Content area
Aim
This study aims to compare the effectiveness of six different teaching methods with traditional teaching approaches in community nursing education for nursing students in China.
BackgroundIn recent years, an increasing number of Chinese nursing educators exploring various pedagogical approaches to enhance the educational outcomes in community nursing. However, there is still no consensus on the superiority of different teaching methods and no direct comparisons of their effectiveness have been made. Therefore, evaluating the impact of six teaching strategies on community nursing education for Chinese nursing students is both necessary and timely.
DesignThis study is a systematic review and network meta-analysis.
MethodsIn November 2024, a comprehensive search was conducted across nine databases to identify studies that met predefined inclusion and exclusion criteria. This process involved screening studies based on set criteria, extracting relevant data, and assessing the quality of the studies before conducting the network meta-analysis. The review protocol of this study was prospectively registered in the PROSPERO (CRD42025635443).
ResultsFrom an initial pool of 3077 articles, 33 were meticulously selected for a network meta-analysis. The findings indicate that PBL significantly outperforms other methods in improving nursing students' final examination scores and practical skills. PAD class approach emerged as the most effective method in fostering self-directed learning capabilities among Chinese nursing students.
ConclusionThe results of this study show that PBL and PAD class were identified as having the greatest potential to enhance Chinese nursing student academic achievement and self-directed learning. Nonetheless, future investigations should employ larger sample sizes and more rigorous methodologies to substantiate these findings.
Tweetable abstractWith the increasing aging of the population in China, it is urgent for nursing educators to think about how to cultivate nursing students' interest in community nursing and improve the teaching quality of community nursing in order to meet the growing demand for primary health services. More and more nursing educators have realized the shortcomings of traditional teaching methods, so they try to use a variety of teaching methods to improve the teaching effect of community nursing. However, there is currently no consensus on the superiority of different teaching methods, and no studies have directly compared their effects. Therefore, evaluating the impact of various teaching strategies on community nursing education is both necessary and meaningful. Hence, we employed network meta-analysis to assess the effectiveness of six common teaching methods used in community nursing education in China, aiming to provide insights for the selection of appropriate teaching models for community nursing education. From the establishment of the database until November 2024, a comprehensive search of nine databases was conducted. According to inclusion and exclusion criteria, 33 literatures on community nursing teaching of nursing students in China were included for analysis. The results ranked PBL as the most effective strategy for improving both final examination scores and practical skills. Furthermore, presentation-assimilation-discussion class (PAD class) approach was identified as having the greatest potential to enhance autonomous learning abilities among these students. Future investigations should employ larger sample sizes and more rigorous methodologies to substantiate these findings.
Community nursing, an interdisciplinary field that integrates public health and nursing theories, plays a vital role in promoting and maintaining the health of community populations. This discipline is pivotal in developing the community nursing capabilities of undergraduate nursing students. Community health service centers, as primary healthcare institutions, are pivotal venues for delivering medical and nursing care to the elderly ( Sun, 2023). The development of community nursing is a crucial approach to adapting to the transformation of medical models and advancing the family doctor system, as well as a significant measure to address population aging. As an essential component of community health services, community nursing also serves as a key factor in ensuring the quality of community healthcare ( Wang et al., 2024). As the aging population in China rapidly increases, its profound impact on community health services is deepening, making the comprehensive, rapid, and standardized development of these services a critical focus of China's healthcare system reform ( Wang et al., 2020b). Consequently, there is an urgent need for nursing educators to engage students in community nursing and enhance the quality of instruction to meet the escalating demand for primary healthcare services.
With the expanding responsibilities of community health services, the imperative to address the suboptimal quality of community nursing in China has become apparent ( Chen et al., 2015). The Chinese government’s "Nursing Development Plan (2021—2025)" includes enhancing community nursing capabilities as a key objective of the "Fourteenth Five-Year Plan" ( Zhuang et al., 2024). Universities play a crucial role in nurturing community nursing talents capable of fulfilling residents' healthcare needs. Currently, the pedagogical approach in the "Community Nursing" course is somewhat monolithic and struggles to effectively integrate theoretical knowledge with practical societal demands, leading to a phenomenon of "disconnected learning" and a deficiency in the comprehensive community nursing skills of students ( Tao et al., 2020), The underlying reason is that the theoretical knowledge acquired by nursing students in classroom settings is often difficult to directly apply to the complex practices of community nursing. Community nursing science not only encompasses foundational nursing knowledge but also integrates multidisciplinary knowledge from public health, epidemiology, sociology, psychology, and health management. Nursing students are required to master a comprehensive knowledge system ranging from individual to population levels and from disease treatment to health promotion. This integrative and interdisciplinary nature may contribute to the perceived heavy academic burden among nursing students.
Moreover, the course is highly practical, covers a broad spectrum of knowledge, and differs significantly from the clinical core courses with which students are familiar. This has led to challenges in traditional teaching methods, including underachievement, lack of student engagement, diminished interest, and poor self-directed learning skills ( Xue et al., 2018; You et al., 2016). The future development of innovative teaching methods in community nursing science will focus on practice orientation and interdisciplinary collaboration, emphasizing the cultivation of students' comprehensive abilities, practical skills, and innovative capabilities. Through the application of innovative approaches such as flipped classroom, presentation-assimilation-discussion class (PAD class), blended learning, and problem-based learning (PBL), community nursing education will become more efficient, flexible, and aligned with practical needs. These innovative methods not only enhance nursing students' academic performance and practical competencies but also provide robust support for the advancement of community nursing. We think that Problem-Based Learning (PBL) will increasingly emphasize interdisciplinary collaboration and the simulation of complex clinical scenarios. This will equip students with the adaptive problem-solving skills necessary to navigate real-world challenges effectively. Additionally, sandwich learning may incorporate virtual reality (VR) and augmented reality (AR) technologies to offer a more immersive learning environment, enhancing students' procedural skills and clinical decision- making abilities. Flipped classroom models are expected to become more personalized and intelligent, leveraging adaptive learning algorithms to tailor content to each student's progress and individual needs. Experiential teaching could integrate artificial intelligence (AI) and big data analytics to generate precision-guided simulations, providing real-time feedback that supports the mastery of complex nursing skills. Furthermore, PAD class may leverage online learning platforms to offer more flexible learning options, enabling students to engage in effective learning and discussions both inside and outside the classroom. Lastly, blended learning systems have the potential to utilize learning analytics to design personalized educational pathways and optimize resource allocation, thereby improving learning outcomes and fostering skill acquisition.
However, there is currently no consensus on the superiority of different teaching methods, and no studies have directly compared their effects. Therefore, evaluating the impact of various teaching strategies on community nursing education is both necessary and meaningful. Network meta-analysis, unlike traditional meta-analysis, allows for the simultaneous assessment and ranking of all interventions within a single evidence body, thereby identifying the optimal approach ( Zeng et al., 2012). This method addresses the limitations of traditional meta-analysis, such as missing evidence chains and scant existing evidence.
In recent years, substantial reforms in China's healthcare system have led to significant changes in medical education. Given the potential impact of educational system and cultural differences between countries on the effectiveness of community nursing education, our study is confined to nursing students in China. We employed network meta-analysis to assess the effectiveness of six most common teaching methods used in community nursing education in China, aiming to provide insights for the selection of appropriate teaching models for community nursing education.
2 Materials and methodsThis study rigorously adheres to the PRISMA-NMA (Preferred Reporting Items for Systematic Reviews and Network Meta-Analyses) guidelines, as outlined in the extended statement ( Hutton et al., 2016).
2.1 Protocol and registrationThe protocol for this review was prospectively registered with the International Prospective
Register of Systematic Reviews (PROSPERO) (CRD42025635443).
2.2 Inclusion and exclusion criteriaThe criteria for inclusion were defined as follows:
a. Participants: Chinese nursing students enrolled in community nursing courses.
b. Interventions: Intervention groups utilized innovative educational approaches, including Problem-Based Learning (PBL), sandwich learning, flipped classroom, presentation- assimilation-discussion class (PAD class), blended learning, and experiential teaching.
c. Comparisons: Control groups underwent traditional teaching methods, such as lecture-based instruction and demonstration teaching, or used the same six innovative methods as the intervention cohorts.
d. Outcomes: Primary outcomes included final examination scores, practical skills, and core competency indicators, such as self-directed learning ability.
e. Study design: Randomized Controlled Trials (RCTs) and quasi-experimental studies were included.
Studies were excluded if they met any of the following conditions:
a. Lack of accessible full texts, incomplete data, or non-extractable information.
b. Duplicate publications.
c. Conference abstracts.
d. Studies in which the intervention group received a combination of two or more instructional methods.
e. Studies exclusively focused on intern nurses.
2.3 Search strategyA comprehensive literature search was conducted across multiple databases, including PubMed, Embase, The Cochrane Library, Web of Science, OVID, CNKI, Wanfang Database, China Biological Literature Database (CBM), and VIP, covering the period from database inception to November 2024. To maximize coverage, both keywords and free-text terms were incorporated into the search strategy. Additionally, citation tracking and manual searches of relevant print sources were performed to identify additional references and ensure a comprehensive retrieval of studies.
The specific search strategy for Web of Science is detailed below:
#1 TS= (Problem-Based Learning OR PBL OR Problem Based Learning OR Problem-Based Curriculum OR Problem Based Curricula OR Experiential Learning OR Active Learning)
#2 TS= (flipped classroom OR Inverted Classroom OR Flipped Class OR Inverted Class OR Flipped Teaching OR Inverted Teaching)
#3 TS= (blended teaching OR mixed teaching OR blended learning OR mixed learning)
#4 TS= (PAD class OR presentation-assimilation-discussion class OR PAD-assisted teaching)
#5 TS= (Sandwich teaching OR Sandwich teaching method)
#6 TS= (Experiential teaching)
#7 TS= (community health nursing OR nursing, community health OR community nursing)
#8 #1 AND #7
#9 #2 AND #7
#10 #3 AND #7
#11 #4 AND #7
#12 #5 AND #7
#13 #6 AND #7
2.4 Literature selectionTitles and abstracts were initially reviewed and duplicates were removed using Endnote X8 software by two researchers (WXY and YLF). Eligible studies were further screened by reading the titles, abstracts, and full texts in succession. Disagreements were resolved through discussion or by consulting a third party (WY).
2.5 Data extractionTwo independent researchers (WXY and DM) systematically extracted data from the included studies and conducted cross-checks to ensure accuracy. Any discrepancies were resolved by consulting a third researcher (WY), who made the final decision. A pre-designed standardized data extraction form was employed to collect the following information:
a. Basic study details: Title, authors, publication year.
b. Study population characteristics: Gender distribution, age, sample size, and intervention details.
c. Key elements for bias assessment: Study design parameters and risk of bias components.
d. Outcome measures: Data on final examination scores, practical skills, and core competency indicators relevant to nursing education.
2.6 Quality assessmentThe methodological quality of Randomized Controlled Trials (RCTs) was independently evaluated by two trained reviewers(WXY and YLF) using the Cochrane Risk of Bias tool (Version 5.1) ( Higgins et al., 2011). The assessment included the following criteria: a) random sequence generation; b) allocation concealment; c) blinding of participants and personnel; d) blinding of outcome assessment; e) completeness of outcome data; f) selective reporting; and g) other potential biases. The quality of quasi-experimental studies was assessed using the JBI Evidence-Based Health Center's quality assessment tool ( Zhou et al., 2018).
2.7 Statistical analysisMeta-analysis and heterogeneity assessments were performed using RevMan 5.4 software. Heterogeneity was evaluated using χ² tests and the I² statistic:
I² < 50 % indicated low heterogeneity, warranting a fixed-effects model.
I² ≥ 50 % suggested substantial heterogeneity, necessitating a random-effects model.
To estimate effect sizes and 95 % confidence intervals (CI), a frequentist framework was applied using Stata 15.0 software ( Salanti, 2012). Due to variability in outcome measure units, Standardized Mean Differences (SMD) were used for effect synthesis.
a. Heterogeneity assessment
Predictive interval plots were generated for each outcome measure ( Riley et al., 2011). If the predictive intervals intersected the null effect line, significant heterogeneity was inferred, requiring a random-effects model. In contrast, a fixed-effects model was implemented in the absence of notable heterogeneity.
b. Similarity testing
Clinical and methodological similarity among studies was rigorously evaluated:
Clinical similarity: Ensured by standardizing participant demographics, intervention strategies, and outcome measures.
Methodological similarity: Only studies rated grade B or higher in quality assessment were included.
c. Consistency testing
To verify agreement between direct and indirect evidence, consistency tests were conducted. If closed loops were present within the evidence network, formal inconsistency tests were performed.
d. Network evidence diagram
A network evidence diagram was constructed to visualize the relationships among various interventions:
a. Node size reflected the total sample size of each intervention.
b. Line thickness represented the number of direct comparative studies.
c. Connected nodes indicated the presence of direct evidence, whereas unconnected nodes suggested a lack of direct comparisons.
e. Publication bias diagram
Funnel plots were examined for asymmetry to assess publication bias and small-study effects ( Chaimani et al., 2013):
a. Symmetrical distribution suggested minimal publication bias.
b. Asymmetry indicated potential small-study effects or significant publication bias.
f. Network meta-analysis
Comparative effectiveness was displayed using league tables and forest plots. To determine the
optimal teaching strategy, the Surface Under the Cumulative Ranking Curve (SUCRA) was utilized
( Mbuagbaw et al., 2017): SUCRA values range from 0 to 1, where 1 indicates the most effective
intervention and 0 the least effective ( Cipriani et al., 2018).
3 Results3.1 Literature search results
An initial search yielded 3077 documents, from which duplicates were removed, leaving 2213 articles to be further screened by title and abstract. A total of 2087 articles were excluded as irrelevant to the research topic, in addition to 19 review articles and 9 articles that combined other teaching methods, resulting in 98 articles eligible for full-text evaluation. Two articles were excluded after full-text screening, 16 did not meet the outcome criteria, 19 were of other research designs, two were classified as clinical practicums, and 25 were deemed low quality. One article was excluded for using the same dataset, ultimately leaving 33 studies for inclusion. The selection process and results are illustrated in Fig. 1.
3.2 Basic characteristics of included studiesTable 1 outlines the main characteristics of the 33 studies ( Bao, 2018; Cai et al., 2022; Chen, 2011; Deng and Sun, 2020; Dong et al., 2021; Feng et al., 2010; Guo, 2010; He and Fan, 2019a, b; He et al., 2019; Li, 2021; Ng, 2023; Qin et al., 2022; Tan et al., 2020; Tan et al., 2019; Wang et al., 2023; Wang and Zhao, 2017; Wang et al., 2020a; Wang, 2020; Wang et al., 2020c; Wang et al., 2018; Xie et al., 2022; Xue, 2018; Yan et al., 2018; Yang and Li, 2018, 2019; Yang, 2019; Zhai et al., 2014; Zhang et al., 2020a; Zhang, 2017; Zhang et al., 2020b; Zhang, 2015; Zou, 2018) included in this research, with a total sample size of 4620 participants, comprising 2344 in the experimental groups and 2276 in the control groups, all of which were two-arm studies. Six teaching methods were assessed, including Problem-Based Learning (PBL), sandwich learning, flipped classroom, PAD class, blended learning, and experiential teaching. The distribution of literature included nine articles on PBL, seven each on blended learning and flipped classroom, two each on experiential and sandwich learning, and six on PAD class, all compared against traditional teaching methods. The studies span publications from 2010 to 2023. Seventeen studies involved undergraduate students, ten involved vocational students, one involved an associate degree program, and five were secondary vocational studies. The research includes five quasi-experimental studies and 28 randomized controlled trials (RCTs).
3.3 Quality assessment of included studiesIn terms of randomization, 25 studies ( Chen, 2011; Deng and Sun, 2020; Feng et al., 2010; Guo, 2010; He and Fan, 2019a, b; He et al., 2019; Li, 2021; Qin et al., 2022; Tan et al., 2020; Tan et al., 2019; Wang et al., 2023; Wang and Zhao, 2017; Wang et al., 2020a; Wang, 2020; Wang et al., 2020c; Xie et al., 2022; Yang and Li, 2018, 2019; Zhai et al., 2014; Zhang et al., 2020a; Zhang, 2017; Zhang et al., 2020b; Zhang, 2015; Zou, 2018) were deemed high risk, including 13( Deng and Sun, 2020; Feng et al., 2010; Guo, 2010; Qin et al., 2022; Tan et al., 2020; Tan et al., 2019; Wang et al., 2023; Wang and Zhao, 2017; Wang, 2020; Wang et al., 2020c; Xie et al., 2022; Zhai et al., 2014; Zhang, 2017) that mentioned randomization without detailing the allocation scheme; two Studies ( Bao, 2018; Wang et al., 2018) were low risk, using a random number table for allocation. None of the 33 studies mentioned allocation concealment or blinding methods. Regarding the completeness of outcome data, all studies reported complete data. All 33 studies were considered low risk for selective reporting of results. Other sources of bias were also considered low risk. The bias risk assessment for the included studies is provided in Table 2.
In quasi-experimental studies, one study ( Yan et al., 2018) did not clearly present baseline comparability between groups, while the other five ( Bao, 2018; Cai et al., 2022; Dong et al., 2021; Ng, 2023; Xue, 2018; Yang, 2019) showed comparable baselines across groups. Four studies ( Dong et al., 2021; Ng, 2023; Yan et al., 2018; Yang, 2019) lacked multidimensional outcome measures both pre- and post-intervention, and two ( Cai et al., 2022; Xue, 2018) demonstrated multidimensional outcomes measures both pre- and post-intervention; other criteria were met as detailed in Table 3.
3.4 Final examination scores3.4.1 Meta analysis
A subgroup analysis was performed on final examination scores. A total of 31 studies ( Bao, 2018; Cai et al., 2022; Chen, 2011; Deng and Sun, 2020; Dong et al., 2021; Feng et al., 2010; Guo, 2010; He and Fan, 2019a, b; He et al., 2019; Li, 2021; Ng, 2023; Qin et al., 2022; Tan et al., 2020; Tan et al., 2019; Wang et al., 2023; Wang and Zhao, 2017; Wang et al., 2020a; Wang, 2020; Wang et al., 2020c; Wang et al., 2018; Xie et al., 2022; Xue, 2018; Yang and Li, 2018, 2019; Yang, 2019; Zhai et al., 2014; Zhang, 2017; Zhang et al., 2020b; Zhang, 2015; Zou, 2018) reported this outcome, encompassing diverse pedagogical approaches, including Problem-Based Learning (PBL), PAD class, flipped classroom, blended learning, experiential teaching, and sandwich teaching. The pooled analysis revealed a statistically significant between-group difference [SMD = 1.37, 95 % CI (1.07, 1.68), P < 0.001], indicating that these six innovative teaching methodologies resulted in superior final examination performance in community nursing education among Chinese nursing students compared to conventional instructional approaches (see Supplement S1).
3.4.2 Evidence networkFinal examination scores were reported in 31 studies ( Bao, 2018; Cai et al., 2022; Chen, 2011; Deng and Sun, 2020; Dong et al., 2021; Feng et al., 2010; Guo, 2010; He and Fan, 2019a, b; He et al., 2019; Li, 2021; Ng, 2023; Qin et al., 2022; Tan et al., 2020; Tan et al., 2019; Wang et al., 2023; Wang and Zhao, 2017; Wang et al., 2020a; Wang, 2020; Wang et al., 2020c; Wang et al., 2018; Xie et al., 2022; Xue, 2018; Yang and Li, 2018, 2019; Yang, 2019; Zhai et al., 2014; Zhang, 2017; Zhang et al., 2020b; Zhang, 2015; Zou, 2018), encompassing six distinct innovative teaching methods and one conventional approach. The evidence network consisted of six direct comparisons with no closed loops, necessitating the application of a consistency model. The network relationships among the interventions are illustrated in Fig. 2.
3.4.3 Network meta-analysisHeterogeneity across studies was substantial, as shown in the prediction interval plots, prompting the use of a random effects model. Network meta-analysis results indicated statistically significant differences in final examination scores between PBL [SMD= -2.37, 95 %CI (-3.10, 1.65)], PAD class [SMD= -1.36, 95 %CI (0.52,2.20)], flipped classroom [SMD= 1.12, 95 %CI (0.29,1.96)], and blended teaching [SMD= 1.00, 95 %CI (0.23,1.76)] when compared to traditional teaching (P < 0.05). PBL teaching method showed statistically significant differences when compared with the other five teaching methods (P < 0.05), but no significant differences were observed between the other five methods when compared pairwise (all P > 0.05), see Fig. 3 and Table 4.
Substantial heterogeneity was observed across studies, as reflected in the prediction interval plots, prompting the implementation of a random-effects model. The network meta-analysis revealed statistically significant differences in final examination scores between PBL [SMD= -2.37, 95 %CI (-3.10, 1.65)], PAD class [SMD= -1.36, 95 %CI (0.52,2.20)], flipped classroom [SMD= 1.12, 95 %CI (0.29,1.96)], and blended teaching [SMD= 1.00, 95 %CI (0.23,1.76)] compared to traditional teaching methods (P < 0.05). Notably, PBL demonstrated a statistically significant difference when compared to the other five active learning strategies (P < 0.05). However, no significant differences were observed in pairwise comparisons among the remaining five teaching methods (all P > 0.05), as depicted in Fig. 3 and detailed in Table 4.
Note: A=PBL; B= Traditional teaching; C= Sandwich teaching; D= PAD class;
E = Flipped classroom; F= Blended teaching; G= Experiential teaching
3.4.4 Ranking of network meta-analysis resultsThe ranking of the six teaching methods for final examination scores from best to worst was: PBL teaching (SUCRA=98.6), PAD class (SUCRA = 67.7), flipped classroom (SUCRA=58.1), blended teaching (SUCRA= 51.5), experiential teaching (SUCRA=35.2), sandwich teaching (SUCRA=31.4), and traditional teaching method (SUCRA=7.6), see Fig. 4.
3.4.5 Publication biasThe funnel plot analysis suggested asymmetry, with a clustering of studies to the left of the central axis, indicating potential publication bias. Moreover, several studies were observed outside the funnel boundaries, suggesting the presence of small-study effects, as illustrated in Fig. 5.
Note: A=PBL; B= Traditional teaching; C= Sandwich teaching; D= PAD class;
E = Flipped classroom; F= Blended teaching; G= Experiential teaching
3.5 Practical skills3.5.1 Meta analysis
A subgroup analysis was conducted to evaluate practical skills, incorporating data from eight studies ( Guo, 2010; Qin et al., 2022; Wang et al., 2023, 2018; Yang and Li, 2019; Yang, 2019; Zhai et al., 2014; Zhang et al., 2020a) that examined the effects of PBL, flipped classroom, and blended teaching methodologies. Two studies ( Yang and Li, 2019; Zhang et al., 2020a) compared the flipped classroom approach with traditional teaching methods, yielding a pooled effect size that did not reach statistical significance [SMD = 1.01, 95 % CI (-0.04, 2.07), P = 0.06]. Similarly, three studies ( Qin et al., 2022; Wang et al., 2023; Yang, 2019) evaluated blended teaching in comparison with traditional methods, also revealing no statistically significant effect [SMD = 0.38, 95 % CI (-1.77, 2.53), P = 0.73]. These findings suggest that neither flipped classroom nor blended teaching demonstrated a statistically significant advantage over traditional methods in terms of enhancing practical skills (see Supplement S2).
3.5.2 Evidence networkPractical skills were reported in eight studies ( Guo, 2010; Qin et al., 2022; Wang et al., 2023, 2018; Yang and Li, 2019; Yang, 2019; Zhai et al., 2014; Zhang et al., 2020a) encompassing three innovative teaching methodologies and one conventional approach. The network structure comprised three direct comparisons without any closed loops, necessitating the implementation of a consistency model. The network of relationships among the interventions is shown in Fig. 6.
3.5.3 Network meta-analysisDue to substantial heterogeneity, as indicated by the prediction interval plots, a random effects model was employed. Network meta-analysis results showed statistically significant differences in practical skills between the PBL teaching method [SMD= 2.84, 95 %CI (1.15,4.54)] and traditional teaching, blended teaching [SMD= -2.46, 95 %CI (-5.04,0.11)], and flipped classroom [SMD= -1.84, 95 %CI (-4.72, 1.04)] (P < 0.05), see Fig. 7 and Table 5.
Given the substantial heterogeneity identified through prediction interval plots, a random-effects model was applied. The network meta-analysis results revealed statistically significant differences in practical skill outcomes between PBL and other instructional methods. Specifically, PBL was associated with significantly higher practical skill performance compared to traditional teaching [SMD= 2.84, 95 %CI (1.15,4.54)], blended teaching [SMD= -2.46, 95 %CI (-5.04,0.11)], and flipped classroom [SMD= -1.84, 95 %CI (-4.72, 1.04)] (P < 0.05). These results suggest a robust advantage of PBL in fostering practical skill. Further details are presented in Fig. 7 and Table 5.
Note: A=PBL; B= Traditional teaching; C= Flipped classroom; D= Blended teaching
3.5.4 Ranking of network meta-analysis resultsThe ranking probability analysis, based on Surface Under the Cumulative Ranking Curve (SUCRA) values, established the following hierarchy of instructional methods for practical skill development, from most to least effective: PBL (SUCRA = 95.5), flipped classroom (SUCRA = 52.9), blended teaching (SUCRA = 34.3), and traditional teaching (SUCRA = 17.3). A visual representation of this ranking is provided in Fig. 8.
3.5.5 Publication biasThe funnel plot analysis revealed an asymmetrical distribution, with a clustering of studies to the right of the central axis, indicating potential publication bias. Additionally, several studies were positioned outside the funnel boundaries, suggesting the presence of small-study effects, as depicted in Fig. 9.
Note: A=PBL; B= Traditional teaching; C= Flipped classroom; D= Blended teaching
3.6 Self-learning ability3.6.1 Meta analysis
A subgroup analysis was conducted on self-learning ability. Seven studies ( Bao, 2018; Qin et al., 2022; Tan et al., 2020, 2019; Yan et al., 2018; Yang and Li, 2019; Zhang et al., 2020b) reported this outcome, involving teaching methods such as PBL teaching, PAD class, flipped classroom, blended teaching, experiential teaching, and sandwich teaching. The six teaching methods were compared with traditional teaching methods, and results showed that PBL [SMD = 0.06, 95 %CI (-0.18, 0.31), P = 0.60], Sandwich teaching [SMD = 0.36, 95 %CI (-0.10, 0.83), P = 0.60], PAD class [SMD = 2.97, 95 %CI (-1.49, 7.43), P = 0.19], flipped classroom [SMD = 2.76, 95 %CI (-1.34, 6.86), P = 0.19] combined effect size was not statistically significant. Experiential teaching [SMD = 0.33, 95 %CI (0.05, 0.62), P = 0.02], blended teaching [SMD = 0.39, 95 %CI (0.12, 0.65), P = 0.004] combined effect size was statistically significant (See, supplement S3).
3.6.2 Evidence networkSeven studies ( Bao, 2018; Qin et al., 2022; Tan et al., 2020, 2019; Yan et al., 2018; Yang and Li, 2019; Zhang et al., 2020b) reported self-learning ability outcomes across five innovative teaching methods and one conventional approach. The network structure comprised six direct comparisons without any closed loops, necessitating the application of a consistency model. The relationships among the interventions are depicted in Fig. 10.
3.6.3 Network meta-analysisPrediction interval plots revealed substantial heterogeneity, warranting the use of a random-effects model. The network meta-analysis demonstrated no statistically significant differences in self-learning ability among the six instructional methodologies (P > 0.05). These findings are illustrated in Fig. 11 and further detailed in Table 6.
Note: A=PBL; B= Traditional teaching; C= Sandwich teaching; D= PAD class;
E = Flipped classroom; F=Experiential teaching；G= Blending teaching
3.6.4 Ranking of network meta-analysis resultsThe ranking of the seven teaching methods for self-learning ability from best to worst was: PAD class (SUCRA=72.6), flipped classroom (SUCRA=71), PBL teaching (SUCRA=52.2), sandwich teaching (SUCRA=41.6), experiential teaching (SUCRA=40.7), blended teaching (SUCRA=40.6), and traditional teaching method (SUCRA=31.4), see Fig. 12.
3.6.5 Publication biasThe funnel plot analysis revealed asymmetry, with a clustering of studies to the left of the central axis, suggesting potential publication bias. Additionally, several studies were observed outside the funnel boundaries, indicating small-study effects, as depicted in Fig. 13.
Note: A=PBL; B= Traditional teaching; C= Sandwich teaching; D= PAD class;
E = Flipped classroom; F=Experiential teaching；G= Blended teaching
4 Discussion4.1 Summary of main findings
To the best of our knowledge, this study represents the first evidence-based evaluation of the pedagogical effectiveness of community nursing education for nursing students in China, as well as the first systematic review and network meta-analysis comparing various instructional methods within this context. Our findings demonstrate that compared to traditional teaching approaches, PBL, PAD class, and flipped classroom significantly enhance final examination performance. Notably, among the six teaching methodologies analyzed, SUCRA rankings indicate that PBL is the most effective in improving both final examination scores and practical skills in Chinese nursing students.
This can be attributed to several factors. Firstly, the core of PBL is to encourage and support students to actively explore, collaborate in problem-solving, and develop interpersonal communication skills, thereby altering passive learning habits and focusing on the formation of knowledge and skills ( Ding and Gu, 2011). Secondly, among the six teaching methods included in the analysis, PBL was the earliest to be implemented in China. The fundamental philosophy of PBL is to foster students' abilities to learn independently and solve problems, which aligns with the ideals and objectives of quality education promoted in China ( Zhang et al., 2010). Consequently, many institutions have adopted the PBL model as a critical component of educational reform initiatives ( Du and Wang, 2015), giving it a clear advantage over other methods in terms of teaching design, faculty training, and changing student attitudes. Additionally, our research suggests that the PAD class method holds the greatest potential for enhancing the self-learning capabilities of Chinese nursing students. Introduced by Professor Zhang Xue Xin in 2014, PAD class model ( Zhang, 2014) maintains certain traditional lecturing elements while integrating discussion-based teaching and dividing class time evenly. It adjusts upon the foundations of flipped classrooms, PBL, and seminar-based teaching methods, retaining a lecture component with a concise lecturing style to engage students effectively, while also providing a quiet and independent off-class learning environment for internalization and absorption. This model integrates the dominant teaching philosophies of "teacher-led" and "student-centered," splitting class time into two halves—one for lecturing and the other for student participation in discussions, with a crucial internalization phase between teaching and discussion. This approach not only encourages students to arrange their post-class studies actively but also shifts from passive to active learning, significantly enhancing students' self-learning abilities. Thus, PAD class emerges as a commendable teaching method, effectively fostering nursing students' autonomy in learning.
5 Strengths and limitations5.1 Strengths of the meta-analysis
This network meta-analysis offers several notable strengths. Most importantly, it is the first study to systematically compare six commonly employed teaching methodologies in community nursing education, providing critical insights for future pedagogical innovations in this evolving field. By conducting a rigorous quality assessment of the included literature, this study ensures a comprehensive evaluation of educational strategies across nursing students with diverse educational backgrounds, thereby establishing a robust reference framework for evidence-based research in community nursing education.
5.2 Limitations of the meta-analysisDespite its contributions, this meta-analysis has certain limitations. First, the methodological quality of the 33 included studies remains moderate, with deficiencies in allocation concealment, blinding procedures, and inadequate reporting of random sequence generation methods, which may introduce potential bias. Second, while low-quality studies were excluded to enhance the credibility of findings, this exclusion may have contributed to publication bias, particularly in relation to outcome measures. Given these limitations, future research should prioritize high-quality randomized controlled trials (RCTs) that directly compare different teaching methodologies, thereby generating more reliable evidence to inform optimal community nursing education models.
5.3 ImplicationsAlthough our study has identified PBL and PAD class as the best methods for improving the educational outcomes of community nursing among Chinese nursing students, future research should aim to improve in several areas: a. Conduct high-quality RCTs to increase the reliability of evidence; b. Develop standardized assessment tools, particularly for professional knowledge and clinical skills; c. Since the success of various teaching implementations largely depends on teachers' preparation and familiarity with teaching methods, future studies should explore factors and interventions affecting the implementation of various teaching methods from the perspective of teaching preparation; d. Investigate how different intervention timings and frequencies affect relevant outcomes, especially in terms of the long-term maintenance and clinical practice translation of core competencies in nursing students; e. Future research should also consider how traditional teaching methods can be integrated with different teaching approaches to enhance clinical core competencies of nursing students while reducing educational costs.
6 ConclusionsThis study systematically evaluated the impact of six teaching methodologies on community nursing education among Chinese nursing students. The results ranked PBL as the most effective strategy for improving both final examination scores and practical skills. Furthermore, PAD class approach was identified as having the greatest potential to enhance autonomous learning abilities among these students. Nonetheless, the substantial heterogeneity and moderate methodological quality of the studies included warrant careful consideration. Future investigations should employ larger sample sizes and more rigorous methodologies to substantiate these findings.
Author contributionsWXY were responsible for conceiving and designing the experiments, collecting and analyzing the data, and writing and revising the manuscript. YLF and DM were responsible for the data interpretation. WY made important intellectual contributions to the research design, provided technical guidance and revised the manuscript. All the authors read and approved the final version of the manuscript.
FundingThis study was supported by teaching project of Hexi College in 2024 ( HXUJXYJ-2025–37) and Ningxia Natural Science Foundation Project ( 2023A1934).
CRediT authorship contribution statementyan Wang xiao: Writing – original draft, Methodology, Formal analysis. feng Yang li: Data curation, Conceptualization. mei Du: Data curation, Conceptualization. yan Wang: Writing – review & editing, Validation, Supervision, Methodology.
Declaration of Competing InterestThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Appendix A Supporting informationSupplementary data associated with this article can be found in the online version at doi:10.1016/j.nepr.2025.104323.
Appendix A Supplementary materialSupplementary material
Supplementary material subgroup analysis of final examination scores
Supplementary material subgroup analysis of practical skills
Supplementary material subgroup analysis of self-learning ability
| Study | Study type | Simple size（T/C） | Gender (female/male) | Age（T/C） | Educational background | Intervention method | Outcome |
| Wang, D. Q 2023 | RCT | 114/125 | / | / | Junior | T: Blending teaching C: Traditional teaching | ①, ② |
| Ng, E. I. L 2023 | q-RCT | 130/141 | T：85/45 C：96/45 | T：18.51 ± 1.59 C：18.33 ± 0.71 | Associate degree | T: Flipped classroom C: Traditional teaching | ① |
| Xie, S. F 2022 | RCT | 64/65 | T：60/4 C：59/6 | / | Junior | T: Experiential teaching C: Traditional teaching | ① |
| Qin, S. X 2022 | RCT | 113/113 | T：105/8 C：106/7 | T：21.50 ± 3.58 C：22.00 ± 3.66 | Junior | T: Blending teaching C: Traditional teaching | ①, ②, ③ |
| Cai, X. Y 2022 | q-RCT | 36/36 | T：25/11 C：27/9 | T：21.17 ± 1.06 C：21.13 ± 1.17 | Undergraduate | T: Sandwich C: Traditional teaching | ① |
| Li, Y 2021 | RCT | 58/40 | T：50/0 C：47/1 | T：17.64 ± 0.23 C：17.51 ± 0.14 | Secondary vocational | T: PAD teaching C: Traditional teaching | ① |
| Dong, Y. J 2021 | q-RCT | 98/90 | T：83/15 C：78/12 | T：21.2 ± 0.77 C：21.3 ± 0.77 | Undergraduate | T: Flipped classroom C: Traditional teaching | ① |
| Zhang, T. L 2020 | RCT | 90/98 | T：80/10 C：86/12 | T：21.34 ± 0.92 C：21.78 ± 1.01 | Undergraduate | T: Experiential teaching C: Traditional teaching | ①, ③ |
| Zhang, F 2020 | RCT | 162/158 | T：156/6 C：154/4 | T：20.58 ± 0.21 C：20.49 ± 0.53 | Junior | T: Flipped classroom C: Traditional teaching | ② |
| Wang, S 2020 | RCT | 39/40 | T：39/0 C：40/0 | 19.2 ± 1.4 | Junior | T: PBL teaching C: Traditional teaching | ① |
| Wang, W, L 2020 | RCT | 48/48 | T：43/5 C：43/5 | T：19.84 ± 0.77 C：19.81 ± 0.73 | Junior | T: Blending teaching C: Traditional teaching | ① |
| Wang, W. C 2020 | RCT | 92/94 | T：43/5 C：43/6 | T：18.32 ± 1.47 C：18.18 ± 1.69 | Undergraduate | T: Blending teaching C: Traditional teaching | ① |
| Tan, X. L 2020 | RCT | 59/59 | T：42/17 C：41/18 | T：21.79 ± 1.53 C：21.57 ± 1.98 | Undergraduate | T: PAD teaching C: Traditional teaching | ①, ③ |
| Deng, W. F 2020 | RCT | 39/36 | T：35/4 C：31/5 | T：20.33 ± 0.98 C：20.06 ± 0.83 | Undergraduate | T: Flipped classroom C: Traditional teaching | ① |
| Yang, J. N 2019 | q-RCT | 180/184 | T：180/0 C：184/0 | / | Junior | T: Blending teaching C: Traditional teaching | ①, ② |
| Yang, D 2019 | RCT | 40/40 | T：32/8 C：30/10 | T：18～20 C：18～21 | Undergraduate | T: Flipped classroom C: Traditional teaching | ①, ②, ③ |
| Tan, X. L 2019 | RCT | 55/59 | T：39/16 C：41/18 | T：18.51 ± 1.59 C：18.33 ± 0.71 | Undergraduate | T: Flipped classroom C: Traditional teaching | ①, ③ |
| He, S 2019 | RCT | 183/181 | / | / | Undergraduate | T: Blending teaching C: Traditional teaching | ① |
| He, H. L 2019a | RCT | 45/45 | T：45/0 C：45/0 | 17 ± 2 | Secondary vocational | T: PAD teaching C: Traditional teaching | ① |
| He, H. L 2019b | RCT | 48/48 | T：48/0 C：48/0 | 17 ± 2 | Secondary vocational | T: PAD teaching C: Traditional teaching | ① |
| Zou, Y. T 2018 | RCT | 40/40 | / | / | Junior | T: PAD teaching C: Traditional teaching | ① |
| Yang, D 2018 | RCT | 40/40 | / | / | Undergraduate | T: Flipped classroom C: Traditional teaching | ① |
| Yan, G. M 2018 | RCT | 125/140 | / | / | Undergraduate | T: PBL teaching C: Traditional teaching | ③ |
| Xue, J. L 2018 | q-RCT | 305/290 | T: 272/31 C: 258/28 | T：20.15 ± 1.26 C：20.03 ± 1.21 | Undergraduate | T: Blending teaching C: Traditional teaching | ① |
| Wang, X. Q 2018 | RCT | 61/61 | T：61/0 C：61/1 | T：20.4 ± 1.3 C：20.2 ± 1.2 | Undergraduate | T: PBL teaching C: Traditional teaching | ①, ② |
| Bao, J. L 2018 | RCT | 46/49 | T：44/2 C：44/5 | T：22.74 ± 0.88 C：22.60 ± 0.81 | Undergraduate | T: PAD teaching C: Traditional teaching | ①, ③ |
| Zhang, L. K 2017 | RCT | 51/50 | / | / | Junior | T: PBL teaching C: Traditional teaching | ① |
| Wang, H 2017 | RCT | 30/29 | T: 29/1 C: 27/2 | / | Undergraduate | T:Sandwich C:Traditional teaching | ① |
| Zhang, X. Y 2015 | RCT | 114/116 | / | / | Secondary vocational | T: PBL teaching C: Traditional teaching | ① |
| Zhai, Z. M 2014 | RCT | 64/63 | / | 19 ～ 22 | Undergraduate | T: PBL teaching C: Traditional teaching | ①, ② |
| Chen, J. H 2011 | RCT | 108/109 | / | / | Secondary vocational | T: PBL teaching C: Traditional teaching | ① |
| Guo, X. Y 2010 | RCT | 40/40 | / | / | Undergraduate | T: PBL teaching C:Traditional teaching | ①, ② |
| Feng, X. Y 2010 | RCT | 79/77 | / | / | Junior | T: PBL teaching C: Traditional teaching | ① |
| Study ID | Random sequence generation | Allocation concealment | Blinding of participants and personnel | Incomplete outcome data | Selective reporting | Other bias |
| Wang, H 2017 | High | Unclear | High | Low | Low | Low |
| Li, Y 2021 | High | Unclear | High | Low | Low | Low |
| Tan, X. L 2020 | High | Unclear | High | Low | Low | Low |
| He, H. L 2019a | High | Unclear | High | Low | Low | Low |
| He, H. L 2019b | High | Unclear | High | Low | Low | Low |
| Zou, Y. T 2018 | High | Unclear | High | Low | Low | Low |
| Bao, J. L 2018 | Low | Unclear | High | Low | Low | Low |
| Xie, S. F 2022 | High | Unclear | High | Low | Low | Low |
| Zhang, T. L 2020 | High | Unclear | High | Low | Low | Low |
| Zhang, F 2020 | High | Unclear | High | Low | Low | Low |
| Deng, W. F 2020 | High | Unclear | High | Low | Low | Low |
| Yang, D 2019 | High | Unclear | High | Low | Low | Low |
| Tan, X. L 2019 | High | Unclear | High | Low | Low | Low |
| Yang, D 2018 | High | Unclear | High | Low | Low | Low |
| Wang, D. Q 2023 | High | Unclear | High | Low | Low | Low |
| Qin, S. X 2022 | High | Unclear | High | Low | Low | Low |
| Wang, W, L 2020 | High | Unclear | High | Low | Low | Low |
| Wang, W. C 2020 | High | Unclear | High | Low | Low | Low |
| He, S 2019 | High | Unclear | High | Low | Low | Low |
| Wang, S 2020 | High | Unclear | High | Low | Low | Low |
| Wang, X. Q 2018 | Low | Unclear | High | Low | Low | Low |
| Zhang, L. K 2017 | High | Unclear | High | Low | Low | Low |
| Zhang, X. Y 2015 | High | Unclear | High | Low | Low | Low |
| Zhai, Z. M 2014 | High | Unclear | High | Low | Low | Low |
| Chen, J. H 2011 | High | Unclear | High | Low | Low | Low |
| Guo, X. Y 2010 | High | Unclear | High | Low | Low | Low |
| Feng, X. Y 2010 | High | Unclear | High | Low | Low | Low |
| Study ID | Item 1 | Item 2 | Item 3 | Item 4 | Item 5 | Item 6 | Item 7 | Item 8 | Item 9 |
| Ng, E. I. L 2023 | Yes | Yes | Yes | Yes | No | Yes | Yes | Yes | Yes |
| Cai, X. Y 2022 | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| Dong, Y. J 2021 | Yes | Yes | Yes | Yes | No | Yes | Yes | Yes | Yes |
| Yang, J. N 2019 | Yes | Yes | Yes | Yes | No | Yes | Yes | Yes | Yes |
| Yan, G. M 2018 | Yes | Unclear | Yes | Yes | No | Yes | Yes | Yes | Yes |
| Xue, J. L 2018 | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| PBL | ||||||
| 2.37 (1.65,3.10) | Traditional teaching | |||||
| 1.89 (0.26,3.51) | −0.49 (−1.94,0.97) | Sandwich teaching | ||||
| 1.01 (−0.09,2.12) | −1.36 (−2.20,−0.52) | −0.87 (−2.55,0.80) | PAD teaching | |||
| 1.25 (0.14,2.35) | −1.12 (−1.96,−0.29) | −0.64 (−2.32,1.04) | 0.24 (−0.95,1.42) | Flipped classroom | ||
| 1.38 (0.33,2.43) | −1.00 (−1.76,−0.23) | −0.51 (−2.15,1.13) | 0.36 (−0.77,1.50) | 0.13 (−1.00,1.26) | Blending teaching | |
| 1.77 (0.17,3.37) | −0.60 (−2.03,0.83) | −0.12 (−2.16,1.92) | 0.76 (−0.90,2.41) | 0.52 (−1.13,2.18) | 0.39 (−1.23,2.01) | Experiential teaching |
| PBL | |||
| 2.84 (1.01,4.68) | Traditional teaching | ||
| 1.84 (−1.04,4.72) | −1.00 (−3.22,1.22) | Flipped classroom | |
| 2.46 (−0.11,5.04) | −0.38 (−2.19,1.43) | 0.62 (−2.25,3.49) | Blending teaching |
| PBL | ||||||
| 0.64 (−5.35,6.63) | Traditional teaching | |||||
| 0.28 (−8.22,8.77) | −0.36 (−6.39,5.66) | Sandwich teaching | ||||
| −2.33 (−9.69,5.02) | −2.97 (−7.24,1.30) | −2.61 (−9.99,4.78) | PAD teaching | |||
| −2.12 (−9.48,5.23) | −2.76 (−7.03,1.51) | −2.40 (−9.78,4.99) | 0.21 (−5.83,6.25) | Flipped classroom | ||
| 0.30 (−8.18,8.79) | −0.33 (−6.35,5.68) | 0.03 (−8.48,8.54) | 2.64 (4.74,10.01) | 2.43 (−4.95,9.80) | Experiential teaching | |
| 0.25(−8.23, 8.74) | −0.39(−6.40, 5.63) | −0.02(−8.53, 8.49) | 2.59(−4.79, 9.96) | 2.38 (−5.00,9.75) | −0.05 (−8.56,8.45) | Blending teaching |
©2025. Elsevier Ltd