1. Bibliographic Information

1.1. Title

A systematic review and sequential explanatory synthesis: Artificial intelligence in healthcare education, a case of nursing

1.2. Authors

• S. Asli Bozkurt PhD, RN, CHSE, Assistant Professor
• Sinan Aydoan PhD, RN, Assistant Professor
• Fatma Dursun Ergezen PhD, RN, Assistant Professor
• Aykut Türkolu PhD, Visiting scholar

Affiliations:

• Massachusetts College of Pharmacy and Health Sciences University, Boston, Massachusetts, USA
• Burdur Mehmet Akif Ersoy University, Burdur, Turkey
• Akdeniz University, Antalya, Turkey
• Boston University, Boston, Massachusetts, USA

1.3. Journal/Conference

The paper was published in the International Nursing Review, a peer-reviewed journal in the field of nursing. This journal typically publishes articles that contribute to the global advancement of nursing knowledge, practice, and policy, suggesting that the work is aimed at a professional nursing audience and has undergone a standard academic review process.

1.4. Publication Year

2025

1.5. Abstract

This systematic review aimed to explore the impact of artificial intelligence (AI) on knowledge acquisition, skills development, and attitudes among nursing students. The authors employed a sequential explanatory mixed-method systematic review design, adhering to PRISMA guidelines. A comprehensive search across five electronic databases identified nine eligible research articles. The methodological quality of these studies was assessed using the mixed-method appraisal tool (MMAT).

The results indicated that quantitative studies largely found positive relationships between AI interventions and improvements in nursing students' knowledge acquisition, skills development, and attitudes toward AI. Qualitative findings corroborated these positive outcomes, highlighting AI's role in fostering learning engagement, self-efficacy, and confidence.

The review concludes that AI holds significant potential for enhancing knowledge acquisition, skills development, and cultivating positive attitudes in nursing students. The implications for nursing practice and policy suggest that AI-driven interventions can improve clinical decision-making, confidence, and knowledge acquisition. The paper recommends developing policies, standardized guidelines, and faculty development programs to ensure ethical integration, equitable access, and adequate training for nurse educators, thereby maximizing AI's benefits in contemporary healthcare education.

1.6. Original Source Link

/files/papers/6939911daf4df128671a0d89/paper.pdf (This appears to be a local or internal file link; the abstract states the DOI: https://doi.org/10.1111/inr.70018, indicating it is an officially published paper).

2. Executive Summary

2.1. Background & Motivation

The core problem the paper aims to address is the need to effectively integrate artificial intelligence (AI) into nursing education to enhance learning experiences and adequately prepare future healthcare professionals for the complexities of modern healthcare. This problem is crucial because AI has demonstrated significant potential to improve patient outcomes, optimize healthcare processes, and support the continuous advancement of medical practices.

Specific challenges or gaps in prior research highlighted by the authors include the lack of an extensive critical literature review that comprehensively summarizes and explains the potential uses and impact of AI interventions on nursing students' knowledge, skills, and attitudes. Despite the recognized benefits of AI, concerns such as potential overreliance, algorithmic biases, data privacy and security issues, and the financial burden of implementation underscore the need for a careful and evidence-based strategy for AI integration in nursing education. This paper seeks to fill this gap by providing a systematic synthesis of existing evidence.

The paper's entry point is to conduct a systematic review, using a mixed-method approach, to meticulously explore and synthesize the current evidence regarding AI's influence on the knowledge acquisition, skills development, and attitudes of nursing students. This approach allows for a comprehensive understanding that integrates both quantitative measures of impact and qualitative insights into student experiences and perceptions.

2.2. Main Contributions / Findings

The paper's primary contributions are:

• Comprehensive Review: It is identified as the first mixed-method systematic review to comprehensively evaluate the impact of artificial intelligence (AI) on knowledge, skills, and attitudes (KSAs) specifically among nursing students. This provides a unique and holistic understanding of AI's role in this educational context.

• Synthesis of Quantitative Evidence: The review systematically gathered and analyzed quantitative data, demonstrating positive relations between AI interventions and improvements in knowledge acquisition, skills development (particularly communication skills), and attitudes toward AI.

• Integration of Qualitative Insights: It incorporated qualitative findings, which revealed that AI positively influences learning engagement, self-efficacy, and confidence among nursing students, providing deeper context and understanding to the quantitative results.

• Identification of Challenges: The review also highlighted challenges and concerns related to AI integration, such as limited realism, lack of interaction, and student anxieties regarding job loss, which are critical for thoughtful implementation.

• Implications for Practice and Policy: The paper provides actionable implications for nursing practice (e.g., enhancing clinical decision-making, confidence, and knowledge acquisition) and nursing policy (e.g., developing standardized guidelines, faculty development programs, and addressing ethical AI integration and equitable access).

  The key conclusions and findings reached by the paper are:

• AI holds significant potential to support knowledge acquisition, skills development, and foster positive attitudes among nursing students.

• AI interventions can create more interactive, personalized, and impactful learning environments.

• The effectiveness of AI in enhancing psychomotor skills requires further investigation, as current data do not provide a definitive conclusion.

• Addressing concerns about interaction issues, realism, and job security is crucial for successful AI integration.

• Strategic development of policies and faculty training is essential to maximize the benefits of AI in nursing education. These findings collectively solve the problem of understanding the multifaceted impact of AI on nursing education, providing a foundation for informed decision-making regarding its implementation and future research directions.

3. Prerequisite Knowledge & Related Work

3.1. Foundational Concepts

To fully understand this paper, a beginner should be familiar with several foundational concepts related to research methodology and the specific domain of artificial intelligence in education.

• Artificial Intelligence (AI): At its core, AI refers to computer software and systems capable of performing tasks that typically require human intelligence, such as learning, problem-solving, decision-making, perception, and understanding language. In the context of healthcare education, AI applications can range from virtual simulations and personalized feedback systems to predictive analytics for student performance. The paper defines AI as "computer software that can assist in making logical and sound judgments."
• Knowledge, Skills, and Attitudes (KSAs): These three components are crucial for competency assessment in many professional fields, including nursing. The paper references Bloom's Taxonomy, a widely used educational framework.
  • Knowledge (Cognitive Domain): Refers to the awareness, comprehension, and expertise gained through education or experience. This includes recalling facts, understanding concepts, and applying principles. In Bloom's Taxonomy, this corresponds to levels like remembering, understanding, and applying.
  • Skills (Psychomotor Domain): Refers to abilities developed through intentional practice and continuous effort to perform tasks. In nursing, this involves clinical procedures, communication techniques, and critical thinking in practice. This aligns with Bloom's psychomotor domain, which deals with physical movement, coordination, and use of motor-skill areas.
  • Attitudes (Affective Domain): Refers to an individual's tendency to react favorably or unfavorably to ideas, individuals, or situations. This encompasses values, beliefs, appreciation, motivation, and engagement. In Bloom's Taxonomy, this relates to the affective domain, which includes feelings, emotions, and motivations.
• Systematic Review: A systematic review is a type of literature review that collects and critically analyzes multiple research studies or papers. It is a rigorous process that uses predefined methods to identify, evaluate, and synthesize all relevant evidence on a particular research question. The goal is to provide a comprehensive and unbiased summary of current knowledge on a topic.
• Mixed-Method Systematic Review: This is a systematic review that integrates both quantitative (numerical data, statistical analysis) and qualitative (non-numerical data, thematic analysis) studies. The paper specifically uses a sequential explanatory mixed-method systematic review design. This means that quantitative findings are initially gathered and analyzed, followed by qualitative findings, with the qualitative data then used to help explain or elaborate on the quantitative results. This sequential approach allows for a deeper and more nuanced understanding of the research phenomenon.
• PRISMA Guidelines (Preferred Reporting Items for Systematic Reviews and Meta-Analyses): These are evidence-based minimum set of items for reporting in systematic reviews and meta-analyses. The aim of PRISMA is to help authors improve the reporting of systematic reviews and meta-analyses. Following these guidelines ensures transparency, completeness, and rigor in the review process.
• PICO Elements: PICO is a mnemonic used in evidence-based practice to formulate a searchable clinical question.
  • P (Population): The group of individuals or patients being studied (e.g., nursing students).
  • I (Intervention): The action or treatment being studied (e.g., AI-based education interventions).
  • C (Comparison): The alternative intervention or standard of care being compared against (e.g., traditional education).
  • O (Outcome): The measurable results or effects of the intervention (e.g., students' knowledge, skills, or attitudes).
• Mixed-Method Appraisal Tool (MMAT): The MMAT is a critical appraisal tool designed to assess the methodological quality of studies included in a systematic review. It can be used for various study designs, including qualitative, quantitative (randomized controlled trials, non-randomized, descriptive), and mixed-methods studies. It helps reviewers evaluate the trustworthiness and rigor of the included research.

3.2. Previous Works

The paper's introduction effectively sets the stage by summarizing prior applications and benefits of AI in nursing education, as well as the challenges.

Benefits and Applications:

• Enhanced Learning Experiences: AI is seen as a way to enrich learning for future nurses.
• Diverse Purposes: AI is employed for simulation, performance assessment, personalized feedback, development of creative educational materials, case studies, and self-directed learning (Baashar et al., 2022; Irwin et al., 2023; Swiecki et al., 2022).
• Transition to Work Settings: AI may help nurses transition into practice, improving their practically applicable knowledge and reducing training costs (Abujaber et al., 2023; Alzahrani et al., 2023).
• Administrative Tasks: AI can automate scheduling, grading, and data management, reducing educator burden (Kilmon et al., 2010).
• Virtual Simulations: AI-powered virtual simulations offer safe, realistic environments for clinical skills and decision-making practice, reducing patient risk (Gillespie et al., 2021; Hameed et al., 2022). These simulations also aid in understanding social determinants of health (Rahman et al., 2022).
• Predictive Analytics: AI-based predictive analytics can identify students at academic or clinical risk early, allowing for proactive intervention and potentially higher retention rates, addressing the nursing shortage (Moseley & Mead, 2008; Rastrollo-Guerrero et al., 2020).
• Overall Impact: Integrating AI can create more efficient, tailored, and impactful learning experiences (Buchanan et al., 2021).

Challenges and Concerns:

• Overreliance on AI: Potential disruption of critical thinking skills development (Glauberman et al., 2023).
• Algorithmic Bias: AI algorithms can amplify biases from training data, leading to disparities in education and patient care (Borg, 2022).
• Data Privacy and Security: The collection and storage of sensitive student and patient data raise ethical and legal concerns (Borenstein & Howard, 2021).
• Financial Burden: High costs of implementing and maintaining AI technologies may divert funds from other essential educational components (De Gagne, 2023).

3.3. Technological Evolution

The field of artificial intelligence has evolved from early rule-based systems to machine learning and deep learning approaches, enabling more sophisticated capabilities such as natural language processing, computer vision, and predictive analytics. This evolution has led to a transition from theoretical concepts to practical applications in various domains, including healthcare and education. Early applications often involved expert systems or basic simulations. More recently, advancements in natural language processing (NLP) and generative AI have led to tools like AI chatbots and virtual patient simulators that offer increasingly realistic and interactive learning experiences. These tools facilitate personalized learning paths, automated feedback, and complex scenario-based training, moving beyond traditional didactic teaching methods. The paper's work fits within this timeline by systematically reviewing the current state of AI application in a specific educational context (nursing) to assess its practical impact on student outcomes, thereby informing future developments and policy.

3.4. Differentiation Analysis

Compared to general reviews on AI in education or healthcare, the core differences and innovations of this paper's approach are:

• Specific Population Focus: This review exclusively focuses on nursing students, providing highly relevant insights for nursing education specifically, rather than healthcare education broadly.
• Comprehensive Outcome Triad: It investigates the combined impact on knowledge acquisition, skills development, and attitudes (KSAs), offering a holistic view of AI's educational effects, which is crucial for competency-based education.
• Mixed-Method Design: The use of a sequential explanatory mixed-method systematic review is a key differentiator. This approach allows for a richer understanding by synthesizing both quantitative (measurable impacts) and qualitative (experiential insights) evidence, providing both "what" and "why" AI affects students. This is noted as the "first review of its kind" to combine these aspects for nursing education.
• Rigorous Methodology: Adherence to PRISMA guidelines and using the MMAT for quality appraisal ensures the review's rigor and transparency.
• Identification of Specific Challenges: While general AI reviews might mention challenges, this paper specifically pinpoints challenges relevant to nursing students, such as the impact on psychomotor skills and student concerns about realism and job security.

4. Methodology

4.1. Principles

The core idea of the method used in this paper is to conduct a sequential explanatory mixed-method systematic review. This approach is rooted in the principle of combining the strengths of both quantitative and qualitative research to gain a more comprehensive understanding of a complex phenomenon, in this case, the impact of AI on nursing students' Knowledge, Skills, and Attitudes (KSAs).

The theoretical basis behind this design is that quantitative data can provide measurable outcomes and statistical relationships (e.g., "AI interventions positively impact knowledge scores"), while qualitative data can offer rich contextual details, experiences, and perceptions that explain how and why those quantitative outcomes occur (e.g., "students feel more engaged and confident due to AI's interactive nature"). The "sequential explanatory" nature means that the quantitative data collection and analysis precede the qualitative phase, with the qualitative findings then used to help explain or elaborate on the quantitative results.

The review also strictly adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, ensuring transparency, rigor, and reproducibility of the systematic review process.

4.2. Core Methodology In-depth (Layer by Layer)

4.2.1. Aim and Research Questions

The overall aim of the review was to explore the impact of AI on knowledge acquisition, skills development, and attitudes among nursing students. To guide this exploration, specific research questions were formulated using the PICO elements (Population, Intervention, Comparison, Outcome).

• Quantitative research question: What is the impact of AI-based interventions (I) on nursing students' KSAs (O), compared with traditional education (C)?
• Qualitative research question: What are nursing students' perceptions (O) of the benefits and challenges (O) of AI-based education interventions (I) on their KSAs (O), compared with traditional education (C)?

4.2.2. Design

This review employed a sequential explanatory mixed-method systematic review design.

• Sequential Explanatory: This means the quantitative data was collected and analyzed first (QUAN), followed by qualitative data collection and analysis (QUAL). The qualitative findings were then used to help explain or elaborate on the quantitative results.

• PRISMA Guidelines: The review was conducted and reported in line with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, which ensure systematic and transparent reporting of literature reviews.

• PROSPERO Registration: The research protocol was registered in PROSPERO with the ID CRD42023428702, indicating a pre-specified plan to reduce bias.

  The systematic review process adhered to seven standard steps:

1. Formulating Qualitative and Quantitative Review Questions: As detailed above using the PICO framework.
2. Establishing Eligibility Criteria: Defining specific criteria for inclusion and exclusion of studies.
3. Conducting a Comprehensive Literature Search: Searching multiple electronic databases.
4. Identifying Potentially Relevant Studies: Initial screening of titles and abstracts.
5. Selecting Relevant Studies Based on Full-Text Review: Detailed evaluation of full-text articles against eligibility criteria.
6. Assessing the Quality of Included Studies: Evaluating the methodological rigor of selected studies.
7. Synthesizing the Findings from the Included Studies: Combining and interpreting the results from both quantitative and qualitative studies.

4.2.3. Search Methods

The search strategy included:

• Search Terms: Relevant Medical Subject Heading (MeSH) words and additional keywords were used. (The specific terms were in Appendix A of the original paper, which was not provided in the input).

• PICO Criteria: Studies were included based on the predefined PICO inclusion and exclusion criteria. These criteria are detailed in the following table:

  The following are the results from Table 1 of the original paper:

  	The inclusion criteria	The exclusion criteria
  Population	Nursing students (undergraduate and graduate)	Studies related to patient education, nurses, or students other than nursing students
  Intervention	Artificial intelligence-based education interventions	Studies that did not include AI-based educational interventions
  Comparison	Traditional education
  Outcomes	Students' knowledge, skills, or attitudes	Studies evaluated other outcomes other than knowledge, skills, or attitudes

• Databases: A comprehensive electronic database search was conducted with the assistance of an experienced librarian. The databases included:

  • CINAHL COMPLETE (Cumulative Index to Nursing and Allied Health Literature)
  • MEDLINE
  • Web of Science
  • PubMed
  • ProQuest Central

• Additional Methods: No additional methods such as citation searching or website identification were used.

• Time Frame: No time frame limitation was applied to the search.

• Search Date: The searches were conducted on December 27, 2023.

• Screening Process: The process involved an initial assessment of titles and abstracts, followed by a full-text review against the PICO inclusion and exclusion criteria.

4.2.4. Search Outcomes

The initial search identified 1922 studies across the selected databases.

• Duplicate records (n = 72) were removed, leaving 1850 studies.

• After reviewing titles and abstracts, 1815 studies were excluded based on the inclusion and exclusion criteria.

• This left 35 full-text studies for evaluation of eligibility.

• Out of these 35, 26 studies were eliminated for various reasons: editorial ($n = 8$), not focusing on AI ($n = 7$), not related to students ($n = 3$), not research ($n = 3$), review ($n = 2$), not related to nursing ($n = 1$), retracted ($n = 1$), and duplication ($n = 1$).

• Ultimately, nine research articles were included in the final review.

  The PRISMA diagram (Figure 1 from the original paper) visually illustrates this selection process:

  该图像是PRISMA流程图，展示了在进行人工智能在护理教育中的影响研究系统评价时，文献筛选的各个阶段，包括识别、筛选和纳入的研究数量。最终纳入的研究数量为9。

As depicted in the PRISMA diagram, the systematic process involved identifying records, screening them based on titles and abstracts, assessing eligibility through full-text review, and finally including the relevant studies. The diagram clearly shows the flow from the initial large number of identified records down to the nine studies that met all inclusion criteria for the review.

4.2.5. Quality Appraisal

• Tool Used: The mixed-method appraisal tool (MMAT) was employed to assess the methodological quality of the included studies. The MMAT is versatile, capable of evaluating qualitative, quantitative, and mixed-method study designs. It consists of five sets of questions tailored to different study types.
• Scoring: Earlier versions of the MMAT assigned an overall score, but the most recent version (which the authors likely used, as they cite the 2018 version) advocates for a holistic assessment rather than a single numerical score. Reviewers answer questions with "yes," "no," or "can't tell."
• Process: Two researchers (A and C) independently reviewed the studies. Any inconsistencies were rechecked and discussed by three researchers (A, B, and C) until a consensus was reached, ensuring reliability in the quality assessment.
• General Findings from Appraisal:

  • All included studies had explicitly stated research questions.

  • The collected data were deemed sufficient to address these questions.

  • For qualitative components, methodologies were well-suited, and data gathering, analysis, and interpretation were rigorous.

  • For quantitative non-randomized studies, participant selection, measurement instruments, and intervention administration were appropriate.

  • For quantitative descriptive studies, sampling strategies, measurement tools, sample representation, and statistical analyses were appropriate.

  • For the mixed-method study, the synthesis of qualitative and quantitative findings was effective, with clear interpretation and appropriate discussion of consistencies and inconsistencies.

    The following are the results from Table 3 of the original paper:

    	First author and year					Rodriguez-Arrastia, 2022	Shorey, 2023	Shorey, 2020	Simsek-Cetinkaya 2023
    	Chang, 2022	Abdelaliem, 2022	Han, 2022	Kwak, Anh, 2022a	Kwak, Seo, 2022b
    Screening questions
    S1. Are there clear research questions?	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
    S2. Do the collected data allow us to address the research questions?	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
    1. Qualitative	Yes					Yes		Yes
    1.1. Are the qualitative data collection methods adequate to address the research question?	Yes					Yes		Yes
    1.3. Are the findings adequately derived from the data?	Yes					Yes
    1.4. Is there coherence between qualitative data sources, collection, analysis, and interpretation?	Yes					Yes		Yes
    3. Quantitative nonrandomized
    3.1. Are the participants representative of the target population?
    3.2. Are measurements appropriate regarding both the outcome and intervention (or exposure)?	Yes		Yes				Yes		Yes
    3.3. Are there complete outcome data?	Yes		Yes				Yes		Yes
    3.4. Are the confounders accounted for in the design and analysis?	Yes		Yes				Yes		Yes
    3.5. During the study period, is the intervention administered (or exposure occurred) as intended?	Yes		Yes				Yes		Yes
    4. Quantitative descriptive
    4.1. Is the sampling strategy relevant to address the research question?		Yes		Yes	Yes
    4.2. Is the sample representative of the target population?		Yes		Yes	Yes
    4.3. Are the measurements appropriate?		Yes		Yes	Yes
    4.4. Is the risk of nonresponse bias low?		Yes		Yes	Yes
    4.5. Is the statistical analysis appropriate?		Yes		Yes	Yes
    5. Mixed methods
    5.1. Is there an adequate rationale for using mixed methods design to address the research question?
    5.2. Are the different components of the study effectively integrated to answer the research question?	Yes
    5.3. Are the outputs of the integration of qualitative and quantitative components adequately interpreted?	Yes
    5.4. Are divergences and inconsistencies between quantitative and qualitative results adequately addressed?	Yes
    5.5. Do the different components of the study adhere to the quality criteria of each tradition of the methods involved?	Yes

4.2.6. Data Extraction

• Duplicate Removal: Endnote X9.3.3 software was initially used to automatically remove duplicate records, with manual verification by two researchers (A and B).
• Initial Screening: Two independent reviewers (A and B) screened titles and abstracts based on predefined eligibility criteria.
• Conflict Resolution: Conflicts or disagreements during screening were resolved through discussion or by consulting a third reviewer (C).
• Full-Text Evaluation: Full-text articles for potentially eligible studies were retrieved and further evaluated, with reasons for exclusion documented.
• Data Management: Eligible articles were imported into ATLAS.ti qualitative analysis software. This software was chosen for its capabilities in efficient organization, categorization, systematic coding, annotation, and data management, which ensured consistency across the research team.
• Standardized Form: A standardized form was used to extract specific data from each included study, covering: (1) author, year, and country; (2) aim/research questions; (3) design; (4) sample size; (5) theoretical framework; and (6) instruments.

4.2.7. Ethical Consideration

Since all data and information were obtained from publicly available literature and databases, no ethical approval from an institutional review board was required for this systematic review.

5. Experimental Setup

This section describes the characteristics of the primary studies included in the systematic review, rather than an experimental setup of the review itself.

5.1. Datasets (Characteristics of Included Studies)

The "datasets" in this context refer to the characteristics of the nine individual studies that were included in the systematic review. These studies represent the primary data sources for the review's synthesis.

• Geographical Distribution:
  • Two studies were from Singapore (Shorey et al., 2020, 2023).
  • Three were from the Republic of Korea (Han et al., 2022; Kwak et al., 2022a, 2022b).
  • One each was from Turkey (Simsek-Cetinkaya & Cakir, 2023), Spain (Rodriguez-Arrastia et al., 2022), Saudi Arabia (Farghaly Abdelaliem et al., 2022), and Taiwan (Chang et al., 2022).
• Sample Size: The sample sizes of the included studies ranged from 24 to 697 participants.
• Theoretical Frameworks: Only one article explicitly stated using a theoretical framework (Simsek-Cetinkaya & Cakir, 2023).
• Study Designs:
  • Two studies were purely qualitative research (Rodriguez-Arrastia et al., 2022; Shorey et al., 2020).
  • One was mixed-method research (Chang et al., 2022).
  • Six were purely quantitative research. Within the quantitative studies:

    • Two were cross-sectional (Farghaly Abdelaliem et al., 2022; Kwak et al., 2022a).

    • Three were experimental (Han et al., 2022; Shorey et al., 2023; Simsek-Cetinkaya & Cakir, 2023).

    • One was a path analysis (Kwak et al., 2022b).

      For a detailed breakdown of each study's characteristics, refer to the "Data Presentation (Tables)" subsection within the "Results & Analysis" section below, where Table 2 from the original paper is transcribed.

5.2. Evaluation Metrics (Outcomes of Included Studies)

As this is a systematic review, it did not employ its own evaluation metrics in the traditional sense. Instead, it synthesized the reported outcomes (which acted as "evaluation metrics" for the primary studies) across the included research articles. The core outcomes evaluated across the studies, as defined by the review's aim and PICO questions, were:

• Knowledge Acquisition: How much students learned or improved their understanding in specific areas. This was measured by various instruments in the primary studies, such as knowledge tests, ethical awareness scales, or assessments of self-directed knowledge.

• Skills Development: Improvements in practical abilities. This included communication skills, clinical reasoning skills, and psychomotor skills like breast self-examination. Instruments varied, from performance scores to self-efficacy scales.

• Attitudes: Students' perceptions, beliefs, satisfaction, anxiety, self-efficacy, and behavioral intentions regarding AI or specific topics. Measures included attitude scales, anxiety inventories, and self-efficacy subscales.

  The paper aimed to identify positive relations or improvements in these KSAs resulting from AI interventions. For qualitative studies, the "metrics" were themes emerging from student perceptions regarding learning enhancement, confidence, user-friendliness, and challenges.

5.3. Baselines (Comparators in Included Studies)

The primary comparator for AI interventions, as specified in the quantitative PICO question, was traditional education. This means that many included studies likely compared an experimental group receiving AI-based interventions with a control group receiving conventional teaching methods. The quantitative studies aimed to determine if AI interventions led to superior or different outcomes compared to these traditional approaches. For qualitative studies, the "comparison" was often implicit, examining student perceptions of AI interventions within their existing educational context, sometimes contrasting them with prior non-AI experiences.

6. Results & Analysis

6.1. Core Results Analysis

6.1.1. Study Characteristics

The nine included studies represented diverse geographical origins and methodological approaches. The sample sizes varied widely, indicating a range of study scopes. A notable observation is the infrequent use of explicit theoretical frameworks in the primary studies, with only one article (Simsek-Cetinkaya & Cakir, 2023) mentioning one. The distribution of study designs included qualitative, mixed-method, and various quantitative approaches (cross-sectional, experimental, path analysis), providing a mixed evidence base for synthesis.

The following are the results from Table 2 of the original paper:

  <2 rowspan="2">Sample size</td>
  <td rowspan="2">Theoretical framework</td>
  <td rowspan="2">Instruments</td>
  <td rowspan="2">Results</td>
</tr>

6.1.2. Quality Appraisal

The quality appraisal using the MMAT generally indicated good methodological quality across the included studies. All studies clearly stated research questions and collected sufficient data. Qualitative methods were appropriate and rigorously executed for qualitative components. Quantitative non-randomized and descriptive studies showed appropriate participant selection, measurement, confounder accounting, and statistical analysis. The mixed-method study demonstrated effective integration and interpretation of its components. This robust quality assessment enhances the trustworthiness of the review's findings.

6.1.3. Results from Quantitative Studies

Seven studies used quantitative methods, and their findings were categorized into three themes: Knowledge, Skills, and Attitude.

6.1.3.1. Theme 1: Knowledge

• Positive Correlations: Farghaly Abdelaliem et al. (2022) found a positive correlation between AI perception and smartphone addiction among nursing students, suggesting that students more accustomed to smart devices might have a more positive view of AI.
• AI Ethics Awareness: Kwak, Ahn, et al. (2022a) identified a positive correlation between AI ethics awareness and AI self-efficacy, implying that understanding ethical aspects of AI can boost confidence in using it.
• Learning Achievement: Chang et al. (2022) demonstrated a statistically significant increase in students' basic vaccine knowledge, infectious diseases knowledge, and decision-making competence using a mobile chatbot-based learning approach. This strongly validates AI's role in knowledge acquisition.
• Self-Directed Knowledge and Interest: $Han et al. (2022)$ observed that while Electronic Fetal Monitor (EFM) knowledge didn't change significantly with an AI teaching tool, students' interest in education and self-directed knowledge levels increased significantly. This highlights AI's potential to foster engagement and autonomy in learning.

6.1.3.2. Theme 2: Skills

• Communication Skills Improvement: Shorey et al. (2023) reported that AI interventions (Virtual Counseling Application) improved nursing students' clinical communication skills. This is a clear positive impact of AI on a crucial professional skill.
• Clinical Reasoning Skills (No Change): Conversely, $Han et al. (2022)$ found that AI did not significantly improve students' clinical reasoning skills. This suggests that the impact of AI on complex cognitive skills might be nuanced or require different AI approaches.
• Psychomotor Skills (Mixed Results): Simsek-Cetinkaya and Cakir (2023) investigated breast self-examination skills. They found that the group using AI-supported interactive screen-based simulation scored significantly lower in performance compared to the standardized patient group. This indicates that for certain psychomotor skills requiring tactile feedback or real human interaction, AI simulations may not be as effective as traditional methods, or current AI tools may lack the necessary realism.

6.1.3.3. Theme 3: Attitude

• Communication Skills and Self-Efficacy: Shorey et al. (2023) observed significant enhancements in attitudes toward communication skills and self-efficacy in communication following an AI-assisted intervention.
• Satisfaction vs. Anxiety: Simsek-Cetinkaya and Cakir (2023) found that participants using AI for breast self-examination reported higher satisfaction but also exhibited increased anxiety levels. This suggests a trade-off or complex emotional response to AI tools in skill practice.
• Attitude, Anxiety, and Ethics: Kwak, Ahn, et al. (2022a) identified a positive correlation between anxiety and negative attitudes toward AI, while also noting a positive association between AI self-efficacy and awareness of AI ethics. This implies that addressing anxiety and promoting ethical understanding are crucial for fostering positive attitudes.
• Perceptions of AI: Abdelaliem et al. (2022) again found a positive correlation between smartphone addiction and perceptions of AI, possibly indicating a readiness or openness to technology among those who use smart devices frequently.
• Behavioral Intention (Acceptance Model): Kwak, Seo, et al. (2022b) conducted a path analysis, revealing that AI behavioral intention (intention to use AI) positively correlated with performance expectancy, effort expectancy, social influence, facilitating conditions, self-efficacy, and a positive attitude toward AI. Conversely, anxiety and negative attitudes toward AI negatively correlated with AI behavioral intention. This highlights that multiple factors, including practical benefits and social context, influence students' acceptance and willingness to use AI.

6.1.4. Results from Qualitative Studies

Three studies adopted qualitative research (Chang et al., 2022; Rodriguez-Arrastia et al., 2022; Shorey et al., 2020), and their findings were categorized into three principal themes: learning enhancement and confidence, user-friendly interface and accessibility, and challenges and concerns.

6.1.4.1. Theme 1: Learning Enhancement and Confidence

Qualitative studies consistently showed that AI fostered positive learning experiences.

• Learning Engagement and Comprehension: Participants noted that AI tools, like mobile chatbot-based learning, made learning more appealing and facilitated comprehension, especially for complex topics such as maternal infectious disease prevention and obstetric vaccination knowledge (Chang et al., 2022).
• In-depth Exploration and Integration: AI approaches enabled deeper exploration and integration of knowledge, enhancing students' ability to educate patients (Chang et al., 2022).
• Confidence and Self-Efficacy: The use of evidence-based information through chatbots augmented confidence in nursing practices and provided quick, user-friendly assistance, thereby fostering self-efficacy and preparedness (Rodriguez-Arrastia et al., 2022).

6.1.4.2. Theme 2: User-Friendly Interface and Accessibility

• Ease of Use: Students reported that AI tools featured user-friendly interfaces, which enhanced accessibility and promoted safe practice environments.
• Increased Interaction and Confidence: Increased interaction with AI-enhanced virtual patients was observed to bolster students' confidence in technical and professional communication skills (Shorey et al., 2020).
• Quick Information Access: Testimonials highlighted the ease of accessing current information through chatbot interactions (Chang et al., 2022) and the utility and user-friendliness of chatbots in emergencies (Rodriguez-Arrastia et al., 2022).

6.1.4.3. Theme 3: Challenges and Concerns

• Interaction Issues: Students noted concerns such as the lack of interaction, confidence, and depth in thinking when using some AI tools (Chang et al., 2022).
• Realism and Emotional Depth: A virtual patient program was critiqued for its limited realism, emotional depth, diversity in responses, and lack of nonverbal cues, affecting its perceived effectiveness (Shorey et al., 2020).
• Professional Acceptance and Fear of the Unknown: One participant suggested that professional acceptance of AI might be hindered more by a fear of the unknown than by age (Rodriguez-Arrastia et al., 2022). This implies a need for better education and exposure to AI.

6.1.5. Synthesis of Quantitative and Qualitative Findings

The sequential explanatory approach involved first synthesizing quantitative results, then qualitative results, and finally integrating them to provide new insights.

• Knowledge and Attitudes: The combined evidence strongly suggests a beneficial impact of AI on students' knowledge and attitudes. Quantitative studies showed increased knowledge, and qualitative insights emphasized AI's positive effect on learning engagement and confidence. However, a negative attitude toward AI was linked to heightened anxiety, which could potentially diminish learning outcomes.
• Psychomotor Skills: The implications of AI on psychomotor skills remained inconclusive. While some skills like communication showed improvement, others like breast self-examination did not, or even performed worse with AI compared to traditional methods, indicating a need for further investigation. This specific area represents a knowledge gap highlighted by the mixed findings.

6.2. Data Presentation (Tables)

The tables from the original paper, providing detailed characteristics of the included studies and their quality appraisal, have been transcribed in the relevant subsections above (Section 4.2.3 and Section 4.2.5 for the Methodology and Section 6.1.1 for Study Characteristics).

6.3. Ablation Studies / Parameter Analysis

The systematic review does not include ablation studies or parameter analysis, as these are typically conducted within primary research studies to evaluate components of a proposed model or the impact of specific parameters. As a secondary research method, this systematic review synthesizes findings from existing studies rather than performing new experimental analyses.

7. Conclusion & Reflections

7.1. Conclusion Summary

This systematic review concludes that Artificial Intelligence (AI) holds significant potential to offer innovative ways to enhance knowledge acquisition, skills development, and attitudes among nursing students. The evidence synthesized demonstrates AI's ability to improve learning outcomes and foster engagement through personalized and interactive approaches. Specifically, AI-driven tools can enhance students' understanding of complex clinical concepts and build their confidence and self-efficacy within controlled, risk-free environments. While AI shows promise, its integration into nursing education is not without challenges, including limited realism in some AI tools, ethical considerations, and accessibility barriers, all of which must be addressed to fully leverage its capabilities. The review emphasizes that by strategically integrating AI and continuously innovating, nursing education can bridge the gap between traditional teaching and the evolving demands of healthcare, ultimately empowering students to become confident, skilled, and adaptable professionals.

7.2. Limitations & Future Work

The authors acknowledged several limitations of their review:

• Limited Generalizability: The results are confined to the samples of the included studies, which may restrict the generalizability of the findings due to potential lack of diversity in nursing students, educational institutions, or clinical settings.

• Small Sample Sizes: The varying and sometimes smaller sample sizes in the included studies could reduce their statistical power, impacting the robustness of some quantitative findings.

  Despite these limitations, the review identified several areas for future work:

• Psychomotor Skill Development: Further research is urgently needed to explore the impact of AI on the psychomotor skill development of nursing students, especially given the mixed and sometimes negative findings in the current review (e.g., breast self-examination skills).

• Student Concerns: More comprehensive studies are required to investigate nursing students' specific concerns about AI, including lack of realism, interaction issues, and the potential loss of jobs.

• Trust Building Strategies: Strategies should be developed to build trust and effectively integrate AI into nursing education and practice.

• Curriculum and Simulation Integration: AI should be integrated into curricula and simulation training to improve positive attitudes toward AI.

• Ethical AI Integration: Policies, standardized guidelines, and faculty development programs need to be developed to promote ethical AI integration, equitable access, and faculty training.

7.3. Personal Insights & Critique

This paper provides a valuable and timely mixed-method systematic review on a critical topic. Its strength lies in its comprehensive approach, combining both quantitative and qualitative evidence to offer a nuanced understanding of AI's impact on nursing students' KSAs. The adherence to PRISMA guidelines and the use of the MMAT ensure methodological rigor, making its findings more trustworthy. The fact that it's identified as the "first of its kind" for nursing students specifically highlights its innovative contribution to the field.

One major insight gained is the dual nature of AI's promise and peril. While AI clearly enhances knowledge acquisition, learning engagement, and communication skills, its effectiveness with psychomotor skills is not consistently positive. This suggests that AI is not a panacea and requires careful, context-specific application. The observed increase in anxiety despite higher satisfaction with certain AI tools is a critical finding, indicating that user experience is complex and needs to be managed proactively.

The methods and conclusions of this paper can be readily transferred to other domains of healthcare education (e.g., medical, allied health) or even broader professional training where practical skills and emotional intelligence are crucial. The framework of analyzing Knowledge, Skills, and Attitudes is universally applicable.

However, a potential area for deeper exploration, even within the scope of a systematic review, could be a more detailed breakdown of what types of AI interventions were most effective for which specific aspects of KSAs. For instance, chatbots seem good for knowledge and self-efficacy, while virtual patients help communication, but current screen-based simulations struggle with complex psychomotor skills. A finer-grained analysis of AI categories and their specific strengths/weaknesses would be beneficial. The paper's use of ChatGPT for language editing is an interesting meta-point, demonstrating the authors' own practical engagement with AI tools in their academic process, which subtly reinforces the paper's themes. The ethical considerations around data privacy and algorithmic bias are mentioned in the introduction but not heavily elaborated in the findings section, suggesting this remains a significant area for future empirical research directly within nursing education.