A convenience sampling method was adopted to recruit 133 standardized residency trainees who received standardized training at Nanjing First Hospital between January 1 and December 31, 2024, among them, 59 (44.36%) were male and 74 (55.64%) were female. The participants’ ages ranged from 23 to 42 years, with a median of 23 years. Participation was entirely voluntary, and informed consent was obtained from all respondents prior to data collection. All procedures performed in this study involving human participants were in accordance with the Declaration of Helsinki. Ethical approval for this study was granted by the Ethics Committee of Nanjing First Hospital (Ethical Approval No.: KY20250728-KS-02).

To ensure data integrity and reliability, all questionnaire items were set as mandatory and collected anonymously. After each questionnaire was completed, the research assistants immediately reviewed it for completeness and logical consistency, focusing on identifying any missing responses, duplicate selections, or clearly contradictory answers. If any issues were detected, the respondents were promptly contacted to provide clarification or corrections. Only questionnaires verified as accurate and complete were collected and subsequently coded and archived. A dual-entry system was used by two independent researchers to enter and verify the data. A total of 133 questionnaires were distributed and collected, all of which were deemed valid, resulting in a 100% valid response rate.

All statistical analyses were performed using SPSS version 26.0 (IBM Corp., Armonk, NY, USA). The Harman single-factor test was used to examine common method bias, with all items from the CD-RISC, CES-D, SAS, and BSI-CV scales in the 133 questionnaires being treated as variables for factor analysis. Since the continuous variables failed to meet normality assumptions as assessed by the Kolmogorov-Smirnov test, they were expressed as medians and interquartile ranges (IQR; P25, P75). Group comparisons were conducted using the Mann-Whitney U test (for two groups) or the Kruskal-Wallis H test (for multiple groups). Spearman’s rank correlation was employed to examine the associations among study variables. Hierarchical regression analysis was conducted to assess the moderating effects of resilience dimensions on the relationships between depression, anxiety, and suicidal ideation. Analytical Steps for Tenacity and Strength: Model 1 (Baseline): Suicidal ideation was regressed on depression/anxiety to examine their direct effect. Model 2 (Main Effects): The moderator (Tenacity or Strength, respectively) was added to Model 1 to assess its main effect on suicidal ideation after controlling for depression/anxiety. Model 3 (Moderation): The interaction term between depression/anxiety and the corresponding moderator (depression/anxiety × Tenacity or depression/anxiety × Strength) was incorporated into Model 2 to test for the moderating effect. Analytical Steps for Optimism: Model 1 (Covariate Model): Monthly household income per capita was entered first as a covariate into the regression equation.

Model 2 (Baseline Model): Depression/Anxiety was then added to Model 1 to examine its direct effect on suicidal ideation after controlling for income. Model 3 (Main Effects Model): The moderator, Optimism, was incorporated into Model 2 to assess its main effect on suicidal ideation, above and beyond the effects of income and depression/anxiety. Model 4 (Moderation Model): Finally, the interaction term between depression/anxiety and Optimism was added to Model 3 to test for the moderating effect. A p-value of less than 0.05 was considered statistically significant.

Harman’s single-factor test uses exploratory factor analysis (EFA) to determine whether there is significant common method bias in the data. When the variance explained by the first factor is less than 40%, it is generally considered that there is no serious common method bias and the measurement results are relatively reliable. However, if the variance explained exceeds 40%, it indicates that a single factor accounts for too much explanatory power, suggesting the possible presence of substantial common method bias that needs to be further controlled or corrected. Harman’s single-factor test revealed 12 factors with eigenvalues greater than 1. The first factor explained 36.52% of variance, far below the 40% threshold, indicating no significant common method bias in the data.

The median (IQR) scores among all participants were as follows: Depression: 25.0 (20.0, 31.0). Anxiety: 41.0 (36.0, 47.0). Suicidal ideation: 4.0 (4.0, 5.0). For resilience dimensions: Tenacity: 27.0 (22.0, 33.0), Strength: 18.0 (16.0, 23.0), Optimism: 9.0 (8.0, 10.0).

When scores for depression, anxiety, suicidal ideation, and resilience (tenacity and strength) were compared across different demographic groups, no statistically significant differences were observed (p > 0.05). However, a significant difference was found in resilience (optimism) scores across groups with different monthly household income per capita (p = 0.020). To control for potential confounding, household income was included as a covariate in the subsequent hierarchical regression analysis (Table 1).

Significant positive correlations were observed between both depression and anxiety with suicidal ideation (r = 0.466 and r = 0.395, respectively; all p < 0.001. Conversely, all three resilience dimensions, including tenacity, strength, and optimism, were found to be negatively correlated with suicidal ideation (r = −0.504, −0.477, and −0.440, respectively; all p < 0.001) (see Table 2).

When tenacity and strength were entered as moderators:

In Model 1, depression was significantly and independently associated with suicidal ideation (β = 0.394, t = 4.905, p < 0.001). In Model 3, after including interaction terms, the explained variance increased to 35.40% and 32.70%, respectively, with additional ΔR² values of 3.10% and 3.40%. Significant negative moderating effects were found for both dimensions (β = –0.196, t = –2.497, p = 0.014; β = –0.204, t = –2.570, p = 0.011) (Table 3 and Table 4).

When optimism was examined as a moderator, and household income was controlled:

Depression remained a significant and independently associated with suicidal ideation (β = 0.394, t = 4.976, p < 0.001). In Model 3, after including the interaction term, the variance explained reached 30.70%, with a ΔR² of 0.035. A significant negative moderating effect of optimism was observed (β = –0.211, t = –2.596, p = 0.011) (Table 5).

The simple slope analysis demonstrated that when different dimensions of resilience were at a low level (Mean − 1 SD [standard deviation]), the slopes were noticeably steeper, indicating that the effect of depression on suicidal ideation was stronger. In contrast, when different dimensions of resilience were at a high level (Mean + 1 SD), the slopes were noticeably flatter, suggesting that the effect of depression on suicidal ideation was weaker (Fig. 2).

When tenacity and strength were entered as moderators:

In Model 1, anxiety was significantly and independently associated with suicidal ideation (β = 0.318, t = 3.844, p < 0.001). In Model 3, after including the interaction terms of anxiety with resilience (tenacity and strength), the model explained 30.50% and 18.355% of the variance in suicidal ideation, respectively. The change in explained variance (ΔR²) was 0.0210 and 0.0270, respectively. The interaction between anxiety and tenacity was not statistically significant (β = –0.162, t = –1.982, p = 0.050), indicating that tenacity was not a significant moderator of the association between anxiety and suicidal ideation. In contrast, the interaction between anxiety and strength was significant (β = –0.178, t = –2.219, p = 0.028), indicating a negative moderating association (Table 6 and Table 7).

When optimism was examined as a moderator, and household income was controlled:

In Model 1, anxiety remained significantly and independently associated with suicidal ideation (β = 0.346, t = 4.23, p < 0.001). After adding the interaction term between anxiety and optimism in Model 3, the model explained 29.3% of the variance in suicidal ideation, with an increase of ΔR² = 2.70%. Optimism significantly moderated the association between anxiety and suicidal ideation in a negative direction (β = –0.188, t = –2.226, p = 0.028) (Table 8).

The simple slope analysis (Fig. 3) showed that when resilience (strength and optimism) was at a low level (Mean − 1 SD), the slopes were steeper, indicating a stronger positive association between anxiety and suicidal ideation. Conversely, when resilience (strength and optimism) was at a high level (Mean + 1 SD), the slopes were flatter, indicating a weaker association between anxiety and suicidal ideation.

Multiple factors have been identified as influencing the occurrence of suicidal ideation in physicians, including demographic characteristics, psychological vulnerabilities, and social stressors. Among these, psychological factors, particularly depression and anxiety, have consistently been linked to elevated suicide risk in medical professionals [7,8]. In the present study, a questionnaire-based survey was conducted to explore how depressive and anxiety symptoms are associated with suicidal ideation among standardized residency trainees. The findings revealed significant associations between both depression and anxiety and suicidal ideation. As standardized residency trainees serve as a crucial reserve workforce in China’s healthcare system, a deeper understanding of the potential pathways linking depression and anxiety with suicidal ideation is of significant importance for improving healthcare quality and building a stronger medical workforce in China.

Individual resilience can be conceptualized along three dimensions: tenacity, strength, and optimism. Tenacity refers to an individual’s ability to remain calm, persistent, and in control in the face of adversity; strength involves the capacity to recover and grow from setbacks or traumatic experiences; optimism reflects a tendency to interpret experiences positively and to trust in both personal and external resources [21,22,23]. This study specifically examined how these three dimensions of resilience moderate the relationships between depression, anxiety, and suicidal ideation. All three dimensions were significantly associated with lower levels of suicidal ideation, particularly in attenuating the positive association between depressive symptoms and suicidal ideation. In other words, higher resilience scores were associated with lower reported levels of suicidal intent, regardless of depressive severity. These findings were consistent with our research hypothesis that psychological distress would be positively associated with suicidal ideation and that resilience—particularly strength and optimism—would buffer these associations. These results were consistent with our research hypothesis. In the present study, tenacity was not found to moderate the relationship between anxiety and suicidal ideation among standardized residency trainees, whereas strength and optimism were associated with weaker positive associations between anxiety and suicidal ideation. Individuals with higher strength and optimism tended to report lower levels of suicidal ideation, even when anxiety symptoms were elevated. Drawing on the buffering hypothesis of the psychological resilience framework [24,25] and the cognitive-behavioral theoretical framework [26,27], several possible interpretations of these associations are proposed.

First, Depression is typically characterized by motivational deficits [28,29]. Individuals with high tenacity may show greater capacity to maintain functioning under depressive states [30]. maintaining basic functioning and preventing total psychological collapse. Those with higher strength may interpret depressive experiences more adaptively, viewing them as opportunities for growth rather than signs of personal failure. Individuals with higher optimism may be more inclined to engage in problem-solving and help-seeking behaviors, which could be associated with lower levels of suicidal ideation [31]. Second, anxiety is characterized by a perceived lack of control and hypervigilance [32,33]. Tenacity, which emphasizes persistence and tolerance, may be less strongly associated with reduced anxiety-related distress. In contrast, individuals with high levels of strength and optimism may perceive anxiety-inducing situations as more manageable and tend to adopt more adaptive emotion-regulation strategies.

This study revealed a significant association between family monthly income per capita and the optimism dimension of resilience among trainees, suggesting that socioeconomic status may be associated with variations in this psychological characteristic. To account for this potential confounder, subsequent analyses controlled for family income. The results showed that even after controlling for income, both depressive and anxiety symptoms remained significantly associated with suicidal ideation. Moreover, optimism continued to exhibit a significant moderating association, indicating that its moderating role remained stable. This suggests that higher optimism may be associated with more positive cognition and a greater sense of hope despite financial pressure, which could relate to lower levels of suicidal ideation.

The differential moderating roles of resilience dimensions highlight the need to move beyond one-size-fits-all approaches. Future studies and training programs may consider developing tailored resilience-building strategies aligned with the specific emotional challenges faced by standardized residency trainees.

While previous research has consistently documented the associative role of psychological resilience in mental health—such as its moderating role in the relationships between stress and anxiety [34] and between adversity and depression [35,36]—the present study extends this body of evidence in two important ways. First, we identified a similar moderating association linking anxiety and depression with suicidal ideation. Second, by moving beyond a unitary construct, this study provided a more nuanced understanding through the distinct moderating roles of the tenacity, strength, and optimism dimensions of resilience. This refined approach offers a more comprehensive depiction of how resilience is associated with emotional distress, thereby enriching the existing literature.

The strengths of this study lie in its pioneering examination of the moderating roles of resilience—across its distinct dimensions of tenacity, strength, and optimism—in the associations between depression, anxiety, and suicidal ideation among medical residents. We applied hierarchical regression analysis with simple slope tests to verify the significance and stability of these moderating associations, thereby enhancing the reliability and validity of our findings. The results provide preliminary theoretical insights that may inform the development of targeted psychological strategies for this population and contribute to a better understanding of the multidimensional associations between resilience and mental health.

However, this study has several limitations. First, the cross-sectional design precludes the determination of causal relationships among the variables. Second, the sample size was limited and primarily drawn from a specific regional cohort of standardized residency trainees, which may restrict the generalizability of the findings. Third, the reliance on self-reported measures introduces the potential for response biases, such as social desirability. Fourth, given resource and feasibility constraints (budget, personnel, and time), this study focuses on individual-level psychological characteristics. Environmental and contextual factors typically require multi-source or longitudinal measurement and larger samples, which exceed the resources available at this stage; accordingly, they were not included in the analyses.

Based on the aforementioned limitations, future research could be explored in the following directions: (1) Optimization of research design: Longitudinal cohort studies or diary sampling methods are recommended to clarify causal relationships and temporal dynamics between variables. (2) Expansion of sample scope: Multi-center, large-scale collaborative surveys should be conducted, covering standardized residency trainees from different regions, hospital tiers, and specialties to enhance the representativeness and generalizability of the findings. (3) Integration of multi-level variables: Beyond focusing on individual psychological traits, future studies should systematically investigate the independent and interactive effects of organizational environmental factors and socio-contextual factors within the pathway linking depression, anxiety, and suicidal ideation.