Prevalence and correlates of perceived stress and psychological health among residents receiving standardized training: a multicenter cross-sectional survey study in China​ Authors: Min Wang, Mengyuan Liu, Jingshuang Wang, Hong Zuo, Bo Wang, Jiachuan Xiong, Weijing Lai Published: 2025 ABSTRACT AbstractBackgroundMedical residents are increasingly conscious of psychological discomfort. However, comprehensive epidemiology on the prevalence and modifiable factors for their perceived stress and psychological health within China’s standardized residency training is lacking.MethodsThe multicenter cross-sectional survey used convenience sampling to recruit participants from 177 teaching hospitals across 28 Chinese regions. Anonymous online questionnaires assessed demographics, training routines, subjective stress levels, and validated measures, including the Chinese Perceived Stress Scale and the Symptom Checklist-90 (SCL-90). A psychological health issue was defined as an SCL-90 total score ≥ 160, ≥ 43 positive items, or ≥ 2 in any symptom dimension.Results Among 2,104 respondents, 49.6% reported a psychological health concern, with no significant differences by gender or training duration (P > 0.05). Modifiable risk factors included staying up late (OR = 1.300; 95% CI: 1.059–1.595), frequent night shifts (OR = 1.304; 95% CI: 1.058–1.608), CPSS score (OR = 1.179; 95% CI: 1.157–1.201), and worsening sleep quality (graded effect: relatively poor sleep quality, OR = 1.503; 95% CI: 1.089–2.074; inferior sleep quality, OR = 2.720; 95% CI: 1.395–5.305). CPSS score ≥ 27 best predicts morbidity (AUC = 0.796, sensitivity = 67.4%, specificity = 75.6%), with interpersonal stress (β = 0.149, P < 0.001) and clinical work stress (β = 0.139, P < 0.001) identified as key stressors.ConclusionApproximately 50% of Chinese medical residents exhibit psychological morbidity, primarily influenced by modifiable stressors. A CPSS score ≥ 27 serves as a preliminary screening threshold, indicating the necessity for training reforms focused on workload modulation and the enhancement of interpersonal skills.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12889-025-25715-4.Keywords: Psychological health, Perceived stress, ​Medical residents, Standardized residency training program Background Residency training represents a vital stage in medical education, during which physicians-in-training develop advanced clinical skills and professional competence. The standardized residency training in China is designed to develop high-quality medical professionals to address the increasing healthcare demands [1]. The standardized training system is essential for enhancing future healthcare delivery, yet it may exacerbate the inherent stressors associated with residency. In addition to striving for excellence in clinical practice, residents frequently encounter long working hours, significant clinical responsibilities, complex patient care situations, and high-stakes examinations, all of which can significantly affect their mental health [2–4]. Recent findings indicate that medical residents participating in standardized training may experience a significant prevalence of psychological health issues [3, 5, 6]. The consequences are complex and substantial, affecting resident education, clinical effectiveness, patient care, and individual well-being, often leading to an increase in medical errors and physician turnover [7–9]. Despite recognizing the burden of resident distress, there is a notable lack of large-scale, rigorous data specifically examining the psychological health status of residents involved in standardized training programs in China. Furthermore, although stress is generally acknowledged as a factor in mental health problems [10–13], the specific prevalence, primary types of psychological distress, and the precise role and relative influence of various stressors within the unique socio-cultural and structural context of Chinese residency training remain insufficiently understood. Identifying key modifiable risk factors and strong predictors is essential for developing targeted interventions. A clear gap remains between the recognized burden of psychological distress among residents and the evidence needed to address it within China’s standardized residency training system. Existing knowledge highlights broad challenges, yet lacks large-scale, detailed data that quantify the prevalence of perceived stress and identify modifiable correlates linked to psychological health outcomes. Such specificity is essential for effective solutions. To address this gap, the present multicenter study aims to: (1) determine the prevalence of perceived stress and specific psychological health problems among medical residents in standardized training across China, and (2) identify the key demographic, occupational, and psychosocial factors most strongly associated with these outcomes. The findings are intended to provide actionable evidence to guide targeted interventions, inform institutional and policy reforms, and support a more sustainable and supportive training environment for the physician workforce. Methods Study design and participants This study was a cross-sectional survey conducted in 2024 using an anonymous online questionnaire via “Questionnaire Star.” Convenient sampling was used as the sampling method. To improve the diversity and potential representativeness of our sample, anonymous participants were recruited from 177 teaching hospitals across 28 provinces in China. The inclusion criteria for medical residents were: (1) at least six months of residency training, (2) voluntary participation, and (3) no organic mental disorders. Participants with any missing key questionnaire data were excluded. The study protocol was approved by the Biomedical Ethics Review Committee of West China Hospital, Sichuan University, in accordance with the Declaration of Helsinki. Informed consent was obtained before participation. Survey instrument An anonymous online survey was conducted using the convenient sampling method through the Questionnaire Star platform (Changsha Ranxing Information Technology Co., Ltd). Demographic information, residency training details, and subjective stress ratings were collected with a newly developed self-administered questionnaire based on qualitative interviews with residents (Qre S, items 1–25). The Chinese Perceived Stress Scale (CPSS) was used to assess perceived stress. The Checklist-90 (SCL-90) was introduced to evaluate mental health. To ensure data quality, two independent researchers were assigned to carefully verify that the questionnaire was completed accurately and completely. CPSS is a translated version of the Perceived Stress Scale (PSS), adapted to the language habits and cultural context of the Chinese population, and was introduced to China in 2003. It is one of the most widely used stress assessment scales in China, demonstrating strong reliability and validity [14], with a Cronbach’s alpha of 0.873 in this study. This scale consists of 14 items, covering two dimensions: sense of tension and sense of loss of control, and is scored on a scale of 1 to 5 for each dimension. The dimension of tension employs a positive scoring system, reflecting the degree of tension and anxiety an individual experiences in response to stress. The loss of control dimension, using ​​reverse scoring, reflects the individual’s sense of control over life events and their ability to cope (Qre S, items 26–39). The SCL-90 is an extensively adopted psychological instrument with well-established reliability and validity across diverse populations [15, 16]. In the present study, it demonstrated excellent internal consistency, with a Cronbach’s alpha of 0.988. The SCL-90 contains 90 items, covering 10 symptom dimensions: somatization (12 items), obsessive-compulsive symptoms (10 items), interpersonal sensitivity (9 items), depression (13 items), anxiety (10 items), hostility (6 items), phobic symptoms (7 items), paranoia (6 items), psychosis (10 items), and others (7 items) (Qre S, items 40–129). Each item is scored on a five-level scale based on severity. If the total score is ≥ 160 points, the number of positive items is ≥ 43, or if the score of any symptom dimension is ≥ 2, it indicates a psychological health problem [17]. Statistical analyses Data that were normally distributed were compared using the mean ± standard deviation, and intergroup comparisons were conducted using the independent samples t-test. Measurement data that did not follow a normal distribution were presented as the median (lower quartile, upper quartile), and intergroup comparisons were performed with the Mann-Whitney U test or the Kruskal-Wallis H test. Categorical data were reported using frequencies and/or percentages; nominal variables were analyzed with the chi-square test, and ordinal variables with the Mann-Whitney U test or the Kruskal-Wallis H test. Binary logistic regression was used to examine risk factors associated with mental health issues, and the ROC curve was used to assess the predictive power of these risk factors. The significance level was set at α = 0.05. Statistical analyses were conducted using SPSS 25.0, R 4.4.2 for Windows, and Prism 9. Results Basic characteristics A total of 2,125 questionnaires were distributed and returned, achieving a 100% response rate. Twenty-one questionnaires were excluded due to logical errors or incomplete key information. Ultimately, 2,104 residents were included in the final analysis, with an average age of 26 years; 765 (36.4%) were male. Participants came from 177 teaching hospitals, with training specialties including internal medicine (49.5%), surgery (28.4%), general medicine (9.3%), medical technology and diagnostics (8.8%), and other disciplines (4.0%). These medical residents represented 20 provinces, 4 autonomous regions, and 4 municipalities in China (Fig. 1). Residents were then categorized into male and female groups. Male residents showed a higher BMI (P < 0.001), a greater tendency to be only children (P < 0.001), and a higher frequency of exercise (P < 0.001) compared to female residents. Yet, they reported poorer sleep quality (P < 0.001). The proportion of female residents specializing in surgery was significantly lower than that of their male counterparts (P < 0.001). No significant differences were observed regarding age, grade, household registration status, marital status, the proportion of graduate students, dormitory accommodation, number of close friends, frequency of staying up late, frequency of night shifts, or parents’ educational level (P > 0. 05) (Table 1). Among respondents, 844 (40.1%) were first-year students, 553 (26.3%) were second-year students, and 707 (33.6%) were third-year students. First-year residents exhibited better sleep quality than their second- and third-year counterparts. The proportion of third-year residents living in dormitories was significantly lower (P = 0.034). As the grade level increased, age and marriage rates rose significantly (P < 0.001), while the frequency of exercise decreased (P = 0.038). Additionally, parents of first- and second-year residents had higher educational attainment than those of third-year residents (P = 0.035). The proportion of second-year residents specializing in surgery was significantly lower compared to residents in the other two grades (P < 0. 001). No significant differences were found across the three grades in terms of gender, BMI, household registration status, being an only child, number of close friends, frequency of staying up late, or frequency of night shifts (P > 0.05) (Table S1). Distribution of respondents by region Table 1Comparison of general information between male and female residentsParametersOverall(n = 2,104)Male(n = 765)Female(n = 1339) P Residency training grades (n, %) First year844 (40.1)296 (38.7)548 (40.9)0.452 Second year553 (26.3)209 (27.3)344 (25.7) Third year707 (33.6)260 (34.0)447 (33.4) Age [(M, Q1-Q3), year]26 (24, 27)26 (25, 27)26 (24, 27)0.075BMI (kg/m2) < 18.5242 (11.5)43 (5.6)242 (11.5)<0.001 18.5–24.91570 (74.6)515 (67.3)1570 (74.6) > 24.9292 (13.9)207 (27.1)292 (13.9)Parents’ educational attainment (n, %) College and higher614 (29.2)240 (31.4)374 (27.9)0.178 High school486 (23.1)170 (22.2)316 (23.6) Middle school or less1004 (47.7)355 (46.4)649 (48.5)Household registration type (n, %) Urban933 (44.3)356 (46.5)577 (43.1)0.126 Rural1171 (55.7)409 (53.5)762 (56.9) Local student (n, %)1522 (72.3)547 (71.5)975 (72.8)0.517Marital status (n, %) Married271 (12.9)104 (13.6)167 (12.5)0.460 Unmarried1833 (87.1)661 (86.4)1172 (87.5) Dormitory accommodation (n, %)1046 (49.7)399 (52.2)647 (48.3)0.090 Only child (n, %)699 (33.2)332 (43.4)367 (27.4)<0.001Types of residents (n, %) Graduate-student resident1195 (56.8)445 (58.2)750 (56.0)0.336 Non-graduate-student resident909 (43.2)320 (41.8)589 (44.0)Training specialties Internal medicine1041 (49.5)259 (33.9)782 (58.4)<0.001 Surgery598 (28.4)355 (46.4)243 (18.1) General medicine196 (9.3)65 (8.5)131 (9.8) Medical technology and diagnostics184 (8.8)51 (6.7)133 (9.9) Other fields85 (4.0)35 (4.6)50 (3.7)Number of close friends (n, %) 0106 (5.0)61(8.0)45 (3.4)0.540 1–2813 (38.6)270 (35.3)543 (40.6) ≥ 31185 (56.3)434 (56.7)751 (56.1)Frequency of exercise (n, %) 0 time/week630 (29.9)156 (20.4)474 (35.4)<0.001 1–2 times/week1199 (57.0)459 (60.0)740 (55.3) 3–4 times/week233 (11.1)121 (15.8)112 (8.3) ≥ 5 times/week42 (2.0)29 (3.8)13 (1.0)Sleep quality (n, %) Very poor110 (5.2)58 (7.6)52 (3.9)<0.001 Relatively poor291 (13.8)129 (16.9)162 (12.1) Moderate​953 (45.3)343 (44.7)611 (45.6) Relatively good621 (29.5)190 (24.8)431 (32.2) Very good129 (6.1)46 (6.0)83 (6.2)Frequency of staying up late (n, %) 0 days/week29 (1.4)20 (2.6)9 (0.7)0.889 1–2 days/week410 (19.5)142 (18.6)268 (20.0) 3–4 days/week631 (30.0)221 (28.9)410 (30.6) ≥ 5 days/week1034 (49.1)382 (49.9)652 (48.7)Frequency of night shifts (n, %) 0–1 times/month261 (12.4)87 (11.4)174 (13.0)0.306 2–4 times/month961 (45.7)349 (45.6)612 (45.7) ≥ 5 times/month882 (41.9)329 (43.0)553 (41.3)BMI Body mass index Comparison of general information between male and female residents BMI Body mass index Subjective stress and perceived stress evaluation Acceptable internal consistency was found for the subjective stress scale (Cronbach’s α = 0.867). The analysis of subjective stress scores by gender indicated that males report significantly elevated levels of economic stress [6(5–9) vs. 6(4–8), P = 0.002], marriage stress [5(2–7) vs. 3(1–6), P<0.001], and academic stress [6(4–8) vs. 6(3–8), P = 0.003] in comparison to females (Fig. 2A). Nonetheless, no significant differences were observed for the other six subjective stress scores referenced above (Table S2). Comparison of CPSS scores revealed that the tension dimension score for males was significantly lower than that for females [13(9–16) vs. 13(10–17), P = 0.003] (Fig. 2B). No significant differences were observed in the total score or the loss of control dimension score of the CPSS scale (P > 0.05). The proportion of residents with a CPSS score of 25 or higher exhibited no significant difference between males and females (Fig. 2C and Table S2). Moreover, we found that the third-year residents experience greater economic and marriage-related stress compared to the first-year residents, while exhibiting lower levels of academic stress. Third-year residents experience greater economic pressure compared to second-year residents (Fig. S1A). There were no significant differences among the three grades in terms of interpersonal stress, scientific research stress, clinical work stress, stress related to pursuing higher degrees, and employment stress. Furthermore, no significant differences were observed in the total score or the tension dimension score on the CPSS scale among residents across the three grades. The loss of control dimension score of the CPSS scale for third-year residents was significantly lower than that for first- and second-year residents (Fig. S1B). More than 60% of residents demonstrated considerable psychological stress, indicated by a total CPSS score of 25 or higher, with a median score of 27 (range 22–31). The median scores for sense of tension and sense of loss of control were 13 (range 10–17) and 14 (range 10–17), respectively (Fig. S1C and Table S3). Subsequent analysis revealed that, compared to first-year residents, the percentage of third-year residents who achieved a CPSS score of 25 or higher was significantly lower (Fig. S1C and Table S3). Distribution of respondents by region Psychological health status evaluation We subsequently analyzed the psychological health status of male and female respondents. We examined the distribution of factor scores ≥ 2 across the ten dimensions of the SCL-90 scale, as a score of ≥ 2 signifies a symptom response [4]. The survey revealed that male residents exhibited a lower prevalence of depression compared to female residents (Fig. 3D), while demonstrating higher rates of hostility, phobic anxiety, paranoid ideation, psychoticism, and other factors such as diet and sleep (Fig. 3F-J). Nonetheless, the remaining four dimensions did not exhibit statistical significance between males and females (Fig. 3A-C, E). The percentages of residents exhibiting a total of ≥ 43 positive symptoms, any factor score of ≥ 2, and a total score of ≥ 160 were 31.3%, 48.7%, and 28.8%, respectively (Table S4). A total of 1,043 residents (49.6%) were identified as having psychological health problems, comprising 362 males (47.3%) and 681 females (50.9%) (Fig. 3L and Table S4). Simultaneously, there was no significant difference in the prevalence of psychological health issues among residents across different grades, with 435 (51.5%) in grade 1, 272 (49.2%) in grade 2, and 336 (47.5%) in grade 3, respectively. The levels of the ten dimensions showed no significant differences across the three grades, nor did the proportions of residents with a positive symptom total of ≥ 43, any factor score of ≥ 2, or a total score of ≥ 160 (Table S5). The psychological health status of the respondents between males and females Factors influencing residents’ psychological status Next, we found significant differences in exercise frequency, sleep quality, late-night activities, number of close friends, average number of night shifts, and CPSS scores between residents with and without mental health problems (P < 0.05). Residents without psychological health issues engaged in exercise more regularly, experienced improved sleep quality, reduced late-night activities, maintained a greater number of close friendships, and exhibited lower scores on the CPSS. Additionally, residents without psychological health issues exhibited a greater proportion of CPSS scores below 25 and recorded lower scores on both CPSS dimensions (tension and loss of control) compared to those with psychological health problems (Table 2). Furthermore, the binary logistic regression analysis revealed that staying up late for five or more days per week (OR = 1.300; p = 0.003), working night shifts five or more times per month (OR = 1.304; P = 0.013), CPSS score (OR = 1.179; P < 0.001), relatively poor sleep quality (OR = 1.503; P = 0.013), and inferior sleep quality (OR = 2.720; P = 0.003) are independent risk factors for mental health issues. In contrast, relatively good sleep quality (OR = 0.740; P = 0.012) and excellent sleep quality (OR = 0.471; P = 0.002) were identified as independent protective factors (Table 3). The ROC curve demonstrated an area under the curve of 0.808 (95% CI: 0.790–0.826) for the combination of the four indices in predicting psychological health issues. The area under the curve for the CPSS score was 0.796 (95% CI: 0.777–9.815), whereas the areas under the curve for the other three indices were below 0.5 (Fig. 4). The ROC curve revealed that the optimal cut-off value for CPSS score is 27, exhibiting a specificity of 75.6% and a sensitivity of 67.4%. Table 2Comparison of general information between residents with or without psychological health problemsParametersAbsence of psychological problems(n=1061Existence of psychological problems(n=1043t/χ2/Z P Age (, year)26.14±2.4926.12±2.520.1740.862Male (n, %)403 (38.0)362 (34.7)2.4390.118BMI (,kg/m2)21.77±2.9321.60±2.911.3230.186Residency training grades (n, %) First year409 (38.5)435 (41.7)2.5260.283 Second year281 (26.5)272 (26.1) Third year371 (35.0)336 (32.2) Only child (n, %)350 (33.0)349 (33.5)0.0530.818Marital status (n, %) Married137 (12.9)134 (12.8)0.6450.724 Unmarried with a romantic partner433 (40.8)409 (32.9) Widowed, divorced, single, or other491 (46.3)500 (47.9)Parents' education attainment College and higher306 (28.8)308 (29.5)−0.0220.982 High school251 (23.7)235 (22.5) Middle school or less504 (47.5)500 (47.9)Residence location Urban471 (44.4)462 (44.3)0.0020.964 Rural590 (55.6)581 (55.7)Frequency of exercise (n, %) 0 time/week248 (23.4)382 (36.6)−6.378<0.001 1–2 times/week654 (61.6)545 (52.3) 3–4 times/week134 (12.6)99 (9.5) ≥ 5 times/week25 (2.4)17 (1.6)Sleep quality (n, %) Very poor12 (1.1)98 (9.4)−14.232<0.001 Relatively poor82 (7.7)209 (20.0) Moderate​​464 (43.7)489 (46.9) Relatively good409 (38.4)214 (20.5) Very good96 (9.0)33 (3.2)Frequency of staying up late (n, %) 0 days/week17 (1.6)12 (1.2)−6.806<0.001 1 - 2 days/week252 (23.8)158 (15.1) 3 - 4 days/week345 (32.5)286 (27.4) ≥ 5 days/week447 (42.1)587 (56.3)Number of close friends (n, %) 023 (2.2)83 (8.0)−8.652<0.001 1–2349 (32.9)464 (44.5) ≥3689 (64.9)496 (47.6) Dormitory living (n, %)552 (49.2)524 (50.2)0.2280.633Frequency of night shifts (n, %) 0 - 1 times/month158 (14.9)103 (9.9)−6.847<0.001 2 - 4 times/month535 (50.4)426 (40.8) ≥ 5 times/month368 (34.7)514 (49.3) Tension (M, Q1-Q3)11 (8–14)15 (13–19)−22.288<0.001 Loss of control (M, Q1-Q3)12 (8–14)15 (13–18)−15.597<0.001Total CPSS score (M, Q1-Q3)24 (19–28)30 (27–36)−23.538<0.001Total CPSS score ≥ 25467 (44.0)873 (83.7)385.193<0.001BMI Body mass index, CPSS Chinese Perceived Stress Scale Comparison of general information between residents with or without psychological health problems BMI Body mass index, CPSS Chinese Perceived Stress Scale Table 3Multivariate logistic regression analysis of psychological health risk factorsVariablesβSEWald χ2 P OR (95%CI)Sleep quality* Very poor1.0010.3418.6230.0032.720 (1.395–5.305) Relatively poor0.4070.1646.1480.0131.503 (1.089–2.074) Relatively good ​​−0.3010.1206.2510.0120.740 (0.584–0.937) Very good−0.7530.2399.9470.0020.471 (0.295–0.752)Frequency staying up late # ≥ 5 days/week0.2620.1046.3140.0121.300 (1.059–1.595)Frequency of night shifts$ ≥ 5 times/month0.2660.1076.1930.0131.304 (1.058–1.608) CPSS core0.1640.010292.372<0.0011.179 (1.157–1.201) Constant−4.6050.272286.824<0.001-CPSS Chinese Perceived Stress Scale*Using “Moderate” as the reference#Using “0–4 days/week” as the reference$Using “0–4 times/month” as the reference The model’s goodness-of-fit was assessed via the Hosmer-Lemeshow test (P = 0.600) Multivariate logistic regression analysis of psychological health risk factors CPSS Chinese Perceived Stress Scale *Using “Moderate” as the reference #Using “0–4 days/week” as the reference $Using “0–4 times/month” as the reference The model’s goodness-of-fit was assessed via the Hosmer-Lemeshow test (P = 0.600) ROC curves for predicting psychological health problems and a CPSS score ≥ 27 Factors influencing perceived stress Considering the predictive significance of the CPSS score for psychological health issues, we conducted a further analysis of the risk factors associated with a CPSS score of 27 or higher. The prevalence of a CPSS score of 27 or higher among all respondents was 53.9%. A CPSS score of 27 or higher, in contrast to a score of less than 27, is associated with an increased frequency of staying up late, a greater occurrence of night shifts, and elevated scores across the eight dimensions of subjective stress (P<0.05). It is correlated with a younger age, reduced exercise frequency, poorer sleep quality, and a smaller number of close friends (P < 0.05) (Table S6). Binary logistic regression analysis revealed that relatively poor sleep quality (OR=1.902; P<0.001), inferior sleep quality (OR༝4.214; P༝0.001), having 0 (OR༝3.969; P<0.001) or 1–2 close friends (OR༝1.659; P<0.001), exercising less than once per week (OR༝1.654; P<0.001), interpersonal stress (OR༝1.161; P<0.001), academic stress (OR༝1.067; P༝0.026), clinical work stress (OR༝1.149; P<0.001), and pursuing higher degrees stress score (OR༝1.050; P༝0.013) were identified as independent risk factors for a CPSS score ≥ 27. In contrast, relatively good sleep quality (OR༝0.597; P<0.001) and inferior sleep quality (OR༝0.590; P༝0.019) were identified as independent protective factors (Table S7). The ROC curve demonstrated that the area under the curve for the combination of the eight indices in predicting a CPSS score ≥ 27 was 0.791 (95% CI: 0.772–0.810). The areas under the curve exceeding 0.7 comprised interpersonal stress (β = 0.149, P < 0.001) and clinical work stress (β = 0.139, P < 0.001) (Fig. 4 and Table S7). Discussion This study examined psychological health issues and perceived stress among Chinese residents who received standardized training. Nearly half of the residents had psychological health issues, regardless of gender or training duration. This suggests a systematic problem with residency training. After further study, sleep quality, late-night stays, night shifts, and the CPSS score were found to be substantially connected with psychological health issues. Preliminarily, with an optimal cut-off value of 27, the CPSS score accurately predicted psychological health issues. Interpersonal and clinical work stress may be severe for persons with a CPSS score of 27 or higher. Since the nationwide launch of the standardized training program for resident physicians in China in 2014 [18], there has been increasing attention to the psychological health issues faced by these physicians [19, 20]. A previous single-center study in China reported that 34.5% of resident physicians scored at or above the conventional clinical threshold (SCL-90 total score ≥ 160) for potential psychological health issues [21]. Our multicenter study used this same cutoff point and found a slightly lower prevalence (28.8%). However, our study discovered that almost 50% of the residents had psychological health issues by combining the SCL-90’s total score, factor scores, and number of positive items. This apparent difference highlights a critical methodological distinction: relying solely on the single SCL-90 total score cutoff provides a limited perspective. The prevalence of psychological health issues disclosed by our study is consistent with or exceeds the rates observed globally in high-pressure medical training environments. A meta-analysis covering all five continents showed that the prevalence of depression or depressive symptoms among medical residents was 28.8% (25.3%−32.5%) [22]. According to some other studies, 36%–43% of resident physicians had depression, 20%–37% had anxiety, and 15%–27% had even mentioned suicidal or death-related thoughts [10, 20, 23, 24]. Therefore, investigating the pertinent risk factors is of utmost importance, and the psychological well-being of resident physicians is a crucial global public health priority. Our findings showed that staying up late ≥ 5 days a week, working night shifts ≥ 5 times a month, and having poor sleep quality were more likely to​ report psychological issues. This ​underscores​ the connection between sleep and mental health among resident physicians. A meta-analysis of 65 studies revealed that sleep improvement significantly affects positive symptoms of psychosis, depression, anxiety, and composite mental health. Moreover, greater improvement in sleep quality leads to a corresponding improvement in mental health [25]. Mechanistically, sleep deprivation causes sustained hyperactivity of the hypothalamic-pituitary-adrenal axis, resulting in elevated cortisol levels. Prolonged elevated cortisol levels adversely affect the plasticity of hippocampus neurons and hinder emotional control [26]. Moreover, insufficient sleep has demonstrated significant genetic correlation with multiple mental health traits. It is linked to the deterioration of brain areas involved in emotional regulation, thereby heightening the risk of mental disorders [27, 28]. Furthermore, exposure to ambient light at night (such as during prolonged wakefulness or nocturnal work shifts) suppresses melatonin release by the pineal gland. It elicits anxiety-like responses [29]. Moreover, inadequate sleep can exacerbate the detrimental effects of unpleasant life events and reduce the beneficial impact of positive occurrences, rendering individuals more susceptible to negative emotions [12]. Consistently staying awake late or working nocturnal shifts might result in reduced sleep duration, hence intensifying insomnia. This behavior not only heightens the likelihood of psychological issues but also elevates the chance of medical errors [30, 31]. To maintain mental health, resident physicians should avoid staying up late and ensure an appropriate frequency of night shifts. According to our findings, night shifts < 5 times per month might be relevant. Moderate stress can enhance work efficiency, strengthen learning ability, and promote personal growth, while excessive stress can disturb both physical and mental health. In our study, 63.7% of the resident physicians experienced excessive perceived stress (CPSS score ≥ 25). Similarly, 50% of residents at a teaching hospital in Nigeria reported that their lives were very stressful [32]. Although stress is one of the most critical factors that can impact mental health, the level of stress that can cause such adverse effects in medical residents remains unclear. According to previous studies in China, a CPSS score of ≥ 25 is associated with weakened psychological resilience, reduced ability to cope with stress and setbacks, and a negative impact on emotions. Therefore, a value of 25 is regarded as the critical threshold for health risks associated with perceived stress [33, 34]. However, the present study found that the optimal cutoff value of the CPSS score associated with psychological health problems in residents receiving standardized training is 27. This difference reflects the population specificity of the perceived stress threshold. The relatively high educational attainment of resident physicians and their long-term adaptation to high-intensity, high-pressure environments may explain why their CPSS threshold is higher than that of the general population. Conducting regular perceived stress questionnaires for resident physicians may help identify individuals under clinically high stress at an early stage, thereby avoiding adverse effects on their mental health. As indicated by previous studies, the stressors for resident physicians include heavy workload, sleep deprivation, complex patients, poor learning environments, and relocation issues, as well as isolation and social problems, financial concerns, cultural and minority issues, information overload, and career planning issues [13]. Our study found that interpersonal stress and clinical work stress had the most significant association with CPSS scores of 27 or higher. The White Paper on the Practice Status of Physicians in China indicated that among physicians who work an average of more than 60 h per week, those aged between 25 and 35 account for 57.27%. This working hour significantly exceeds the international warning line (56 h per week) [35]. Our results once again emphasized the negative impact of clinic workload on the perceived stress of residents. Regulating the workload may play a crucial role in improving the psychological well-being of resident physicians in China. Meanwhile, interpersonal communication is also an essential factor in stress regulation for resident physicians. In China, the doctor-patient relationship has been tense for years [36], and resident physicians who directly face patients are bound to bear more interpersonal pressure from patients. Heavy clinical workloads also pose a significant obstacle to effective interpersonal communication for them. The other interpersonal stress experienced by resident physicians may arise from conflicts with colleagues and supervisors, as well as estrangement from their own families and friends due to demanding work schedules [11]. Altogether, establishing training programs to enhance positive interpersonal interaction may help alleviate the stress load of resident physicians. Previous studies have shown that females experience more stress and have a higher risk of psychological problems compared to males [37]. In our study, although no gender differences were found in the total CPSS score or the overall rate of mental health problems (as defined by the SCL-90), males had lower scores on the tension dimension of the CPSS and a lower positive rate for the depression factor of the SCL-90. However, males experienced higher subjective stress scores in economic stress, marriage stress, and stress related to pursuing higher degrees. Consistently, higher levels of education are associated with improved mental health [38]. In recent years, the educational opportunities and social status of Chinese women have undergone significant improvements. The respondents in our study, both male and female, all hold a bachelor’s degree or higher, which has helped narrow the differences in mental health issues associated with educational levels. Furthermore, the proportion of married people among the included population is very low. Therefore, the gender differences resulting from family contributions and stressors such as pregnancy and lactation were minimal. Additionally, within the context of Chinese culture, men are often expected to support their families and fulfill both family and social responsibilities [39, 40]. In the present study, economic stress, marriage stress, and the pursuit of higher degrees were significantly greater among men than among women, further confirming this view. Since these stress loads of the residents were at a medium level, they might not yet be sufficient to significantly impact an individual’s overall perceived stress and mental health. Several limitations may be associated with this study. First, it is impossible to conclude causality from the observed associations due to the cross-sectional design, and the hypotheses generated from these cross-sectional findings necessitate future longitudinal research for validation. Second, the use of convenience sampling raises the possibility of selection bias, which may impact the representativeness of the sample and the generalizability of our findings. Third, self-report questionnaires, which are susceptible to social desirability effects, recall bias, and measurement error, were the primary tool used to measure stress and mental health. While these tools are necessary to assess subjective experiences, future studies would benefit from incorporating objective measures. Fourth, the subjective stress questionnaire used in this study to determine specific stress domains was newly developed for this project based on qualitative interviews with residents. Although it demonstrated good internal consistency in our sample, it has not undergone comprehensive validation, including tests for convergent and discriminant validity. Therefore, findings related to these specific stress domains should be interpreted with caution. Future studies are needed to validate this instrument fully. Fifth, several potential confounders, such as socioeconomic status, social support, mental health resources, pre-existing mental health vulnerabilities, etc., were not measured in this study. Although we controlled for key demographic and work-related variables, the influence of these unmeasured factors cannot be ruled out, and they may have contributed to the observed associations. Our exclusion of residents with organic mental disorders may underestimate the prevalence of psychological distress and limit the generalizability of results. Furthermore, the prevalence of psychological health issues reported in this study was based on a comprehensive SCL-90 criterion, which may contribute to the wide variation in estimates across the literature. This choice, while aimed at capturing a broader spectrum of significant distress, affects the direct comparability of our prevalence rates with studies using different thresholds. In conclusion, our results show that Chinese medical residents receiving standardized training have a remarkably high rate of psychological health issues. Stress and sleep disruptions significantly correlated with various mental health issues. CPSS demonstrated the strongest correlation, with interpersonal and clinical work stress being major contributors. These findings indicate that prioritizing the physical health of Chinese medical residents and reforming the residency training system should be considered. Supplementary Information Supplementary Material 1. Acknowledgements The authors would like to thank all the students who participated in the study, as well as the Questionnaire Star platform. Abbreviations Chinese perceived stress scale Symptom checklist-90 Body mass index Odds ratio Confidence interval Area under the curve Receiver operating characteristic curve Authors’ contributions BW, JCX and WJL contributed to concept and design of the study. All authors were involved in data collection, analysis and interpretation. MW, MYL and WJL wrote the original draft. JCX and WJL involved in revision of the paper. All authors have read and approved the submission. Funding This work was supported by the Graduate Education and Teaching Reform Project of Chengdu Medical College [YJG202428] and the Youth Doctoral Program of the Second Affiliated Hospital of Army Medical University. Figures were created with R-4.4.2-win, Microsoft Office PowerPoint 2019, and Prism 9. Data availability The datasets used and analyzed during this study are available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate This study, which was conducted in accordance with the Declaration of Helsinki, was approved by the Biomedical Ethics Review Committee of West China Hospital, Sichuan University, China (No.1442, 2024). Informed consent was obtained from all subjects involved in the study. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Min Wang and Mengyuan Liu contributed equally to this work. Contributor Information Jiachuan Xiong, Email: xiongjc@tmmu.edu.cn. Weijing Lai, Email: yylaiweijing@163.com. References Associated Data Supplementary Materials Supplementary Material 1. Data Availability Statement The datasets used and analyzed during this study are available from the corresponding author on reasonable request.