Mental Health of Pharmaceutical Sciences Students: A Repeated Cross‐Sectional Study (2018–2025) Authors: Juliana Alvares Duarte Bonini Campos, Livia Oliveira Santos, Adrielly Santos, João Marôco, Lucas Arrais Campos Published: 2026 ABSTRACT ABSTRACTBackground and AimsUniversity students enrolled in pharmaceutical sciences programs are increasingly vulnerable to mental health issues, yet temporal data on symptom patterns and psychological constructs remain scarce. This study aimed to estimate the prevalence and severity of depressive, anxiety, and stress symptoms, psychological distress, loneliness, and mood state among pharmaceutical sciences students; assess associations between these symptoms and demographic characteristics; identify central symptoms within the symptom networks; and analyze trends from 2018 to 2025.MethodsThis repeated cross‐sectional study analyzed data from 11 independent survey waves conducted between 2018 and 2025 among pharmacy students at a public university in Southeast Brazil. Data were collected using self‐administered questionnaires (DASS‑21, IES‑R, BRUMS, and UCLA‑3). Prevalence and mean scores were estimated with 95% confidence intervals. Associations were examined using multiple logistic regression, with adjusted odds ratios pooled across waves using meta‐analysis. Network analysis identified central symptoms, and joinpoint regression evaluated temporal trends.ResultsParticipants were predominantly women (64.0%–77.8%), and 30.1%–41.3% reported a lifetime diagnosis of mental disorder. Depressive symptoms increased by 20.6% and stress by 7.9% over the study period, with a modest post‐pandemic decline. Anxiety remained consistently high, with a 1.25% average increase across waves. Although loneliness and psychological distress declined slightly, prevalence remained elevated. In multiple logistic regression analyses, female students (pooled ORs ≥ 1.56) and those with a lifetime mental disorder diagnosis (pooled ORs ≥ 1.98) showed higher odds of anxiety, stress, and psychological distress, whereas no statistically significant associations were observed for age. Network analysis identified stress, tension, and hyperarousal as central nodes.ConclusionMental health‐related symptoms were highly prevalent and were increasing before the pandemic among pharmaceutical sciences students. Loneliness persisted post‐pandemic, and social determinants played a key role in shaping mental health‐related symptoms. Universities should implement sustained, targeted prevention and intervention strategies, particularly for vulnerable groups. Introduction Over recent decades, youth mental health has gained prominence in public health agendas due to the rising prevalence of mental disorder symptoms and diagnoses [1, 2, 3, 4]. The World Health Organization estimates that one in seven adolescents experiences a mental disorder, including depression, anxiety, and behavioral disorders [5]. In the United States, diagnoses among adolescents rose 35% between 2016 and 2023 [6]. This trend is further supported by data from the Kaiser Family Foundation, which highlights a significant deterioration in emotional well‐being indicators among youth [7]. In Brazil, this rise is also substantial. In 2017, the country reported the highest global rates of anxiety and depression [8, 9]. The National Health Survey (PNS) reported 31.6% of 18‐ to 24‐year‐olds diagnosed with anxiety and 14.1% with depression [2]. A comparative analysis of the 2013 and 2019 PNS editions (2024) showed that the prevalence of depressive symptoms nearly doubled during this period, from 5.6% to 10.9%, with the most pronounced increase observed among 18‐ to 20‐year‐old individuals. Additionally, a 2013–2014 study [2] involving 74,589 adolescents from large Brazilian cities found that approximately one‐third of participants exhibited symptoms of common mental disorders (CMDs). University students are a key group in youth mental health research due to psychosocial challenges during the transition to adulthood. This period involves career pressures, unstable job markets, and challenges in remaining in higher education [10, 11, 12]. Mental distress often intensifies during this phase, and many students lack the psychological resources or social support needed to cope with daily academic and personal demands [13, 14]. Therefore, university students are considered a vulnerable group for developing symptoms of CMDs [12]. Pharmaceutical sciences students represent a population exposed to distinctive academic and professional demands that may increase vulnerability to mental health problems [10, 15]. Pharmacy education typically involves a dense, cumulative curriculum that combines theoretical coursework, laboratory‐based training, and the need to balance academic activities with internships or early professional experiences. In addition, students are frequently exposed to high academic workload, frequent assessments, and increasing expectations regarding professional responsibility throughout their training [10]. Studies based on data collected before the COVID‐19 pandemic have consistently documented that mild to high levels of depressive, anxiety, stress, and burnout symptoms were already present among pharmacy students, highlighting a substantial mental health burden within this population [15, 16, 17]. COVID‐19 pandemic worsened these vulnerabilities by altering relationships and teaching‐learning processes in higher education [13, 18, 19, 20]. Social isolation and abrupt remote learning disrupted routines and intensified psychological impacts [13, 19, 20, 21]. Students faced difficulties adapting to online education [22], cognitive overload [13, 23], and increased stress, burnout, and emotional exhaustion levels [11]. Common outcomes included stress, anxiety, depression, psychological distress, loneliness, and deteriorated sleep quality [13, 15, 18, 21, 24, 25]. Traumatic events like the COVID‐19 pandemic can trigger dysfunctional routines and post‐traumatic symptoms alongside other mental disorders. Loneliness, intensified by the social distancing, was a key factor. Remote learning reduced academic interaction, increasing the feeling of isolation [13, 24, 26, 27]. This disruption altered psychosocial perceptions and caused mood fluctuations [28], likely worsened by tension, anger, and confusion during a period of widespread instability [29]. Prolonged isolation, remote education, and uncertainty due to the pandemic context contributed to cognitive and emotional fatigue and overload [29]. Evidence suggests that the psychological impact of the COVID‑19 pandemic did not resolve immediately after the easing of social restrictions. Cross‐sectional studies in Germany [30], Ireland [25], and Italy [31] report worsening students' mental health persisting after restrictions were relaxed. Persistent symptoms may increase mental disorder risk, impair learning, and disrupt a functional and healthy routine [32]. Specifically regarding pharmacy students, they experienced a substantial deterioration in mental health following the onset of the COVID‑19 pandemic, reflecting both the acute impact of the crisis and the amplification of pre‑existing academic stressors [15]. During the early phases of the pandemic, increases in symptoms of stress, anxiety, emotional distress, and difficulties with coping were observed among undergraduate pharmacy students [15, 33]. Evidence from the first year following the onset of the COVID‑19 pandemic indicates that mental health symptoms among pharmacy students remained prevalent and relatively stable over time, with no substantial reduction observed during this period [34]. In this context, updated post‑pandemic data are needed to determine whether mental health indicators among pharmacy students improved, stabilized, or further deteriorated as academic and social environments progressively transitioned beyond the acute phase of the COVID‑19 pandemic. Such data allow for the monitoring of mental health before, during, and after the pandemic. They can inform educational, preventive, and intervention strategies aimed at strengthening socioemotional skills and coping resources. They may also contribute to assessing the psychological impact of the pandemic in the context of ongoing psychosocial demands and to supporting the restoration of mental health and well‑being. This rationale motivated the present study. We hope the present study may contribute to addressing psychological distress among pharmaceutical sciences students by providing evidence and critical reflection to support the planning and implementation of health‐promoting initiatives within academic institutions. Such initiatives may foster the development of support networks involving the broader academic community. We also aimed to investigate whether gender and a lifetime mental disorder diagnosis increase vulnerability to anxiety, depression, stress, psychological distress, and loneliness, as previous research suggests [21, 26, 30, 35, 36, 37]. The objectives of this study were to: (i) identify the prevalence and/or scores of symptoms related to depression, anxiety, stress, psychological distress, and loneliness, as well as mood state among pharmaceutical sciences students; (ii) assess the association between these symptoms and demographic characteristics; (iii) identify the relationships and strength of associations among different indicators contributing to the maintenance of symptom networks related to mental disorders; and (iv) evaluate the patterns of prevalence and/or mean scores related to symptoms of depression, anxiety, stress, psychological distress, loneliness, as well as mood state among pharmaceutical sciences students from 2018 to 2025. Methods Study Design and Sampling This observational cross‐sectional study included eleven time points. The first two (March–May 2018; August–November 2019) preceded COVID‐19 and assessed depression, anxiety, and stress in pharmacy students at a public university in Southeast Brazil. Six subsequent points occurred during the pandemic, every 6 months (Phases 1–6: May–June and November–December, 2020–2022), adding psychological distress, mood states, and, from Phase 2, loneliness. In 2023, data collection became annual; from 2024 onward, only depression, anxiety, and stress were monitored. All pharmacy and biotechnology undergraduates were invited (Pharmacy full‐time: N = 350; evening: N = 180; Bioprocess and Biotechnology Engineering: N = 200). Minimum sample size was calculated based on psychometric analyses requirements: five participants per item in the factor model [38]. Considering that the largest scale includes 24 items, the minimum required sample size was estimated at 120 participants. Only participants who provided complete responses to all items of the measurement scales were included in the analyses. This study was designed and reported in accordance with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) recommendations [39], and the completed STROBE checklist is provided in Supporting Information S1. Demographic and academic variables included age (in years), gender (male, female, nonbinary, prefer not to disclose), academic program (Pharmacy, Bioprocess and Biotechnology Engineering), period of course (full‐time or evening), and employment (yes/no). From Phase 1 (third time point), additional data included monthly household income (in Brazilian reais) and health‐related variables. These included the presence or absence of health conditions, COVID‐19‐related symptoms (none, anxiety, anguish, shortness of breath, fear, panic, tachycardia, insomnia, and/or others), mental disorder diagnosis (type specified), and mental health care practices (yes/no). Those who answered affirmatively to the last question were asked to indicate the type of care (e.g., medication use, psychotherapy, lifestyle monitoring, and/or others). Those who mentioned lifestyle‐related strategies were further prompted to describe the specific approaches they adopted. Measurement Scales Symptoms of depression, anxiety, and stress were assessed using the Portuguese version of the Depression, Anxiety and Stress Scale (DASS‐21) [40], translated by Vignola and Tucci [41] and later adapted by Martins et al. [42]. DASS‐21 includes 21 items across three factors (depression, anxiety, and stress) rated on a 4‐point Likert scale (0 = never; 1 = sometimes; 2 = very often; and 3 = almost always). Factor scores were calculated by summing the responses to the items and multiplying by 2 [40]. Based on these scores, participants were categorized by severity level (Depression: Normal [0–9], Mild [10–13], Moderate [14–20], Severe [21–27], Extremely Severe [≥ 28]; Anxiety: Normal [0–7], Mild [8–9], Moderate [10–14], Severe [15–19], Extremely Severe [≥ 20]; Stress: Normal [0–14], Mild [15–18], Moderate [19–25], Severe [26–33], and Extremely Severe [≥ 34]) [40]. Psychological distress was measured using the Portuguese version of the Impact of Event Scale–revised (IES‐R) [43]. The scale has 22 items across three factors (avoidance, intrusion, and hyperarousal) rated on a 5‐point Likert scale (0 = not at all, 1 = slightly, 2 = moderately, 3 = very, and 4 = extremely). Factor scores are calculated by summing the responses. A total psychological distress score can also be computed by summing all item responses. Based on this total score, participants are classified into severity levels as follows: Normal (0–23), Mild (24–32), Moderate (33–36), and Severe (≥ 37) [37]. Mood states were assessed by the Portuguese version of the Brunel Mood Scale (BRUMS) [44]. It consists of 24 mood‐related descriptors rated on a 5‐point Likert scale (0 = not at all, 1 = a little, 2 = moderately, 3 = quite a bit, and 4 = extremely). The scale comprises six factors (anger, confusion, depression, fatigue, tension, and vigor). Mean scores for each factor can be calculated to estimate overall levels of each mood dimension. Loneliness was measured using the short version of the UCLA Loneliness Scale (UCLA‐3) (2004). It includes three items rated on a 3‐point Likert scale (1 = hardly ever, 2 = some of the time, and 3 = often). However, considering that the UCLA‐3 items were derived from the original UCLA Loneliness Scale [45], which uses a 4‐point response format (1 = never, 2 = rarely, 3 = sometimes, and 4 = often), the latter response scale format was adopted in the present study to allow for greater response variability [46]. The translation and adaptation version of the UCLA Loneliness Scale into Portuguese for the Brazilian context, which includes the UCLA‐3 items, was presented by Kuznier et al. [47]. Scores are calculated by summing responses to the items, with higher scores indicating greater levels of loneliness [48]. To classify loneliness levels, percentile cutoffs were applied [47]: scores< 2.5 indicate no loneliness, between 2.5 and 3.0 reflect moderate loneliness, and ≥ 3.0 indicate exacerbated loneliness. Procedures and Ethical Aspects Table 1 presents the procedures used for data collection. Before the pandemic (2018–2019), students completed the demographic questionnaire and the DASS‐21 using a self‐administered paper‐and‐pencil method. Following the onset of the pandemic, data collection shifted to an online format. In 2020, Google Forms was used, and from 2021 onward, LimeSurvey was adopted. 1 Phases 1–4 (2020–2021) included a link sent via institutional email with demographic and pandemic‐related questions, plus DASS‐21, IES‐R, BRUMS, and UCLA‐3 (added in Phase 2). In May/June 2022 (Phase 5), with the return to in‐person classes, paper‐based collection resumed. The same scales were used in 2022–2023 and, in 2024 and 2025, only DASS‐21 was applied. Data were collected semiannually (2020–2022) and annually from 2023. Table 1Summary of data collection procedures.Data collectionYearMonthData collection methodScales used20182018March/MayIn‐person, Paper‐and‐pencilDASS‐2120192019August/NovemberIn‐person, Paper‐and‐pencilDASS‐21Phase 12020May/JuneOnline, Google FormsDASS‐21, IER‐S, BRUMSPhase 22020November/DecemberOnline, Google FormsDASS‐21, IER‐S, BRUMS, UCLA‐3Phase 32021May/JuneOnline, Lime SurveyDASS‐21, IER‐S, BRUMS, UCLA‐3Phase 42021November/DecemberOnline, Lime SurveyDASS‐21, IER‐S, BRUMS, UCLA‐3Phase 52022May/JuneIn‐person, Paper‐and‐pencilDASS‐21, IER‐S, BRUMS, UCLA‐3Phase 62022November/DecemberIn‐person, Paper‐and‐pencilDASS‐21, IER‐S, BRUMS, UCLA‐320232023May/JuneIn‐person, Paper‐and‐pencilDASS‐21, IER‐S, BRUMS, UCLA‐320242024May/JuneIn‐person, Paper‐and‐pencilDASS‐2120252025May/JuneIn‐person, Paper‐and‐pencilDASS‐21John Wiley & Sons, Ltd. Summary of data collection procedures. Recruitment occurred in classrooms (pre‐pandemic and post‐2022) or via email (2020–2021), in collaboration with the Course Council and Student Union. The invitation email included a link to access the survey. On the first page of the survey, whether paper‐based or online, all participants reviewed and agreed to the Informed Consent Form before proceeding to the survey. As participation was voluntary and anonymous, reasons for nonparticipation could not be assessed. The data collections conducted prior to the pandemic were approved by the Research Ethics Committee on Human Subjects of the School of Pharmaceutical Sciences–UNESP (CAAE: 63553516.4.0000.5426 and 11735719.0.0000.5426). Authorization for the post‐pandemic phase of the study was granted by the National Commission for Ethics in Research of the Brazilian Ministry of Health (CONEP) (CAAE: 30604220.4.0000.0008). Data Validity and Reliability Confirmatory factor analysis (CFA) using the Weighted Least Squares Means and Variance‐adjusted (WLSMV) robust estimator was used to assess the model fit for the measurement scales. Fit indices included comparative fit index (CFI), Tucker–Lewis Index (TLI), root mean square error of approximation (90% confidence interval [CI] RMSEA), and standardized root mean square residual (SRMR). Model fit was considered acceptable when CFI and TLI ≥ 0.90, RMSEA ≤ 0.10, and SRMR ≤ 0.08 [38, 49]. Reliability was assessed using the ordinal alpha coefficient, with values above 0.70 considered acceptable. To enable comparisons across different data collection phases, metric and scalar invariance were tested between samples [50], using changes in CFI (ΔCFI). Absolute differences < 0.01 indicate model invariance. Analyses were conducted in R 4.2.1 (R Core Team 2022), using the “lavaan” package [51] and “semTools” [52]. The factor model of the scales demonstrated good fit and strict measurement invariance across data collection rounds (Supporting Information S2), supporting the validity and reliability of the data across both time points and administration modes, and allowing for direct comparisons over time. Statistical Analysis The statistical analysis was organized into four components: (1) descriptive analyses, prevalence estimates, mood‐state scores, and temporal trend analysis; (2) wave‐specific logistic regression analyses; (3) meta‐analytic pooling of estimates across survey waves; and (4) symptom network analysis. Descriptive Analyses, Prevalence Estimates, Mood‐State Scores, and Temporal Trend Analysis Prevalence estimates for symptoms of depression, anxiety, stress, psychological distress, and loneliness, as well as mean scores of mood change, were calculated with 95% CIs. Analyses were conducted using IBM SPSS Statistics, version 22 (IBM Corp., Armonk, NY, USA). Joinpoint regression analyzed temporal changes in prevalence and scores, detecting inflection points via Monte Carlo permutation testing [53]. Weighted Bayesian information criterion (WBIC) guided model selection. Annual percentage change (APC) was calculated with 95% CIs, and the average annual percentage change (AAPC) was computed to summarize the overall direction and magnitude of change across the full study period. Analyses were conducted using Joinpoint software version 5.0.2 (Statistical Research and Applications Branch, National Cancer Institute 2023). Wave‐Specific Logistic Regression Analyses Multiple logistic regression models assessed associations between symptoms (dependent variables) and predictors (independent variables). Each symptom (depression, anxiety, stress, psychological distress, avoidance, intrusion, and hyperarousal) was analyzed independently and considered present when the factor score met the threshold for at least moderate severity (Supporting Information S3). Independent variables included gender (reference category [rc]: male), lifetime history of a medical diagnosis of a mental disorder (rc: absent), and age (in years). Independent variables were selected a priori based on theoretical relevance and evidence from previous studies [13, 15, 54]. Odds ratios (ORs) with 95% CIs were reported. Analyses were conducted using two‐sided tests, with a significance level of α = 0.05 in IBM SPSS Statistics, version 22 (IBM Corp., Armonk, NY, USA). Meta‐Analytic Pooling of Estimates Across Survey Waves A meta‐analysis pooled ORs and 95% CI (pooled effect size) for significant predictors across phases, treating samples as independent due to the following: (1) cross‐sectional design without systematic participant tracking; (2) low participant overlap across data collections (from 1.5% to 17.3%); (3) approximately 20% of the sample was renewed annually with incoming students; (4) response rates varied between 20.9% and 60.9%, reducing the likelihood of repeated participation; and (5) natural student turnover due to graduation, dropout, leave of absence, or course withdrawal. This methodological decision was further supported by a leave‐one‐out sensitivity analysis using the DerSimonian–Laird random‐effects model. Data heterogeneity was assessed using the Residual Heterogeneity Test (Q‐test, α = 0.05) and residual funnel plots. Fixed‐effects models (t‐test) were used when no heterogeneity was found; otherwise, random‐effects models with Knapp–Hartung adjustment were applied. Forest plots (Supporting Information S4 and S5) illustrate weighted estimates per data collection (study weights). We clarify that the meta‐analysis was conducted as an internal quantitative synthesis of repeated cross‐sectional survey waves and does not constitute a systematic review. Accordingly, the meta‐analytic procedures were designed and reported in line with the relevant PRISMA 2020 items applicable to statistical synthesis and reporting of meta‐analyses [55], with the corresponding checklist available in Supporting Information S6. Analyses were performed using JASP (version 0.16; JASP Team 2021). Symptom Network Analysis Network analysis examined associations among indicators from each measurement scale (DASS‐21, IES‐R, BRUMS, and UCLA‐3) using a Gaussian graphical model [56] with least absolute shrinkage and selection operator (LASSO) regularization method. Extended Bayesian information criterion (EBIC, hyperparameter = 0.50) guided model selection. Centrality indices (node strength, closeness, and betweenness) and expected influences were estimated to identify the influential nodes in maintaining symptom networks. Additionally, a network was estimated to assess the relationships among factor scores from each scale across individual data collections and the overall sample. This analysis included only data from Phase 2 through 2023, as all scales (DASS‐21, IES‐R, BRUMS, and UCLA‐3) were administered. Network accuracy and stability, following the guidelines proposed by Epskamp et al. [57], were verified using nonparametric bootstrapping and case‐dropping, respectively, and were deemed adequate (Supporting Information S7). Analyses were performed using JASP (version 0.16; JASP Team 2021). Results Table 2 presents the sample characteristics. Most participants were young, female, full‐time pharmacy students from middle‐income families and not engaged in paid work. These characteristics suggest that the sample is representative of the student population enrolled in the pharmacy program at the university. Overall, participation rates were higher during in‐person data collection phases. Table 2Sample characteristics at each data collection point. a n (%)Characteristic2018 (n = 330)2019 (n = 268)Phase 1 (n = 153)Phase 2 (n = 254)Phase 3 (n = 275)Phase 4 (n = 243)Phase 5 (n = 444)Phase 6 (n = 360)2023 (n = 442)2024 (n = 408)2025 (n = 409)Age [mean(standard deviation)]20.4 (2.4)21.1 (2.5)21.8 (3.3)21.3 (2.2)21.6 (2.4)21.2 (2.7)20.8 (2.2)20.6 (2.1)21.0 (2.3)21.0 (2.4)20.6 (2.8)GenderMale111 (26.2)79 (26.9)34 (22.2)67 (26.5)96 (34.9)81 (33.3)143 (32.4)103 (28.8)136 (31.1)108 (26.5)107 (26.3)Female313 (73.8)215 (73.1)119 (77.8)186 (73.5)176 (64.0)158 (65.0)293 (66.3)246 (68.7)294 (67.1)294 (72.0)295 (72.5)Nonbinary————2 (0.7)2 (0.8)5 (1.1)7 (2.0)6 (1.4)6 (1.47)5 (1.2)Prefer not to answer————1 (0.4)2 (0.8)1 (0.2)2 (0.6)2 (0.5)——Period of courseFull‐time program319 (77.8)203 (69.3)124 (81.0)212 (83.8)218 (79.3)183 (75.3)337 (76.1)267 (74.6)325 (74.0)293 (71.8)309 (75.7)Evening program91 (22.2)90 (30.7)29 (18.9)41 (16.2)57 (20.7)60 (24.7)106 (23.9)91 (25.4)114 (26.0)115 (28.2)98 (24.1)Academic programPharmacy330 (77.8)294 (100.0)123 (80.4)182 (71.9)216 (78.5)193 (79.4)327 (73.8)269 (75.1)357 (81.3)317 (77.7)278 (68.3)Bioprocess and Biotechnology Engineering94 (22.2)—30 (19.6)71 (28.1)59 (21.5)50 (20.6)116 (26.2)89 (24.9)82 (18.7)91 (22.3)129 (31.7)In addition to studying, do you engage in any employment activities (work)?No398 (94.3)247 (84.3)116 (75.8)199 (78.6)221 (80.4)197 (81.1)376 (87.4)304 (89.1)350 (83.3)——Yes24 (5.7)46 (15.7)37 (24.2)54 (21.4)54 (19.6)46 (18.9)54 (12.6)37 (10.9)70 (16.7)——Monthly household income (R$)# ≤ 1254——8 (5.3)11 (4.5)9 (3.3)11 (4.6)24 (5.6)16 (4.7)12 (2.9)7 (1.7)18 (4.4)1255–2004——21 (13.7)37 (15.3)43 (15.7)36 (14.9)60 (14.0)43 (12.7)55 (13.2)44 (10.9)47 (11.6)2005–8640——90 (58.8)133 (55.0)151 (55.1)143 (59.4)237 (55.4)185 (54.6)245 (58.8)223 (55.3)195 (48.1)8641–11,261——19 (12.4)32 (13.2)36 (13.1)29 (12.0)63 (14.7)56 (16.5)58 (13.9)64 (15.9)81 (20.0)≥ 11,262——15 (9.8)29 (12.0)35 (12.8)22 (9.1)44 (10.3)39 (11.5)47 (11.2)65 (16.1)64 (15.8)a2018: March to June 2018; 2019: August to November 2019; Phase 1: May to June 2020; Phase 2: November to December 2020; Phase 3: May to June 2021; Phase 4: November to December 2021; Phase 5: May to June 2022; Phase 6: November to December 2022; Phase 7: May to June 2023. #R$: Brazilian reais, 1USD = 5.57 BRL(R$) (18/07/2025). Accessed on www.bcb.gov.br.John Wiley & Sons, Ltd. Sample characteristics at each data collection point. a 2018: March to June 2018; 2019: August to November 2019; Phase 1: May to June 2020; Phase 2: November to December 2020; Phase 3: May to June 2021; Phase 4: November to December 2021; Phase 5: May to June 2022; Phase 6: November to December 2022; Phase 7: May to June 2023. #R$: Brazilian reais, 1USD = 5.57 BRL(R$) (18/07/2025). Accessed on www.bcb.gov.br. Table 3 highlights that almost all participants reported at least one symptom since the pandemic began, with anxiety being most prevalent, followed by distress, fear, and insomnia. About one‐third had a lifetime diagnosis, mainly anxiety and depression. Except for Phases 1 (early pandemic), 3, and 4 (2021), around two‐thirds engaged in mental health self‐care, mainly via lifestyle strategies. Attending therapy use was high in Phase 1 (beginning of the pandemic), dropped in Phase 2 (second half of 2022), rose slightly in Phases 3–4 (most critical periods of the pandemic in Brazil), and declined from 2022 onward. Medication use ranged from 15.7% to 24.7%. Common lifestyle strategies included exercise, healthy eating, and leisure. Table 3Students' responses to health‐related questions following the onset of the COVID‐19 pandemic. n (%)Phase 1 (n = 153)Phase 2 (n = 254)Phase 3 (n = 275)Phase 4 (n = 243)Phase 5 (n = 444)Phase 6 (n = 360)2023 (n = 442)2024 (n = 408)2025 (n = 409)Do you have any health problems?No112 (73.2)212 (83.5)226 (82.1)207 (85.2)370 (83.9)296 (82.7)354 (81.6)——Yes41 (26.8)42 (16.5)49 (17.9)36 (14.8)71 (16.1)62 (17.3)80 (18.4)——Since the beginning of the COVID‐19 pandemic, have you experienced any of the following symptoms?I had no symptoms3 (2.0)6 (2.4)7 (2.5)8 (3.3)23 (5.2)9 (2.5)19 (4.3)——Yes150 (98.0)248 (97.6)268 (97.5)235 (96.7)421 (94.8)351 (97.5)423 (95.7)——Anxiety137 (89.5)237 (93.3)256 (93.1)223 (91.8)404 (91.0)333 (92.5)403 (91.2)——Anguish111 (72.5)189 (74.4)210 (76.4)179 (73.7)315 (70.9)268 (74.4)286 (64.7)——Shortness of breath49 (32,0)64 (25.2)77 (28.0)81 (33.3)130 (29.3)105 (29.2)121 (27.4)——Fear90 (58.8)164 (64.6)197 (71.6)165 (67.9)265 (59.7)216 (60.0)277 (62.7)——Panic35 (22.9)58 (22.8)69 (25.1)66 (27.2)110 (24.8)84 (23.3)104 (23.5)——Tachycardia47 (30.7)89 (35.0)113 (41.1)99 (40.7)161 (36.3)131 (36.4)152 (34.4)——Insomnia105 (68.6)184 (72.4)183 (66.5)159 (65.4)260 (58.6)194 (53.9)232 (52.5)——Others14 (9.2)4 (1.6)23 (8.4)11 (4.5)4 (0.9)————Have you ever received a medical diagnosis related to a mental health disorder during your life?No87 (59.2)174 (69.9)189 (68.7)165 (67.9)302 (68.2)238 (66.1)285 (65.1)239 (58.7)268 (66.0)Yes60 (40.8)75 (30.1)86 (31.3)78 (32.1)141 (31.8)122 (33.9)153 (34.9)168 (41.3)138 (34.0)If yes, what medical diagnosis was given?Generalized anxiety disorder50 (32.7)64 (25.2)72 (26.2)64 (26.3)111 (25.0)102 (28.3)122 (27.6)138 (33,8)117 (28.6)Panic disorder14 (9.2)9 (3.5)13 (4.7)21 (8.6)26 (5.9)15 (4.2)23 (5.2)29 (7.1)21 (5.1)Depressive disorders26 (17.0)31 (12.2)47 (17.1)40 (16.5)67 (15.1)55 (15.3)62 (14.0)69 (16.9)59 (14.4)Bipolar disorder1 (0.7)2 (0.8)4 (1.5)1 (0.4)2 (0.5)6 (1.7)7 (1.6)10 (2.5)5 (1.2)Schizophrenia————1 (0.2)——1 (0.2)—Phobic disorders—4 (1.6)3 (1.1)5 (2.1)5 (1.1)6 (1.7)6 (1.4)11 (2.7)5 (1.2)Obsessive‐compulsive disorder1 (0.7)6 (2.4)3 (1.1)5 (2.1)1 (0.2)5 (1.4)6 (1.4)17 (4.2)—Others4 (2.6)3 (1.2)4 (1.5)6 (2.5)16 (3.6)7 (1.9)17 (3.8)—21 (5.1)Do you take care of your mental health?No72 (47.1)72 (28.3)129 (46.9)111 (45.7)111 (25.0)90 (25.0)97 (21.9)127 (31.3)57 (14.0)Yes81 (52.9)182 (71.7)146 (53.1)132 (54.3)333 (75.0)270 (75.0)341 (77.1)279 (68.7)351 (86.0)If yes, how do you take care of your mental health?Therapy24 (29.6)18 (9.7)26 (18.6)24 (18.4)16 (4.8)10 (3.7)10 (3.0)36(12.6)8 (2.2)Lifestyle strategies30 (37.0)123 (66.5)64 (45.7)53 (40.8)210 (63.1)164 (60.7)208 (61.4)134 (46.8)197 (55.2)Medication use8 (9.9)7 (3.8)9 (6.4)7 (5.4)7 (2.1)8 (3.0)9 (2.6)14 (4.9)10 (2.8)Therapy + Lifestyle strategies7 (8.6)15 (8.1)16 (11.4)20 (15.4)51 (15.3)46 (17.0)54 (15.9)55 (19.2)66 (18.5)Therapy + Medication use7 (8.6)7 (3.8)7 (5.0)13 (10.0)6 (1.8)5 (1.9)4 (1.2)19 (6.6)7 (2.0)Lifestyle strategies + Medication use2 (2.5)3 (1.6)3 (2.1)7 (5.4)20 (6.0)18 (6.7)19 (5.6)11 (3.8)38 (10.6)Terapia + Lifestyle strategies + Medication use3 (3.7)12 (6.5)15 (10.7)6 (4.6)23 (6.9)19 (7.0)35 (10.3)17 (5.9)31 (8.7)If you use lifestyle strategies for mental health care, which one(s) do you practice?Physical exercise—123 (66.5)70 (25.5)63 (25.9)201 (45.3)159 (44.2)224 (50.7)—230 (56.2)Healthy eating—129 (69.7)75 (27.3)61 (25.1)199 (44.8)165 (45.8)222 (50.2)—228 (55.7)Meditation—38 (20.5)25 (9.1)18 (7.4)46 (10.4)34 (9.4)58 (13.1)—38 (9.3)Religious/spiritual practice—48 (25.9)26 (9.5)23 (9.5)98 (22.1)80 (22.2)100 (22.6)—132 (32.3)Leisure activities—130 (70.3)72 (26.2)66 (27.2)235 (52.9)184 (51.1)232 (52.5)—265 (64.8)Arts (dance, painting, crafts, music…)—99 (53.5)41 (14.9)40 (16.5)126 (28.4)100 (27.8)122 (27.6)—144 (35.2)Self‐help—19 (10.3)12 (4.4)11 (4.5)41 (9.2)39 (10.8)43 (9.7)—32 (7.8)Alternative or complementary medicine techniques—30 (16.2)17 (6.2)13 (5.3)36 (8.1)23 (6.4)31 (7.0)—31 (7.6)John Wiley & Sons, Ltd. Students' responses to health‐related questions following the onset of the COVID‐19 pandemic. Table 4 presents the prevalence of symptoms of depression, anxiety, stress, psychological distress, avoidance, intrusion, hyperarousal, and loneliness, as well as mood state scores reported by participants across different data collections. The analysis emphasizes the identification of inflection points using joinpoint regression. The most significant increase in symptom prevalence was observed for depressive symptoms, which rose by 20.6% from the baseline (pre‐pandemic). Although this prevalence has slightly declined since the onset of the pandemic (by 7.4%), it remains elevated. Stress levels also increased between 2018 and the second half of 2021, with a rise of 7.99%. Table 4Prevalence of symptoms (at least moderate severity) among participants across data collections and joinpoint regression analysis.20182019Phase 1Phase 2Phase 3Phase 4Phase 5Phase 6202320242025JoinpointTime pointAPC [95% CI]AAPC [95% CI]Prevalence [95% CI]Depression45.0 [40.4–49.8]51.2 [45.4–57.0]75.7 [68.4–82.0]71.3 [65.5–76.6]74.2 [68.8–79.1]71.9 [66.0–77.3]59.1 [54.4–63.7]63.0 [57.9–68.0]54.3 [49.5–59.0]50.5 [45.6–55.3]43.3 [38.6–48.2]2018–Ph220.57 [7.82–55.14]* 0.25 [−2.43 to 4.05]Ph2–2025−7.37 [−12.56 to −4.35]* Anxiety48.8 [44.1–53.6]52.9 [47.2–58.6]54.6 [46.7–62.4]52.8 [46.6–58.8]54.2 [48.3–60.0]55.8 [49.5–62.0]55.4 [50.7–60.1]57.4 [52.1–62.5]53.3 [48.5–58.0]57.4 [52.5–62.1]54.4 [49.6–59.2]2018–Ph15.23 [0.22–10.89]* 1.25 [0.17–2.21]* Ph1–20250.28 [−4.51 to 2.62]Stress48.8 [43.7–53.9]59.3 [53.1–65.4]65.6 [57.0–73.6]70.0 [63.6–75.8]70.5 [64.4–76.2]71.6 [65.0–77.6]66.9 [61.8–71.7]70.6 [65.2–75.6]62.5 [57.4–67.5]55.9 [51.0–60.7]49.1 [44.3–54.0]2018–Ph47.99 [2.94–22.84]* −0.18 [−2.88 to 3.06]Ph4–2025−7.73 [−17.50 to −3.30]* Avoidance——62.3 [54.4–69.7]58.3 [52.1–64.2]59.5 [53.6–65.2]54.7 [48.4–60.9]47.1 [42.4–51.8]43.8 [38.6–49.1]41.5 [35.4–47.9]——Ph2–Ph7−7.03 [−8.90 to −5.24]* −7.03 [−8.90 to −5.24]* Intrusion——58.1 [49.3–66.5]64.0 [57.4–70.3]63.0 [56.7–69.1]49.8 [43.1–56.5]41.2 [36.2–46.4]39.9 [34.3–45.6]38.6 [32.2–45.4]——Ph2–Ph7−9.83 [−14.95 to −5.33]* −9.83 [−14.95 to −5.33]* Hyperarousal——67.9 [59.7–75.4]64.8 [58.3–71.0]65.0 [58.8–70.9]60.9 [54.3–67.3]48.8 [43.7–53.9]47.2 [41.5–52.8]40.8 [34.2–47.7]——Ph2–Ph7−8.16 [−11.76 to −4.79]* −8.16 [−11.76 to −4.79]* Psychological distress——53.3 [45.3–61.2]56.3 [50.2–62.3]57.3 [51.4–63.1]49.4 [43.4–55.7]36.5 [31.9–41.3]37.0 [31.9–42.3]34.4 [28.4–40.7]——Ph2–Ph7−9.54 [−16.15 to −3.47]* −9.54 [−16.15 to −3.47]* Loneliness———78.0 [72.6–82.8]70.4 [64.9–75.6]69.0 [63.0–74.6]63.9 [59.4–68.3]67.4 [62.4–72.1]61.8 [57.2–66.3]——Ph2–Ph7−3.96 [−7.51 to −0.17]* −3.96 [−7.51 to −0.17]* Mean score [95% CI]Anger——1.79 [1.63–1.96]1.86 [1.72–2.00]2.01 [1.89–2.14]1.64 [1.51–1.77]1.48 [1.38–1.58]1.65 [1.54–1.75]1.42 [1.33–1.51]——Ph1–Ph7−4.34 [−9.27 to 0.54]−4.34 [−9.27 to 0.54]Mental confusion——2.18 [2.03–2.34]2.15 [2.27–2.04]2.20 [2.09–2.31]2.09 [1.97–2.21]1.99 [1.89–2.08]2.05 [2.15–1.94]1.97 [1.87–2.06]——Ph1–Ph7−1.81] [−2.69 to −0.94]* −1.81 [−2.69 to −0.94]* Depressed mood——2.07 [1.91–2.23]2.02 [1.90–2.15]2.29 [2.17–2.41]1.81 [1.67–1.95]1.56 [1.46–1.66]1.73 [1.62–1.83]1.53 [1.43–1.63]——Ph1–Ph7−5.55 [−9.71 to −1.53]* −5.55 [−9.71 to −1.53]* Fatigue——2.25 [2.09–2.41]2.67 [2.56–2.79]2.79 [2.68–2.89]2.72 [2.59–2.84]2.68 [2.59–2.77]2.77 [2.67–2.87]2.45 [2.35–2.54]——Ph1–Ph310.90 [3.31–18.70]* 1.66 [−0.64 to 3.67]Ph3–Ph7−2.67 [−8.19 to −0.26]* Tension——2.34 [2.19–2.48]2.31 [2.19–2.42]2.48 [2.38–2.6]2.10 [1.98–2.21]2.19 [2.10–2.28]2.25 [2.16–2.35]2.19 [2.10–2.27]——Ph1–Ph7−1.36 [−3.21 to 0.55]−1.36 [−3.21 to 0.55]Vigor——1.00 [0.90–1.09]1.11 [1.03–1.18]1.07 [1.01–1.13]1.37 [1.29‐1.45]1.78 [1.72–1.84]1.71 [1.65–1.77]1.84 [1.77–1.90]——Ph1–Ph712.00 [8.38–16.11]* 12.00 [8.38–16.11]* * p < 0.05.John Wiley & Sons, Ltd. Prevalence of symptoms (at least moderate severity) among participants across data collections and joinpoint regression analysis. p < 0.05. Anxiety symptoms showed relatively stable prevalence across data collections (Table 4); however, the annual growth rate was statistically significant, warranting attention. In contrast, psychological distress and its dimensions (avoidance, intrusion, and hyperarousal) have shown a downward trend. Feelings of loneliness also declined, though the overall prevalence remains notably high among students. Regarding mood states, anger and tension remained stable from the beginning of the pandemic through the last data collection. Mental confusion and depressed mood showed slight decreases since Phase 1, by 1.81% and 5.55%, respectively. Fatigue increased substantially from early 2020 to 2021 (10.9%), followed by a modest decline of 2.67% between early 2021 and mid‐2023 (Phase 7), indicating that students have not yet returned to pre‐pandemic levels. Conversely, vigor increased by 12%. In the logistic regression analysis, there was no statistically significant association between age and the likelihood of experiencing symptoms of depression, anxiety, stress, or psychological distress, including its dimensions. Consequently, this variable was excluded from the model. As shown in Table 5, female students and those with a lifetime medical diagnosis of a mental disorder were generally more likely to report symptoms of anxiety, stress, and psychological distress. Sensitivity analysis revealed that the variations in the recalculated estimates, based on the sequential exclusion of individual data collections, were minimal for both gender and mental disorder diagnosis. These variations did not alter the direction or statistical significance of the effects, indicating that the meta‐analysis results are robust, consistent, and stable, even when individual studies are excluded. No statistically significant association was observed between gender and depressive symptoms, intrusive thoughts, or hyperarousal. Table 5Odds ratios (ORs) and 95% confidence intervals [95% CIs] for the presence of symptoms (absent/present) according to gender and lifetime medical diagnosis of a mental disorder (mental disorder diagnosis).CharacteristicData collectionSymptom OR [95% CI]a (pooled effect size, %)DepressionAnxietyStressPsychological distressAvoidanceIntrusionHyperarousalGender (reference category: male)Phase 11.33 [0.53–3.37] (1.70)2.25 [0.91–5.61] (1.86) 2.70 [1.02–7.15] (2.15) 1.98 [0.81–4.84] (3.05) 2.38 [1.01–5.60] (52.09) 0.71 [0.27–1.83] (16.95)0.66 [0.24–1.77] (17.57)Phase 21.06 [0.56–1.99] (6.71)1.53 [0.85–2.75] (11.40)1.71 [0.89–3.28] (14.12)1.24 [0.69–2.21] (21.46)1.44 [0.81–2.58] (14.03)1.63 [0.86–3.10] (8.22)1.47 [0.77–2.81] (9.88)Phase 30.98 [0.55–1.74] (9.69) 1.97 [1.18–3.31] (9.07) 2.27 [1.24–4.16] (9.46) 1.86 [1.11–3.11] (12.39) 1.77 [1.06–2.96] (12.18) 1.28 [0.72–2.27] (17.17)1.65 [0.93–2.93] (10.28)Phase 41.27 [0.70–2.31] (5.29) 2.35 [1.32–4.15] (5.14) 3.40 [1.71–6.72] (3.21) 1.83 [1.05–3.19] (10.82) 1.80 [1.04–3.10] (10.36) 1.71 [0.93–3.14] (8.45) 1.85 [1.01–3.39] (7.26) Phase 51.13 [0.74–1.71] (14.58) 2.20 [1.44–3.37] (11.05) 2.44 [1.51–3.96] (13.44) 1.66 [1.05–2.63] (19.86)1.51 [1.00–2.28] (26.84) 1.93 [1.19–3.15] (10.74) 1.54 [0.97–2.43] (19.30)Phase 61.48 [0.92–2.41] (6.18) 2.57 [1.57–4.22] (5.86) 4.38 [2.45–7.81] (2.81) 1.53 [0.91–2.58] (17.77)1.48 [0.91–2.42] (19.28)1.19 [0.70–2.03] (23.32)1.32 [0.78–2.22] (19.84)20230.88 [0.57–1.36] (21.98) 1.97 [1.27–3.05] (12.99) 2.81 [1.71–4.61] (9.59) 1.40 [0.75–2.59] (14.64)1.39 [0.78–2.48] (15.21)1.25 [0.67–2.32] (15.15)1.16 [0.61–2.22] (15.87)20240.95 [0.60–1.49] (17.32)1.48 [0.93–2.34] (20.70) 2.03 [1.28–3.21] (21.65) ———20250.98 [0.62–1.53] (16.56)1.43 [0.90–2.27] (21.93) 1.93 [1.21–3.06] (23.57) ———Pooled Estimateb 1.03 [0.81–1.25] 1.79 [1.41–2.16] 2.24 [1.71–2.77] 1.56 [1.12–2.00] 1.56 [1.14–1.97] 1.29 [0.89–1.69]1.31 [0.91–1.71]Mental disorder diagnosis (reference category: absent)Phase 1 3.64 [1.47–9.03] (0.91) 4.81 [2.28–10.13] (1.63) 5.62 [2.18–14.4] (0.73) 3.00 [1.47–6.15] (3.78) 1.17 [0.58–2.36] (10.36) 2.60 [1.18–5.70] (5.47) 3.43 [1.49–7.88] (4.19) Phase 2 2.98 [1.45–6.14] (2.37) 2.33 [1.31–4.17] (12.26) 2.73 [1.27–5.88] (5.16) 2.14 [1.19–3.83] (11.88) 1.37 [0.77–2.44] (11.77) 2.13 [1.09–4.14] (12.01) 2.63 [1.32–5.22] (11.26) Phase 3 2.18 [1.13–4.22] (5.46) 2.48 [1.42–4.32] (11.92) 2.54 [1.24–5.19] (7.03) 1.98 [1.13–3.45] (15.39) 1.52 [0.88–2.64] (10.60) 2.08 [1.11–3.90] (14.35) 2.46 [1.29–4.67] (14.99) Phase 4 2.73 [1.36–5.51] (3.03) 3.76 [2.00–7.08] (3.89) 4.71 [1.95–11.3] (1.23) 2.14 [1.21–3.77] (12.64) 1.54 [0.88–2.72] (9.70) 2.21 [1.19–4.12] (13.01) 3.16 [1.62–6.15] (8.34) Phase 5 2.43 [1.55–3.81] (10.20) 2.93 [1.87–4.61] (13.36) 2.93 [1.69–5.10] (9.43) 1.99 [1.28–3.09] (25.28) 1.24 [0.82–1.86] (30.36)2.22 [1.39–3.57] (23.51) 3.17 [2.00–5.04] (18.53) Phase 6 2.09 [1.27–3.45] (10.96) 2.27 [1.38–3.72] (18.31) 2.75 [1.41–5.36] (7.03) 1.94 [1.20–3.14] (22.01) 1.27 [0.80–2.01] (22.43) 2.08 [1.25–3.46] (22.88) 2.14 [1.30–3.54] (34.12) 2023 1.84 [1.21–2.82] (20.10) 3.00 [1.93–4.66] (13.45) 5.27 [2.93–9.50] (2.54) 2.42 [1.34–4.37] (9.02) 2.21 [1.26–3.88] (4.78) 2.76 [1.50–5.07] (8.77) 3.38 [1.78–6.25] (8.57) 2024 1.89 [1.26–2.83] (21.14) 2.93 [1.91–4.50] (14.95) 2.30 [1.51–3.49] (27.98) ————2025 1.64 [1.07–2.49] (25.84) 3.33 [2.12–5.25] (10.24) 1.90 [1.24–2.92] (38.86) ————Pooled estimateb 1.98 [1.55–2.40] 2.80 [2.21–3.38] 2.40 [1.83–2.98] 2.09 [1.52–2.66] 1.36 [1.00–1.72] 2.22 [1.56–2.88] 2.68 [1.86–3.50] Note: Cells in bold indicate p < 0.05. a Estimate obtained from the logistic regression model. b Meta‐analysis (adjusted OR values accounting for variability and the contribution of each data collection) is presented in Supporting Information S4 and S5. Random‐effects model.John Wiley & Sons, Ltd. Odds ratios (ORs) and 95% confidence intervals [95% CIs] for the presence of symptoms (absent/present) according to gender and lifetime medical diagnosis of a mental disorder (mental disorder diagnosis). Note: Cells in bold indicate p < 0.05. Estimate obtained from the logistic regression model. Meta‐analysis (adjusted OR values accounting for variability and the contribution of each data collection) is presented in Supporting Information S4 and S5. Random‐effects model. Table 6 presents the standardized node estimates from the network derived from the relationships among factor scores of the DASS‐21, IES‐R, BRUMS, and UCLA‐3 scales. It also includes standardized node estimates from a second network, constructed based on the analysis of individual indicators (items) comprising each of the measurement scales. Hyperarousal, stress, and tension emerged as the central nodes sustaining the symptom network, except in Phase 6, where depressed mood became the primary node. In the overall sample, depressive symptoms showed strong associations (node strength) with multiple other symptoms. Stress exhibited the highest number of indirect connections (node closeness), suggesting it may be quickly influenced by or have influence across the network. Depressed mood functioned as a bridging symptom (node betweenness), linking various symptom pairs. Table 6Centrality table and plot using standardized node measures (network analysis) (data collection: Phases 1–6 and 2023).ScaleScoresData collection time point (expected influence)Total sample (n = 2018)Phase 2Phase 3Phase 4Phase 5Phase 62023BetweennessClosenessStrengthExpected influenceDASS‐21Depression−0.298−0.512−0.4180.3060.127−0.0590.0660.792 1.233 −0.077Anxiety−0.002−0.2100.065−0.1430.2020.137−0.9590.4360.0180.018Stress0.7800.876 1.223 1.038 0.8220.904−0.618 1.297 0.708 1.237 IES‐RAvoidance−0.849−0.783−0.594−0.617−0.638−0.463−0.959−1.647−1.606−0.644Intrusion0.4870.7310.738 1.038 0.6480.7250.920−0.6340.1350.632Hyperarousal 1.383 1.222 0.8940.716 0.975 0.957 1.262 −0.3190.369 1.074 UCLA‐3Loneliness−0.983−0.791−0.452−0.897−1.268−1.286−0.959−1.751−1.795−1.017BRUMSAnger−0.377−0.135−0.321−0.266−0.553−0.4501.0911.0560.137−0.555Mental Confusion0.4560.840.3670.2030.1350.212−0.4470.078−0.2690.519Depressed mood0.6070.6450.3320.680 1.154 0.836 1.774 1.171 1.424 0.514Fatigue−0.384−0.243−0.433−0.255−0.3630.052−0.789−0.288−1.005−0.392Tension 1.360 0.781 1.158 0.7971.0100.9010.5780.4640.9091.020Vigor−2.179−2.419−2.559−2.600−2.253−2.467−0.959−0.655−0.259−2.329Sparsity0.3210.3080.3080.3720.3210.3330.128 Note: Cells in bold represent key indicators (central nodes in the network). Analysis based on the individual indicators (items) of each instrument (DASS‐21, IES‐R, and BRUMS).John Wiley & Sons, Ltd. Centrality table and plot using standardized node measures (network analysis) (data collection: Phases 1–6 and 2023). Note: Cells in bold represent key indicators (central nodes in the network). Analysis based on the individual indicators (items) of each instrument (DASS‐21, IES‐R, and BRUMS). In the general sample (Table 6), the most prominent indicators from the DASS‐21 were “…panic” (Item 15) and “…down‐hearted and blue” (Item 13); from the IES‐R, “…waves of strong feelings” (Item 16); and from the BRUMS, “exhausted” (Item 8) and “worried” (Item 14). However, notable differences were observed across the time points of data collection. For instance, in the IES‐R, Item 11 (“…I try not to think…”) was central in Phase 1, while Item 9 (“…pictures about it pop into my mind”) played a key role in 2023. Discussion This study reveals a concerning level of subjective distress among university students. About one‐third reported a medical diagnosis of a mental disorder, despite being in an early life stage. Since before COVID‐19, symptoms of depression and stress have increased notably, while anxiety remained consistently high, showing a slow and sustained rise, suggesting possible chronicity. Network analysis identified tension, depressed mood, stress, and hyperarousal as central nodes, likely reflecting broader social determinants such as dissolution of spatial‐temporal boundaries, accelerated lifestyles, reduced face‐to‐face interaction, and overstimulation from digital technologies [1, 13, 23, 24, 58, 59, 60, 61, 62]. These factors may contribute to psychological distress and loneliness. Findings point to the need for institutional mental health initiatives, including structured policies, continuous support programs, and context‐sensitive interventions. Repeated cross‐sectional studies like this one are essential to track trends and guide decision‐making in educational and social settings. In Brazil, new workplace regulations require companies to assess psychosocial risks [63], highlighting the urgency of institutional mental health policies. Although aimed at the private sector, this rationale should extend to public and academic settings. University students, though not formally employed, engage in demanding intellectual work under pressure, making them vulnerable to distress. Our findings support the psychosocial risk assessments in universities and targeted intervention strategies aligned with emerging workplace mental health frameworks. Such efforts may reduce mental health‐related leave, as noted in the 2025 Statistical Yearbook of the Brazilian Ministry of Labour and Employment [64], and promote healthier psychological development by expanding emotional and cognitive resources [65, 66, 67]. The weakening of meaningful interpersonal bonds, lack of listening spaces, and a sense of non‐belonging have made mental health a socially recognized way to interpret and manage distress [62]. When community ties, shared environments, and everyday conversations are eroded, distress becomes more visible, not only as clinical symptoms but also as a means of expressing emotional vulnerability. This dynamic reflects the individualization of social suffering and a search for recognition and care. It calls for reimagining institutional support systems and rebuilding meaningful connections within universities. Our findings suggest that the increasing symptom prevalence cannot be explained solely by COVID‐19 but reflects broader sociocultural changes of the postmodern era [1, 61, 62]. The pandemic acted as a turning point, legitimizing expressions of distress and expanding dialog on mental health. High symptom rates at the pandemic's onset likely reflect abrupt disruptions to routine and adaptation to a threatening context [15, 54]. However, the persistence of symptoms and the consistently high prevalence rates suggest a more complex sociopolitical landscape. Although psychotherapy demand initially rose in the early months of the health crisis, its later decline may indicate barriers such as cost or perceptions of limited benefit [68]. A particularly concerning finding is the medication use among students, ranging from 15.7% to 24.7%. Despite their pharmaceutical training, behavior diverged from theoretical understanding and practical decision‐making, raising concerns about how principles of rational drug use, taught in pharmacy programs, may be overlooked. This pattern aligns with broader critiques of medicalization. Rowland [69] argues that society increasingly treats subjective distress as a problem to be quickly managed, often through medication. Similarly, Maté and Maté [70] and Arboleda [71] argue that contemporary culture pathologizes natural emotional responses to adversity while normalizing subjective distress caused by performance‐driven, emotionally disconnected lifestyles. Our findings reflect this tension and should be interpreted as a collective warning. Despite the clear influence of social determinants, subjective distress is often framed as an individual issue to be silenced. Medication becomes a quick fix, allowing students to maintain productivity without pausing to reflect on their habits or confront their distress [71]. Mood changes, particularly fatigue, increased from Phase 1 to Phase 2 (first to second semester of 2020), coinciding with remote learning implementation. Online education can heighten cognitive overload through screen‐mediated interaction, constant attention demand, overlap of academic and personal tasks, and sensory challenges such as excessive visual stimuli, limited nonverbal cues [13, 58, 59, 60]. Fatigue remained high until 2022, declining only in 2023, possibly due to prolonged remote learning and difficulties readjusting to in‐person classes. Cognitive overload may have impaired self‐regulation, sustaining fatigue even after resuming in‐person activities [13, 58, 59, 60]. Interestingly, vigor rose immediately after returning to campus. Although remote learning served as an important emergency approach during the pandemic and expanded opportunities for pedagogical innovation, it should be viewed as a complementary teaching strategy. Exclusive online teaching risks compromising mental health, well‐being, and engagement, reinforcing the importance of in‐person education. Women's higher likelihood of experiencing psychological symptoms and distress compared to men may be linked to the specific pressures they face in patriarchal societies such as Brazil. These pressures often include the burden of multiple roles, combining professional responsibilities with domestic and caregiving duties. Additionally, the broader sociocultural acceptance of emotional expression among women may facilitate greater self‐recognition and reporting of psychological distress [13, 54, 72]. However, no statistically significant differences were observed between gender and depressive symptoms, intrusive thoughts, or hyperarousal. This suggests the need for a deeper investigation of social determinants that may help explain these findings. For instance, both men and women are currently exposed to similar levels of information overload, academic pressure, uncertainty about the future, and weakened social ties. These shared stressors may contribute to the development and persistence of symptoms such as intrusive thoughts and hyperarousal, regardless of gender [73]. The observed vulnerability among students with a history of psychiatric diagnosis aligns with existing literature [14, 21, 26, 30, 36, 37]. These findings highlight the pressing need for awareness campaigns, mental health literacy programs, and targeted interventions. Institutional policies should be systematically and continuously developed to promote both individual well‐being and a supportive university environment. These initiatives should be embedded in academic life and prioritized alongside teaching, research, and community engagement. Integrating mental health care and well‐being into these core activities of higher education can help foster an institutional culture that values collective care, strengthens social connections, and supports the development of emotional and cognitive skills suited to the demands of contemporary life. This integration has the potential to reduce future individual and societal burdens while promoting healthier and more sustainable academic and professional trajectories. Moreover, such efforts are essential for supporting students' holistic development, preventing or alleviating subjective distress, and strengthening resilience, critical thinking, and adaptability. Therefore, the present study has relevant practical implications within this broader institutional framework, highlighting the need for the implementation of concrete mental health support strategies. One such strategy involves establishing counseling services for students experiencing mental health concerns [74, 75]. Although barriers to access and utilization of these services may exist, evidence suggests that structured university counseling services can effectively reduce psychological distress while improving student well‐being [74, 75]. In addition, the persistence of anxiety, stress, and loneliness across survey waves, including moments before and after the COVID‐19 pandemic, indicates that mental health monitoring should not be restricted to crisis periods. The development of mental health screening programs should rely on validated and reliable instruments capable of identifying students at risk of mental health problems and facilitating early identification of those problems [76]. Such screening approaches should be combined with appropriate referral pathways, including the counseling services described above, to ensure timely support and continuity of care [76, 77]. Although optimal screening intervals for university populations remain heterogeneous and no single consensus has been established, our findings suggest that institutions should define the timing and frequency of mental health screening according to their academic context. In this regard, screenings may be particularly relevant during periods of increased academic demand, such as examination preparation, with mental health screening conducted at least once per year. This study has some limitations that should be considered when interpreting its findings. Although data collection was repeated over time, the cross‐sectional design does not allow for causal inferences. Additionally, while not directly measurable, selection bias is a plausible concern. Students with a greater interest in mental health or higher levels of engagement may have been more likely to participate consistently. Another limitation involves external validity. Since the data were collected from a single Brazilian public university, generalizing the results to other institutions or sociocultural contexts may be limited. Despite these limitations, the repeated cross‐sectional design, with seven to nine data collection points over several years, is a significant strength. This design enables the identification of temporal patterns in subjective distress among university students and supports the development of etiological hypotheses. It also provides information for the planning of preventive, supportive, and intervention strategies in similar educational settings. In a global context marked by the persistent rise in mental health issues among young adults, a trend that was accelerated but not initiated by the COVID‐19 pandemic, the present findings emphasize the urgent need for evidence‐based, context‐sensitive institutional policies. By highlighting recurring patterns of subjective distress, their social determinants, and their persistence over time, this study offers strategic insights for universities worldwide to reassess their approaches to mental health support, prevention, and promotion. Ultimately, this study contributes to the international discourse on mental health care in higher education and supports the development of more integrated, effective, and equitable strategies. Conclusion A high prevalence of psychological symptoms and mental health disorders was observed among students, with a rising trend already emerging before the COVID‐19 pandemic. Loneliness persisted at high levels following the COVID‐19 pandemic, and anxiety symptoms have continued to increase progressively over the years. Logistic regression and network analysis of the symptoms suggest that social determinants play a critical role in shaping both the development and maintenance of mental health conditions. Sustained prevention, education, and intervention strategies within the university setting, especially those targeting the most vulnerable subgroups, are vital to alleviating subjective distress, enhancing students' emotional well‐being, and reducing the incidence of mental disorders throughout their academic journey. Author Contributions Juliana Alvares Duarte Bonini Campos: conceptualization, formal analysis, funding acquisition, project administration, supervision, visualization, writing – original draft. Livia Oliveira Santos: conceptualization, investigation, writing – review and editing. Adrielly Santos: conceptualization, investigation, writing – review and editing. João Marôco: formal analysis, resources, writing – review and editing. Lucas Arrais Campos: conceptualization, data curation, formal analysis, investigation, visualization, writing – original draft. Conflicts of Interest The authors declare no conflicts of interest. Transparency Statement Lucas Arrais Campos affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned have been explained. Supporting information Supporting File 1 Supporting File 2 Supporting File 3 Supporting File 4 Supporting File 5 Supporting File 6 Supporting File 7