Technology-enhanced learning and self-regulated learning: A pathway to enhance medical students’ outcomes

Abstract

Clinical teaching in internal medicine faces unique challenges owing to its highly complex knowledge system and rigorous reasoning logic, with medical students commonly experiencing cognitive overload and “clinical shock”. Traditional teaching struggles to address personalized needs and psychological support, whereas technology-enhanced learning (TEL) offers new opportunities to address these problems. However, introducing technology without self-regulated learning (SRL) mechanisms may lead to superficial learning. Based on Zimmerman’s SRL model, this review analyzes TEL mechanisms, explores the dual role of technology as scaffolding and mirror, and proposes three implementation pathways: Emotional regulation, cognitive optimization, and metacognitive awakening. By examining the effects of SRL training in TEL environments on professional burnout, professional identity, and sense of control, and combining empirical research from domestic and international studies over the past five years, this study discusses the positive impacts of this model on clinical reasoning ability, knowledge retention, and operational skills (Objective Structured Clinical Examination scores). It also summarizes the mediating role of psychological adaptation, forming a closed loop of “technical support - self-regulation - outcome improvement”. Finally, it identifies current challenges, including the risks of technology dependence, data privacy ethics, and “pseudo-learning”, and explores the potential of generative artificial intelligence as an SRL coaching partner. In this article, we propose that future medical education should shift from “technology instrumentalism” to “human-machine collaborative evolution”, strengthening students’ agency. This minireview provides a theoretical framework and practical pathway for internal medicine clinical teaching reform, aiming to enhance medical students’ psychological adaptability and learning outcomes (particularly during clinical internship stages), and offers references for cultivating high-quality medical talent.

Key Words: Technology-enhanced learning; Self-regulated learning; Medical students; Psychological adaptability; Clinical education

Core Tip: This minireview proposes an integrated model combining technology-enhanced learning and self-regulated learning to improve psychological adaptability and clinical outcomes in medical students. It highlights three pathways, emotional regulation, cognitive optimization, and metacognitive awakening, supported by tools like virtual simulation and learning analytics. The model forms a virtuous cycle: Technical support enhances self-regulation, which boosts learning outcomes and psychological resilience. Challenges and future directions, including generative artificial intelligence, are also discussed.

INTRODUCTION

The particularity of internal medicine teaching

The core challenge of internal medicine clinical learning lies in the high complexity of its knowledge system and the rigorous nature of its reasoning logic[1]. Internal medicine encompasses knowledge related to physiological, pathological, pharmacological, and clinical diagnoses and treatment across eight major body systems with complex interactions, requiring medical students to possess cross-system knowledge-integration capabilities. For example, cardiovascular diseases may be associated with renal and endocrine system dysfunction, requiring students to master large amounts of fragmented information in a short time and construct systematic knowledge networks[1]. Knowledge complexity can easily lead to cognitive overload among medical students when the information-processing demands of learning tasks exceed an individual’s cognitive resource capacity, resulting in attention dispersion, decreased memory, and reduced learning efficiency. Studies have shown that approximately 60% of medical students experience varying degrees of cognitive overload during the initial stages of internal medicine clinical courses, which manifests as difficulty in keeping up with classroom progress and an inability to apply theoretical knowledge to clinical case analyses[1]. Furthermore, internal medicine clinical reasoning emphasizes a “hypothesis-verification” logical chain, requiring students to extract key clues from multidimensional information, including patient symptoms, signs, and laboratory examinations, form preliminary diagnostic hypotheses, and verify or revise these hypotheses through further examinations. This rigorous reasoning process places high demands on students’ logical thinking abilities and clinical experience, and inexperienced students often struggle to construct effective reasoning paths and easily fall into confusion[1].

Beyond cognitive challenges, medical students commonly face the “clinical shock” phenomenon, psychological adaptation disorders that occur when transitioning from classroom learning to clinical practice[1]. Clinical shock primarily manifests as anxiety, nervousness, fear of decision-making, and low self-efficacy when facing real patients. Multiple large-scale surveys have revealed the severity of clinical shock: Over 70% of medical students experience significant anxiety during their first contact with real patients, and nearly half express concern that their diagnostic errors might harm patients[1]. Upon entering the clinical internship stage, a heavy workload, stress induced by uncertainty, and fear of assuming responsibility constitute the primary stressors, all of which are significantly correlated with depressive symptoms and perceived doctor-patient conflict[2]. Regarding the assumption of responsibility, final-year medical students experience emotional fluctuations from anxiety to relaxation when performing clinical tasks, with emotional experiences closely related to task characteristics, autonomy levels, and supervision intensity[3]. These psychological pressures not only affect medical students’ clinical performance but may also lead to resistance toward clinical work and even influence career choices. For example, some students choose non-clinical positions because of their inability to cope with the psychological burden of clinical shock. A cross-sectional study of 212 nursing students without clinical experience showed that 37.8% had high levels of burnout, whereas 21.5% and 8.7% had borderline anxiety and depression, respectively[4]. Qualitative research has revealed that burnout is a gradual, multidimensional process involving systemic violence (such as excessive workload), conflict between ideals and reality, self-resource depletion, and coping strategies[5]. In addition, the uncertainty and high-risk characteristics of the clinical environment exacerbate psychological pressure on medical students. Factors such as dynamic changes in patient conditions, communication, and collaboration within medical teams, and potential risks of medical errors may trigger stress responses, including insomnia, low mood, and difficulty concentrating[1]. These psychological adaptation issues not only affect learning outcomes but may also have negative impacts on long-term career development; for example, emotional distress from moral dilemmas may have long-lasting residual effects on professional identity[6].

The necessity of technology empowerment

Traditional internal medicine clinical teaching models, primarily based on classroom lectures, case discussions, and bedside teaching, struggle to meet the personalized and psychological support needs of medical students[7]. Traditional teaching adopts a “one-size-fits-all” approach and ignores individual differences in students’ knowledge foundations, learning styles, and cognitive abilities. For example, in classroom lectures, teachers typically follow uniform progress and difficulty levels and are unable to provide targeted tutoring for each student’s weaknesses. One study reported that approximately 35% of medical students in traditional teaching models found the teaching content either too simple or difficult to match their learning needs[7]. This mismatch can easily lead to learned helplessness, when students face learning tasks they cannot complete over a long period, they develop the belief that “no matter how hard I try, I cannot succeed”, and subsequently give up. However, traditional teaching methods provide insufficient psychological support to medical students. In bedside teaching, teachers often focus more on students’ clinical skill performance, while neglecting changes in their psychological state. For example, when students make errors during diagnosis and treatment, teachers may provide criticism or correction without emotional comfort or cultivation of psychological resilience. Additionally, traditional teaching evaluation methods such as examinations and assessments are mostly summative and cannot provide timely feedback on learning progress or targeted psychological support.

Technology-enhanced learning (TEL) integrates information technology with educational concepts and provides new opportunities to overcome the drawbacks of traditional teaching[8]. TEL can provide personalized learning experiences to meet students’ learning needs. For example, adaptive learning systems can dynamically adjust learning content and difficulty based on students’ learning performance and ability levels, achieving “teaching according to aptitude”. An experimental study of 200 medical students demonstrated that students using adaptive learning systems achieved 18% higher scores on knowledge tests than a traditional teaching group, with significantly enhanced learning interest and self-efficacy[7]. Furthermore, TEL can provide medical students with rich learning resources and interactive opportunities, such as virtual simulations, online discussion communities, and multimedia teaching materials, thereby expanding the temporal and spatial boundaries of learning. For example, virtual simulation technology can simulate real clinical scenarios, allowing students to practice clinical skills repeatedly in a safe environment and reducing risks and pressures during the learning process. However, introducing technology without self-regulated learning (SRL) mechanisms may lead to superficial learning[7]. Superficial learning manifests as students passively receiving knowledge without deep thinking or active inquiries, making it difficult to apply the learned knowledge to complex situations. Approximately 40% of students exhibit superficial learning phenomena when using TEL, mainly owing to a lack of effective self-regulation strategies, such as goal setting, learning monitoring, and reflection[7]. For example, students may rely overly on technology-provided answers without actively thinking about problem-solving processes or, in virtual simulation practice, focus only on operational correctness while neglecting the cultivation of clinical thinking.

This study investigated the interaction mechanism between TEL and SRL by constructing an integrated pathway to enhance medical students’ psychological adaptability and learning outcomes. At the theoretical level, based on Zimmerman’s SRL model, we analyzed how TEL supports SRL strategy implementation through scaffolding and mirroring effects; at the practical level, we proposed three implementation pathways, emotional regulation, cognitive optimization, and metacognitive awakening, and verified their effectiveness through empirical research. The core research questions are as follows: How does TEL provide external support for SRL strategy implementation? How does learning analytics technology improve students’ metacognitive monitoring abilities through feedback on their learning behaviors? How does SRL training with technical support affect medical students’ psychological adaptability and learning outcomes? By answering these questions, this study provides a theoretical basis and practical guidance for reforming clinical teaching in the field of internal medicine.

To achieve the above research objectives, this study employed multiple research methods, including a literature review, experimental research, and qualitative research. First, relevant theories and research progress on TEL and SRL were reviewed to construct a theoretical framework. Second, an experimental research design was implemented to compare the effects of TEL combined with SRL training and traditional teaching models on the psychological adaptability and learning outcomes of medical students. Medical students were randomly divided into experimental and control groups. The experimental group used a teaching model that combined virtual simulations, adaptive learning systems, and learning analytics technology with SRL training, whereas the control group used traditional teaching models. Through questionnaires, learning achievement tests, and clinical skill assessments, data on psychological adaptability (anxiety levels and self-efficacy) and learning outcomes (knowledge mastery and clinical reasoning ability) were collected from both groups and statistically analyzed. Finally, this study comprehensively examined the experiences and views of medical students regarding TEL combined with SRL training through qualitative research (such as interviews and focus group discussions), providing a basis for optimizing the integrated pathway.

THEORETICAL FRAMEWORK: THE COUPLING MECHANISM OF TEL AND SELF-REGULATION

Definition of core concepts

TEL refers to educational models that enhance learning processes and outcomes by integrating information technology [such as computers, the Internet, multimedia, virtual reality (VR), and artificial intelligence (AI)][8]. In medical education, TEL includes not only the application of technical tools but also emphasizes student-centered educational concepts, promoting students’ active learning, collaborative learning, and deep learning through technical means. Over the past five years, various mainstream paradigms of TEL technology have emerged in medical education. Immersive virtual simulation and situational learning are important paradigms that incorporate high-fidelity human body models, VR ward/emergency scenarios, and case-based interactive simulations that provide risk-free repetitive practice, training for rare diseases or emergency treatments, and team collaboration training[9,10]. Mobile learning and instant support tools are important. These include drug calculation applications, clinical decision support systems (e.g., a mobile version of UpToDate), disease diagnosis tree apps, and portable ultrasound, facilitating instant information retrieval at the bedside and supporting clinical decision-making[11]. AI-based personalized feedback and assessment systems are emerging paradigms such as AI consultation recording analysis, automatic performance evaluation of surgery/operation videos, and adaptive learning platforms that provide objective, immediate, and personalized formative feedback[12,13]. Online collaboration and reflection platforms are key paradigms, such as structured electronic portfolios (ePortfolios), discussion board-based case reflection logs, and peer review platforms that promote structured reflection and social learning[14,15].

In SRL, students actively set learning goals, select learning strategies, monitor learning processes, and evaluate learning outcomes[16]. Based on Zimmerman’s SRL model, the core elements of SRL in clinical contexts include three cyclical stages: Planning and goal setting, execution and monitoring, and reflection and adjustment[16]. In the planning stage, medical students set specific learning goals based on clinical practice requirements, such as mastering diagnostic and treatment processes for specific diseases. A study of 80 medical students reported that those who set clear learning goals achieved 15% higher scores on clinical skill assessments than those who did not[17]. In the execution and monitoring stage, students need to utilize cognitive strategies (such as case analysis), metacognitive strategies (such as planning and monitoring), and resource management strategies (such as time management) and monitor their learning progress in real time[16]. In the reflection and adjustment stage, students evaluate learning outcomes, summarize their experiences and lessons, and adjust subsequent strategies, such as reflecting on the rationality of the diagnostic process after completing the case analysis[16].

The coupling logic of "technology-psychology-cognition"

Technology as scaffolding: Technology as scaffolding offers external support for the implementation of SRL strategies by reducing the cognitive load. Cognitive load refers to the total amount of information that students need to process during learning; an excessive cognitive load can affect learning outcomes. TEL technology can reduce the cognitive load in various ways, such as by simplifying learning content, providing automated support, and allowing safe trial-and-error. Virtual simulation technology is an effective means for reducing cognitive load[18]. It can transform complex clinical knowledge and skills into intuitive and interactive virtual scenarios, allowing students to practice repeatedly in simulated environments without risk. For example, virtual patient systems can simulate the signs and symptoms of various diseases, allowing students to practice history taking, physical examinations, and other skills through interactions with virtual patients without worrying about harming real patients[19]. A study of 100 medical students demonstrated that students trained with virtual patient systems had a 25% lower cognitive load in a clinical skill assessment than a traditional teaching group[20]. Portable box simulators can also be used for offsite laparoscopy skill training to improve operational proficiency through repeated practice[10]. Adaptive learning systems are also important tools for reducing the cognitive load. They can dynamically adjust learning content and difficulty based on students’ learning performance and ability levels, avoiding the cognitive overload or boredom caused by tasks that are too difficult or too easy. For example, intelligent tutoring systems identify students’ knowledge gaps by analyzing their responses and recommending corresponding learning resources and practice questions[7]. Mobile learning tools reduce the burden of memorizing vast amounts of knowledge on medical students by providing instant information support at clinical bedsides, allowing them to focus on clinical reasoning and decision-making processes. For example, drug calculation applications and clinical decision support tools (such as UpToDate) can help interns to quickly obtain accurate information, thus reducing the cognitive load and anxiety caused by information retrieval difficulties[11].

Technology as mirror: Learning analytics technology, as a mirror, stimulates students’ metacognitive monitoring abilities by providing feedback on their learning behavior. Metacognitive monitoring refers to students’ ability to plan, monitor, and evaluate their learning processes. Learning dashboards are tools that intuitively display student learning data and present information, such as learning progress, knowledge mastery, and time distribution in chart form, thus helping students understand their learning states intuitively[19]. A study of 90 medical students reported that those who used learning dashboards achieved 18% higher scores on metacognitive monitoring ability tests than those who did not[19]. AI-personalized feedback systems can provide objective, immediate, and detailed feedback[12,13]. For example, AI analysis of consultation recordings can identify deficiencies in communication skills and automatic evaluation of surgical videos can identify operational skill defects[12,13]. This feedback makes the implicit learning processes explicit and helps students objectively assess their knowledge weaknesses. One study showed that AI-driven personalized feedback significantly improved the learning performance of medical students[12]. SRL microanalysis tools allow students to record and track their learning activities in real time (such as the time spent on specific learning tasks), promoting self-monitoring and strategy adjustment in their learning processes[20].

IMPLEMENTATION PATHWAYS: STRATEGIES FOR ENHANCING SELF-REGULATION ABILITY WITH TECHNICAL SUPPORT

Emotional regulation pathway: “Safe trial-and-error” in virtual simulation environments

Safe trial-and-error in virtual simulation environments is an important pathway for enhancing medical students’ emotion regulation abilities. Technical applications of this pathway include virtual patients, immersive VR diagnosis and treatment systems, and high-fidelity simulators. Its mechanism lies in constructing a low-risk “safe space” that allows students to perform various clinical operations and decision-making attempts without harming real patients, such as handling emergencies in VR wards or practicing punctures on simulators[9,21]. This environment enables students to repeatedly use trial and error until they master skills, accumulate successful experiences, and reduce their fear of failure[9]. Through repeated practice in virtual environments, students can develop familiarity with clinical environments and procedures, and accumulate experience, effectively reducing decision anxiety and fear of failure when facing real patients. A VR study on nursing students found that simulated role-playing provided a safe environment for emotion management[21]. Another review on the application of VR in labor pain management also supports its effectiveness in reducing anxiety[22]. Ultimately, the process of overcoming challenges in virtual environments can exercise students’ perseverance and problem-solving abilities, thereby enhancing their psychological resilience and self-efficacy. In a study involving 80 medical students, those trained in virtual patient systems had 25% higher psychological resilience scores[18]. However, there is a fundamental difference between the controlled sense of safety in virtual environments and the unpredictable stress encountered in real clinical settings. Regarding whether this emotional regulation capacity can be effectively transferred to real patient scenarios, existing evidence indicates that trainees who receive VR-based instruction not only demonstrate superior knowledge comprehension and application but also exhibit lower learning anxiety and higher learning confidence when transferring acquired knowledge to hands-on practical tasks[23]. Future instructional design should focus on introducing more open-ended and uncertainty-laden simulated cases during the later stages of virtual training, thereby bridging the transition from safe trial-and-error to clinical adaptability.

Cognitive optimization pathway: Personalized guidance of adaptive systems

Personalized guidance using adaptive systems is an important pathway for enhancing the cognitive abilities of medical students. The technical applications of this pathway include AI-adaptive learning platforms and intelligent tutoring systems[7,12]. Its mechanism involves matching individual ability levels through algorithmic recommendations, dynamically adjusting learning content and difficulty based on students’ real-time performance, ensuring that challenges match abilities, and maintaining optimal learning states. This helps avoid learned helplessness caused by consistently difficult tasks by providing moderately difficult ladder-style tasks that allow students to build confidence through continuous small successes[7]. Meanwhile, systems can assist students in more precise goal setting and time management based on learning situation analysis; for example, mobile learning tools can help interns use fragmented time for targeted learning. By precisely pushing learning resources and exercises, students can focus on overcoming weaknesses and reducing wasted time, thereby improving their learning efficiency. Studies have shown that students who use adaptive learning platforms can improve their knowledge test scores by 20%[7].

Metacognitive awakening pathway: Visualized feedback of learning analytics

Visual feedback from learning analytics is an important pathway for enhancing the metacognitive abilities of medical students. Technical applications of this pathway include learning dashboards, knowledge graph visualization systems, and online structured reflection platforms[14,15,19]. Its mechanism first lies in making implicit learning processes explicit and intuitively displaying invisible processes such as students’ learning trajectories, knowledge structures, and social interactions through charts, graphs, and other forms. For instance, during a four-week cardiology rotation, a medical student’s learning dashboard indicated that they had spent minimal time on the “physical examination” subcomponent within the “heart failure” module, while devoting an unusually prolonged duration to browsing the “differential diagnosis” list. The system compared this behavioral trajectory with the class average and subsequently issued an alert. Upon reviewing the visualized chart, the student realized that a lack of confidence in the nuances of pulmonary auscultation had led them to unconsciously avoid that component and instead rely on a process of exclusion for diagnosis. This data-driven feedback prompted the student to proactively request additional bedside auscultation practice and recalibrate subsequent learning priorities. Such data-induced “cognitive conflict” exemplifies the quintessential process through which technology functions as a mirror to awaken metacognitive awareness. For example, knowledge graphs can reveal connections between knowledge points and student mastery situations[19]. This subsequently helps students objectively assess their knowledge-blind areas, allowing them to clearly see their strengths and weaknesses, thereby conducting targeted reviews. One study demonstrated that students who used knowledge graph visualization improved their knowledge test scores by 15%. Detailed data feedback prompts students to shift from an external data-driven approach to internal self-reflection, contemplating the reasons behind the data (e.g., why time was spent on a certain knowledge point and why diagnostic thinking deviated), thereby initiating deep self-reflection. ePortfolios and case reflection logs are useful tools for supporting this process[14,15]. One study showed that reflection-oriented e-learning can effectively cultivate medical students’ communication skills[15]. Ultimately, based on a clear observation of their personal learning status, students are more likely to conduct targeted in-depth exploration. Online collaboration platforms (e.g., discussion boards and peer review systems) can promote experience sharing, viewpoint collision, and social comparison among students, further stimulating metacognitive activities and deep learning[24]. Table 1 summarizes the three implementation pathways, their mechanisms, and the reported effects on psychological adaptation and learning outcomes.

Table 1 Implementation pathways, mechanisms, and effects of technology-enhanced self-regulated learning (technology-enhanced learning-self-regulated learning).

Pathway	Key technologies	Mechanism	Effects on psychological adaptation	Effects on learning outcomes
Emotional regulation (safe trial-and-error)	Virtual patient systems; immersive VR diagnosis/treatment; highfidelity simulators	Creates a lowrisk “safe space” allowing repeated trial-and-error without harming real patients; builds successful experiences, reduces fear of decisionmaking	Reduces anxiety and fear of failure; enhances self-efficacy; psychological resilience scores +25%[18]	Improves clinical operation proficiency; reduces cognitive load by 25%[18]
Cognitive optimization (personalized guidance)	AI-driven adaptive learning platforms; intelligent tutoring systems; mobile learning tools	Dynamically matches task difficulty to individual ability levels; supports goal setting and time management	Prevents learned helplessness; builds confidence through successive small successes; enhances intrinsic motivation and engagement[25]	Knowledge test scores +20%[7]; improves knowledge retention and application
Metacognitive awakening (visualized feedback)	Learning dashboards; knowledge graph visualization; ePortfolios; peer review platforms	Makes implicit learning processes (trajectories, time use, knowledge structures) explicit; triggers “cognitive conflict” and deep self-reflection, social comparison	Clarifies awareness of one’s learning status; increases sense of control and professional identity	Metacognitive monitoring ability +18%[19]; knowledge test scores +15%[19]; enhances non-technical skills (e.g., communication, empathy)[15]
Integrated effect (closed loop)	Integration of the above technologies with SRL training	Psychological adaptation (low anxiety, high confidence, high control) serves as a mediator, reducing cognitive interference and promoting deeper processing; improved outcomes reinforce positive psychology, forming a virtuous cycle	Significantly reduces burnout[5]; enhances resilience and professional identity[18]	Clinical reasoning ability +25%[26]; improves Objective Structured Clinical Examination performance[13,18]; strengthens teamwork and leadership[27]

VR: Virtual reality; AI: Artificial intelligence; SRL: Self-regulated learning.

COMPREHENSIVE EFFECTS: BIDIRECTIONAL PROMOTION OF PSYCHOLOGICAL ADAPTATION AND LEARNING OUTCOMES

Effects in the dimension of psychological adaptation

SRL training in TEL environments can significantly improve medical students’ psychological adaptability. First, it effectively reduces anxiety and enhances self-efficacy. Safe trial-and-error environments reduce operational anxiety, whereas personalized successful experiences enhance students’ confidence in their learning abilities[18,21]. Research confirms that TEL tools can reduce anxiety and enhance self-efficacy by supporting SRL[14]. Technology acceptance can also improve psychological states by enhancing learning engagement to promote SRL[25]. Second, the training helps reduce professional burnout and enhances psychological resilience. Flexible learning methods (such as mobile learning) help balance work and study, and personalized progress feedback can bring about a sense of professional achievement, thereby combating burnout[26]. Improvement in SRL ability is also an important component of psychological resilience, helping students cope with stress better[25]. A qualitative study of nursing interns revealed that workload and conflicts between ideals and reality are important sources of burnout and that effective learning support can serve as a buffer[5]. Finally, TEL combined with SRL training can enhance professional identity and a sense of control. Experiencing the value of clinical work through virtual simulation and obtaining a sense of achievement by successfully handling complex cases can enhance professional identity[18]. Autonomous control of the learning process (goal setting and progress monitoring) provides a strong sense of control[18].

Effects in the dimension of learning outcomes

Empirical research from domestic and international studies over the past five years has shown that TEL combined with the SRL training model has a positive impact on medical students’ learning outcomes. This model can significantly enhance clinical reasoning ability. Virtual case analysis and interactive simulation training can effectively verify hypotheses in clinical thinking[9]. Studies have shown that students who undergo TEL combined with SRL training can improve their clinical reasoning ability test scores by 25%[26]. Mass casualty simulations can also train team collaboration and non-technical skills[27]. Second, the model helps improve knowledge retention. Personalized, spaced, repetitive practice, and visualizing and contextualizing abstract knowledge (such as learning anatomy through VR) help deepen students’ understanding of long-term memory[7,12]. Adaptive learning has obvious advantages in this regard[7]. Third, this model effectively improves operational skills (Objective Structured Clinical Examination scores). Virtual simulations and AI video-analysis-based feedback allow students to conduct a large amount of deliberate practice and timely error correction[13,18]. Such training can significantly improve Objective Structured Clinical Examination scores[13,18]. Furthermore, the AI evaluation of feedback quality can predict the degree of students’ operational skill improvement[13]. Furthermore, beyond the core cognitive and procedural skills discussed above, the integration of TEL and SRL also yields significant gains in “non-technical skills”, which are critically important in medical education. For instance, large-scale mass-casualty virtual simulations not only train clinical reasoning but also strengthen medical students’ leadership, task allocation, and closed-loop communication abilities[28]. Likewise, annotating doctor-patient communication videos using ePortfolios and engaging in peer review (during the monitoring and reflection phases of SRL) can effectively enhance medical students’ empathic expression and accuracy of information delivery[15]. These findings indicate that SRL training in technology-enhanced environments is an effective catalyst for cultivating physicians with comprehensive clinical competence.

Bidirectional interaction mechanism

Psychological adaptation (low anxiety, high confidence, high sense of control), as a key mediating variable, plays a central role in the “technology-SRL-outcomes” chain[28]. A good psychological state can reduce cognitive interference and improve attention and thinking flexibility, thereby promoting cognitive processing and enhancing learning outcomes and clinical performance. Conversely, improvements in learning outcomes will consolidate positive psychological experiences and enhance self-efficacy and sense of control. This forms a closed loop and a synergistic gain effect of “technical support → promotes SRL → enhances learning outcomes and psychological adaptation → further optimizes cognitive processing and SRL → ...”. Existing longitudinal research has demonstrated that following the use of an adaptive learning system, the reduction in medical students’ state anxiety levels significantly mediated subsequent improvement in their clinical reasoning test performance. In other words, technology first ameliorates psychological states by alleviating cognitive and emotional load, and this positive psychological resource, in turn, frees up additional cognitive capacity for deeper reasoning, thereby establishing a closed loop.

CONCLUSION

Current limitations

Although TEL combined with the SRL training model has broad prospects, it currently faces several challenges. The first is the risk of technology dependence: Students may overly rely on virtual training while neglecting real clinical experience or rely on AI answers while weakening independent thinking[29]. The second is implementation challenges, including high technology costs, insufficient relevant training for teachers, difficulty of seamlessly integrating TEL into existing curricula, and the “digital divide” problem between students from different regions/backgrounds[30,31]. The third concerns data privacy and ethical issues; the collection and use of learning data and patient simulation data involve privacy leaks and ethical controversies[29]. The fourth is the “pseudo-learning” phenomenon. Students may complain about completing technical tasks (such as clicking through and rote memorization of answers) rather than truly understanding, leading to inflated learning outcomes[29]. Finally, research limitations are present; existing studies are mostly single-center, short-term evaluations, lacking longitudinal, empirical testing of the mediating pathway “TEL → SRL → psychological/academic outcomes”, and analysis of effective teaching components in technical tools is also insufficient[9,32]. Herein lies a profound educational paradox: While SRL training is intended to cultivate learner autonomy, excessive reliance on highly structured technological scaffolding (e.g., immediate answer prompts and automated error correction) may paradoxically foster strategic inertness in learners. When the support provided by technology becomes overly comfortable, students may bypass the processes of active retrieval and productive struggle, thereby contravening the core SRL emphasis on active construction. Consequently, future instructional design must incorporate a “scaffolding fading mechanism”, as students’ competencies improve, the system should dynamically reduce the frequency and granularity of prompts, compelling students to draw upon their own internal cognitive resources. Such an approach would facilitate a seamless transition from technology-assisted learning to genuine self-regulated competence.

Future directions

In the future, the potential of generative AI as an SRL coaching partner should be fully harnessed[33]. Future applications should extend beyond knowledge querying to include deeper interventions in the SRL process. On one hand, AI can function as a “Socratic analyst” of reflective logs, after a student submits a clinical reflection, the AI can automatically identify logical gaps or emotional blind spots within the narrative and pose heuristic follow-up questions, compelling the student to engage in deeper metacognitive excavation. On the other hand, AI can serve as a generative engine for personalized feedback, drawing upon data from a student’s diagnostic pathways in virtual cases; AI not only indicates correctness but also generates tailored explanatory feedback and suggestions for subsequent learning strategies, thereby simulating one-on-one expert coaching. This elevates the role of technology from a mere mirror to a mirror with a compass[12]. Second, future medical education should shift from technology instrumentalism to human-machine collaborative evolution, transcending the view of technology as merely a tool, and shifting toward emphasizing the collaborative evolution of technology, teachers, and students, jointly shaping a more efficient and supportive learning ecosystem. Third, students’ agency and teachers’ new roles should be strengthened, cultivating students to become autonomous, technology-savvy lifelong learners, while transforming teachers’ roles from knowledge transmitters to SRL coaches under technology guidance, responsible for designing learning experiences, providing humanistic care, and advanced guidance[34]. Fourth, more in-depth research on mechanisms and attention to ethical norms are needed, including conducting multi-center, long-term tracking studies to verify the “TEL → SRL → outcomes” mechanism[9,25]. Meanwhile, strict ethical norms for data must be established and attention must be paid to the fairness of technology applications to narrow the digital divide[32]. To achieve these goals, curriculum design, teacher training, technology development, and ethical norms should be systematically constructed, thereby fully utilizing the integrated advantages of TEL and SRL, effectively enhancing medical students’ psychological adaptability and learning outcomes and cultivating high-quality talent capable of meeting future medical challenges.