusion and hinder decision-making, it may enhance the response to action cues when users perceive the information to be structured, offering clear guidance and motivation. In mobile health applications, the phenomenon of information overload is widespread, with patients often confronting a plethora of health advice, preventive measures, and treatment options. This not only imposes a cognitive burden but also provides opportunities to motivate changes in health behaviors. Information overload might inspire patients to adopt more proactive information search and processing strategies, identifying more action cues. Therefore, this study proposes the following hypothesis:

Perceived severity, defined as an individual’s psychological assessment of the seriousness of a health threat, is a core factor that influences medical decisions and behaviors. For instance, studies during the COVID-19 pandemic found that high perceived severity of infection prompted patients to be more likely to engage in self-isolation behaviors (21, 55). There is empirical support that perceived severity is positively correlated with the overutilization of medical services. Driven by concerns about the potential consequences of diseases, patients may tend to opt for more diagnostic tests and treatment measures, even though these additional medical activities might not be necessary objectively (56). This highlights the impact of perceived severity on health-related decision-making and actions, potentially leading to an increase in healthcare resource utilization based on subjective health threat assessments rather than objective medical necessity.

In the digital health environment, particularly through mobile health applications, patients are exposed to an abundance of information regarding disease descriptions, treatment options, and risk factors. When patients have a high perceived severity of disease, this could exacerbate their concerns about health issues, making them more receptive to additional treatment measures. Consequently, they might adopt more aggressive medical behaviors, increasing the risk of overutilization of medical services. Therefore, this study proposes the following hypothesis:

Perceived susceptibility, defined as the degree to which an individual believes they are susceptible to a particular disease or health issue, is influenced by various factors, including personal or family medical history, the volume of health information encountered, and media coverage (57, 58). Higher perceived susceptibility can motivate individuals to pay closer attention to health risks and take steps to mitigate those risks (59, 60), leading to increased focus on threatening information (61). This heightened focus can motivate them to take proactive measures against health threats. However, it may also trigger behaviors associated with the overutilization of medical services, as patients’ increased anxiety and concern drive them to seek out more health interventions, potentially exceeding what is medically necessary.

In the mobile health app environment, patients with high perceived susceptibility may use the app more frequently to obtain information about symptoms, preventive measures, and treatment methods. This frequent exposure to information might exacerbate their concerns, prompting them to seek additional medical interventions to alleviate anxiety. The self-diagnosis and health monitoring features provided by mobile health apps could be overly relied upon by highly susceptible patients, leading to unnecessary anxiety and the behavior of overutilizing medical services. Moreover, the online consultation and appointment scheduling features of these apps might be frequently used by highly susceptible patients, increasing unnecessary doctor visits and medical examinations, thereby intensifying the risk of medical service overutilization. Therefore, this study proposes the following hypothesis:

The overutilization of healthcare services is indeed a complex issue, and this is mirrored in the design of mobile health (mHealth) applications. Developers often include an extensive amount of information and offer multiple solutions within these apps to demonstrate their comprehensiveness (62, 63). However, these applications may lack credibility, applicability, personalization, and accessibility, rendering the content they provide not always able to meet the specific needs or preferences of individual patients (64). Moreover, patients’ unrealistic expectations regarding treatment outcomes are a significant factor leading to the overuse of healthcare services (45). Patients may erroneously believe that more or costlier treatments yield better outcomes, thus seeking or consenting to unnecessary treatments (5). Uncertainty in the diagnostic process is another key factor, as medicine is not always an exact science (65). This uncertainty can leave patients who rely on information from mobile health applications feeling confused. To eliminate the potential risks of missed or incorrect diagnoses, patients may lean towards seeking additional examinations and treatments, even when these measures are not necessary in some cases. Under these interrelated factors, patients’ positive perceptions of treatment benefits may further promote the occurrence of healthcare service overutilization. Patients’ beliefs and expectations largely determine their acceptance of treatments (66). When patients firmly believe that a certain treatment or diagnostic procedure will bring significant benefits, they are more likely to consent to these treatments, even if they are not medically strictly necessary. Therefore, this study proposes the following hypothesis:

Treatment barriers refer to the various difficulties individuals encounter when accessing specific treatments, such as cost, time, accessibility, and other factors. Typically, perceived treatment barriers would deter patients from pursuing or continuing treatment. However, in the context of mobile healthcare, this relationship exhibits different characteristics. In this context, patients’ perception of treatment barriers may positively correlate with their tendency to overuse healthcare services. This phenomenon can be explained through psychological and socioeconomic mechanisms. When patients face perceived obstacles, they may seek faster or more significant treatment outcomes in order to ease health concerns. From a behavioral economics perspective, patients facing higher perceived barriers might prefer actions that offer immediate relief from anxiety or symptoms, especially when experiencing severe discomfort. They may opt for quicker symptom relief methods rather than longer-term but milder treatment options (67). Additionally, socio-psychological factors also play a crucial role (68). For instance, patients may pursue treatments perceived to provide quick solutions due to social and cultural pressures. In physician-patient communication, the doctor’s advice is often seen as authoritative guidance. When faced with high perceived treatment barriers, patients may be more inclined to unconditionally accept doctors’ treatment suggestions, even if these recommendations could lead to the overutilization of healthcare services.

While mobile health applications offer convenient access to health information and advice, reducing physical barriers, they may concurrently elevate cognitive and emotional obstacles for patients, such as excessive health anxiety and unrealistic expectations for treatment outcomes. When encountering significant perceived treatment barriers, patients might be more inclined to seek additional medical interventions, like unnecessary diagnostic tests or treatments. Concerns regarding uncertainties or side effects could also lead patients to pursue more medical opinions or alternative therapies, thereby heightening the risk of healthcare service overutilization. Accordingly, this study puts forth the following hypothesis:

Self-efficacy, the belief in one’s capabilities to organize and execute the courses of action required to manage prospective health behaviors and decisions, is a significant psychological factor influencing health behavior (69). In the context of mobile health, users receive health information and feedback through applications, an interaction that can enhance their sense of self-efficacy, thereby motivating them to more actively participate in self-management and medical decision-making processes. For instance, studies have shown that heart failure patients using the HeartMapp mobile application experienced enhanced confidence and user experience (70), suggesting that mobile applications can encourage individuals to actively engage in self-management, thereby boosting their self-efficacy to achieve desired health outcomes.

Self-efficacy and response efficacy play pivotal roles in individuals’ acceptance of mobile health services, influencing their perceptions of the usability and effectiveness of applications (71). Patients with high self-efficacy are more confident, a trait that propels them to use mobile health applications more frequently and to proactively explore medical options, including non-essential treatments. Believing in their ability to manage complex medical information, these patients tend to adopt proactive healthcare measures, such as excessive monitoring, diagnostics, and interventions, which may lead to an overreliance on medical interventions. Additionally, patients with high self-efficacy may hold optimistic expectations about treatment outcomes (72). They might overly focus on the benefits described in treatment plans while underestimating the likelihood of side effects (73), thereby overlooking the potential risks associated with medical interventions. Therefore, this study proposes the following hypothesis:

In the Health Belief Model, “cues to action” are internal or external factors that motivate individuals to change their behavior. These cues can come from various channels, including public health information, doctors’ advice, and more. An increase in cues to action is believed to potentially enhance the likelihood that individuals will adopt healthy behaviors. For instance, one study demonstrated that reminder postcards developed based on elements of the Health Belief Model significantly increased vaccination rates (74). However, in the context of mobile health applications, cues to action that are unfiltered or lack appropriate medical guidance may increase the risk of overutilizing medical services. This is because patients might interpret each cue to action as a signal that immediate medical intervention is needed, prompting them to excessively use medical services.

Existing empirical research has begun to focus on such relationships. Patients who receive an abundance of cues to action, such as widespread health screening campaigns, excessive attention to potential health risks, and frequent recommendations about treatment from family and friends, are more likely to request or agree to additional medical examinations and treatments (75). Mobile health applications provide an abundance of cues to action through health notifications, reminders, and suggestions. These excessive prompts may lead to patients’ overreaction and exacerbate the risk of overutilizing medical services. Frequent cues can heighten patients’ concerns about health issues, prompting them to seek additional diagnostics or treatments to alleviate anxiety. This anxiety can lead patients to have exaggerated expectations of the need for and effectiveness of treatment without sufficient evidence. Therefore, this study proposes the following hypothesis:

This study has designed a questionnaire aimed at assessing the relationship between information overload in mHealth applications and users’ tendency to overuse medical services. The items in the questionnaire employ a 5-point Likert scale, with all entries being based on existing research and suitably adjusted and adapted for the context of this study. To ensure the completeness of the questionnaire, all questions are set as mandatory, and submissions of incompletely answered questionnaires will not be accepted. In recent years, digital data collection methods have become increasingly popular due to their high efficiency and wide coverage. These methods enable researchers to access diverse populations and efficiently collect large amounts of data, with significant advantages especially in fields such as social and management sciences (76). From the respondents’ perspective, Monday and Gever found through their research that online surveys have a 16% higher response rate than in-person surveys, with the majority of respondents preferring to participate via digital platforms (77). From the point of view of researchers, Gever et al. have demonstrated that digital data collection methods have the advantages of collecting large-scale and diverse data, cost effectiveness, and timely collection, which are well suited for contemporary research (78, 79). Therefore, this study takes the online survey tool Questionnaire Star to survey users of digital media platforms of medical treatment.

Following the questionnaire’s development, experts in the field of public health were invited to review the survey instrument to ensure its content validity and relevance. Moreover, 60 undergraduate students participated in a pilot test of the questionnaire to verify its comprehensibility and logical structure. The constructs in the questionnaire were specifically adapted from the following sources: the mobile health information overload section was adapted from Cao et al. (35); perceived severity (PSEV) and perceived susceptibility (PSUS) were adapted from Walrave et al. (80); perceived treatment benefits (PTB) from Liu et al. (81); perceived treatment barriers (PBA) from Adiyoso et al. (82); self-efficacy (SE) from Wu et al. (83); and cues to action (CTA) Arabyat et al. (84). Given the lack of empirically validated scales for overuse of medical services in the medical field, this study referred to Ding et al. (85) scale on smartphone overuse and, combined with the pilot test results, developed a scale for medical service overuse suitable for the context of this study. The complete scale is included in Appendix A of the document.

Before conducting the survey, this study received approval from the university’s ethics committee to ensure the ethical conduct of the research. All participants were clearly informed about the following aspects before completing the questionnaire: the anonymity of the survey, the content and purpose of the research, the voluntary nature of their participation, the absence of personal privacy information in the survey, and the incentives provided upon completion of the questionnaire.

Baidu Index is a data analysis platform based on Baidu’s massive Internet user behavior data, which is widely used in market research, brand analysis, trend prediction and other fields. It reflects the public’s attention to a specific topic by counting the frequency of users’ keyword searches in Baidu’s search engine (86). As Baidu is one of the most commonly used search engines by Chinese users (87), the search data of Baidu Index has strong representativeness and authority, and has high reference value. Based on data from Baidu Index¹, there was a high level of interest in digital healthcare in Guangdong Province. In addition, the purpose of this study is to explore the user behavior characteristics in areas with high attention and understand the relationship between information overload and overuse of medical services, rather than the national general behavior. Therefore, the target population for this study was identified as mobile health app users in Guangdong Province. Using a random sampling method, questionnaire was distributed from 2024.2.4 to 2.20. A total of 1,494 responses to the questionnaire were collected, of which 1,137 were valid responses, resulting in a valid response rate of 76.1% (Figure 2).

Table 1 presents the demographic information of the 1,137 respondents. The proportion of female respondents (N = 693, 60.95%) is higher than that of male respondents (N = 444, 39.05%). The age of the respondents is predominantly between 30 to 49 years old (N = 745, 65.52%). The monthly income of most respondents falls between 9,000 RMB to 11,999 RMB (N = 543, 47.76%).

In the present investigation, the proposed model incorporated eight distinct variables. Subsequently, an assessment of multivariate normality was executed on the amassed data, utilizing an online computational tool² to ascertain the data’s distribution profile. The findings delineated Mardia’s multivariate skewness (β = 28.623, p > 0.05) and kurtosis (β = 1051.143, p < 0.001), indicating a deviation from multivariate normality (88). This investigation is characterized as exploratory in nature. In essence, the applicability of this study extends to data analyses employing Partial Least Squares Structural Equation Modeling (PLS-SEM).

In this study, we examine latent psychological variables that cannot be directly measured. To validate the proposed hypotheses, the study employed Structural Equation Modeling (SEM) to analyze the collected data.

Structural Equation Modeling (SEM) is broadly categorized into two types: Covariance-Based Structural Equation Modeling (CB-SEM) and Partial Least Squares Structural Equation Modeling (PLS-SEM). PLS-SEM, as a second-generation multivariate data analysis method, is primarily suited for exploratory theoretical research, allowing for the analysis while ensuring the integrity of all relationships between independent and dependent variables (89).

PLS-SEM offers several significant advantages over CB-SEM: (1) PLS-SEM is more suitable for complex models involving more than six variables (90). (2) PLS-SEM can effectively handle small sample data (90). (3) PLS-SEM is applicable to data that are not normally distributed (90). Given these factors, PLS-SEM has more adaptability than CB-SEM in the theory development stage and has been proven to effectively replace CB-SEM in most social science research contexts (90).

Secondly, Asogwa et al. ‘s research shows that digital software is significantly superior to manual methods in terms of analytical efficiency, subject recognition, and visualization techniques, especially when dealing with complex data sets (91). Therefore, in this study, in order to improve the efficiency of qualitative data analysis, we choose PLS-SEM and its supporting digital software SmartPLS 4.0 for data analysis.

Common method bias is a prevalent issue in questionnaire surveys. Harman’s single factor analysis, introduced by Harman in 1976, is widely employed in social science research to detect common method bias (92). This method suggests extracting a single factor, and if the variance is less than 40%, it indicates minimal influence of common method bias on the survey data (92). The Harman analysis conducted in this study revealed a proportion of 30.49% for the extracted variables (below 40%).

Furthermore, we conducted a Full-Variance Inflation Factor (Full-VIF) Test on the data to further examine common method bias. Research suggests that applying the Full-VIF Test to all variables in the model, including dummy variables, can help detect potential common method bias. If the Variance Inflation Factor (VIF) exceeds 3.3, it indicates a possible influence of common method bias on the model. However, if all VIF values obtained from the Full-VIF Test are equal to or below 3.3, it can be inferred that the model is not significantly affected by common method bias (93). This method has been widely applied across various research domains (88). The test results of this study found that all VIF values were below 3.3. Considering the results of both methods used to test for common method bias, it can be concluded that common method bias is not a significant concern in this study.

A paired t-test was employed to scrutinize the potential nonresponse bias within the demographic data of the initial and concluding 25 participants. The outcome revealed no statistically significant disparities, thereby indicating that nonresponse bias did not pose a substantive issue within the context of this investigation.

The quality of the model was assessed through the evaluation of composite reliability (CR), average variance extracted (AVE), discriminant validity, and outer loading. As shown in Table 2, the composite reliability and Cronbach’s alpha of each variable exceed 0.7, indicating satisfactory internal consistency of the data in this study. Additionally, the AVE values of each variable surpass 0.5, with outer loadings exceeding 0.7, confirming the acceptable convergent validity in this study (90).

Table 3 presents the results of Fornell and Larcker’s Test and the Heterotrait-Monotrait ratio (HTMT) Test, used to assess discriminant validity. The HTMT values between variables are below the threshold of 0.85, and the square root of the AVE for each variable exceeds its correlation with other variables, as suggested by (90). A comprehensive analysis indicates that this study exhibits good levels of composite reliability, convergent validity, and discriminant validity.

Before conducting structural model measurement, we assessed collinearity, and the VIF values for each variable were below 3. Hence, collinearity is not a significant concern in this study. After ensuring the reliability and validity of the model, we used the structural model to validate the hypotheses. The research results show that information overload caused by the use of mobile medical software has significant positive correlation with perceived severity, perceived susceptibility, perceived treatment benefits, perceived cure obstacles, self-efficacy and action cues, and H1, H2, H3, H4, H5, and H6 are all supported. Perceived severity and perceived susceptibility were positively correlated with overuse of medical services, that is, H7 and H8 were supported. However, the perceived benefit of treatment showed no significant effect on overuse of medical services, and H9 was not supported. Finally, perceived healing barriers, self-efficacy, and action cues were positively correlated with medical overuse, that is, H10, H11, and H12 were all supported. Additionally, control variables do not have a significant impact on medical overuse (Table 4).

Finally, we used the Standardized Root Mean Square Residual (SRMR) to test the goodness of fit (GOF) of the model. The SRMR value for the model is 0.054, which is below the threshold of 0.08. Thus, the fit of the model is satisfactory (94).