Development and validation of a machine learning-based predictive model for compassion fatigue in nursing interns: A cross-sectional study with latent profile analysis

doi:10.21203/rs.3.rs-4709842/v1

Download PDF

Research Article

Development and validation of a machine learning-based predictive model for compassion fatigue in nursing interns: A cross-sectional study with latent profile analysis

https://doi.org/10.21203/rs.3.rs-4709842/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background

Exposure to compassion fatigue during internships can significantly impact on nursing students’ future career trajectories and their intention to stay in the nursing profession. Accurately identifying nursing students at high risk of compassion fatigue is vital for timely interventions. However, existing assessment tools often fail to account for within-group variability and lack predictive capabilities. To develop and validate a predictive model for detecting the risk of compassion fatigue among nursing students during their placement.

Design:

A cross-sectional study design.

Methods

Data from 2256 nursing students in China between December 2021 and June 2022 were collected on compassion fatigue, professional identity, self-efficacy, social support, psychological resilience, coping styles, and demographic characteristics. The latent profile analysis was performed to classify compassion fatigue levels of nursing students. Univariate analysis, least absolute shrinkage and selection operator regression analysis were conducted to identify potential predictors of compassion fatigue. Eight machine learning algorithms were selected to predict compassion fatigue, and the performance of these machine learning models were evaluated using calibration and discrimination metrics. Additionally, the best-performing model from this evaluation was selected for further independent assessment.

Results

A three-profile model best fit the data, identifying low (55.73%), moderate (32.17%), and severe (12.10%) profiles for compassion fatigue. The area under the curve values for the eight machine learning models ranged from 0.644 to 0.826 for the training set and from 0.651 to 0.757 for the test set. The eXtreme Gradient Boosting performed best, with area under the receiver operating characteristic curve values of 0.840, 0.768, and 0.731 in the training, validation, and test sets, respectively. SHAP analysis clarified the model’s explanatory variables, with psychological resilience, professional identity, and social support being the most significant contributors to the risk of compassion fatigue. A user-friendly, web-based prediction tool for calculating the risk of compassion fatigue was developed.

Conclusions

The eXtreme Gradient Boosting classifier demonstrates exceptional performance, and clinical implementation of the online tool can provide nursing managers with an effective means to manage compassion fatigue.

Nursing student

Internship

nonmedical

Compassion fatigue

Prediction model

Latent profile analysis

Machine learning

Nursing is a profession defined by its commitment to compassion, but it is also accompanied by significant stress. Nurses frequently witness patients’ suffering and distress while providing care and support. Prolonged exposure to such traumatic experiences can result in psychological distress, impairing their capacity to deliver optimal care (1–3). This phenomenon, known as compassion fatigue, was first defined by Joinson in 1992, and it encompasses physical, psychological, and social dysfunction arising from continuous exposure to patient suffering or traumatic events (4). Compassion fatigue comprises two elements: burnout and secondary traumatic stress (5, 6). This emotional and psychological exhaustion leads to various health issues among nurses, including musculoskeletal disorders, sleep disturbances, depression, and reduced professional identity and work engagement (7–10).

Compassion fatigue is not exclusive to registered nurses; it also affects nursing students (11–14). In China, pre-licensure nursing education includes both associate (three or five-year programs) and baccalaureate (four-year programs) tracks. These programs emphasize a combination of theoretical knowledge and practical skills. After completing on-campus theoretical studies, students are required to undertake at least eight months of clinical placement in secondary or higher-level hospitals. During this clinical placement period, these students are referred to as nursing interns. A survey of 972 Chinese nursing interns revealed that 97.8% experienced moderate burnout, and 55.3% suffered from secondary traumatic stress (15). High levels of compassion fatigue or burnout are associated with a higher intention to drop out, increasing actual attrition rates (16–20). Additionally, burnout developed during educational programs may persist after graduation, impacting the health and retention of new nurses (20, 21). Thus, compassion fatigue or burnout adversely affects the career development of nursing students. Early identification and intervention for those at risk of compassion fatigue are crucial to prevent the worsening of symptoms.

Current assessments of compassion fatigue primarily rely on surveys (22). Researchers such as Figley and Stamm have developed several measurement tools, including the Compassion Fatigue Self-Test (23), the Compassion Satisfaction and Fatigue Scale (24), the Compassion Fatigue Scale (25), the Compassion Fatigue Short Scale (26) and the Professional Quality of Life Scale (27). While these tools are essential for assessing compassion fatigue, they have limitations. For instance, some tools consider compassion fatigue as a combination of burnout and secondary traumatic stress or focus on a single dimension, complicating the assessment process (22–25, 27). Additionally, the Compassion Fatigue Short Scale, despite providing a straightforward measurement method, lacks clear cutoff values, making it difficult to distinguish between low-risk and high-risk individuals (26, 28). More importantly, the aforementioned tools typically assess compassion fatigue among nursing interns using total scores, which overlooks individual differences and are primarily designed for assessment rather than prediction(15, 16, 29, 30). Consequently, there is an urgent need for a new method to effectively assess and predict the risk of compassion fatigue among nursing interns.

Several psychosocial factors, such as social support, coping strategies, self-efficacy, psychological resilience, and professional identity, are closely related to compassion fatigue among nursing students (15, 16, 31, 32). Studies have shown that educational background and career-related factors, such as academic major, program length, previous student leadership, and career intentions, are significantly associated with compassion fatigue or burnout risk (15, 31, 33, 34). Demographic factors like gender, residence, only-child status, and monthly expenses, as well as internship-related characteristics such as the level of the internship hospital and the frequency of night shifts, are also associated with compassion fatigue or burnout (32, 33, 35–37). These factors should be considered when assessing nursing students at potential risk.

Compassion fatigue is a complex and dynamic phenomenon. Traditional “one-size-fits-all” approaches are insufficient to convey its impact on different psychological profiles. Existing cross-sectional surveys on compassion fatigue among nursing students typically use variable-centered approaches, measuring compassion fatigue through total scores from scales or subscales, neglecting within-group variability (15, 16, 31, 32, 38). To reveal the heterogeneity of compassion fatigue among nursing students more comprehensively and precisely and to identify characteristics of different profiles, combining latent profile analysis (LPA) with machine learning could be a potential solution.

LPA is a person-centered statistical method that classifies nursing students into different risk groups based on the measurement data of compassion fatigue, providing a basis for tailored interventions (39, 40). Unlike traditional cutoff-based methods, LPA identifies unobserved heterogeneity within the population based on individuals’ responses to continuous variables, grouping those with similar response patterns into homogeneous subgroups (41). This approach addresses the issue of neglecting within-group variability when using total scores, making LPA a superior choice (41, 42). However, LPA lacks predictive capabilities, necessitating its combination with other methods for predictive functions. Machine learning, a technique for learning patterns from data and making predictions, possesses strong data processing and prediction capabilities (43). In compassion fatigue research, machine learning can construct predictive models based on LPA classification results and other related variables, enabling accurate predictions of individual risk levels. Therefore, combining LPA with machine learning can provide a more comprehensive assessment and prediction of compassion fatigue risk among nursing interns.

The main objectives of this study are: (1) to identify potential classifications of compassion fatigue among nursing interns using LPA; (2) to develop and validate machine learning models for predicting individual risk levels; and (3) to develop an online prediction tool for practical application.

Study design

This cross-sectional study was conducted from December 2021 to June 2022 among nursing interns at 10 public junior colleges in Hunan Province, China. This design was chosen to provide a comprehensive overview of the population at a specific time point. The study adheres to guidelines outlined in the STROBE and TRIPOD statements.

Participants

Participants were recruited using a convenience sampling method. Inclusion criteria were: (1) enrollment in a three- or five-year associate nursing program, (2) active participation in a clinical internship lasting at least eight months in a secondary-level or above hospital, and (3) willingness to participate with informed consent. Nursing interns whose clinical practice was in clerical management or administration, without direct patient contact, were excluded.

Sample size

Larger sample sizes are typically used in survey research to achieve more accurate and stable results (44). However, there is no standardized method for calculating sample size in survey studies using machine learning techniques (45). For LPA, a sample size of more than 500 cases is recommended to ensure accuracy (46). Furthermore, a review by Spurk et al. revealed that 53.4% of studies used sample sizes greater than 500, further supporting the appropriateness of this rule of thumb (47).

Data collection

Full-time student counselors responsible for managing nursing students during their internships, were trained as research assistants. They provided standardized explanations of the purpose, risks, and benefits of this study. Data were collected via online surveys created on the WenJuanXing platform (https://www.wjx.cn), shared through WeChat groups. Each IP address was restricted to one submission. To reduce missing data and ensure data quality, the online survey required all questions to be answered. Participants’ anonymity and voluntary participation were emphasized, and participants could withdraw at any time. Surveys with identical or patterned responses were excluded. After removing 149 invalid responses, the final sample consisted of 2256 participants, resulting in an effective response rate of 93.8%.

Instruments

General demographic information collection table

A self-designed table collected sociodemographic data including gender, academic major, program length, place of residence, only-child status, monthly expenditure, previous experience as a student leader, hospital level during internship, number of night shifts per month, and career intentions.

The Compassion Fatigue Short Scale

The Compassion Fatigue Short Scale, developed by Adams et al., comprises 13 items measuring secondary traumatic stress (5 items) and job burnout (8 items) (26). Responses range from 1 (never) to 10 (very often), yielding a total score ranging from 13 to 130, where higher scores indicate greater levels of compassion fatigue. The Chinese version, translated and validated by Sun (2015) (28), demonstrated good reliability and validity, with Cronbach’s alpha coefficients ranging from 0.87 to 0.95. In this study, the overall Cronbach’s alpha coefficient of the Chinese version was 0.92.

The Professional Identity Scale

Professional identity was measured using the Professional Identity Scale by Brown et al., consisting of 10 items rated on a 5-point scale from 1 (never) to 5 (always) (48). Higher scores indicate stronger professional identity. The Chinese version, translated and validated by Lu et al., achieved a Cronbach's alpha coefficient of 0.82 (49). In this study, the overall Cronbach's alpha coefficient of the Chinese version was 0.80.

The General Self-efficacy Scale

The General Self-Efficacy Scale, developed by Schwarzer and Jerusalem (1995) (50) and adapted for Chinese population by Zeng et al. (51), includes 10 items rated from 1 (not true at all) to 4 (exactly true), with total scores ranging from 10 to 40, where higher scores indicate greater self-efficacy. Zeng et al.'s (51) research demonstrated that the scale exhibits good internal consistency and criterion validity, with an overall Cronbach's alpha coefficient of 0.89 in this study.

The Perceived Social Support Scale

The Perceived Social Support Scale, developed by Zimet et al. (52), translated by Jiang (53) and modified by Yan et al. (54), includes 12 items across three dimensions: family support (items 3, 4, 8, 11), friend support (items 6, 7, 9, 12), and other support (items 1, 2, 5, 10). Responses range from 1 (strongly disagree) to 7 (strongly agree), resulting in a total score ranging from 12 to 84. Higher scores indicate higher levels of perceived social support. In Yan et al.'s study, the scale demonstrated internal consistency and test-retest reliability coefficients of 0.87 and 0.85, respectively (54). In the present study, the Chinese version of the scale achieved an overall Cronbach’s alpha coefficient of 0.92.

The 10-item Connor-Davidson Resilience Scale

The 10-item Connor-Davidson Resilience Scale is a refinement by Campbell-Sills of the original scale developed by Connor and Davidson (55, 56). This unidimensional scale comprises 10 items, each rated on a 5-point scale (0–4), with higher scores indicating greater resilience. In this study, we utilized the Chinese version of the scale, translated and validated by Ye in 2016 (57). In a sample of Chinese nursing students, this scale accounted for 48.641% of the total variance, with a Cronbach's α coefficient of 0.851, demonstrating good reliability and validity (34). In the present study, the overall Cronbach's α coefficient for the scale was 0.93.

Simple Coping Style Questionnaire

The simple coping style questionnaire was adapted and translated by Xie et al. based on Folkman and Lazarus' Ways of Coping Questionnaire (58, 59). The 20-item scale comprises two dimensions: positive coping styles (items 1–12) and negative coping styles (items 13–20). Each item uses a Likert four-point scale, ranging from 1 (do not take) to 4 (often take). Higher scores in the positive coping style dimension indicate a greater likelihood of adopting positive coping strategies, while higher scores in the negative coping style dimension indicate a greater likelihood of adopting negative coping strategies. The scale has demonstrated good reliability and validity among Chinese populations, with a Cronbach’s α coefficient of 0.90 (58). In the present study, the scale achieved an overall Cronbach’s α coefficient of 0.87.

Statistical analyses

Descriptive and univariate analysis

Descriptive statistics summarized the general characteristics of the participants. Continuous variables, which did not follow a normal distribution, were represented as median and inter-quartile range, while categorical variables were shown as numbers and percentages. Univariate analysis, using the Mann-Whitney U test, chi-squared test, or Fisher’s exact test as appropriate, identified predictive factors for compassion fatigue or burnout. The candidate factors considered included demographic characteristics, professional identity, self-efficacy, social support, psychological resilience, and coping styles. Analyses were performed using SPSS 26.0 (IBM Corp., Armonk, NY, USA), with a significance threshold of P < 0.05 for two-sided test.

Latent profile analysis

LPA was conducted using the 13 items from the Compassion Fatigue Short Scale as indicators, employing robust maximum likelihood estimation to identify subgroups of compassion fatigue symptoms. Various models were compared based on entropy, the Lo-Mendell-Rubin likelihood ratio test (LMR), the bootstrap likelihood ratio test (BLRT), Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), and adjusted Bayesian Information Criterion (aBIC) (60–62). Specifically, an entropy value > 0.80 indicated that the latent classes were highly discriminative, and significant p-values for the LMR and BLRT suggested that the k-class model was preferable to the k-1 class model (60). Additionally, lower values of AIC, BIC, and aBIC indicated better model fit, while the “turning point” in the scree plot for aBIC suggested the appropriate number of classes (61, 63). Ultimately, the number of latent profiles was determined by a combination of these fit criteria. Analyses were conducted using Mplus version 8.3 (64).

To facilitate the construction of the subsequent predictive model, following recommendation from previous studies (64, 65), we assigned individuals who belonged to the latent profile representing the lowest level of symptoms or risks as “non-cases,” while other individuals were considered “cases.”

Features selection

After determining the optimal model and defining classifications, we assessed differences in some sociodemographic characteristics, professional identity, self-efficacy, social support, psychological resilience, and coping styles between the “non-cases” and “cases” groups. Significant variables from the univariate analysis (P ≤ 0.05) were further analyzed using the Least Absolute Shrinkage and Selection Operator (LASSO) regression, a technique that penalizes regression coefficients to optimize the model by retaining only significant predictors(66). This approach handles complex covariance structures and improves predictive accuracy. LASSO regression was performed using Lasso CV with 10-fold cross-validation.

Development and evaluation of machine learning models learning

Eight machine learning algorithms were used to develop and compare models for predicting the risk of compassion fatigue, including logistic regression, support vector machine, random forest, multi-layer perceptron, extreme gradient boosting (XGBoost), gradient boosting decision trees, Gaussian naive Bayes, and adaptive boosting. All models were implemented using Python 3.7, with the "xgboost 2.0.1" package for XGBoost and the "scikit-learn 1.1.3" package for the remaining algorithms. Models were trained and validated using bootstrap resampling, with a 7:3 training-to-testing ratio. Performance was evaluated using metrics such as area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, specificity, positive predictive value, negative predictive value, and F1 score. Calibration was assessed by comparing predicted and observed incidence of “cases.”

Model optimization and evaluation

To ensure robustness and mitigate overfitting, 10-fold cross-validation assessed predictive performance. The dataset was randomly divided into ten equal parts, with training and validation repeated five times. Model discrimination was evaluated using receiver operating characteristic (ROC) analysis and quantified by AUC. Calibration plots assessed the agreement between predicted probabilities and actual outcomes. Decision curve analysis (DCA) estimated the clinical utility and net benefit. Feature importance was assessed using SHAP analysis, with higher absolute SHAP values indicating greater impact on predictions (67). Additionally, we explored the distribution of feature values and their relationship with model predictions to gain further insights into model behavior. SHAP analysis was conducted using the “shap 0.43.0” package.

Descriptive statistics

Data were collected from 2256 nursing interns. Most participants were female (2077, 92.07%), came from rural areas (1784, 79.08%), and were non-only children (1962, 86.97%). Regarding economic status, 69.77% had a monthly expenditure between 1000 and 2000 RMB (1574, 69.77%). A significant majority were enrolled in nursing programs (1660, 73.58%), with most of them in three-year programs (1736, 76.95%). Approximately half had experience serving as student leaders (1172, 51.95%).

In terms of internship conditions, 75.71% interned at tertiary hospitals (1708, 75.71%), and most worked night shifts 3 to 4 times per month (1037, 45.96%). Career intention surveys showed that 86.13% intended to pursue a nursing career (1943, 86.13%). Psychological assessments revealed median scores of 39.00 for professional identity, 2.800 for self-efficacy, 58.00 for social support, 23.00 for psychological resilience, and 34.00 for coping style. Further details of participants’ characteristics are presented in Table 1.

Table 1

Univariate analysis of influencing factors of compassion fatigue of nursing interns
Variable	N (%)	Non-case ^a	Case ^b	Statistics	p
Academic major, n (%)				15.358	< 0.001
Nursing	1660 (73.58)	973 (76.80)	687 (69.46)
Midwifery	596 (26.42)	294 (23.20)	302 (30.54)
Length of schooling, n (%)				1.452	0.228
3-year	1736 (76.95)	963 (76.01)	773 (78.16)
5-year	520 (23.05)	304 (23.99)	216 (21.84)
Gender, n (%)				14.759	< 0.001
Male	179 (7.93)	125 (9.87)	54 (5.46)
Female	2077 (92.07)	1142 (90.13)	935 (94.54)
Residence, n (%)				0.167	0.683
Urban	472 (20.92)	269 (21.23)	203 (20.53)
Rural	1784 (79.08)	998 (78.77)	786 (79.47)
Only child or not, n (%)				0.988	0.320
Yes	294 (13.03)	173 (13.65)	121 (12.24)
No	1962 (86.97)	1094 (86.35)	868 (87.76)
Monthly expenses, n (%)				0.447	0.800
<1000 RMB	420 (18.62)	242 (19.10)	178 (18.00)
1000–2000 RMB	1574 (69.77)	879 (69.38)	695 (70.27)
2001–3000 RMB	262 (11.61)	146 (11.52)	116 (11.73)
Whether as a student cadre, n (%)				0.035	0.851
Yes	1172 (51.95)	656 (51.78)	516 (52.17)
No	1084 (48.05)	611 (48.22)	473 (47.83)
Hospital level during internship, n (%)				6.498	0.011
Tertiary	1708 (75.71)	985 (77.74)	723 (73.10)
Secondary	548 (24.29)	282 (22.26)	266 (26.90)
Frequency of night-shift per month, n (%)				8.868	0.031
0–2	803 (35.60)	468 (36.94)	335 (33.87)
3–4	1037 (45.96)	590 (46.57)	447 (45.20)
5–6	278 (12.32)	145 (11.44)	133 (13.45)
>6	138 (6.12)	64 (5.05)	74 (7.48)
Career intention, n (%)				79.820	< 0.001
Yes	1943 (86.13)	1164 (91.87)	779 (78.77)
No	313 (13.87)	103 (8.13)	210 (21.23)
Professional identity, median (IQR)	39.00 (34.00, 43.00)	41.00 (37.00, 44.00)	36.00 (31.00, 40.00)	17.774	< 0.001
Self-efficacy, median (IQR)	2.800 (2.50, 3.000)	2.80 (2.60, 3.00)	2.70 (2.40, 3.00)	10.160	< 0.001
Social support, median IQR)	58.00 (50.00, 66.00)	61.00 (53.00, 69.00)	54.00 (48.00, 61.00)	13.058	< 0.001
Psychological resilience, median (IQR)	23.00 (20.00, 29.00)	26.00 (21.00, 30.00)	20.00 (18.00, 25.00)	16.793	< 0.001
Coping style, median (IQR)	34.00 (27.00,4 0.00)	34.00 (27.00, 40.00)	33.00 (27.00, 39.00)	2.691	0.007
Note: IQR, Interquartile spacing; CF, compassion fatigue. a:"non-cases" refers to nursing interns with low levels of compassion fatigue; b: "cases" refers to nursing interns with moderate to severe levels of compassion fatigue.

Latent profiles of nursing students’ compassion fatigue

LPA was performed with one to four latent classes, and the fit indices for these models are shown in Table 2. All classifications had entropy values exceeding 0.9. The BLRT was statistically significant across all models, while the LMR was significant only for the one to three-class models. With an increase in the number of classes, the AIC, BIC, and aBIC values decreased, and the scree plot of aBIC flattened after the three-class model (Fig. 1). Considering statistical criteria, interpretability, and parsimony, the three-class model was selected as the optimal model.

Table 2

Goodness of Fit Statistics for 1 to 4 Class Models *Note*: AIC, Akaike Information Criterion; BIC, Bayesian Information Criterion; aBIC, adjusted Bayesian Information Criterion; LMR, Lo–Mendell–Rubin; BLRT, bootstrap likelihood ratio test; n.a., not applicable.
Class	AIC	BIC	aBIC	Entropy	LMR(p)	BLRT(p)	Sample proportion (%) per class
1	135405.166	135553.921	135471.315	n.a.	n.a.	n.a.	n.a.
2	125201.556	125430.410	125303.324	0.935	< 0.001	< 0.001	71.37/28.63
3	122468.063	122777.016	122605.449	0.908	< 0.001	< 0.001	55.73/32.17/12.10
4	121328.755	121717.807	121501.759	0.916	0.3132	< 0.001	25.67/55.70/8.23/10.40

Table 3 shows the average latent class probabilities for the most likely latent class membership in the three-class model (0.968, 0.940, and 0.960), indicating reasonable classification and good distinction. The distribution and conditional means of the Compassion Fatigue Short Scale items for each class in the three-class model are illustrated in Fig. 2. Based on the conditional means of the items for each class, Class 1 (n = 1257, 55.73%) was defined as the “low compassion fatigue” group, Class 2 (n = 726, 32.17%) as the “moderate compassion fatigue” group, and Class 3 (n = 273, 12.10%) as the “severe compassion fatigue” group.

Table 3

Average latent class probabilities for most likely latent class membership by latent class
Latent class	Latent class membership
Latent class	1	2	3
1	0.968	0.032	0.000
2	0.043	0.940	0.018
3	0.000	0.040	0.960

Influencing factors of nursing students’ compassion fatigue

To explore influencing factors and develop a risk prediction model for compassion fatigue among nursing interns, participants in the “low compassion fatigue” group identified by LPA were categorized as “non-cases” (i.e., without compassion fatigue), and those in the “moderate compassion fatigue” and “severe compassion fatigue” groups were categorized as “cases” (i.e., potentially experiencing compassion fatigue). Univariate analysis revealed statistically significant differences between the “non-case” and “case” groups in terms of major (χ² = 15.358, P < 0.001), gender (χ² = 14.759, P < 0.001), level of internship hospital (χ2 = 6.498, P = 0.011), frequency of night shifts per month (χ² = 8.868, P = 0.031), career intention (χ² = 79.820, P < 0.001), professional identity (z = 17.774, P < 0.001), self-efficacy (z = 10.160, P < 0.001), social support (z = 13.058, P < 0.001), psychological resilience (z = 16.793, P < 0.001), and coping style (z = 2.691, P = 0.007) (see Table 1). These 10 variables were included in the LASSO model. Using 10-fold cross-validation, we identified 9 key features for developing the prediction model: major, gender, frequency of night shifts per month, career intention, professional identity, self-efficacy, social support, psychological resilience, and coping style (see Fig. 3).

Comparison of multiple classification models

We compared the performance of eight machine learning classification models for predicting the risk of compassion fatigue among nursing students during their internship. The XGBoost model demonstrated high stability and accuracy in both the training and validation sets, outperforming others with its AUC, accuracy, and F1 scores. In contrast, the Random Forest model showed perfect performance in the training set but significant overfitting, resulting in a notable decline in the validation set. Other models, such as Logistic Regression, Support Vector Machine, and Multi-Layer Perceptron, showed relatively weaker performance across various metrics. Results are summarized in Table 4, and Fig. 4 provides a comprehensive visualization of the ROC curves for all models in the training and validation sets (Figs. 4A and 4B), the calibration plots (Fig. 4C), and the forest plot with the AUC score results (Fig. 4D).

Table 4

Predictive performance of the eight machine learning techniques in the training and validation sets for compassion fatigue of nursing interns
Models	AUC	Accuracy	Sensitivity	Specificity	PPV	NPV	F1score
Training set
XGBoost	0.826	0.748	0.771	0.731	0.697	0.797	0.732
LR	0.762	0.700	0.663	0.730	0.663	0.737	0.659
RF	1.000	0.999	1.000	1.000	1.000	0.999	1.000
AdaBoost	0.788	0.709	0.731	0.694	0.657	0.768	0.688
GBDT	0.778	0.707	0.711	0.699	0.665	0.742	0.687
GNB	0.744	0.686	0.712	0.667	0.631	0.746	0.667
MLP	0.719	0.654	0.703	0.617	0.587	0.729	0.639
SVM	0.644	0.611	0.677	0.562	0.548	0.688	0.605
Validation set
XGBoost	0.750	0.679	0.643	0.730	0.613	0.738	0.627
LR	0.772	0.707	0.710	0.709	0.665	0.747	0.684
RF	0.749	0.682	0.696	0.684	0.708	0.673	0.701
AdaBoost	0.757	0.677	0.668	0.722	0.608	0.751	0.635
GBDT	0.749	0.679	0.624	0.753	0.638	0.712	0.630
GNB	0.738	0.671	0.661	0.711	0.603	0.738	0.630
MLP	0.704	0.637	0.696	0.610	0.577	0.707	0.630
SVM	0.651	0.612	0.694	0.552	0.548	0.685	0.611

Note: AUC, area under the curve; PPV, positive predictive value; NPV, negative predictive value; LR, logistic regression; SVM, support vector machine; RF, random forest; GBDT, gradient boosting decision tree; MLP, multi-layer perceptron; XGBoost, extreme gradient boosting; GNB, Gaussian naive Bayes; AdaBoost, adaptive boosting.

Model optimization

The XGBoost model, identified as the best performer in predicting compassion fatigue risk, was further optimized through hyper-parameter tuning and selected 9 pre-identified variables as input features. A 10-fold cross-validation strategy was employed, dividing participants into 10 groups, with each group serving as the test set in turn, while the remaining 9 groups were used for training and validation. Figures 5A-C display the ROC curves and AUC values of the XGBoost model in the training, validation, and test sets, which were 0.840, 0.768, and 0.731, respectively, indicating good predictive capability (Table 5 and Fig. 5A-C). Before evaluating the model's accuracy, we generated a learning curve. As shown in Fig. 5D, the error between the training set and the validation set gradually converged as the training samples increased, indicating no significant overfitting. We then assessed the model's accuracy using a calibration plot (Fig. 5E), which demonstrated that the predicted probabilities of the XGBoost model were highly consistent with the actual incidence of "cases." Finally, decision curve analysis (DCA) (Fig. 5F) revealed that the model provided significant net benefits within a risk threshold below 70% compared to the “treat all” or “treat none” strategies, further demonstrating its effectiveness and practicality in real-world applications.

Table 5

Performance of the XGBoost classification model for predicting the risk of compassion fatigue of nursing interns
Models	AUC	Accuracy	Sensitivity	Specificity	PPV	NPV	F1 score
Training set	0.840	0.761	0.720	0.793	0.730	0.788	0.723
Validation set	0.768	0.704	0.653	0.789	0.660	0.741	0.653
Test set	0.731	0.656	0.770	0.581	0.607	0.704	0.679
Note: AUC, area under the curve; PPV, positive predictive value; NPV, negative predictive value. XGBoost, extreme gradient boosting.

Model interpretation

Figure 6 illustrates the impact of various features on the prediction outcomes of the model. As shown in Fig. 6A, the SHAP summary plot, indicates that psychological resilience, professional identity, and social support are the features with the greatest influence on the model’s output. Figure 6B, the bar chart of mean absolute SHAP values, further confirms this finding. Figures 6C and 6D decompose the feature contributions for two sample students, demonstrating the individual-level impact of these features on risk prediction. These figures clearly identify which feature values significantly affect the final risk prediction scores and cause them to fluctuate.

Development of a web-based application for predicting compassion fatigue

We developed a web-based application based on the final model’s predicted risks (available at http://www.xsmartanalysis.com/model/list/predict/model/html?mid= 15239&symbol = 21Dx71Ab62wF580Bl642), facilitating accurate individual risk assessment (Fig. 7).

Latent profile analysis of compassion fatigue among nursing students during the internship

This study utilized LPA to categorize compassion fatigue levels among nursing interns. Based on model fit criteria, the optimal model identified three profiles: “low compassion fatigue,” “moderate compassion fatigue,” and “severe compassion fatigue.”

The “low compassion fatigue” group, comprising 55.73% of the participants, had average scores of 18.74 for burnout and 10.14 for secondary traumatic stress, with an overall average score of 28.874 on the Compassion Fatigue Short Scale. Although these nursing interns showed low overall compassion fatigue, they scored higher on item 2 (“I feel I have not accomplished much in my life”) and item 7 (“I often feel weak, tired, or exhausted as a caregiver”). These scores reflect their emotional and psychological adaptation to new environments and work challenges, likely due to the high intensity of internship tasks, academic pressure, and the fear of incompetence during the transition to their new roles (68, 69). Despite the overall low level of compassion fatigue, the higher scores on specific items warrant attention to prevent progression to moderate or severe levels of fatigue.

The “moderate compassion fatigue” group included 32.17% of the participants, with average scores of 37.78 for burnout and 19.14 for secondary traumatic stress, totaling 56.92 on the Compassion Fatigue Short Scale. This group also scored highest on items 2 and 7. Besides, item 9 (“I feel frustrated with my job”) showed the largest score difference, indicating professional confusion, fatigue, and emotional distress. This may stem from heightened sensitivity to patient suffering and death, leading to increased burnout and frustration (70). Additionally, the inability to establish stable relationships with supervisors and staff during a short internship exacerbates their emotional burden (71). Therefore, mindfulness training to enhance self-awareness and emotional management, along with a well-established professional and peer support system to provide continuous guidance and emotional support, is crucial to prevent the development of higher levels of compassion fatigue (72, 73).

The “severe compassion fatigue” group, representing 12.10% of the participants, had average scores of 55.92 for burnout and 32.83 for secondary traumatic stress, with a total score of 88.75. This group scored highest on items 2 and 7, and showed the largest score differences on item 8 (“I have intrusive thoughts about the traumatic situations I have encountered”) and item 10 (“I relive traumatic experiences when helping those in crisis”), indicating rapidly increasing psychological trauma and distress. These interns lack effective coping strategies and psychological resilience, leading to higher secondary traumatic stress and compassion fatigue (74, 75). Long-term high-pressure environments and cumulative emotional burdens further aggravate these issues (75). Therefore, psychological counseling, support groups, stress management and coping skills training, and crisis peer support workshops are crucial for managing these high-pressure environments and traumatic situations(76).

Influencing factors of compassion fatigue among nursing students during the internship

Demographic characteristics

Our study identified several demographic factors associated with higher compassion fatigue: being female, specializing in midwifery, having frequent night shifts, and intending to switch to another profession after graduation.

Female students reported higher compassion fatigue than males. Sacco et al. (77) found similar results among critical care nurses. Female students generally possess higher emotional intelligence and empathy(78), making them emotionally engaged with patients, leading to emotional exhaustion and compassion fatigue over time (79, 80). Additionally, cultural norms in China discourage emotional expression in males, making male students less likely to engage deeply emotionally, reducing their risk of compassion fatigue (78).

Midwifery students reported higher compassion fatigue than nursing students. Despite working in generally positive obstetric environments, midwifery students frequently encounter complex or critical childbirth situations and neonatal deaths, leading to significant emotional strain (81, 82). They also form strong empathetic bonds with patients, increasing their risk of secondary traumatic stress (76, 83). Early emphasis on normal physiological birth may mislead students about the challenges of adverse events, resulting in increased emotional pressure and fatigue (84).

Students with frequent night shifts reported higher compassion fatigue scores, consistent with findings among hospice nurses (85). Night shifts increase clinical knowledge and independence but also cause constant stress, emotional exhaustion, and exploitation or bullying, leading to anxiety and depression (85).

Students intending to leave the profession after graduation reported higher compassion fatigue. This lack of intrinsic motivation and misalignment between expectations and reality during internships leads to emotional exhaustion and moral distress (86). Educators should encourage reflection on the nursing role to establish realistic career expectations and provide necessary support (87).

Psychological resilience

Psychological resilience was found to be a protective factor against compassion fatigue. Higher psychological resilience was associated with a lower risk of developing compassion fatigue, aligning with previous research (15, 88). Psychological resilience enhances nursing students’ psychosocial functioning and professional performance (88, 89), avoids emotion-centered coping, and builds confidence in managing workplace stress, thereby mitigating compassion fatigue and facilitating adaptation to the clinical environment (90). As advocated by the American Association of Colleges of Nursing, nursing education should emphasize resilience development to improve overall well-being of nursing students (91).

Professional identity

Higher professional identity was associated with lower compassion fatigue. This negative correlation between compassion fatigue and professional identity has also been validated in similar groups (92). Professional identity acts as a psychological resource, helping nursing students maintain a positive attitude during high-pressure and high-risk internship tasks and reducing negative emotional impact (93). Decreased professional identity leads to reduced work enthusiasm and satisfaction, increasing susceptibility to compassion fatigue (94, 95).

Social support

Greater social support was linked to lower risk of compassion fatigue, consistent with findings reported by a study involving 307 intern nursing and midwifery students (14). Social support enhances resilience, alleviates stress, loneliness, and anxiety (96–98). According to the stress coping model, social support is a crucial external resource for coping with stress (99). When students receive support from family, friends, and mentors, they may gain confidence and courage to employ positive coping strategies, thus mitigating negative emotions and preventing compassion fatigue (100). Therefore, it is essential to develop and refine social skills development strategies related to support networks for nursing students, guiding them to actively seek social support when facing internship pressures.

Coping style

Negative coping strategies significantly and positively predict higher compassion fatigue, while positive coping strategies negatively predict lower compassion fatigue. Similar findings have also been validated in Rui's study on clinical nurses (101). Coping strategies influence occupational stress and mental health (102). The demanding internship tasks and academic pressures consume a substantial amount of energy, and as resources are continuously depleted, negative emotions gradually emerge (103). Nursing students who adopt negative coping strategies are more likely to develop negative thoughts and avoidance behaviors, thereby increasing the risk of compassion fatigue (104). In contrast, positive coping strategies help nursing students maintain a positive attitude and mental health by effectively managing stressful events and mobilizing resources, thereby mitigating the impact of compassion fatigue (102, 104).

Self-efficacy

Higher self-efficacy was associated with lower compassion fatigue, consistent with findings in other nurse populations (105, 106). Self-efficacy fosters a positive perception of nursing, stronger professional identity, and better career preparation (107, 108), enhancing job satisfaction and reducing burnout (105, 109). Therefore, nursing educators should guide students in seeking meaningful experiences and learning opportunities to build self-efficacy, improve adaptability, and reduce the incidence of compassion fatigue. Abusubhiah et al. (2023) advocate for the use of multimodal interventions, including flipped classrooms, simulations, debriefing, and role-playing, to better enhance nursing students’ self-efficacy (110).

Development and application of a machine learning-based predictive model for compassion fatigue among nursing students during internships

To our knowledge, this study is the first to develop a user-friendly, personalized model to predict compassion fatigue among nursing interns. Utilizing machine learning techniques, we confirmed the accuracy of this model, which contributes to optimizing clinical management. Additionally, our study identified nine independent risk factors for compassion fatigue: gender, specialty, frequency of night shifts, career intention, resilience scores, professional identity scores, social support scores, coping style scores, and self-efficacy scores. Recently, a study based on 21 tertiary hospitals in China developed a risk prediction model for compassion fatigue among emergency department nurses using a logistic regression model (111). This model included only seven factors: job position, job satisfaction, dietary habits, daily sleep duration, occupational stress, physical harassment, and workplace violence levels. In contrast, our model demonstrated higher predictive efficiency, covering a broader range of influencing factors, and providing a more comprehensive assessment of the risk of compassion fatigue. However, our model has not yet been validated in external populations, so its external validity may need further confirmation.

Our prediction model can identify nursing students at high risk of compassion fatigue before symptoms manifest. In clinical applications, the online measurement tool can be combined with existing compassion fatigue scales to provide a comprehensive assessment of nursing students. For instance, a student currently showing no symptoms of compassion fatigue may score zero on the scale, but this does not mean they will not experience compassion fatigue in the future. Our online tool calculates the probability of each student developing compassion fatigue, serving as a powerful supplement to existing scales and assisting nursing educators and clinical managers in devising appropriate interventions to support students' health and professional development.

Strengths and limitation

This study has yielded some promising findings, primarily due to the following advantages. First, there are few existing machine learning prediction models specifically addressing compassion fatigue in nursing students during the internship, and our study fills this gap. Second, this study innovatively combines LPA and machine learning to identify different risk groups among nursing students through LPA and to predict individual risk levels using machine learning models. The application of machine learning algorithms enhances the predictive power of the model, offering advantages over traditional risk prediction methods, and aids school and hospital administrators in identifying high-risk individuals for compassion fatigue and implementing targeted interventions.

However, this study also acknowledges several limitations. First, the cross-sectional design may limit causal inference and introduce selection bias. Second, the relatively small sample size and lack of external validation may affect the robustness of the model. Therefore, further prospective studies with larger and more diverse samples are needed to evaluate the diagnostic sensitivity and specificity of the model. Third, the selection of input data and model parameters may vary when encountering new data, necessitating further research. Finally, the routine collection of psychosocial parameters in clinical settings may be challenging, potentially hindering the practical application of the model. Therefore, developing predictive models that include only objective parameters with high predictive performance would be a beneficial approach.

This study developed a predictive model integrating demographic characteristics, work-related, and psychosocial variables, demonstrating excellent calibration and predictive capability for compassion fatigue among nursing students. The resulting online tool can screen students at moderate to high risk during internships, aiding nursing managers in optimizing prevention strategies and reducing incidence.

LPA, latent profile analysis

LMR, Lo-Mendell-Rubin likelihood ratio test

BLRT, bootstrap likelihood ratio test

AIC, Akaike Information Criterion

BIC, Bayesian Information Criterion

aBIC, adjusted Bayesian Information Criterion

LASSO, the Least Absolute Shrinkage and Selection Operator

AUC, area under the receiver operating characteristic curve

ROC, receiver operating characteristic

DCA, Decision curve analysis

Ethics approval and consent to participate

Institutional Review Board approval was granted for this study by the institution review board of Hunan Traditional Chinese Medical College, with an approval number of YX202212001. Informed consent was obtained from all participants included in the study.

Consent for publication

Not applicable.

Availability of data and materials

The datasets used and/or analysed during the current study are available from the corresponding author[LJY] on reasonable request.

Competing interests

The authors declare that they have no competing interests.

Funding

The study was funded by 2024 Hunan Provincial Universities Ideological and Political Work Quality Project (Practice Education)[ grant number:24JP039].

Authors’ contributions

LIY and XT conceived and designed the study. LJY, TS, and XT contributed to the acquisition, analysis, and interpretation of data. LJY, YL, JJZ, and MFJH were involved in investigation, methodology, and data curation. LJY drafted the manuscript. TX revised subsequent drafts. All authors reviewed and approved the final manuscript.

Acknowledgements

Thank you to all nursing interns who volunteered to participate in this study and also to the full-time student counselors for their great support in data collection. At the same time, we are grateful for the hard work of the editors and the valuable suggestions of the reviewers.

Cheng J, Cui J, Yu W, Kang H, Tian Y, Jiang X. Factors influencing nurses' behavioral intention toward caring for COVID-19 patients on mechanical ventilation: A cross-sectional study. PLoS ONE. 2021;16(11):e0259658.
Lombardo B, Eyre C. Compassion fatigue: a nurse's primer. Online J Issues Nurs. 2011;16(1):3.
Zhang YY, Zhang C, Han XR, Li W, Wang YL. Determinants of compassion satisfaction, compassion fatigue and burn out in nursing: A correlative meta-analysis. Med (Baltim). 2018;97(26):e11086.
Joinson C. Coping with compassion fatigue. Nursing. 1992;22(4):116.
Figley CR. New York: Brunner Mazel [Internet]1995. [cited 7].
Figley CR. Compassion fatigue: psychotherapists' chronic lack of self care. J Clin Psychol. 2002;58(11):1433–41.
Phillips CS, Becker H, Gonzalez E. Psychosocial Well-Being: An Exploratory Cross-Sectional Evaluation of Loneliness, Anxiety, Depression, Self-Compassion, and Professional Quality of Life in Oncology Nurses. Clin J Oncol Nurs. 2021;25(5):530–8.
Lee LJ, Wehrlen L, Ding Y, Ross A. Professional quality of life, sleep disturbance and health among nurses: A mediation analysis. Nurs Open. 2022;9(6):2771–80.
Zhang Y, Peng Q, Dong W, Hou C, Chen C. Professional identity and sense of coherence affect the between compassion fatigue and work engagement among Chinese hospital nurses. BMC Nurs. 2023;22(1):472.
Tuna R, Eskin Bacaksiz F, Kahraman B. The Relationship Between Musculoskeletal Discomfort and Professional Quality of Life Among Nurses. Pain Manag Nurs. 2022;23(1):62–7.
Xie W, Chen L, Feng F, Okoli CTC, Tang P, Zeng L, et al. The prevalence of compassion satisfaction and compassion fatigue among nurses: A systematic review and meta-analysis. Int J Nurs Stud. 2021;120:103973.
Garnett A, Hui L, Oleynikov C, Boamah S. Compassion fatigue in healthcare providers: a scoping review. BMC Health Serv Res. 2023;23(1):1336.
Han B, R,;, Chen H, Xie J. Current status of compassion fatigue level among nursing students at the late stage of internship and its influencing factors. Chin J Mod Nurs. 2017;23(21):5.
Li JN, Jiang XM, Zheng QX, Lin F, Chen XQ, Pan YQ, et al. Mediating effect of resilience between social support and compassion fatigue among intern nursing and midwifery students during COVID-19: A cross-sectional study. BMC Nurs. 2023;22(1):42.
Cao XY, Wang L, Wei SY, Li J, Gong S. Prevalence and predictors for compassion fatigue and compassion satisfaction in nursing students during clinical placement. Nurse Educ Pract. 2021;51:102999.
Chachula KM. Professional Quality of Life Factors and Relationships in Nursing and Psychiatric Nursing Students: An Exploratory Study. SAGE open Nurs. 2021;7:2377960821994394.
Chi DL, Randall CL, Hill CM. Dental trainees' mental health and intention to leave their programs during the COVID-19 pandemic. Journal of the American Dental Association (1939). 2021;152(7):526 – 34.
Drummond JR, Duguid R. Student drop-out from UK dental schools. Br Dent J. 1997;182(9):347–9.
Marôco J, Assunção H, Harju-Luukkainen H, Lin SW, Sit PS, Cheung KC, et al. Predictors of academic efficacy and dropout intention in university students: Can engagement suppress burnout? PLoS ONE. 2020;15(10):e0239816.
Rudman A, Gustavsson P, Hultell D. A prospective study of nurses' intentions to leave the profession during their first five years of practice in Sweden. Int J Nurs Stud. 2014;51(4):612–24.
Rudman A, Gustavsson JP. Burnout during nursing education predicts lower occupational preparedness and future clinical performance: A longitudinal study. Int J Nurs Stud. 2012;49(8):988–1001.
Sinclair S, Raffin-Bouchal S, Venturato L, Mijovic-Kondejewski J, Smith-MacDonald L. Compassion fatigue: A meta-narrative review of the healthcare literature. Int J Nurs Stud. 2017;69:9–24.
Bride B, Radey M, Figley C. Measuring Compassion Fatigue. Clin Soc Work J. 2007;35:155–63.
Figley CR. Compassion fatigue: Coping with secondary traumatic stress disorder in those who treat the traumatized. Philadelphia, PA,USA: Brunner/Mazel; 1995.
Gentry JE, Baranowsky AB, Dunning K. The accelerated recovery program (ARP) for compassion fatigue. Treating compassion fatigue. New York, US: Brunner-Routledge; 2002. pp. 123–38.
Adams RE, Boscarino JA, Figley CR. Compassion fatigue and psychological distress among social workers: a validation study. Am J Orthopsychiatry. 2006;76(1):103–8.
Stamm BH. The Concise ProQOL Manual: The concise manual for the Professional Quality of Life Scale, 2 nd Edition 2010. http://www.proqol.org/uploads/ProQOL_ Concise_2ndEd_12-2010.pdf.
Sun B, Hu M, Yu S, Jiang Y, Lou B. Validation of the Compassion Fatigue Short Scale among Chinese medical workers and firefighters: a cross-sectional study. BMJ Open. 2016;6(6):e011279.
Bouchard L, Rainbow J. Compassion fatigue, presenteeism, Adverse Childhood Experiences (ACES), and resiliency levels of Doctor of Nursing Practice (DNP) students. Nurse Educ Today. 2021;100:104852.
Liu H, Zhang Z, Chi C, Tao X, Zhang M. Exploring the Impact of the COVID-19 Pandemic on Academic Burnout Among Nursing College Students in China: A Web-Based Survey. Med Sci Monit Basic Res. 2023;29:e940997.
Jing T, Peng Y, Wu S. Study on the status and influencing factors of empathy fatigue among nursing interns under the background of public health emergencies. Health Vocat Educ. 2024;42(4):131–5.
Chen C, Li Y, Huang D, Liu X, Xu Z, Li J. Trends and influencing factors of empathy fatigue in nursing interns: A longitudinal study[J/OL]. J Adv Nurs. 2023.
Wang M, Guan H, Li Y, Xing C, Rui B. Academic burnout and professional self-concept of nursing students: A cross-sectional study. Nurse Educ Today. 2019;77:27–31.
Yang H-J. Factors affecting student burnout and academic achievement in multiple enrollment programs in Taiwan’s technical–vocational colleges. Int J Educational Dev. 2004;24(3):283–301.
Ye L. Analysis of the current status and influencing factors of empathy fatigue among nursing interns. In Proceedings of the 2022 Shanghai International Nursing Conference; Shanghai2022. pp. 35 – 6.
Liu Z, Xie Y, Sun Z, Liu D, Yin H, Shi L. Factors associated with academic burnout and its prevalence among university students: a cross-sectional study. BMC Med Educ. 2023;23(1):317.
Miranda-Ackerman RC, Barbosa-Camacho FJ, Sander-Möller MJ, Buenrostro-Jiménez AD, Mares-País R, Cortes-Flores AO, et al. Burnout syndrome prevalence during internship in public and private hospitals: a survey study in Mexico. Med Educ Online. 2019;24(1):1593785.
Wang X, Cai HJ, Ye HF, Wang Y, Li QQ. Analysis of the current situation and related factors of empathy fatigue in undergraduate nursing interns. Chin High Med Educ. 2023(3):19–21.
Williams GA, Kibowski F. Latent class analysis and latent profile analysis. Handbook of methodological approaches to community-based research: Qualitative, quantitative, and mixed methods. 2016;15:143 – 51.
Roeser RW, Eccles JS, Sameroff AJ. Academic and emotional functioning in early adolescence: Longitudinal relations, patterns, and prediction by experience in middle school. Dev Psychopathol. 1998;10(2):321–52.
Mathew A, Doorenbos AZ. Latent profile analysis–an emerging advanced statistical approach to subgroup identification. Indian J Continuing Nurs Educ. 2022;23(2):127–33.
Ferguson SL, Moore G, Hull EW. Finding latent groups in observed data: A primer on latent profile analysis in Mplus for applied researchers. Int J Behav Dev. 2019;44(5):458–68.
Huettmann F, Craig EH, Herrick KA, Baltensperger AP, Humphries GRW, Lieske DJ, et al. Use of Machine Learning (ML) for Predicting and Analyzing Ecological and ‘Presence Only’ Data: An Overview of Applications and a Good Outlook. In: Humphries G, Magness DR, Huettmann F, editors. Machine Learning for Ecology and Sustainable Natural Resource Management. Cham: Springer International Publishing; 2018. pp. 27–61.
Andrade C. Sample Size and its Importance in Research. Indian J Psychol Med. 2020;42(1):102–3.
Riley RD, Ensor J, Snell KIE, Harrell FE, Martin GP, Reitsma JB, et al. Calculating the sample size required for developing a clinical prediction model. BMJ. 2020;368:m441.
Nylund-Gibson K, Garber AC, Carter DB, Chan M, Arch DAN, Simon O, et al. Ten frequently asked questions about latent transition analysis. Psychol Methods. 2023;28(2):284–300.
Spurk D, Hirschi A, Wang M, Valero D, Kauffeld S. Latent profile analysis: A review and how to guide of its application within vocational behavior research. J Vocat Behav. 2020;120:103445.
Brown J, Suls J. Psychological perspectives on the self. Lawrence Erlbaum Associates Hillsdale, NJ; 1993.
Lu H, While AE, Barriball KL. Job satisfaction and its related factors: a questionnaire survey of hospital nurses in Mainland China. Int J Nurs Stud. 2007;44(4):574–88.
Schwarzer R, Jerusalem M, Weinman J, Wright S, Johnston M. Measures in health psychology: A user’s portfolio. Causal and control beliefs. Causal control beliefs. 1995;1(011):35–7.
Zeng G, Fung S-f, Li J, Hussain N, Yu P. Evaluating the psychometric properties and factor structure of the general self-efficacy scale in China. Curr Psychol. 2020:1–11.
Zimet GD, Powell SS, Farley GK, Werkman S, Berkoff KA. Psychometric characteristics of the Multidimensional Scale of Perceived Social Support. J Pers Assess. 1990;55(3–4):610–7.
Jiang Q. Perceived social support scale. Chin J Behav Med Sci. 2001;10(10):41–3.
Yan B, Zheng X, Zhang X. The influence mechanism of social support on subjective well-being of college students: the mediating role of self-control and depression. Psychol Sci. 2011;34(02):471–5.
Connor KM, Davidson JR. Development of a new resilience scale: the Connor-Davidson Resilience Scale (CD-RISC). Depress Anxiety. 2003;18(2):76–82.
Campbell-Sills L, Stein MB. Psychometric analysis and refinement of the Connor-davidson Resilience Scale (CD-RISC): Validation of a 10-item measure of resilience. J Trauma Stress. 2007;20(6):1019–28.
Ye Z, Ruan X, Zeng Z, Xie Q, Cheng M, Peng C, Qiu H. Psychometric properties of 10-item Connor-Davidson resilience scale among nursing students. J Nurs. 2016;23:9–13.
Xie Y-N. A preliminary study on the reliability and validity of the simple coping style scale. Chin J Clin Psychol. 1998;6(2):114–5.
Folkman S, Lazarus RS. Ways of coping questionnaire. 1988.
Tein JY, Coxe S, Cham H. Statistical Power to Detect the Correct Number of Classes in Latent Profile Analysis. Struct Equ Model. 2013;20(4):640–57.
Fonseca-Pedrero E, Ortuño-Sierra J, de Álbeniz AP, Muñiz J, Cohen AS. A latent profile analysis of schizotypal dimensions: Associations with psychopathology and personality. Psychiatry Res. 2017;253:110–5.
Marsh HW, Hau K-T, Wen Z. In Search of Golden Rules: Comment on Hypothesis-Testing Approaches to Setting Cutoff Values for Fit Indexes and Dangers in Overgeneralizing Hu and Bentler's (1999) Findings. Structural Equation Modeling: A Multidisciplinary Journal. 2004;11(3):320 – 41.
Li JB, Wu AMS, Feng LF, Deng Y, Li JH, Chen YX, et al. Classification of probable online social networking addiction: A latent profile analysis from a large-scale survey among Chinese adolescents. J Behav Addict. 2020;9(3):698–708.
Fu H, Si L, Guo R. What Is the Optimal Cut-Off Point of the 10-Item Center for Epidemiologic Studies Depression Scale for Screening Depression Among Chinese Individuals Aged 45 and Over? An Exploration Using Latent Profile Analysis. Front Psychiatry. 2022;13.
Wu Y, Dai Z, Xiao W, Wang H, Huang Y, Si M et al. Perceived stigma among discharged patients of COVID-19 in Wuhan, China: A latent profile analysis. Front Public Health. 2023;11.
Vasquez MM, Hu C, Roe DJ, Chen Z, Halonen M, Guerra S. Least absolute shrinkage and selection operator type methods for the identification of serum biomarkers of overweight and obesity: simulation and application. BMC Med Res Methodol. 2016;16(1):154.
Covert I, Lundberg SM, Lee S-I. Understanding global feature contributions with additive importance measures. Adv Neural Inf Process Syst. 2020;33:17212–23.
Lopes AR, Nihei OK. Burnout among nursing students: predictors and association with empathy and self-efficacy. Rev Bras Enferm. 2020;73(1):e20180280.
Valero-Chillerón MJ, González-Chordá VM, López-Peña N, Cervera-Gasch Á, Suárez-Alcázar MP, Mena-Tudela D. Burnout syndrome in nursing students: An observational study. Nurse Educ Today. 2019;76:38–43.
Zheng R, Guo Q, Dong F, Gao L. Death Self-efficacy, Attitudes Toward Death and Burnout Among Oncology Nurses: A Multicenter Cross-sectional Study. Cancer Nurs. 2022;45(2):E388–96.
Velando-Soriano A, Suleiman-Martos N, Pradas-Hernández L, Membrive-Jiménez MJ, Ramírez-Baena L, Gómez-Urquiza JL. Cañadas-De La Fuente GA. Factors related to the appearance and development of burnout in nursing students: a systematic review and meta-analysis. Front Public Health. 2023;11:1142576.
Melissant HC, Hendriks RRA, Bakker EJM, Kox JHAM, Rietveld N, Miedema HS, et al. Interventions that support novice nurses' transition into practice: A realist review. Int J Nurs Stud. 2024;157:104785.
Hegney DG, Rees CS, Eley R, Osseiran-Moisson R, Francis K. The contribution of individual psychological resilience in determining the professional quality of life of Australian nurses. Front Psychol. 2015;6:1613.
Barleycorn D. Awareness of secondary traumatic stress in emergency nursing. Emerg Nurse. 2019;27(5):19–22.
Davies S, Coldridge L. No Man׳s Land': An exploration of the traumatic experiences of student midwives in practice. Midwifery. 2015;31(9):858–64.
Sacco TL, Ciurzynski SM, Harvey ME, Ingersoll GL. Compassion Satisfaction and Compassion Fatigue Among Critical Care Nurses. Crit Care Nurse. 2015;35(4):32–43. quiz 1p following.
Deng X, Chen S, Li X, Tan C, Li W, Zhong C, et al. Gender differences in empathy, emotional intelligence and problem-solving ability among nursing students: A cross-sectional study. Nurse Educ Today. 2023;120:105649.
Yu H, Qiao A, Gui L. Predictors of compassion fatigue, burnout, and compassion satisfaction among emergency nurses: A cross-sectional survey. Int Emerg Nurs. 2021;55:100961.
Borges EMN, Fonseca CINS, Baptista PCP, Queirós CML, Baldonedo-Mosteiro M, Mosteiro-Diaz MP. Fadiga por compaixão em enfermeiros de urgência e emergência hospitalar de adultos. Rev Latinoam Enferm. 2019;27:e3175.
Weeks F, Pantoja L, Ortiz J, Foster J, Cavada G, Binfa L. Labor and Birth Care Satisfaction Associated With Medical Interventions and Accompaniment During Labor Among Chilean Women. J Midwifery Womens Health. 2017;62(2):196–203.
Saxbe D, Horton KT, Tsai AB. The Birth Experiences Questionnaire: A brief measure assessing psychosocial dimensions of childbirth. J Fam Psychol. 2018;32(2):262–8.
Curtis P, Ball L, Kirkham M. Why do midwives leave? (Not) being the kind of midwife you want to be. Br J Midwifery. 2006;14(1):27–31.
McKenna L, Rolls C. Undergraduate midwifery students’ first experiences with stillbirth and neonatal death. Contemp Nurse. 2011;38(1–2):76–83.
Shi H, Shan B, Chen Q, Guo F, Zhou X, Shi M, Liu Y. Prevalence and predictors of compassion satisfaction, secondary traumatic stress, and burnout among Chinese hospice nurses: A cross-sectional study. Appl Nurs Res. 2023;69:151648.
Kandil F, El Seesy N, Banakhar M. Factors affecting students’ preference for nursing education and their intent to leave: A cross-sectional study. Open Nurs J. 2021;15(1).
Kotera Y, Cockerill V, Chircop J, Kaluzeviciute G, Dyson S. Predicting self-compassion in UK nursing students: Relationships with resilience, engagement, motivation, and mental wellbeing. Nurse Educ Pract. 2021;51:102989.
Karabey T. Compassion fatigue and psychological resilience levels of nursing final students: A descriptive, cross-sectional, and relational study. Palliat Support Care. 2023:1–7.
Garcia-Dia MJ, DiNapoli JM, Garcia-Ona L, Jakubowski R, O'Flaherty D. Concept analysis: resilience. Arch Psychiatr Nurs. 2013;27(6):264–70.
Yu F, Raphael D, Mackay L, Smith M, King A. Personal and work-related factors associated with nurse resilience: A systematic review. Int J Nurs Stud. 2019;93:129–40.
Nursing AAoCo. well-being and resilience 2020 [ https://www.aacnnursing.org/our-initiatives/well-being-resilience.
Ruan TT. Analysis of nursing students' professional identity and influencing factors under the epidemic of public health emergencies. Contemp Nurse: Late Ed. 2021;28(12):4.
Geoffrion S, Morselli C, Guay S. Rethinking Compassion Fatigue Through the Lens of Professional Identity: The Case of Child-Protection Workers. Trauma Violence Abuse. 2016;17(3):270–83.
Zhou X, Yang H, Yang Y, Zhang Y, Yang H. Mediating effect of professional identity on professional benefit and work engagement among nurses. J Nurs Sci. 2019;34:51–4.
Xie Y, Zhao H, Huang Y, Liu C. The correlation between empathy fatigue and the sense of professional benefits in intern nurses. Contemp Nurse (Mid-Month Edition). 2020;27(10):153–6.
Jeong YJ, Koh CK. Female nursing graduate students' stress and health: the mediating effects of sense of coherence and social support. BMC Nurs. 2021;20(1):40.
Labrague LJ. COVID-19 phobia, loneliness, and dropout intention among nursing students: the mediating role of social support and coping. Curr Psychol. 2023:1–9.
Onieva-Zafra MD, Fernández-Muñoz JJ, Fernández-Martínez E, García-Sánchez FJ, Abreu-Sánchez A, Parra-Fernández ML. Anxiety, perceived stress and coping strategies in nursing students: a cross-sectional, correlational, descriptive study. BMC Med Educ. 2020;20(1):370.
Lazarus RS, Folkman S. Stress, appraisal, and coping. Springer publishing company; 1984.
Zhang J, Wang X, Chen O, Li J, Li Y, Chen Y, et al. Social support, empathy and compassion fatigue among clinical nurses: structural equation modeling. BMC Nurs. 2023;22(1):425.
Jiao R, Li J, Cheng N, Liu X, Tan Y. The mediating role of coping styles between nurses' workplace bullying and professional quality of life. BMC Nurs. 2023;22(1):459.
Zhou Y, Guo X, Yin H. A structural equation model of the relationship among occupational stress, coping styles, and mental health of pediatric nurses in China: a cross-sectional study. BMC Psychiatry. 2022;22(1):416.
Visier-Alfonso ME, Sarabia-Cobo C, Cobo-Cuenca AI, Nieto-López M, López-Honrubia R, Bartolomé-Gutiérrez R, et al. Stress, mental health, and protective factors in nursing students: An observational study. Nurse Educ Today. 2024;139:106258.
Ding Y, Yang Y, Yang X, Zhang T, Qiu X, He X, et al. The Mediating Role of Coping Style in the Relationship between Psychological Capital and Burnout among Chinese Nurses. PLoS ONE. 2015;10(4):e0122128.
Zhang J, Wang X, Xu T, Li J, Li H, Wu Y, et al. The effect of resilience and self-efficacy on nurses' compassion fatigue: A cross-sectional study. J Adv Nurs. 2022;78(7):2030–41.
Hao X, Chen X, Sun C, Zhu Z, Qiao C, Ji M. The mediating effect of general self-efficacy of outpatient nurses on the relationship between perception of work environment and compassion fatigue. Chin Nurs Manage. 2023;23(3):461–6.
Fangonil-Gagalang E. Association of self-efficacy and faculty support on students' readiness for practice. J Prof Nurs. 2024;52:30–9.
Liu Y, Chong MC, Han Y, Wang H, Xiong L. The mediating effects of self-efficacy and study engagement on the relationship between specialty identity and career maturity of Chinese nursing students: a cross-sectional study. BMC Nurs. 2024;23(1):339.
Xu R, Du W, Liu Z. Correlation analysis of general self-efficacy and nursing quality among neurology nurses in Lanzhou. Chin J Practical Nurs. 2016;32(19):1445–50.
Abusubhiah M, Walshe N, Creedon R, Noonan B, Hegarty J. Self-efficacy in the context of nursing education and transition to practice as a registered practitioner: A systematic review. Nurs Open. 2023;10(10):6650–67.
Xie W, Liu M, Okoli CTC, Zeng L, Huang S, Ye X, et al. Construction and evaluation of a predictive model for compassion fatigue among emergency department nurses: A cross-sectional study. Int J Nurs Stud. 2023;148:104613.

No competing interests reported.

Download PDF

Editorial decision: Revision requested
09 Sep, 2024
Reviews received at journal
03 Sep, 2024
Reviewers agreed at journal
23 Aug, 2024
Reviews received at journal
08 Aug, 2024
Reviewers agreed at journal
07 Aug, 2024
Reviewers invited by journal
05 Aug, 2024
Editor invited by journal
16 Jul, 2024
Editor assigned by journal
15 Jul, 2024
Submission checks completed at journal
15 Jul, 2024
First submitted to journal
09 Jul, 2024

You are reading this latest preprint version

Development and validation of a machine learning-based predictive model for compassion fatigue in nursing interns: A cross-sectional study with latent profile analysis

Status:

Version 1

Abstract

Background

Design:

Methods

Results

Conclusions

Figures

Background

Methods

Study design

Participants

Sample size

Data collection

Instruments

General demographic information collection table

The Compassion Fatigue Short Scale

The Professional Identity Scale

The General Self-efficacy Scale

The Perceived Social Support Scale

The 10-item Connor-Davidson Resilience Scale

Simple Coping Style Questionnaire

Statistical analyses

Descriptive and univariate analysis

Latent profile analysis

Features selection

Development and evaluation of machine learning models learning

Model optimization and evaluation

Results

Descriptive statistics

Latent profiles of nursing students’ compassion fatigue

Influencing factors of nursing students’ compassion fatigue

Comparison of multiple classification models

Model optimization

Model interpretation

Development of a web-based application for predicting compassion fatigue

Discussion

Latent profile analysis of compassion fatigue among nursing students during the internship

Influencing factors of compassion fatigue among nursing students during the internship

Demographic characteristics

Psychological resilience

Professional identity

Social support

Coping style

Self-efficacy

Strengths and limitation

Conclusion

Abbreviations

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1