Introduction
The crisis caused by COVID-19 has tested health systems in the vast majority of countries worldwide.Reference Almaguer, Alvarez and Santos1 The pandemic has had a significant impact on both the physical and emotional health of the population, largely due to the virus itself and the social isolation measures imposed by authorities.Reference Pollán, Pérez-Gómez and Pastor-Barriuso2
During this health crisis, university healthcare students have also been considered a vulnerable population, as their mental health has been seriously affected by the pandemic, online learning, and the confinement measures implemented by governments. These factors have severely impacted students’ educational and socio-emotional development.Reference Cobo-Rendón, Vega-Valenzuela and García-Álvarez3 Moreover, substantial adjustments were made to the original training plans and learning contexts. These changes were implemented rapidly, generating a high level of uncertainty due to a lack of information and widespread fear of contagion.Reference Sánchez-Teruel, Robles-Bello and Valencia-Naranjo4 Reports indicate that the prevalence of anxiety, depression, and stress among Spanish university students was 21.34%, 34.19%, and 28.14%, respectively. These figures reflect reported symptoms rather than formal clinical diagnoses, and occurred during a period when major modifications were made to academic programs and clinical training.Reference Odriozola-González, Planchuelo-Gómez and Irurtia5 Resilience plays a crucial role in shaping adaptive responses to stress caused by the pandemic.Reference Sánchez-Teruel, Robles-Bello and Lara-Cabrera6, Reference Sánchez-Teruel and Robles-Bello7 For nursing students—who may be exposed to clinical environments—these adaptive responses are particularly relevant.
Currently, mental health professionals and researchers are striving to understand the psychological impact of the COVID-19 pandemic, as well as the factors that may mitigate its effects on mental well-being. However, knowledge gaps remain regarding how nursing students responded to this adverse shift in their learning environment.Reference Mendoza Bernal, Sánchez-Teruel and Robles-Bello8 This study explores the potential effects of the COVID-19 pandemic on nursing students’ mental health (specifically, anxiety and depression) in relation to resilience, across two distinct time points during the pandemic, in order to gain a deeper understanding of the global phenomenon and its specific characteristics in Spain. The study follows a research design involving 2 time points: March 2020 and October 2020.
The aim of this study is to evaluate the level of resilience in Spanish nursing students at two different points in time, in order to determine whether differences exist between an initial stage of the pandemic (March 2020) and a later stage several months afterward (October 2020). This allows for intra-group comparisons and the identification of potential changes. In addition, the study seeks to analyze which psychosocial variables are most predictive of high resilience. To achieve this, the following specific objectives are established: first, to compare the psychological variables—depression, anxiety, and resilience—among nursing students during 1 (mandatory confinement, first wave) and the same group during 2 (the “new normal” or third wave). Second, to examine the relationships between protective variables (self-efficacy, dispositional optimism, and emotional intelligence) and risk variables (depression and anxiety), in relation to resilience.
Method
Participants
The sample consisted of 361 nursing students (46% men, 54% women) with a mean age of 20 years (SD = 3.90), who were assessed during the first and second waves of the COVID-19 pandemic. Inclusion criteria were: being a nursing student, being enrolled in a university degree program, and providing informed consent (Table 1).
Table 1. Sociodemographic data of nursing students
First Wave = March 2020; Second Wave = October 2020; η2 = eta square; ns = not significant; Power = power of contrast; t = Student-T; * = P <.05; ** = P <.01.
The same students were assessed during both the first and second waves of the pandemic, with all participants completing the questionnaire except for 30. In March 2020, they were in their second year of nursing studies, and by October 2020, they had progressed to the third year.
Instruments
Sociodemographic datasheet: sex, age, locality, and whether the participant was living alone at the time.
Wong and Law’s Emotional Intelligence Scale (WLEIS)Reference Wong and Law9 consists of 16 items and is based on the definition of emotional intelligence proposed by Mayer and Salovey.Reference Mayer, Salovey, Caruso, Sternberg and Kaufman10 It includes 4 dimensions: a) Perception of one’s own emotions; b) Perception of others’ emotions; c) Use of emotions; and d) Emotion regulation. Each subscale contains 4 items. The response format is a 7-point Likert-type scale (1 = completely disagree, 7 = completely agree), with higher scores indicating greater emotional intelligence. The original study reported internal consistency indices ranging from 0.83 to 0.90. Participants completed the Spanish version of the scale, which has demonstrated good validity and reliability in Spanish populations.Reference Extremera Pacheco, Rey and Sánchez-Álvarez11 In this study, Cronbach’s alpha for the student sample was 0.89.
Connor-Davidson Resilience Scale (CD-RISC)Reference Connor and Davidson12 has been translated into Spanish.Reference Notario-Pacheco, Solera-Martínez and Serrano-Parra13 The abbreviated version of the Connor-Davidson Resilience Scale (CD-RISC-10) consists of 10 items (numbers 1, 4, 6, 7, 8, 11, 14, 16, 17, 19) from the original scale, focusing exclusively on the ability to adapt in the face of adversity—considered the core component of resilience. This includes the capacity to cope with challenges such as change, personal problems, illness, pressure, failure, and emotional distress. Each item is rated on a 5-point Likert-type scale from 0 (strongly disagree) to 4 (strongly agree), with a total score ranging from 0 to 40. The scale was validated by the original authors in both general and clinical populations, showing strong reliability indices: Cronbach’s alpha = .89 and test-retest reliability = .87. In this study, Cronbach’s alpha for students was .86.
Life Orientation Test-Revised (LOT-R)Reference Scheier, Carver and Bridges14 is used to assess dispositional optimism. It is a shortened, revised version of the Life Orientation Test.Reference Scheier and Carver15 We used the Spanish adaptation of the LOT-R.Reference Ferrando, Chico and Tous16 This instrument includes 10 items: 3 assessing optimism (items 1, 4, and 10), 3 assessing pessimism (items 3, 7, and 9), and 4 filler items (2, 5, 6, and 8) that are not scored. Respondents rate their agreement on a 5-point Likert-type scale from 0 (strongly disagree) to 4 (strongly agree). Higher total scores indicate greater dispositional optimism. The LOT-R shows good internal consistency; Cronbach’s alpha for the Spanish version was .70 for optimism and .69 for pessimism, with test-retest correlations ranging from .68 to .79. In this study, Cronbach’s alpha for students was .75.
General Self-Efficacy Scale,Reference Schwarzer, Jerusalem, Weinman, Wright and Johnston17 which was translated into Spanish,Reference Suárez, García and Moreno18 assesses individuals’ beliefs in their ability to manage various life situations. It is positively correlated with self-esteem, optimism, and job satisfaction and negatively correlated with anxiety, depression, and physical symptoms. The scale consists of 10 items rated on a 4-point Likert-type scale, where 1 = “not true” and 4 = “completely true." The total score ranges from 10 to 40, with no established cut-off points; higher scores indicate greater perceived self-efficacy. The original version demonstrated good internal consistency (Cronbach’s alpha between .76 and .90). The Spanish version showed an internal consistency of .84.Reference Suárez, García and Moreno18 In this study, Cronbach’s alpha for students was .90.
Hospital Anxiety and Depression Scale (HADS) by Zigmond and SnaithReference Zigmond and Snaith19–Reference Herrero, Blanch and Peri21 includes 14 items divided into 2 subscales: anxiety (HADS-A) and depression (HADS-D), each containing 7 items. These subscales assess the frequency and intensity of anxiety and depression symptoms using a 4-option Likert-type scale, with varying response formats. Respondents are asked to reflect on their experiences over the past few weeks. Each subscale yields a total score between 0 and 21: a score of 0 to 7 indicates no anxiety/depression, 8 to 10 suggests possible clinical concerns, and scores above 10 indicate probable clinical problems. The scale has demonstrated strong reliability, internal consistency, and validity in both clinicalReference Tejero, Guimerá, Farré and Peri20 and nonclinicalReference Terol, López-Roig and Rodríguez-Marín22 Spanish populations. Cronbach’s alpha for both subscales is consistently acceptable, generally above .70 and in most studies exceeding .80.Reference Terol-Cantero, Cabrera-Perona and Martín-Aragón23 In this study, Cronbach’s alpha for students was .70.
Research Design and Procedure
Data collection was conducted in 2 phases: the first between March and May, and the second between September and October 2020. During the first data collection period, a self-administered questionnaire was distributed to participants via email. The purpose, methodology, and instructions were clearly explained at the beginning of the Google Form.
At that time, the entire Spanish population was under strict lockdown due to the declaration of a state of alarm—the least severe of the 3 exceptional legal states in Spain (state of alarm, state of emergency, and state of siege).24 This measure, intended to manage the health crisis caused by COVID-19, was declared in mid-March and remained in effect until the end of June.
During the second data collection period, the population had returned to a phase known as the new normal. However, by late October, a new state of alarm was declared, albeit with less restrictive measures. These included a nationwide curfew, a ban on inter-regional travel between autonomous communities (territorial entities made up of one or more provinces),25 and limits on gatherings of non-cohabiting individuals.
For data collection, an online survey was used, which included an informed consent section. Participants were informed that their data would remain confidential and would be processed in accordance with EU Regulation 2016/679 of the European Parliament and of the Council (April 27, 2016), as well as Organic Law 3/2018 of December 5 on the Protection of Personal Data and Guarantee of Digital Rights.
The study was approved by the Ethics Committee of the University of Jaén (Spain) (Code: ABR.20/4.PRY) and adhered to the principles outlined in the Declaration of Helsinki.
Data Analysis
First, preliminary analyses were conducted to calculate descriptive statistics and internal consistency, as well as to compare the distributions of three variables—depression, anxiety, and resilience—across 2 time points (March and October). Given that the assumption of normality was not met, the nonparametric Mann-Whitney U test was used. To examine the relationships between protective variables (self-efficacy, dispositional optimism, and emotional intelligence) and risk variables (depression and anxiety) with resilience at both time points, Spearman’s correlation coefficient was employed. This also allowed for identifying which subdimension of emotional intelligence was most strongly associated with resilience. Finally, hierarchical multiple regression analyses were conducted using the protective and risk variables at both time points to determine which factors best predicted resilience in the sample. All analyses were performed using IBM SPSS Statistics, version 28.0.
Results
Descriptive Analysis
The descriptive results of the sample indicate that resilience at time point 1 was slightly higher than at time point 2. Conversely, anxiety and depression levels were higher at time point 2 compared to time point 1 (Table 2).
Table 2. Descriptive sample of nursing students
As (ET) = Asymmetric (standard error); C (ET) = Curtosis (typical error); K–S = Kolmogorov–Smirnov; M = median; SD = standard deviation; S–W = Shapiro–Wilk.
The values obtained from the Mann-Whitney U test for the three variables in the study sample showed no significant differences in resilience (z = –1.173, P = .241) or depression (z = –0.583, P = .560) between the 2 time points. However, anxiety did show a significant difference (z = –2.079, P = .038), with higher levels observed at time point 2 (Table 2).
Relationships between Variables
As shown by Spearman’s correlation coefficient at time point 1 (Table 3) and time point 2 (Table 3), there is a significant positive correlation between resilience and dispositional optimism, self-efficacy, and emotional intelligence, as well as a significant negative correlation with depression and anxiety. At time point 1, the strongest positive correlation is observed between self-efficacy and resilience, while the strongest negative correlation is with depression. At time point 2, self-efficacy again shows the strongest positive correlation with resilience, whereas the strongest negative correlation is with anxiety.
Table 3. Correlation of psychosocial variables of nursing students at time points 1 and 2
CD-RISC = Resilience; GSES = self-efficacy; HAD-A = anxiety subdimension; HAD-D = depression subdimension; LOT-R = dispositional optimism; WLEIS = Wong and Law’s emotional intelligence scale.
* The correlation is significant at the .05 level (bilateral).
** The correlation is significant at the .01 level (bilateral).
To determine the relationship between the subdimensions of the intelligence scale and resilience, Pearson’s correlation coefficient indicates a positive and significant relationship. Among the subdimensions, “emotion regulation” shows the strongest correlation with resilience (Table 4).
Table 4. Correlation of intelligence scale subdimensions with resilience in nursing students
CD-RISC = Resilience; OEA = evaluation of others’ emotions; ROE = regulation of emotions; SEA = evaluation of one’s own emotions; UOE = use of emotions.
* The correlation is significant at the .05 level (bilateral).
** The correlation is significant at the .01 level (bilateral).
Predictors of Resilience
To identify the variables that predicted higher levels of resilience at different time points during the pandemic, a hierarchical multiple regression analysis was conducted. Prior to the analysis, assumptions for the application of this statistical test were verified. Residuals were normally distributed, and the Kolmogorov–Smirnov (K–S) test yielded a non-significant value in the study sample (K–S = .052). Therefore, it can be assumed that the residuals of the dependent variable followed a normal distribution after prediction. Additionally, no autocorrelation was detected among the variables, thereby satisfying the assumption of error independence (Durbin–Watson = 1–3). Specifically, the Durbin–Watson values for university healthcare students were within the acceptable range (moment 1 = DW = 1.81; moment 2 = DW = 1.75), supporting the generalizability of the findings. Variance inflation factor (VIF) values were examined to assess multicollinearity among the predictor variables. At moment 1, VIF ranged from 1 to 1.723, and at moment 2, VIF was 1. Tolerance values were consistently close to 1 and did not approach the critical threshold of 10, indicating the absence of multicollinearity.
Table 5 presents the results: at time point 1 (TP1), the most predictive model of resilience in nursing students was model 2, in which the inclusion of both demographic and protective variables accounted for 67.1% of the variance in resilience (R2c = .624; F = 52.81; P <.01). Specifically, the most salient predictors of higher resilience were living alone, high levels of self-efficacy, optimism, and, to a lesser extent, the use of emotions component of emotional intelligence. At time point 2 (TP2), model 2 again emerged as the most predictive of resilient outcomes, explaining 71.1% of the variance in resilience (R2c = .698; F = 39.11; P <.01). In this case, the most influential predictors were not living alone, being optimistic, and possessing a high capacity for emotional regulation. Effect sizes were large, and statistical power was high.
Table 5. Multiple regression at each time point
1-β = statistical power; β = regression result or beta equation; CI = confidence intervals; DW = Durbin-Watson Test; Emotional intelligence (OEA) = emotional intelligence (sub-dimension perception of the emotions of others); Emotional intelligence (ROE) = emotional intelligence (sub-dimension regulation of emotions); Emotional intelligence (UOE) = emotional intelligence (sub-dimension use of emotions); F = test statistic (ANOVA); LL = lower limit; ns = not significant; R2c = corrected coefficient of determination; SE = standard error; TP1 = time point 1; TP2 = time point 2; t = test statistic for predictor variables; UL = upper limit; * p <.05 ** p <.01; ƒ2 = effect size.
Discussion
The objective of this study was to assess resilience levels in Spanish nursing students at two distinct time points characterized by varying degrees of adversity, in order to determine whether significant differences existed between the initial stage of the COVID-19 pandemic (March 2020, during the first wave of mandatory nationwide confinement in Spain) and a subsequent stage several months later (October 2020, the so-called “new normal”). The study involved within-group comparisons and identification of relevant differences. Furthermore, we analyzed which demographic and psychological risk variables (anxiety and depression) and protective variables (emotional intelligence, optimism, and self-efficacy) were the most predictive of high resilience at each time point.
This study offers robust empirical evidence on the psychosocial dynamics experienced by Spanish nursing students during two critical phases of the COVID-19 pandemic. The longitudinal design enabled a nuanced understanding of the evolution of resilience and its associated protective and risk factors over time. The observed increase in anxiety levels and the slight decline in resilience during the later stage (October 2020) suggest that prolonged uncertainty, academic demands, and clinical responsibilities during the “new normal” contributed to emotional fatigue. Interestingly, depression levels remained relatively stable, indicating a potentially delayed onset or more complex manifestation of depressive symptoms compared to anxiety. This stability might also reflect the high levels of familial stress experienced during confinement, arising from economic, social, or health-related challenges, including hospitalization or bereavement due to the loss of loved ones.Reference Balluerka-Lasa, Gómez Benito and Hidalgo Montesinos26
At time point 2, nursing students exhibited higher levels of anxiety, coinciding with the relaxation of restrictive measures but also with one of the most academically demanding periods, marked by upcoming examinations and the continuation of remote learning.
Strong positive correlations were found between resilience and the variables of self-efficacy, optimism, and emotional intelligence, while negative associations emerged with anxiety and depression. These findings reinforce existing theoretical frameworks concerning psychological protection and adaptation in young adults.Reference Masten27 Notably, emotional regulation emerged as the subdimension of emotional intelligence most predictive of resilience, highlighting the critical role of emotional management in sustaining psychological well-being under chronic stress conditions. These results are consistent with previous studies, such as that conducted by Espinosa-Ferro,Reference Espinosa Ferro, Mesa Trujillo and Díaz Castro28 which identified self-efficacy as one of the key dimensions protecting students from developing psychological disorders during confinement.
Recent research has also confirmed the association between emotional intelligence (EI) and resilience, showing that individuals with higher EI demonstrate greater emotional self-regulation and adaptive capacity in the face of stress. EI facilitates the use of coping strategies that diminish negative emotions and maintain positive affect, thus functioning as a protective factor.Reference Meléndez, Delhom and Satorres29
Moreover, hierarchical regression models revealed subtle differences across the 2 time points. While self-efficacy and use of emotions were stronger predictors of resilience during the initial lockdown, optimism and emotional regulation were more salient during the second phase. This shift suggests a transition from action-oriented coping mechanisms to more cognitive-emotional regulation strategies as the pandemic progressed, reflecting the adaptability of psychological processes in response to changing contextual demands.
Finally, the variables most predictive of resilience differed between T1 and T2. Specifically, during periods of heightened adversity and elevated emotional intensity (T2), it becomes essential to strengthen both sociodemographic and protective variables to foster effective adaptation. Protective factors such as having a partner, cohabitating with others, being optimistic, employing problem-focused emotional regulation, and enhancing self-efficacy were more predictive of resilient rather than psychopathological outcomes. These findings are consistent with prior research.Reference Mohammed, Bady and Abdelhamid30
Understanding the protective factors that support resilience in nursing students during pandemics is of substantial clinical importance. The present findings, emphasizing the beneficial influence of self-efficacy, dispositional optimism, and emotional intelligence—particularly emotional regulation—suggest actionable targets for intervention. From a clinical standpoint, resilience training programs integrated into nursing curricula may alleviate psychological distress and cultivate adaptive coping strategies in high-pressure environments such as clinical placements or healthcare crises. Additionally, structured emotional intelligence training, focusing on emotional awareness, application, and regulation, may enhance students’ capacity to navigate interpersonal stressors and the emotional demands inherent in healthcare work.
Clinically, fostering self-efficacy through experiential learning, mentorship, and reflective practice may empower students to perceive themselves as competent and in control even amid adversity, thereby reducing vulnerability to anxiety and depressive symptoms. Similarly, promoting dispositional optimism could positively shape students’ outcome expectations, encouraging a proactive rather than avoidant coping style.
Interventions that promote resilience in nursing education are not only critical for safeguarding mental health but may also reduce burnout, improve quality of care, and enhance workforce sustainability in times of crisis. These findings are aligned with broader literature indicating that resilience training effectively reduces psychological morbidity among healthcare professionals during pandemics.Reference Masten27
Limitations and Recommendations for Future Research
This study has some limitations. First, preexisting disorders were not assessed before the variables were measured at either time point. This may have affected the results. However, longitudinal collection of data from the same participants at 2 different time points (March and October 2020) may be an important methodological advantage over cross-sectional studies. Second, the study sample was limited exclusively to nursing students and does not reflect the diversity of the Spanish university community. This group has been found to have suffered significant emotional impact during the pandemic, so the data from our study may be of interest to academic authorities for implementing psychosocial prevention measures. Another important limitation is that the sample represents a percentage of the population with a high level of education, a key factor for resilience.Reference Repo, Herkama and Salmivalli31 The participants were all young university students. However, psychosocial literature has shown that university students should be a priority group for implementation of psychological prevention, assessment, and intervention actions due to their high levels of distress and suicide-related behaviors.Reference Coentre and Góis32–Reference Sánchez-Teruel, Robles-Bello and Camacho-Conde34 All of these aspects may lead to some bias in the results; hence any generalizations should be undertaken with caution and the limitations should be taken into account.
Conclusion
Although numerous studies have analyzed the detrimental effects of the COVID-19 pandemic on nursing students, few have addressed the protective factors that could mitigate these outcomes. This study offers insight into the psychological impact of the pandemic, emphasizing the potential role of self-efficacy, optimism, and emotional regulation in fostering resilience. The stability of both risk and protective factors over time suggests the need for sustained preventive interventions within Spanish universities. Despite the existence of psychosocial support services in some countries, many higher education institutions lack such structures or provide limited access. These services should be considered essential components of university systems. Consistent with prior research, the findings reaffirm that university students are especially susceptible to mental health disorders, even more so than their non-university peers. This highlights the imperative for academic institutions and policymakers to prioritize not only intellectual development but also students’ psychological well-being.
Data availability statement
The data that support the findings of this study are available upon reasonable request from the corresponding author.
Acknowledgements
Funding for open access charge: University of Jaén/CBUA
Author contribution
María Auxiliadora Robles-Bello and David Sánchez-Teruel were responsible for the study conception and design. Irhomis Mendoza supervised the whole thesis. All authors prepared the first draft of the manuscript. All authors did the data analysis, made critical revisions to the paper for important intellectual content, and supervised the study. All authors read and approved the final manuscript.
Funding statement
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Competing interests
The authors have no conflict of interest to declare.
Informed consent
The informed consent form was completed by all the participants.
Ethical approval
The study was approved by the ethics committee of the University of Jaén (code: ABR.20/4.PRY) and followed the principles of the Declaration of Helsinki.
Open access
