Heterogeneity in clinical nurses’ core competency and its associations with self-efficacy and self-directed learning ability: a cross-sectional latent profile analysis

Abstract

Objectives

To identify latent profiles of core competency among hospital clinical nurses and to examine the associations of self-efficacy and self-directed learning ability with profile membership.

Design

A cross-sectional survey reported in accordance with Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) principles.

Setting

Two general hospitals in Yanji, Jilin Province, China.

Participants

A total of 605 registered clinical nurses who were directly engaged in clinical nursing work were recruited using convenience sampling from January to March 2026.

Methods

Latent profile analysis was conducted using the seven Competency Inventory for Registered Nurses (CIRN) dimension scores as continuous indicators to identify core competency profiles. Univariable analyses were used to compare demographic, occupational and psychological characteristics across profiles. Multivariable multinomial logistic regression was subsequently performed with profile membership as the outcome to examine associated factors.

Main outcome measures

Core competency measured using the seven dimensions of the CIRN, general self-efficacy, self-directed learning ability and latent core competency profile membership.

Results

Three profiles were identified: low-level core competency (266/605, 43.97%), moderate-level core competency (254/605, 41.98%) and high-level core competency (85/605, 14.05%). After adjustment for demographic and occupational variables, self-directed learning ability was positively associated with membership of the moderate-level profile (OR 1.119, 95% CI 1.077 to 1.163) and high-level profile (OR 1.142, 95% CI 1.078 to 1.210) rather than the low-level profile. General self-efficacy showed similar positive associations with the moderate-level profile (OR 1.161, 95% CI 1.113 to 1.213) and high-level profile (OR 1.578, 95% CI 1.452 to 1.715).

Conclusions

Clinical nurses’ core competency showed clear heterogeneity, with a relatively small high-level subgroup. Self-directed learning ability and self-efficacy were modifiable correlates of more favourable competency profiles and may inform stratified competency development strategies.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• This study used a person-centred latent profile analysis based on seven dimensions of core competency to identify heterogeneity within the clinical nursing workforce.

• The analysis included self-directed learning ability and general self-efficacy, two potentially modifiable factors that may inform stratified competency development.

• The cross-sectional design precludes causal inference about the relationships between self-directed learning ability, self-efficacy and profile membership.

• Participants were recruited by convenience sampling from two general hospitals in one city, which may limit generalisability to other regions, hospital types and nursing populations.

• Profile assignment was based on maximum posterior probability and key variables were measured by self-report scales, so classification error and common method bias cannot be excluded.

Introduction

As the scope of clinical nursing continues to expand and care environments become increasingly complex, nurses are required to integrate knowledge, technical skills and clinical judgement across diverse practice situations to provide safe, continuous and high-quality care.^1–3 The quality of nursing care therefore depends not only on the performance of routine tasks but also on nurses’ broader ability to respond to complex clinical scenarios. Core competency among clinical nurses captures this integrated capability in practice and has become an important issue for understanding variation in nursing quality and optimising nursing workforce development.^{4 5}

Core competency is not a single technical skill. Rather, it reflects an integrated set of professional knowledge, clinical skills, attitudes and developmental capacities that nurses apply in clinical practice.⁶ It commonly includes dimensions such as clinical care, critical thinking, communication and collaboration, legal and ethical practice, education and consultation, leadership and professional development.⁷ Development of these competencies is unlikely to be uniform across nurses, and higher-order competencies may depend on continuing learning, supportive environments and accumulated clinical experience.⁸

Core competency is shaped by multiple factors. In addition to demographic and occupational characteristics such as age, years of nursing experience and professional title, psychological and learning-related factors warrant attention. General self-efficacy refers to an individual’s confidence in their ability to complete tasks and is rooted in Bandura’s self-efficacy theory.⁹ Previous nursing research has shown that self-efficacy is closely related to clinical capability and learning performance.¹⁰ Self-directed learning ability reflects the tendency to learn actively and pursue continuing development, and is also associated with professional growth.^{11 12} Self-efficacy and self-directed learning ability therefore represent distinct but complementary individual characteristics related to confidence and learning orientation.

Recent studies have illuminated the relationships between competency and psychological-learning attributes. One cross-sectional study of newly graduated nurses explored the relationships among core competency, self-efficacy and transition shock.¹³ Meanwhile, a multicentre study of clinical nurses reported a positive correlation between self-directed learning ability and nursing competency.¹¹ Furthermore, a Swedish multicentre study identified differences in professional competency and self-efficacy across groups with varying work experience.¹⁴ Collectively, existing research indicates that competency is associated with professional experience, confidence and learning orientation. Nevertheless, these studies predominantly analysed competency using total scale scores, and person-centred evidence remains relatively scarce. Finnish researchers utilised cluster analysis to identify low-, moderate- and high-level profiles of preceptorship competency among nurse clinical preceptors.¹⁵ In addition, recent latent-profile nursing studies have addressed competencies for intravenous therapy,¹⁶ core competencies in palliative care¹⁷ and specialist competencies in gerontological nursing.¹⁸ While these studies demonstrate the value of subgroup identification, they all focused on competency within a specific role or specialty rather than general core competency for routine hospital nursing practice.

Existing studies have examined factors associated with nurses’ core competency, but many have used variable-centred analyses that focus on total scores or associations with individual demographic variables. Such approaches tend to assume within-sample homogeneity and may obscure meaningful subgroups. By contrast, person-centred approaches can identify latent subgroups and describe their distinctive characteristics.¹⁹ Latent profile studies in nursing have examined topics such as intravenous therapy competency,¹⁶ palliative care competency¹⁷ and geriatric nursing competencies,¹⁸ but evidence on general core competency among hospital clinical nurses remains limited.

Latent profile analysis (LPA) is a person-centred statistical method that uses multiple continuous indicators to identify unobserved homogeneous subgroups within a sample.¹⁹ This study therefore aimed to identify latent profiles of core competency among hospital clinical nurses and to examine whether self-directed learning ability and general self-efficacy were associated with profile membership. Identifying competency profiles and their modifiable correlates may help nurse managers design more targeted competency development strategies and support improvements in nursing practice quality and patient safety.

Methods

Study design and participants

This cross-sectional study was conducted from January to March 2026 in two general hospitals in Yanji, Jilin Province, China. Convenience sampling was used to recruit clinical nurses who met the eligibility criteria. Inclusion criteria were: holding a registered nurse licence; being directly engaged in clinical nursing work for at least 12 months; and providing informed consent. Exclusion criteria were: not being directly involved in clinical nursing work; and being unable to participate during data collection because of leave, external training or other reasons. A total of 605 nurses were included, which was considered sufficient for the planned analyses.

Instruments

Demographic and occupational questionnaire: the questionnaire collected sex, age, marital status, monthly income, educational level, professional title, years of nursing experience and specialist nurse status.

Competency Inventory for Registered Nurses (CIRN): core competencies were assessed using the CIRN developed by Liu et al,²⁰ with reference to validation work on the Chinese version.²¹ The scale comprises 40 items across seven dimensions: critical thinking and research (five items), clinical care (six items), leadership (eight items), interpersonal relationships (six items), legal and ethical practice (five items), professional development (four items) and education and consultation (six items). Cronbach’s α for the total scale was 0.98 in the original study and 0.99 in the present study. Items were rated on a 5-point Likert scale, with response options ranging from ‘no competency’ to ‘high competency’. For analysis, these categories were coded from 1 to 5, corresponding to the 0–4 scores shown in the questionnaire. Each dimension score was calculated by summing the scores of its constituent items, and the possible total scale score ranged from 40 to 200, with higher scores indicating greater core competency. The seven dimension sum scores were used as continuous indicators in the LPA. The complete questionnaire, including item wording and response options, is provided in online supplemental file 1.

General Self-Efficacy Scale (GSES): general self-efficacy was measured using the Chinese version of the GSES revised and validated by Wang et al.²² The scale includes 10 items rated on a 4-point Likert scale from 1 (completely incorrect) to 4 (completely correct), with total scores ranging from 10 to 40; higher scores indicate higher general self-efficacy. The Cronbach’s alpha of the Chinese version was 0.87, and the Cronbach’s alpha in this study was 0.95.

Adult Learning Traits Scale: self-directed learning ability was assessed using the Chinese version of the self-directed learning instrument validated by Shen et al.²³ The scale includes nine items rated on a 5-point Likert scale from 1 (completely inconsistent) to 5 (completely consistent), with total scores ranging from 9 to 45; higher scores indicate stronger self-directed learning ability. The original scale had a Cronbach’s alpha of 0.90, and the Cronbach’s alpha in this study was 0.95.

Data collection and quality control

Data were collected through the Wenjuanxing online survey platform. After permission was obtained from the nursing departments of the participating hospitals, a questionnaire QR code was distributed to clinical nurses through hospital nursing WeChat groups, together with a study description and an estimated completion time of 10–15 min. The first page of the questionnaire emphasised anonymous completion, academic use only and the absence of standard answers to reduce social desirability bias and common method bias. To ensure data quality, all items were set as mandatory, a minimum completion time of 120 s was applied, IP address verification was used to avoid duplicate submissions, and two researchers independently screened completed questionnaires to exclude responses with obvious repetitive patterns. The final analysis included 605 questionnaires that were completed and passed quality control; this number represents the final valid sample rather than an overall response rate. The study was reported in accordance with the basic principles of Strengthening the Reporting of Observational Studies in Epidemiology (STROBE).²⁴

Patient and public involvement

Patients and/or the public were not involved in the design, conduct, reporting or dissemination plans of this research.

Statistical analysis

Data were analysed using IBM SPSS Statistics V.28.0 and Mplus V.8.3. Statistical tests were two-sided and alpha was set at 0.05.

Categorical variables, including sex, age, marital status, monthly income, educational level, professional title, years of nursing experience and specialist nurse status, were described using frequencies and percentages. Continuous variables, including self-directed learning ability, general self-efficacy and the CIRN total, mean item and seven dimension scores, were assessed using Shapiro-Wilk tests and Q–Q plots. To describe the distributions more fully, CIRN scores are presented as both mean (SD) and median (Q1–Q3), together with their possible ranges in table 1.

Table 1. Core competency dimension scores.

Measure	Score range	Mean (SD)	Median (Q1–Q3)
Total core competency score	40–200	142.30 (33.03)	145.00 (120.00–160.00)
Mean item score	1–5	3.56 (0.83)	3.63 (3.00–4.00)
Critical thinking and research	5–25	17.33 (4.33)	18.00 (15.00–20.00)
Clinical care	6–30	20.78 (5.34)	21.00 (18.00–24.00)
Leadership	8–40	28.10 (6.90)	29.00 (24.00–32.00)
Interpersonal relationships	6–30	21.38 (5.23)	22.00 (18.00–24.00)
Legal and ethical practice	5–25	18.90 (4.41)	20.00 (15.00–22.00)
Professional development	4–20	14.29 (3.53)	15.00 (12.00–16.00)
Education and consultation	6–30	21.53 (5.38)	23.00 (18.00–24.00)

Q1–Q3 (25th and 75th percentiles).

LPA was conducted in Mplus V.8.3 using the seven CIRN dimension scores as continuous observed indicators. Models were estimated using robust maximum likelihood estimation. Within each profile, covariances among the seven indicators were fixed at zero, assuming conditional independence. The variance of each indicator was constrained to be equal across profiles, whereas indicator means were freely estimated across profiles. For models with two or more profiles, 100 sets of random starting values were used in the initial stage, with 10 retained for final optimisation (STARTS=100 10). Models with one to five profiles were fitted sequentially.²⁵ Model fit was evaluated using the Akaike information criterion (AIC),²⁶ Bayesian information criterion (BIC),²⁷ sample-size adjusted BIC (aBIC), Lo-Mendell-Rubin likelihood ratio test/Vuong-Lo-Mendell-Rubin likelihood ratio test (LMR/VLMR),²⁸ bootstrap likelihood ratio test (BLRT),²⁹ entropy, the smallest class proportion and substantive interpretability. A smallest class proportion below 10% was prespecified as a potential indicator of limited model stability. Model selection was based on the balance of statistical fit, classification stability and clinical interpretability rather than a single fit index.^{30 31}

Differences in demographic, occupational and psychological characteristics across profiles were examined using χ² tests or Fisher’s exact tests for categorical variables. For continuous variables, one-way analysis of variance was used for approximately normally distributed variables and the Kruskal-Wallis H test for skewed variables.

Multinomial logistic regression was used to examine factors associated with latent profile membership, with the low-level core competency profile as the reference category. Independent variables were selected based on theoretical assumptions, previous evidence and univariable findings, and included age, marital status, monthly income, professional title, years of nursing experience, self-directed learning ability and general self-efficacy. Variance inflation factors were calculated to assess multicollinearity among age, professional title and years of nursing experience; all variance inflation factors were below 5, indicating no marked multicollinearity. SEs, ORs, 95% CIs and p values were reported. Subsequent regression analysis used profile assignment based on maximum posterior probability, a common classify-analyse approach; the potential risk of classification error is addressed in the limitations.^{32 33}

Results

Participant characteristics

A total of 605 clinical nurses were included. Most participants were women (590/605, 97.5%), and the largest age group was 31–40 years (314/605, 51.9%). Most participants had a bachelor’s degree (490/605, 81.0%), held a junior professional title (282/605, 46.6%) and had a moderate length of nursing experience (285/605, 47.1%). Detailed characteristics are shown in online supplemental table S1.

Core competency scores

The mean total core competency score was 142.30 (SD 33.03), with a mean item score of 3.56 (SD 0.83) that corresponded most closely to the response category ‘adequate competency’. Among the seven dimensions, leadership had the highest mean score (28.10, SD 6.90), whereas professional development had the lowest mean score (14.29, SD 3.53). Scores for critical thinking and research, clinical care, interpersonal relationships, legal and ethical practice, and education and consultation are shown in table 1.

Latent profile analysis

Considering the AIC, BIC, aBIC, LMR/VLMR and BLRT results, entropy, the smallest class proportion and substantive interpretability, the three-profile model was selected as the optimal solution (table 2). This model had an entropy of 0.97 and a smallest class proportion of 14.1%, suggesting high classification accuracy and acceptable stability. Class proportions in the model fit table are presented in the original model output order. The three profiles were subsequently renamed according to their scores across dimensions as low-level, moderate-level and high-level core competency profiles. Although the AIC, BIC and aBIC continued to decrease in the four- and five-profile models, both models included at least one small class below 10% (9.75%), and the additional classes mainly represented further splitting of the existing gradient without a distinct clinical pattern. Therefore, the three-profile model was retained.

Table 2. Fit indices for latent profile models.

Model	LL	AIC	BIC	aBIC	Entropy	VLMR/LMR (p value)	BLRT (p value)	Class proportions (%)
One profile	−12754.635	25 537.269	25 598.943	25 554.496				100.00
Two profiles	−11187.127	22 418.253	22 515.168	22 445.324	0.968	<0.001	<0.001	47.11/52.89
Three profiles*	−10599.084	21 258.169	21 390.326	21 295.083	0.970	0.012	<0.001	43.97/41.98/14.05
Four profiles	−9760.674	19 597.348	19 764.747	19 644.106	0.985	<0.001	<0.001	9.75/38.51/37.69/14.05
Five profiles	−9596.231	19 284.462	19 487.102	19 341.064	0.988	<0.001	<0.001	8.93/1.16/37.36/14.05/38.51

^*Selected model.

aBIC, sample-size adjusted BIC; AIC, Akaike information criterion; BIC, Bayesian information criterion; BLRT, bootstrap likelihood ratio test; LL, log likelihood; LMR/VLMR, Lo-Mendell-Rubin/Vuong-Lo-Mendell-Rubin likelihood ratio test.

The seven CIRN dimensions increased across the low-, moderate- and high-level profiles. Based on the most likely profile assignments, these groups included 266 (43.97%), 254 (41.98%) and 85 (14.05%) nurses, respectively. Table 3 presents observed mean item scores ±SD within each group; figure 1 shows the corresponding observed means.

Table 3. Observed CIRN mean item scores by assigned profile.

CIRN dimension	Low-level (n=266)	Moderate-level (n=254)	High-level (n=85)
Critical thinking and research	2.75±0.57	3.78±0.45	4.76±0.39
Clinical care	2.69±0.56	3.83±0.39	4.80±0.36
Leadership	2.74±0.50	3.87±0.33	4.86±0.29
Interpersonal relationships	2.78±0.50	3.92±0.34	4.95±0.13
Legal and ethical practice	3.04±0.68	4.16±0.41	4.97±0.12
Professional development	2.80±0.56	3.94±0.39	4.89±0.26
Education and consultation	2.79±0.57	3.98±0.37	4.91±0.27

CIRN, Competency Inventory for Registered Nurses.

Figure 1. Observed Competency Inventory for Registered Nurses mean item scores across the three assigned profiles.

Differences across profiles

Univariable analyses showed statistically significant differences across the three profiles in age, marital status, monthly income, professional title, years of nursing experience, self-directed learning ability and general self-efficacy (all p<0.05). Differences in sex, educational level and specialist nurse status were not statistically significant (p>0.05). Overall, the differences were concentrated in indicators of career development stage and in psychological and behavioural variables. Self-directed learning ability and general self-efficacy increased stepwise from the low-level to the high-level profile, suggesting that these modifiable factors may be particularly useful for differentiating competency levels. Detailed comparisons are shown in online supplemental table S2.

Multinomial logistic regression

Using the low-level core competency profile as the reference category, variables selected based on theory and univariable findings were entered into the multinomial logistic regression model. After adjustment for age, marital status, monthly income, professional title and years of nursing experience, self-directed learning ability and general self-efficacy were both positively associated with membership of the moderate-level and high-level profiles. The association for general self-efficacy was particularly strong for membership of the high-level profile.

Specifically, each one-point increase in self-directed learning ability was associated with higher odds of being in the moderate-level profile (OR 1.119) or high-level profile (OR 1.142) rather than the low-level profile. Each one-point increase in general self-efficacy was associated with corresponding ORs of 1.161 and 1.578. Among demographic and occupational variables, only intermediate professional title was positively associated with the high-level profile, whereas short nursing tenure was negatively associated with the high-level profile (table 4).

Table 4. Multinomial logistic regression for latent profile membership, with the low-level core competency profile as the reference category.

Variable	Moderate vs low				High vs low
	SE	P value	OR	95% CI	SE	P value	OR	95% CI
Age (reference: >50 years)
20–30 years	0.906	0.614	1.580	0.268 to 9.318	1.473	0.538	2.478	0.138 to 44.468
31–40 years	0.812	0.747	0.769	0.157 to 3.777	1.282	0.948	1.088	0.088 to 13.422
41–50 years	0.713	0.975	0.978	0.242 to 3.955	1.037	0.911	1.123	0.147 to 8.577
Marital status (reference: married)
Unmarried	0.318	0.082	0.576	0.309 to 1.073	0.578	0.801	0.865	0.279 to 2.684
Monthly income (reference: ≥¥6000)
<¥4000	0.563	0.924	1.055	0.350 to 3.183	0.898	0.304	0.398	0.069 to 2.310
¥4000–¥5999	0.496	0.962	1.024	0.387 to 2.709	0.753	0.105	0.295	0.067 to 1.293
Professional title (reference: senior title)
Junior title	0.556	0.081	0.379	0.127 to 1.126	0.928	0.267	2.800	0.454 to 17.275
Intermediate title	0.503	0.267	0.572	0.214 to 1.532	0.833	0.019	7.062	1.380 to 36.136
Years of nursing experience (reference: long tenure)
Short tenure	0.625	0.537	0.680	0.200 to 2.314	1.050	0.043	0.119	0.015 to 0.933
Moderate tenure	0.539	0.765	1.174	0.408 to 3.379	0.893	0.118	0.247	0.043 to 1.423
Self-directed learning ability	0.020	<0.001	1.119	1.077 to 1.163	0.030	<0.001	1.142	1.078 to 1.210
General self-efficacy	0.022	<0.001	1.161	1.113 to 1.213	0.042	<0.001	1.578	1.452 to 1.715

Discussion

Principal findings and implications for stratified competency development

The three profiles identified in this study primarily exhibited overall level differences with synchronous increases across the seven dimensions. No distinct competency combinations showing divergent strengths and weaknesses were detected. Accordingly, the expectation of identifying profiles characterised by varied competency combinations was not supported in the present sample. Nevertheless, this result alone cannot prove that core competency is a unidimensional trait. Shared self-evaluation tendencies may lead nurses to give similar ratings for different competency dimensions. Furthermore, comparable training and work environments may contribute to the synchronous development of various competencies. In addition, the use of aggregated dimension scores and model variance specifications may influence the resulting profile patterns.

Insufficient core competencies may limit nurses’ assessment and judgement, risk recognition, communication and coordination, health education and evidence-based practice in complex clinical situations, potentially affecting nursing decision-making, implementation of nursing interventions, patient safety and professional development.^{34 35} Recent latent profile or latent class studies have also identified heterogeneity in transition shock, psychological burden and clinical teaching-related competencies among nurses.^36–38 Unlike studies focused on a single specialised competency, the present study examined general core competencies among hospital clinical nurses and found an overall gradient across all seven dimensions rather than a profile characterised by one isolated dimension. Nurse managers may use these profiles to establish an assessment-intervention-reassessment pathway: strengthening foundational competencies in the low-level profile, supporting capability expansion in the moderate-level profile and preparing nurses in the high-level profile for clinical teaching, quality improvement and specialist development.

Professional experience and competency development

The three profiles differed significantly in age, marital status, monthly income, professional title and years of nursing experience, suggesting that competency profiles are related to career stage and accumulated clinical experience. In univariable analyses, the moderate- and high-level profiles included higher proportions of older nurses, married nurses, nurses with higher income, nurses with intermediate or senior professional titles, and nurses with longer clinical experience. These variables may jointly reflect job stability, exposure to clinical situations, professional responsibility and access to organisational resources. Differences in sex, education and specialist nurse status were not statistically significant, possibly because most participants were women and had bachelor’s degrees and because internal training and role expectations were relatively homogeneous across the participating hospitals.

Previous studies on nurses’ professional competence have suggested that work experience, professional title, educational support and baseline individual capability influence competency development.^{8 14 39} In the multinomial regression model, intermediate professional title was positively associated with the high-level competency profile, whereas short nursing tenure was negatively associated with the high-level profile. These findings suggest that career maturity, reflected partly by professional title and nursing tenure, remains an important background factor in the development of high-level competency and is consistent with Benner’s novice-to-expert framework.⁴⁰ However, age, marital status and income were no longer significant in the adjusted model, indicating that their associations may be partly explained by professional title, nursing tenure and psychological or behavioural factors.

Self-directed learning ability and continuing progression

Self-directed learning ability increased stepwise across the low-, moderate- and high-level profiles. Regression analysis showed that higher self-directed learning ability was positively associated with membership of both the moderate- and high-level profiles, suggesting that it is a potentially modifiable factor supporting competency progression. In the context of digital nursing, intelligent healthcare and rapid development of artificial intelligence, nursing knowledge, clinical evidence, technical tools and service models are constantly changing. Nurses with insufficient self-directed learning ability may experience delayed knowledge updating, limited technological adaptation and weaker evidence-based practice capacity, all of which may affect role competence.^41–43

For nurses in the low-level profile, weak foundational capability, limited clinical experience and unclear learning goals may contribute to passive learning and poor translation of learning into practice. These nurses may benefit from structured foundational knowledge support, standardised skills training, mentorship and staged feedback.⁴⁴ Nurses in the moderate-level profile may already have clinical experience but still need support in identifying clinical problems, applying evidence and engaging in reflective practice. Case-based learning, evidence-based nursing training and reflective supervision may help them move from experience-based practice to improvement-oriented practice.^{45 46} Nurses in the high-level profile may require opportunities for advanced development, including participation in research, clinical teaching, specialist nursing development and digital nursing practice.^{15 42} These findings suggest that self-directed learning should be embedded across all stages of competency development rather than treated as an isolated training target.

Self-efficacy and high-level competency profile membership

General self-efficacy differed significantly across the three profiles and was positively associated with membership of the moderate- and high-level profiles. The association was particularly strong for the high-level profile, suggesting that confidence in task performance may be an important psychological resource for movement from lower to higher competency levels. In clinical practice, nurses with low self-efficacy may lack confidence when managing complex patient care, interprofessional communication, patient education, risk identification and quality improvement tasks, which may in turn affect the development and application of core competency.⁴⁷ Previous nursing education research has also shown that training in core nursing skills can improve self-efficacy, clinical competence and practice satisfaction.⁴⁸

For nurses in the low-level profile, limited successful experiences may weaken confidence in handling complex tasks and lead to avoidance in learning and practice. Rather than relying solely on assessment pressure, nurse managers should use modelling, simulation, staged goal setting and positive feedback to help these nurses accumulate visible success experiences.^{49 50} Nurses in the moderate-level profile may have basic role competence but require confidence-building exposure to more complex specialist nursing, teaching or quality improvement tasks. Appropriate delegation, case review, peer support and participation in specialist practice may help them move beyond their current level. Nurses in the high-level profile may benefit from leadership roles, clinical teaching, professional development platforms and high-quality supervision to sustain self-efficacy and professional growth.^{51 52} Taken together, both general self-efficacy and self-directed learning ability were positively associated with membership in the moderate- and high-level core competency profiles. Nevertheless, these associations allow for alternative directional interpretations. Stronger self-efficacy may motivate nurses to embrace clinical challenges, and self-directed learning ability may also facilitate the acquisition and application of new knowledge. Conversely, greater core competency and successful clinical experiences may boost nurses’ confidence and help them identify learning needs more clearly, resulting in higher self-reported self-directed learning ability. Shared factors including training opportunities, work environments and professional experiences may also correlate with all three variables simultaneously.

Strengths, limitations and future research

This study used LPA to identify latent profiles of core competency among clinical nurses and examined self-directed learning ability, general self-efficacy and demographic and occupational factors associated with profile membership. The findings provide empirical evidence for stratified competency development in clinical nursing. Nevertheless, several limitations should be noted. First, the cross-sectional design means that statistical associations cannot be interpreted as causal relationships. Second, participants were recruited by convenience sampling from two general hospitals in Yanji, Jilin Province, which may limit generalisability to other regions, hospital levels and nursing populations. Third, subsequent regression analysis used modal class assignment based on maximum posterior probability, and classification error may have influenced estimates. Fourth, the main variables were measured using self-report scales, and common method bias cannot be excluded. Future studies should include multicentre samples, longitudinal designs, mixed methods or intervention studies, and objective indicators such as competency assessment, job performance and patient safety outcomes to further examine the stability of competency profiles and the mechanisms through which self-directed learning ability and self-efficacy support competency development.

Conclusion

Among clinical nurses from two general hospitals in Yanji, Jilin Province, this study identified three latent profiles of core competency: low-level, moderate-level and high-level profiles. The results indicate clear heterogeneity in core competency and a relatively small high-level subgroup, suggesting substantial room for improvement. After adjustment for demographic and occupational variables, self-directed learning ability and general self-efficacy remained significantly associated with membership of higher competency profiles, whereas professional title and years of nursing experience showed more limited associations. Nursing management strategies should consider profile-based assessment and differentiated support, with self-directed learning ability and self-efficacy as important targets for competency development. These findings require further validation in larger and more diverse samples.

Data availability statement

Data are available upon reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information.