Pre-event psychiatric states predict trajectories of post-traumatic stress symptoms during the COVID-19 pandemic

Pre-event psychiatric states predict trajectories of post-traumatic stress symptoms during the COVID-19 pandemic | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Pre-event psychiatric states predict trajectories of post-traumatic stress symptoms during the COVID-19 pandemic Toshinori Chiba, Takatomi Kubo, Fumiya Nakai, Taiki Oka, Nao Kobayashi, and 8 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3865066/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Recent time-dependent analyses of stress-related disorders have identified heterogeneity of trajectories and their modifying factors. While psychiatric patients are vulnerable to stress events, it is unclear how psychiatric conditions in the general population modulate subsequent stress responses. Using our longitudinal online survey from before the COVID-19 pandemic to post-pandemic follow-ups (n = 3815 Japanese adults), here we identified four trajectories of post-traumatic stress symptoms (PTSS) a latent growth mixture model; resilient, chronic, mild chronic, and early response. The depression/anxiety were identified as specific risk factors for the early-response trajectory. In contrast, general psychiatric burden and social withdrawal were identified as common risk/protective factors. Further, we estimated “baseline” PTSS to determine the predictability of the PTSS prognosis from pre-pandemic states. The chronic group showed significantly higher baseline PTSS scores than the mild-chronic and early-response groups, both of which were significantly higher than the resilient group. We concluded that prior psychiatric conditions significantly affect the PTSS trajectory. These results suggest that prior psychiatric conditions may be considered for the prevention and treatment of maladaptive stress responses. Health sciences/Diseases/Trauma Health sciences/Diseases/Psychiatric disorders/Post-traumatic stress disorder Health sciences/Health care/Public health/Population screening Health sciences/Diseases/Psychiatric disorders/Depression Health sciences/Diseases/Psychiatric disorders/Anxiety Figures Figure 1 Figure 2 INTRODUCTION The COVID-19 pandemic has been a global calamity, and it offered a unique test case to investigate psychiatric resilience and effects of stress exposure. 1 Exposure to a stressful/traumatic event can cause severe mental problems, including post-traumatic stress symptoms (PTSS), which were indeed induced by the COVID-19 pandemic. 1,2,3,4,5 Recently, we found the strong temporal specificity of the association between the event-related post-traumatic stress disorder (PTSD) and the suicide rate. 6 The symptoms could thus be considered as reliable surrogate endpoints for increased suicide. 7,8,9,10 Recent research has therefore focused on analyzing trajectories of PTSS after traumatic events. 11,12,13,14,15,16,17 Trajectories of PTSS are usually classified into four types: resilient, chronic, recovery, and delayed-onset. 11, 13, 14, 16 The resilient type shows stable low or no PTSD severity. The chronic type shows consistently high severity. The recovery type shows rapid exacerbation that subsequently improves. The delayed-onset type shows low-level symptoms in the initial phase, which become exacerbated several months after the trauma exposure. 18 Heterogeneity of PTSS trajectories is partly explained by risk/protective factors, such as personality, demographics, and stress-coping strategies. 12,19,20 Psychiatric conditions prior to stress exposure are also among the most important risk factors. 21,22 In a limited number of studies that have dealt with prior psychiatric conditions, the importance of data from before the trauma is emphasized. These studies have identified new prognostic trajectories that could not be identified solely with post-trauma data, highlighting the significance of prior psychiatric conditions. 23 However, comprehending how prior psychiatric conditions affect later stress responses presents a formidable challenge due to the intricate interplay between them. While PTSS is modulated by other psychiatric conditions, those conditions are also exacerbated by stress events. Therefore, in order to overcome this challenge, information on psychiatric conditions must be collected prior to traumatic events, which is usually impossible because traumatized patients visit clinics or hospitals only after they have been traumatized. Here we tackled this problem using our longitudinal data of general mental health before and after the COVID-19 pandemic. The intricate interplay between PTSD and other psychiatric conditions should also be considered in terms of their possible shared mechanisms. PTSD exhibits a substantial co-occurrence with various psychiatric disorders, including depression, anxiety disorders, and substance disorders. In addition, these disorders display highly overlapping genetic foundations. These indicate the presence of shared underlying mechanisms across PTSD and other psychiatric conditions. PTSD symptoms can be decomposed into those attributed to psychiatric states prior to traumatic events and those associated with exacerbation of psychological conditions triggered by the traumatic event. As a proof-of-concept study, we estimated “baseline” PTSS. This was estimated using other psychiatric conditions, which we assume correspond to the former component. In this study, we examined the effects of psychiatric states and demographics before the COVID-19 pandemic on PTSS trajectories during the pandemic. Further, using baseline PTSS, we provide an intuitive interpretation of PTSS prognosis from the baseline. RESULTS We conducted a large online longitudinal survey to investigate mental health of the general Japanese adult population. The original survey was conducted in December 2019 (T0) before the identification of the first COVID-19 case in Japan, and five follow-up surveys of T0 participants were conducted from August 2020 (T1) to December 2021 (T5). The number of participants who contributed to the data at T0 was 3815 (51.3% male and 48.7% female), and the mean age was 46.3 (45.9–46.6 for 95% CI) years (see Table 1 ). Questionnaires included 9 types of validated questionnaires of psychiatric disorders at T0 survey, and additional questionnaires for PTSS to COVID-19 pandemic were appended since T1 survey. The statistical test revealed a significant decrease of PTSS severity from T1 to T2-T5, from T2 to T5 and from T3 to T5 (one-way ANOVA: p < .001, F 4,12594 = 60.69, η 2 = 0.02, post-hoc Tukey test: all p < .001). Latent growth mixture model analysis A latent growth mixture analysis (LGMM) was used to identify the latent class trajectories of PTSS. For the LGMM analysis, all pairs of model functions and numbers of latent classes were tested and evaluated by statistical indicators. As a result, all three indicators (AIC, BIC, and adjusted-BIC) showed that a quadratic function with a four-class model offered the best fit (after removal of classes that had no individuals assigned, see table S2). Consistent with previous studies, 24,25,26 four types of trajectory were named: resilient, chronic, mild chronic, and early response (Fig. 1 ). The resilient group comprised 2712 participants [71.1%], the chronic group consisted of 120 participants [3.2%], the mild-chronic group included 623 participants [16.3%], and the early response group consisted of 360 participants [9.4%]. In the resilient group, a significant decrease was observed at T2 from T1 ( p < .001, t = 5.25, df = 4418) and T5 from T4 ( p < .001, t = 4.02, df = 2805). No significant change was observed in the chronic group at any time pair. In the mild chronic group, a significant decrease was observed at T4 from T3 ( p < .001, t = 4.51, df = 740) and at T5 from T4 ( p < .001, t = 4.77, df = 602). Finally, in the early response group, a significant decrease was observed at T2 from T1 ( p < .001, t = 28.09, df = 579) and at T3 from T2 ( p < .001, t = 5.66, df = 433), but a significant increase was also observed at T5 from T4 ( p < .001, t = 5.19, df = 306). Multinomial logistic regression analysis Psychiatric scores at T0 were compressed by principal component analysis (PCA) and the top 4 components (25.3% for PC1, 13.9% for PC2, 11.2% for PC3, 10.9% for PC4 and 61.2% for all, Figure S1 ) were used in the sequential modifying factor analysis. Identified factors were consistent with the previous study 5 and each PC was named in the same way: PC1 as a general psychiatric burden, PC2 as social withdrawal, PC3 as alcohol-related problems, and PC4 as depression/anxiety. Risk factors for each class were identified by multinomial logistic regression (MLR) with the resilient group as a reference (multiple comparison was corrected by the Bonferroni method, Fig. 2 ). As a result, a general psychiatric burden component (PC1) was identified as a significant risk factor for all groups(chronic: b = 2.16; 95% CI [1.89–2.42], p < .001, t = 15.87, df = 9234, mild chronic: b = 1.27; 95% CI [1.14–1.40], p < .001, t = 19.13, df = 9234, early response: b = 1.24; 95% CI [1.08–1.39], p < .001, t = 15.62, df = 9234). Also, a social withdrawal component (PC2) was identified as a significant protective factor for all groups(chronic: b = -0.70; 95% CI [-0.90-0.49], p < .001, t = -6.57, df = 9234, mild chronic: b = -0.51; 95% CI [-0.62-0.40], p < .001, t = -8.88, df = 9234, early response: b = -0.43; 95% CI [-0.56-0.30], p < .001, t = -6.39, df = 9234). An alcohol-related problem component (PC3) was identified as a significant protective factor for the chronic group( b = -0.30; 95% CI [-0.49-0.10], p < .01, t = -3.02, df = 9234) and the mild chronic group ( b = -0.23; 95% CI [-0.34-0.13], p < .001, t = -4.30, df = 9234). Depression/anxiety component (PC4) was also identified as a significant risk factor for the early response group ( b = 0.23; 95% CI [0.10–0.35], p = .053, t = 3.55, df = 9234). Young age and unmarried status also showed a tendency as a risk factor for the chronic group, but no significant difference was identified. Baseline PTSS estimation In order to provide an intuitive interpretation of PTSS prognosis from the baseline, we estimated “baseline PTSS'', which would have arisen even if the pandemic had not occurred. We also tested the estimation validity of baseline PTSS by using leave-one-time-point-out cross-validation analysis. The mean adjusted- R 2 score of baseline PTSS estimation model showed 0.525 at training and 0.436 at testing. The leave-one-time-point-out cross-validation analysis showed higher than 0.5 adjusted- R 2 score for all Tx conditions at training and higher than 0.4 adjusted- R 2 score for all Tx conditions at test (see details in Supplementally). The baseline PTSS score of the chronic group (31.8; 95% CI [29.5–34.1]) was significantly higher than those of the mild chronic (22.4; 95% CI [21.3–23.6]), early-response (22.8; 95% CI [22.0-23.6]) and resilient groups (10.0; 95% CI [9.7–10.4]), and scores of mild chronic and early-response groups were significantly higher than that of the resilient group (one-way ANOVA, p < .001, F 3,3811 = 512.9, η 2 = 0.29; post-hoc Tukey test chronic versus resilient, mild chronic and early-response p < .001; resilient versus mild chronic, early response p < .001). No significant difference was found between mean scores of the mild chronic and early-response groups. DISCUSSION In this study, we conducted an online longitudinal survey including COVID19-related PTSS over a period of 1.5 years since before the pandemic. Our LGMM trajectory analysis identified four PTSS trajectory types: resilient, chronic, mild chronic, and early response. We then demonstrated how these prognoses are modulated by psychiatric conditions prior to the pandemic. For demonstration purposes, we further estimated “baseline” PTSS before the pandemic, which we assume to correspond to the PTSS score that would have arisen during the time of the pandemic if this psychiatric exacerbation had not occurred. Generally, higher baseline PTSS was associated with higher PTSS during the pandemic, with more pronounced exacerbation in PTSS. Some psychiatric conditions before the pandemic were identified as risk or protective factors for specific trajectory patterns, while others were common to all trajectories. This fact may lead to two possible strategies for prevention and treatment of stress responses, namely, those tailored to an individual trajectory type and those common to all types. The depression/anxiety component was identified as a risk factor specific to the early response group. It may be feasible to assume that PTSS in this type reflects temporal depression and anxiety which resemble a normal stress response in healthy individuals. As with adjustment disorders and acute stress disorders, symptomatic psychotherapy and pharmacotherapy may be more effective than curative pharmacotherapy, such as SSRIs. It may also be effective to allay individuals’ anxieties through psychoeducation that promotes an awareness that the response is common, normal response to stressful or traumatic events. 27,28 In contrast to a specific risk factor, the social withdrawal/general psychiatric burden component was identified as a common protective/risk factor for all trajectories. While most previous studies underlined the adverse effects of quarantine policies on mental health through inhibition of social communication, 29,30 effects of such policies may have varied according to individual social propensity. While individuals with high sociability may have experienced greatly exacerbated status during the quarantine, those with low sociability may have improved their psychological conditions by not having to participate in unwanted social activities in which they were forced to participate before the pandemic. Quarantine policies, thus, should be flexibly set according to individual propensities for social activities. After the COVID-19 pandemic, we should also shift to a lifestyle that accepts diverse social communication styles rather than simply returning to the lifestyle before the pandemic. The general psychiatric burden component was also identified as a common risk factor. This finding aligns with established observations indicating the vulnerability of psychiatric patients to stress. 7,31 From this perspective, those with high general psychiatric burden components may have experienced a more pronounced exacerbation of their psychological states. On the other hand, from the perspective that many psychiatric disorders arise from a single common factor, 32 stress-induced psychiatric symptoms may be attributable to emotional malaise that existed before the event. Individuals may have interpreted their malaise that existed before the pandemic rather than that was induced by the pandemic. In such a scenario, the individual’s holistic problem, rather than the effect of the trigger event, should be resolved to treat PTSS. To identify which perspective better explains the mechanism of PTSS development, it is necessary to separate psychiatric problems existing prior to the event from those exacerbated during the event. As a proof-of-concept, we estimated “baseline“ PTSS. This was calculated using pre-pandemic psychiatric conditions based on the correspondence between PTSS and psychiatric conditions during the pandemic. Here, we assume that baseline PTSS corresponds to symptoms attributed to psychiatric conditions prior to traumatic events. Leave-one-time-out regression suggested that PTSS scores could be estimated from scores of other psychiatric disorders at other time points (mean adjusted-R 2 = 0.411). The ANOVA conducted on the mean baseline PTSS score for trajectory types revealed three levels: high for the chronic group, low for the resilient group, and intermediate for the others. This result suggests a general trend in which higher baseline PTSS levels correspond to higher levels of PTSS during the pandemic. Under our assumptions, our results suggest mechanisms by which stress symptoms evolved, especially in the context of long-lasting stressful events like the COVID-19 pandemic. Higher baseline PTSS levels contribute to an elevation in PTSS during the initial phase of the pandemic. The increased PTSS, serving as the new baseline, further intensified PTSS during the middle phase of the pandemic. This cyclic change of exacerbation continued throughout the pandemic. If this hypothesis is confirmed, it is crucial to make concerted efforts to reduce levels of “baseline” PTSS in the early phase of the pandemic. While precise interpretation of baseline PTSS requires further discussion and clarification, our proposal represents an initial step toward addressing the intricate relationship between stress vulnerability and exacerbations triggered by stress. There are two limitations in our study. First, while most results were consistent with previous studies, the delayed-onset trajectory was not identified in our study. This may be explained by the difference in observation period, interval, and the event type. 12 Note that delayed-onset PTSD would have developed after our observation period (24 months). 18 Second, it is also important to verify whether the results of our study can be generalized to other types of traumatic experience; therefore, it is necessary to apply the same analysis to other cases. In this study, we identified four types of stress response with different modifying factors. In addition, we also found that these four trajectories were partly predictable, even before the onset of the traumatic event. These findings may help to develop personalized-care according to the psychiatric states underlying their pathogenesis. Declarations Authors Contributions : FN, TK, and TC had full access to all of the data in the study and take responsibility for the integrity of the work. Concept and design : FN, TK, TO, TC. Data acquisition : TO, TH, NK, MH, MM, TC. Statistical analyses : FN, TK, TO, TC. Interpretation of data : FN, TK, TO, AC, TC. Drafting the first version of the manuscript : FN, TK, AC, TC. Revised and approved the final version of the manuscript : All authors. Conflicts of Interest : NK, TH, MH, and YM are employees of KDDI Research Inc. NK, TH, MH, and YM were involved in the data acquisition and revised the manuscript. For the remaining authors, there were no conflicts of interest. Funding/ Support : This research was supported by a KDDI collaborative research contract. This work was also supported by AMED under Grant Number JP20dm0307008. Role of Funder/Sponsor : The sponsor, KDDI, reviewed the design, the collection, and the analysis. The funder had no role in the conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. Disclaimers : All information and materials are original. Data Sharing : The main summary statistical data that support the findings of this study are available in the Supplementary Data. Owing to company cohort data sharing restrictions, individual-level data cannot be publicly posted. However, data are available from the authors upon request and with permission of KDDI Corporation. Additional Contributions : We thank Rumi Yorizawa (The Department of Decoded Neurofeedback, Computational Neuroscience Laboratories, Advanced Telecommunications Research Institute International, Kyoto), and Miho Nagata (The Department of Decoded Neurofeedback, Computational Neuroscience Laboratories, Advanced Telecommunications Research Institute International, Kyoto) for data collection and organization. References Barzilay R, Moore TM, Greenberg DM, DiDomenico GE, Brown LA, White LK, Gur RC, Gur RE. Resilience, COVID-19-related stress, anxiety and depression during the pandemic in a large population enriched for healthcare providers. Transl Psychiatry. 2020;10(1):291. doi: 10.1038/s41398-020-00982-4 . PMID: 32820171; PMCID: PMC7439246. Brülhart M, Klotzbücher V, Lalive R, Reich SK. Mental health concerns during the COVID-19 pandemic as revealed by helpline calls. Nature. 2021;600(7887):121–126. doi: 10.1038/s41586-021-04099-6 . Epub 2021 Nov 17. PMID: 34789873; PMCID: PMC9973557. Pierce M, McManus S, Hope H, Hotopf M, Ford T, Hatch SL, John A, Kontopantelis E, Webb RT, Wessely S, Abel KM. Mental health responses to the COVID-19 pandemic: a latent class trajectory analysis using longitudinal UK data. Lancet Psychiatry. 2021;8(7):610–619. doi: 10.1016/S2215-0366(21)00151-6 . Epub 2021 May 6. PMID: 33965057; PMCID: PMC9764381. Oka T, Hamamura T, Miyake Y, Kobayashi N, Honjo M, Kawato M, Kubo T, Chiba T. Prevalence and risk factors of internet gaming disorder and problematic internet use before and during the COVID-19 pandemic: A large online survey of Japanese adults. J Psychiatr Res. 2021;142:218–225. doi: 10.1016/j.jpsychires.2021.07.054 . Epub 2021 Aug 4. Erratum in: J Psychiatr Res. 2022;148:275–276. PMID: 34385071. Oka T, Kubo T, Kobayashi N, Nakai F, Miyake Y, Hamamura T, Honjo M, Toda H, Boku S, Kanazawa T, Nagamine M, Cortese A, Takebayashi M, Kawato M, Chiba T. Multiple time measurements of multidimensional psychiatric states from immediately before the COVID-19 pandemic to one year later: a longitudinal online survey of the Japanese population. Transl Psychiatry. 2021;11(1):573. doi: 10.1038/s41398-021-01696-x . PMID: 34759293; PMCID: PMC8581018. Chiba T, Ide K, Murakami M, Kobayashi N, Oka T, Nakai F, Yorizawa R, Miyake Y, Hamamura T, Honjo M, Toda H, Kanazawa T, Boku S, Kubo T, Hishimoto A, Kawato M, Cortese A. Event-related PTSD symptoms as a high-risk factor for suicide—Longitudinal observational study. Nature Mental Health. 2023;1(12):1013–22.. Harris TB, Carlisle LL, Sargent J, Primm AB. Trauma and diverse child populations. Child Adolesc Psychiatr Clin N Am. 2010;19(4):869 – 87. doi: 10.1016/j.chc.2010.07.007 . PMID: 21056351. Brown GW, Bifulco A, Harris T, Bridge L. Life stress, chronic subclinical symptoms and vulnerability to clinical depression. J Affect Disord. 1986 Jul-Aug;11(1):1–19. doi: 10.1016/0165-0327(86)90054-6 . PMID: 2944923. Judkins JL, Moore BA, Collette TL, Hale WJ, Peterson AL, Morissette SB. Incidence Rates of Posttraumatic Stress Disorder Over a 17-Year Period in Active Duty Military Service Members. J Trauma Stress. 2020;33(6):994–1006. doi: 10.1002/jts.22558 . Epub 2020 Jun 29. PMID: 32598575. Nagamine M, Giltay EJ, Shigemura J, van der Wee NJ, Yamamoto T, Takahashi Y, Saito T, Tanichi M, Koga M, Toda H, Shimizu K, Yoshino A, Vermetten E. Assessment of Factors Associated With Long-term Posttraumatic Stress Symptoms Among 56 388 First Responders After the 2011 Great East Japan Earthquake. JAMA Netw Open. 2020;3(9):e2018339. doi: 10.1001/jamanetworkopen.2020.18339 . PMID: 32990742; PMCID: PMC7525349. Andersen SB, Karstoft KI, Bertelsen M, Madsen T. Latent trajectories of trauma symptoms and resilience: the 3-year longitudinal prospective USPER study of Danish veterans deployed in Afghanistan. J Clin Psychiatry. 2014;75(9):1001-8. doi: 10.4088/JCP.13m08914 . PMID: 25188895. Bonanno GA, Mancini AD. Beyond resilience and PTSD: Mapping the heterogeneity of responses to potential trauma. Psychological trauma: Theory, research, practice, and policy. 2012;4(1):74. doi: 10.1037/a0017829 Bonanno GA, Mancini AD, Horton JL, Powell TM, Leardmann CA, Boyko EJ, Wells TS, Hooper TI, Gackstetter GD, Smith TC; Millennium Cohort Study Team. Trajectories of trauma symptoms and resilience in deployed U.S. military service members: prospective cohort study. Br J Psychiatry. 2012;200(4):317 – 23. doi: 10.1192/bjp.bp.111.096552 . Epub 2012 Feb 23. PMID: 22361018. Dickstein BD, Suvak M, Litz BT, Adler AB. Heterogeneity in the course of posttraumatic stress disorder: trajectories of symptomatology. J Trauma Stress. 2010;23(3):331-9. doi: 10.1002/jts.20523 . PMID: 20564365. Orcutt HK, Erickson DJ, Wolfe J. The course of PTSD symptoms among Gulf War veterans: a growth mixture modeling approach. J Trauma Stress. 2004;17(3):195–202. doi: 10.1023/B:JOTS.0000029262.42865.c2 . PMID: 15253091. Karstoft KI, Armour C, Elklit A, Solomon Z. The role of locus of control and coping style in predicting longitudinal PTSD-trajectories after combat exposure. J Anxiety Disord. 2015;32:89–94. doi: 10.1016/j.janxdis.2015.03.007. Epub 2015 Mar 28. PMID: 25890288. Saito T, van der Does FHS, Nagamine M, van der Wee NJ, Shigemura J, Yamamoto T, Takahashi Y, Koga M, Toda H, Yoshino A, Vermetten E, Giltay EJ. Risk and resilience in trajectories of post-traumatic stress symptoms among first responders after the 2011 Great East Japan Earthquake: 7-year prospective cohort study. Br J Psychiatry. 2022 Feb 22:1–8. doi: 10.1192/bjp.2022.2. Epub ahead of print. PMID: 35191369. Andrews B, Brewin CR, Philpott R, Stewart L. Delayed-onset posttraumatic stress disorder: a systematic review of the evidence. Am J Psychiatry. 2007;164(9):1319-26. doi: 10.1176/appi.ajp.2007.06091491 . PMID: 17728415. Dohrenwend BP, Turner JB, Turse NA, Adams BG, Koenen KC, Marshall R. The psychological risks of Vietnam for U.S. veterans: a revisit with new data and methods. Science. 2006;313(5789):979–82. doi: 10.1126/science.1128944 . PMID: 16917066; PMCID: PMC1584215. Pan L, Xu Q, Kuang X, Zhang X, Fang F, Gui L, Li M, Tefsen B, Zha L, Liu H. Prevalence and factors associated with post-traumatic stress disorder in healthcare workers exposed to COVID-19 in Wuhan, China: a cross-sectional survey. BMC Psychiatry. 2021;21(1):572. doi: 10.1186/s12888-021-03589-1 . PMID: 34781901; PMCID: PMC8591965. Bomyea J, Risbrough V, Lang AJ. A consideration of select pre-trauma factors as key vulnerabilities in PTSD. Clin Psychol Rev. 2012;32(7):630–41. doi: 10.1016/j.cpr.2012.06.008 . Epub 2012 Jul 20. PMID: 22917742; PMCID: PMC3444146. Bonanno GA, Chen S, Galatzer-Levy IR. Resilience to potential trauma and adversity through regulatory flexibility. Nature Reviews Psychology. 2023 Sep 19:1–3. Steenkamp MM, Dickstein BD, Salters-Pedneault K, Hofmann SG, Litz BT. Trajectories of PTSD symptoms following sexual assault: is resilience the modal outcome? J Trauma Stress. 2012;25(4):469–74. doi: 10.1002/jts.21718 . Epub 2012 Jul 13. PMID: 22807251. deRoon-Cassini TA, Mancini AD, Rusch MD, Bonanno GA. Psychopathology and resilience following traumatic injury: a latent growth mixture model analysis. Rehabil Psychol. 2010;55(1):1–11. doi: 10.1037/a0018601 . PMID: 20175629. Murphy D, Smith KV. Treatment Efficacy for Veterans With Posttraumatic Stress Disorder: Latent Class Trajectories of Treatment Response and Their Predictors. J Trauma Stress. 2018;31(5):753–763. doi: 10.1002/jts.22333 . Epub 2018 Oct 19. PMID: 30338573; PMCID: PMC6456014. Cohen JA, Mannarino AP. Trauma-focused Cognitive Behavior Therapy for Traumatized Children and Families. Child Adolesc Psychiatr Clin N Am. 2015;24(3):557–70. doi: 10.1016/j.chc.2015.02.005 . Epub 2015 Apr 22. PMID: 26092739; PMCID: PMC4476061. Bersani FS, Delle Chiaie R. The end method: Normalization. Empathy, Normalization and De-escalation: Management of the Agitated Patient in Emergency and Critical Situations. 2021:57–64. van der Velden PG, Hyland P, Contino C, von Gaudecker HM, Muffels R, Das M. Anxiety and depression symptoms, the recovery from symptoms, and loneliness before and after the COVID-19 outbreak among the general population: Findings from a Dutch population-based longitudinal study. PLoS One. 2021;16(1):e0245057. doi: 10.1371/journal.pone.0245057 . PMID: 33411843; PMCID: PMC7790276. Brunoni AR, Suen PJ, Bacchi PS, Razza LB, Klein I, Dos Santos LA, de Souza Santos I, Valiengo LD, Gallucci-Neto J, Moreno ML, Pinto BS. Prevalence and risk factors of psychiatric symptoms and diagnoses before and during the COVID-19 pandemic: findings from the ELSA-Brasil COVID-19 mental health cohort. Psychological medicine. 2023;53(2):446–57. Goh C, Agius M. The stress-vulnerability model how does stress impact on mental illness at the level of the brain and what are the consequences? Psychiatr Danub. 2010;22(2):198–202. PMID: 20562747. Caspi A, Moffitt TE. All for One and One for All: Mental Disorders in One Dimension. Am J Psychiatry. 2018;175(9):831–844. doi: 10.1176/appi.ajp.2018.17121383 . Epub 2018 Apr 6. PMID: 29621902; PMCID: PMC6120790. Chiba T, Oka T, Hamamura T, Kobayashi N, Honjo M, Miyake Y, Kubo T, Toda H, Kanazawa T, Boku S, Hishimoto A. PTSD symptoms related to COVID-19 as a high risk factor for suicide-Key to prevention. medRxiv. 2020 Dec 18:2020–12. doi: 10.1101/2020.12.15.20246819 . Weiss DS. The impact of event scale: revised. Cross-cultural assessment of psychological trauma and PTSD. 2007:219 – 38. Radloff LS. The CES-D scale: A self-report depression scale for research in the general population. Applied psychological measurement. 1977;1(3):385–401. Foa EB, Kozak MJ, Salkovskis PM, Coles ME, Amir N. The validation of a new obsessive–compulsive disorder scale: The Obsessive–Compulsive Inventory. Psychological assessment. 1998;10(3):206. Meerkerk GJ, Van Den Eijnden RJ, Vermulst AA, Garretsen HF. The compulsive internet use scale (CIUS): some psychometric properties. Cyberpsychology & behavior. 2009;12(1):1–6. Kessler RC, Adler L, Ames M, Demler O, Faraone S, Hiripi EV, Howes MJ, Jin R, Secnik K, Spencer T, Ustun TB. The World Health Organization Adult ADHD Self-Report Scale (ASRS): a short screening scale for use in the general population. Psychological medicine. 2005;35(2):245–56. Baron-Cohen S, Wheelwright S, Skinner R, Martin J, Clubley E. The autism-spectrum quotient (AQ): Evidence from asperger syndrome/high-functioning autism, malesand females, scientists and mathematicians. Journal of autism and developmental disorders. 2001;31:5–17. Baker SL, Heinrichs N, Kim HJ, Hofmann SG. The Liebowitz social anxiety scale as a self-report instrument: a preliminary psychometric analysis. Behaviour research and therapy. 2002;40(6):701–15. Spielberger CD. State-Trait Anxiety Inventory for Adults. 1983. DOI: 10.1037/t06496-000 . Saunders JB, Aasland OG, Babor TF, De la Fuente JR, Grant M. Development of the alcohol use disorders identification test (AUDIT): WHO collaborative project on early detection of persons with harmful alcohol consumption-II. Addiction. 1993;88(6):791–804. Muthén B. Latent variable analysis. The Sage handbook of quantitative methodology for the social sciences. 2004;345(368):106-9. Muthén B, Muthén LK. Integrating person-centered and variable-centered analyses: growth mixture modeling with latent trajectory classes. Alcohol Clin Exp Res. 2000;24(6):882–91. PMID: 10888079. Proust-Lima C, Philipps V, Diakite A, Liquet B, Proust MC. Package ‘lcmm’. CRAN R. 2023 Feb 20. Wardenaar K. Latent Class Growth Analysis and Growth Mixture Modeling using R: A tutorial for two R-packages and a comparison with Mplus. Nylund KL, Asparouhov T, Muthén BO. Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural equation modeling: A multidisciplinary Journal. 2007;14(4):535–69. doi: 10.1080/10705510701575396 . Wold S, Esbensen K, Geladi P. Principal component analysis. Chemometrics and intelligent laboratory systems. 1987;2(1–3):37–52. doi: 10.1016/0169-7439(87)80084-9 . Vapnik V, Golowich S, Smola A. Support vector method for function approximation, regression estimation and signal processing. Advances in neural information processing systems. 1996;9. Bishop CM, Nasrabadi NM. Pattern recognition and machine learning. New York: springer; 2006 Aug 17. Methods Procedures and outcomes The online survey was conducted by Macromill Inc.(Japan) as a large, longitudinal survey to investigate mental health of the general Japanese adult population (see our previous studies 4,5,33 for details). The original panel survey was conducted in December 2019 (T0) before identification of the first COVID-19 case in Japan (January 2020). In response to the outbreak of COVID-19, follow-up surveys of T0 participants were conducted in August 2020 (T1), December 2020 (T2), April 2021 (T3), August 2021 (T4), and December 2021 (T5). Invitations for the original survey were sent to 5955 individuals, of which 478 were excluded due to inconsistencies or contradictions in their answers. An additional 481 individuals were also excluded because of unreliable answers. They answered using only the maximum or minimum answers in questionnaires including reverse items. The original and follow-up research designs were approved by the Ethics Committees of the Advanced Telecommunications Research Institute International (ATR) (approval No. 21–195 for the original study & 21–749 for the follow-up study). All participants read a full explanation of the study and gave informed consent before each survey. Questionnaires were constructed from questions for psychiatric, demographic and COVID-19 related items. Psychiatric questions included 10 types of validated questionnaires for psychiatric disorders; PTSS (IES-R 34 ), major depressive disorder (CES-D 35 ), obsessive-compulsive disorder (OCI 36 ), internet-related problems (CIUS 37 ), attention-deficit/hyperactivity disorder (ADHD) (ASRS 38 ), autistic spectrum disorder (AQ 39 ), social anxiety (LSAS-fear/avoid 40 ), general anxiety (STAI-Y-state 41 ), and alcohol-related problems (AUDIT 42 ). Collected demographic data included sex (women and men), age, job status (self-employed, employed, unemployed and other), marital state and household income per year (lowest; less than four million yen, 2nd; four ~ six million yen, 3rd; six ~ eight million yen, 4th; eight ~ ten million yen, Highest; more than ten million yen, and Missing) and they were analyzed as pre-COVID demographics. The temporal change of PTSS score was tested by one-way ANOVA with post-hoc Tukey-HSD test. Trajectory analysis using the Latent Growth Mixture Model In this study, we assumed that the global heterogeneity of PTSS trajectories can be explained by a set of homogeneous trajectories. Consistent with previous studies, the latent growth mixture model 43,44 (LGMM) was used to identify latent class trajectories of PTSS measured by IES-R from T1 to T5, during the COVID-19 pandemic. The “ lcmm ” package 45 in R version. 4.1.0 was used to identify the latent class. 46 The classification was performed under all parameter conditions (6×3: the number of latent classes (1 to 6 class) × the model function (linear, quadratic and exponential)). The optimal condition was determined according to the indicators. Commensurate with recommendations in the previous study 47 , we relied on three types of indicators to determine the best model for clustering; AIC (Akaike information criterion), BIC (Bayesian information criterion), sample-size-adjusted BIC. A grid search approach with 100 iterations was used to estimate optimal values of model parameters and 100 repetitions were applied to achieve stable results. After repetition, the optimal parameter was determined using the maximum likelihood method. After estimating optimal model parameters, the membership probability was calculated for each participant to assign a class label. Each trajectory was also analyzed by unpaired t-test to confirm the temporal change (multi-comparison was adjusted using the Bonferroni method). Multinomial logistic regression analysis We used a multinomial logistic regression model to identify risk/protective factors for each trajectory class. Prior to apply the model, the nine psychiatric scores of all participants were compressed into four dimensional data using principal component analysis (PCA) on T0 data. 5,48 Then, according to the estimated loading of the top four PCs, orthogonal transformation was applied and scores of each participant at each time point were converted into four-dimensional orthogonal scores.Then, logistic regression was performed using class assignment as a dependent variable and compressed psychiatric scores at T0 and other demographics (sex, age, income, employment status, and marital status) as independent variables. The relative risk b from the resilient group for each independent variable was estimated for each trajectory group, and p-values were adjusted by Bonferroni correction. We fitted the model using the Statistical Machine Learning Toolbox in matlab . Baseline PTSS estimation We estimated “baseline PTSS" by using multiple psychiatric scores collected prior to COVID-19 in multiple ridge regression analysis. 49,50 First, we analyzed the linear association of PTSS score with nine other psychiatric scores at T1 to T5 during the pandemic. All data from T1 to T5 were randomly split in half to create training and test sets. To obtain reliable results, this random sampling was repeated 100 times. For each data set, both training and test data were normalized based on the mean and variance of the training dataset. Training data were further divided into training data and validation data with a 10-fold cross validation approach to tune the regularization parameter. Specifically, the optimal penalty coefficient was searched from 0 to 1 for every 0.001. In each cross-validation, regression coefficients and the penalty coefficient were determined based on minimization of the cost function that was estimated with the validation dataset. Successively, for each repetition, optimal parameters were selected based on minimization of the cost function among the folds of cross-validation. Each model was evaluated using the test dataset based on the adjusted- R 2 metrics and the model for the baseline PTSS estimation was evaluated by the mean score of adjusted- R 2 metrics for all models. Averages of the coefficients were used later in estimation of baseline PTSS. Second, we evaluated the applicability of the model across time points with a similar approach. Specifically, we performed leave-one-time-point-out cross validation. In this validation, the model was trained on scores at four time points, except fixed time point Tx (selected between T1 to T5), and the scores at Tx were used as test data. Finally, based on the estimated linear association, baseline PTSS scores of each subject were calculated from nine other psychiatric scores at T0. The difference in the mean baseline PTSS score for each group was tested by one-way ANOVA, with post-hoc Tukey test. Table 1 Table 1 is available in the Supplementary Files section. Additional Declarations Yes there is potential Competing Interest. The sponsor, KDDI, reviewed the design, the collection, and the analysis. The funder had no role in the conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. Supplementary Files floatimage3.png Table 1. Demographics of the survey. PTSStrajnatureMHSupplementally.docx Supplemental Figures and Tables Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-3865066","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":269106122,"identity":"e946fc2d-6aa6-4c03-a0c3-d6f7209a9cbc","order_by":0,"name":"Toshinori Chiba","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA4UlEQVRIiWNgGAWjYHACNobEBgYefmbmgw+APB4+Qup5oFrkJNvbkg1AAmxEaWFsYDA2OHNGTQJiKQFgz8D87MHDHXaJG27ksFV+zbGTYWNgfvjoBl5b2MwNEs8kJ868kXvstuy2ZKDD2IyNc/Bq4WGTSGxjTuy7kZd2W3IbM1ALD5s0EVrqExtu5JgVS26rJ1rLYWOBM2fMGD9uO0yElsNsZhKJZ46DA1macdtxHjZmAn5hb29+JvlzRzU4Kj/+3FZtz8/e/PAxPi0MzMhsHnQRgoDxBymqR8EoGAWjYMQAAGjSQp0vsMR0AAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0002-2173-4519","institution":"ATR Computational Neurosciences Laboratories","correspondingAuthor":true,"prefix":"","firstName":"Toshinori","middleName":"","lastName":"Chiba","suffix":""},{"id":269106123,"identity":"00e0c708-ea8e-4865-afb3-717fcb204265","order_by":1,"name":"Takatomi Kubo","email":"","orcid":"https://orcid.org/0000-0003-0192-7177","institution":"ATR Computational Neurosciences Laboratories","correspondingAuthor":false,"prefix":"","firstName":"Takatomi","middleName":"","lastName":"Kubo","suffix":""},{"id":269106124,"identity":"7f9687f4-acd7-4ea1-913e-b5641fd72573","order_by":2,"name":"Fumiya Nakai","email":"","orcid":"","institution":"ATR Computational Neurosciences Laboratories","correspondingAuthor":false,"prefix":"","firstName":"Fumiya","middleName":"","lastName":"Nakai","suffix":""},{"id":269106125,"identity":"1054f945-d138-4d3f-a9bc-007df9efff3b","order_by":3,"name":"Taiki Oka","email":"","orcid":"","institution":"ATR Computational Neurosciences Laboratories","correspondingAuthor":false,"prefix":"","firstName":"Taiki","middleName":"","lastName":"Oka","suffix":""},{"id":269106126,"identity":"eb40f100-e9e4-41f1-b6be-e30d0b6d9388","order_by":4,"name":"Nao Kobayashi","email":"","orcid":"","institution":"Neuro Science Project, AI Division, KDDI Research, Inc.","correspondingAuthor":false,"prefix":"","firstName":"Nao","middleName":"","lastName":"Kobayashi","suffix":""},{"id":269106127,"identity":"14570a0b-c418-471a-b299-b6666ebd381e","order_by":5,"name":"Masaaki Tanichi","email":"","orcid":"","institution":"Self-Defense Forces Hanshin Hospital","correspondingAuthor":false,"prefix":"","firstName":"Masaaki","middleName":"","lastName":"Tanichi","suffix":""},{"id":269106128,"identity":"baa0604f-3e00-4b38-b076-5a42ed50ef8d","order_by":6,"name":"Misa Murakami","email":"","orcid":"","institution":"ATR Computational Neurosciences Laboratories","correspondingAuthor":false,"prefix":"","firstName":"Misa","middleName":"","lastName":"Murakami","suffix":""},{"id":269106129,"identity":"4cfcc162-9ca5-4abc-85bb-98c3c101063d","order_by":7,"name":"Toshitaka Hamamura","email":"","orcid":"","institution":"KDDI Research, Inc.","correspondingAuthor":false,"prefix":"","firstName":"Toshitaka","middleName":"","lastName":"Hamamura","suffix":""},{"id":269106130,"identity":"79027359-2a35-4c62-86e6-1415706a760c","order_by":8,"name":"Masaru Honjo","email":"","orcid":"","institution":"Healthcare Medical Group, KDDI Research, Inc.","correspondingAuthor":false,"prefix":"","firstName":"Masaru","middleName":"","lastName":"Honjo","suffix":""},{"id":269106131,"identity":"28c30ad5-06c9-4f73-b04a-db6ed163850b","order_by":9,"name":"Yuka Miyake","email":"","orcid":"","institution":"Neuro Science Project, AI Division, KDDI Research, Inc.","correspondingAuthor":false,"prefix":"","firstName":"Yuka","middleName":"","lastName":"Miyake","suffix":""},{"id":269106132,"identity":"99cfcb11-f221-4b54-b371-afb5be8f3d2f","order_by":10,"name":"Kentarou Ide","email":"","orcid":"","institution":"ATR Computational Neurosciences Laboratories","correspondingAuthor":false,"prefix":"","firstName":"Kentarou","middleName":"","lastName":"Ide","suffix":""},{"id":269106133,"identity":"28836295-d490-47b1-b87e-08f6f4e4791d","order_by":11,"name":"Aurelio Cortese","email":"","orcid":"https://orcid.org/0000-0003-4567-0924","institution":"ATR Institute International","correspondingAuthor":false,"prefix":"","firstName":"Aurelio","middleName":"","lastName":"Cortese","suffix":""},{"id":269106134,"identity":"2366b104-f788-443a-8803-b6d62aa9be10","order_by":12,"name":"Masanori Nagamine","email":"","orcid":"","institution":"National Defense Medical College Research Institute","correspondingAuthor":false,"prefix":"","firstName":"Masanori","middleName":"","lastName":"Nagamine","suffix":""}],"badges":[],"createdAt":"2024-01-15 01:35:35","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-3865066/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3865066/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":50225569,"identity":"8b68e30c-f46f-4157-87c4-14d410746c86","added_by":"auto","created_at":"2024-01-26 17:52:57","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":290364,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe trajectories of PTSS scores for each class and PC trajectories for each class\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea.\u003c/strong\u003eThick lines show mean PTSS trajectories of the 4 groups estimated by LGMM, and thin lines show their 95% confidence intervals. PTSS scores were normalized at T1, and T0 scores were estimated by Ridge regression. Each time transition was tested by unpaired t-test ( * shows significant difference with Bonferroni correction (n=5) ). The blue bars represent a moving average in back-and-forth seven days of new daily cases of COVID-19 per 100,000 Japanese residents.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eb.\u003c/strong\u003ePopulation ratio of four estimated groups.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ec.\u003c/strong\u003eThick lines show mean trajectories of psychiatric items estimated by PCA for each of the four groups, and thin lines show their 95% confidence intervals. Each time transition was tested by ANOVA ( * shows significant difference adjusted by Bonferroni correction (n=5) ).\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-3865066/v1/c65088c37a12c7821bb4e525.png"},{"id":50225571,"identity":"51f067b0-48cb-4c3a-a1d1-f3553e5da265","added_by":"auto","created_at":"2024-01-26 17:52:57","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":113851,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMultinomial logistic regression analysis showing risk/protective factors for each PTSS class\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMultinomial logistic regression analysis showing risk/protective factors for each PTSS class. Psychiatrics items were estimated by PCA of the scores of nine other psychiatric symptoms. Demographic items were selected as pre-existing demographics before the trigger event of PTSS. Each dot shows coefficients of the MLR and the line shows the 95%-ile. Scores of the resilient group were used as a reference and * shows a significant difference from the reference (adjusted by Bonferroni correction (n=3) ).\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-3865066/v1/f31b21e4820c4f58d49f4d06.png"},{"id":56273291,"identity":"af839f65-3bce-4666-bf22-519552137ec0","added_by":"auto","created_at":"2024-05-10 18:51:42","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":901914,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3865066/v1/091f2ef9-dda2-4d77-b358-d7c7d15d30cb.pdf"},{"id":50226183,"identity":"6899f709-2d0e-4b5f-ba5a-eb895348ffd3","added_by":"auto","created_at":"2024-01-26 18:00:57","extension":"png","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":27787,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTable 1.\u003c/strong\u003e \u003cstrong\u003eDemographics of the survey.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-3865066/v1/86e159256ea68e915837cb80.png"},{"id":50225572,"identity":"8c562de0-d07f-4f8e-b3e6-00d33da3f409","added_by":"auto","created_at":"2024-01-26 17:52:58","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":250252,"visible":true,"origin":"","legend":"\u003cp\u003eSupplemental Figures and Tables\u003c/p\u003e","description":"","filename":"PTSStrajnatureMHSupplementally.docx","url":"https://assets-eu.researchsquare.com/files/rs-3865066/v1/6075a4f9c8bea25c8d276f3e.docx"}],"financialInterests":"\u003cb\u003eYes\u003c/b\u003e there is potential Competing Interest.\nThe sponsor, KDDI, reviewed the design, the collection, and the analysis. The funder had no role in the conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.","formattedTitle":"Pre-event psychiatric states predict trajectories of post-traumatic stress symptoms during the COVID-19 pandemic","fulltext":[{"header":"INTRODUCTION","content":"\u003cp\u003eThe COVID-19 pandemic has been a global calamity, and it offered a unique test case to investigate psychiatric resilience and effects of stress exposure.\u003csup\u003e1\u003c/sup\u003e Exposure to a stressful/traumatic event can cause severe mental problems, including post-traumatic stress symptoms (PTSS), which were indeed induced by the COVID-19 pandemic.\u003csup\u003e1,2,3,4,5\u003c/sup\u003e Recently, we found the strong temporal specificity of the association between the event-related post-traumatic stress disorder (PTSD) and the suicide rate.\u003csup\u003e6\u003c/sup\u003e The symptoms could thus be considered as reliable surrogate endpoints for increased suicide.\u003csup\u003e7,8,9,10\u003c/sup\u003e Recent research has therefore focused on analyzing trajectories of PTSS after traumatic events.\u003csup\u003e11,12,13,14,15,16,17\u003c/sup\u003e Trajectories of PTSS are usually classified into four types: resilient, chronic, recovery, and delayed-onset.\u003csup\u003e11, 13, 14, 16\u003c/sup\u003e The resilient type shows stable low or no PTSD severity. The chronic type shows consistently high severity. The recovery type shows rapid exacerbation that subsequently improves. The delayed-onset type shows low-level symptoms in the initial phase, which become exacerbated several months after the trauma exposure.\u003csup\u003e18\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eHeterogeneity of PTSS trajectories is partly explained by risk/protective factors, such as personality, demographics, and stress-coping strategies.\u003csup\u003e12,19,20\u003c/sup\u003e Psychiatric conditions prior to stress exposure are also among the most important risk factors.\u003csup\u003e21,22\u003c/sup\u003e In a limited number of studies that have dealt with prior psychiatric conditions, the importance of data from before the trauma is emphasized. These studies have identified new prognostic trajectories that could not be identified solely with post-trauma data, highlighting the significance of prior psychiatric conditions.\u003csup\u003e23\u003c/sup\u003e However, comprehending how prior psychiatric conditions affect later stress responses presents a formidable challenge due to the intricate interplay between them. While PTSS is modulated by other psychiatric conditions, those conditions are also exacerbated by stress events. Therefore, in order to overcome this challenge, information on psychiatric conditions must be collected prior to traumatic events, which is usually impossible because traumatized patients visit clinics or hospitals only after they have been traumatized. Here we tackled this problem using our longitudinal data of general mental health before and after the COVID-19 pandemic.\u003c/p\u003e \u003cp\u003eThe intricate interplay between PTSD and other psychiatric conditions should also be considered in terms of their possible shared mechanisms. PTSD exhibits a substantial co-occurrence with various psychiatric disorders, including depression, anxiety disorders, and substance disorders. In addition, these disorders display highly overlapping genetic foundations. These indicate the presence of shared underlying mechanisms across PTSD and other psychiatric conditions. PTSD symptoms can be decomposed into those attributed to psychiatric states prior to traumatic events and those associated with exacerbation of psychological conditions triggered by the traumatic event. As a proof-of-concept study, we estimated \u0026ldquo;baseline\u0026rdquo; PTSS. This was estimated using other psychiatric conditions, which we assume correspond to the former component.\u003c/p\u003e \u003cp\u003eIn this study, we examined the effects of psychiatric states and demographics before the COVID-19 pandemic on PTSS trajectories during the pandemic. Further, using baseline PTSS, we provide an intuitive interpretation of PTSS prognosis from the baseline.\u003c/p\u003e"},{"header":"RESULTS","content":"\u003cp\u003eWe conducted a large online longitudinal survey to investigate mental health of the general Japanese adult population. The original survey was conducted in December 2019 (T0) before the identification of the first COVID-19 case in Japan, and five follow-up surveys of T0 participants were conducted from August 2020 (T1) to December 2021 (T5). The number of participants who contributed to the data at T0 was 3815 (51.3% male and 48.7% female), and the mean age was 46.3 (45.9\u0026ndash;46.6 for 95% CI) years (see Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Questionnaires included 9 types of validated questionnaires of psychiatric disorders at T0 survey, and additional questionnaires for PTSS to COVID-19 pandemic were appended since T1 survey. The statistical test revealed a significant decrease of PTSS severity from T1 to T2-T5, from T2 to T5 and from T3 to T5 (one-way ANOVA: \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003eF\u003c/em\u003e\u003csub\u003e4,12594\u003c/sub\u003e = 60.69, \u003cem\u003eη\u003c/em\u003e\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.02, post-hoc Tukey test: all \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001).\u003c/p\u003e\n\u003ch3\u003eLatent growth mixture model analysis\u003c/h3\u003e\n\u003cp\u003eA latent growth mixture analysis (LGMM) was used to identify the latent class trajectories of PTSS. For the LGMM analysis, all pairs of model functions and numbers of latent classes were tested and evaluated by statistical indicators. As a result, all three indicators (AIC, BIC, and adjusted-BIC) showed that a quadratic function with a four-class model offered the best fit (after removal of classes that had no individuals assigned, see table S2). Consistent with previous studies,\u003csup\u003e24,25,26\u003c/sup\u003e four types of trajectory were named: resilient, chronic, mild chronic, and early response (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The resilient group comprised 2712 participants [71.1%], the chronic group consisted of 120 participants [3.2%], the mild-chronic group included 623 participants [16.3%], and the early response group consisted of 360 participants [9.4%]. In the resilient group, a significant decrease was observed at T2 from T1 (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003et\u003c/em\u003e\u0026thinsp;=\u0026thinsp;5.25, df\u0026thinsp;=\u0026thinsp;4418) and T5 from T4 (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003et\u003c/em\u003e\u0026thinsp;=\u0026thinsp;4.02, df\u0026thinsp;=\u0026thinsp;2805). No significant change was observed in the chronic group at any time pair. In the mild chronic group, a significant decrease was observed at T4 from T3 (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003et\u003c/em\u003e\u0026thinsp;=\u0026thinsp;4.51, df\u0026thinsp;=\u0026thinsp;740) and at T5 from T4 (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003et\u003c/em\u003e\u0026thinsp;=\u0026thinsp;4.77, df\u0026thinsp;=\u0026thinsp;602). Finally, in the early response group, a significant decrease was observed at T2 from T1 (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003et\u003c/em\u003e\u0026thinsp;=\u0026thinsp;28.09, df\u0026thinsp;=\u0026thinsp;579) and at T3 from T2 (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003et\u003c/em\u003e\u0026thinsp;=\u0026thinsp;5.66, df\u0026thinsp;=\u0026thinsp;433), but a significant increase was also observed at T5 from T4 (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003et\u003c/em\u003e\u0026thinsp;=\u0026thinsp;5.19, df\u0026thinsp;=\u0026thinsp;306).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eMultinomial logistic regression analysis\u003c/h2\u003e \u003cp\u003ePsychiatric scores at T0 were compressed by principal component analysis (PCA) and the top 4 components (25.3% for PC1, 13.9% for PC2, 11.2% for PC3, 10.9% for PC4 and 61.2% for all, Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e) were used in the sequential modifying factor analysis. Identified factors were consistent with the previous study\u003csup\u003e5\u003c/sup\u003e and each PC was named in the same way: PC1 as a general psychiatric burden, PC2 as social withdrawal, PC3 as alcohol-related problems, and PC4 as depression/anxiety. Risk factors for each class were identified by multinomial logistic regression (MLR) with the resilient group as a reference (multiple comparison was corrected by the Bonferroni method, Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). As a result, a general psychiatric burden component (PC1) was identified as a significant risk factor for all groups(chronic: \u003cem\u003eb\u003c/em\u003e\u0026thinsp;=\u0026thinsp;2.16; 95% CI [1.89\u0026ndash;2.42], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003et\u003c/em\u003e\u0026thinsp;=\u0026thinsp;15.87, df\u0026thinsp;=\u0026thinsp;9234, mild chronic: \u003cem\u003eb\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1.27; 95% CI [1.14\u0026ndash;1.40], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003et\u003c/em\u003e\u0026thinsp;=\u0026thinsp;19.13, df\u0026thinsp;=\u0026thinsp;9234, early response: \u003cem\u003eb\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1.24; 95% CI [1.08\u0026ndash;1.39], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003et\u003c/em\u003e\u0026thinsp;=\u0026thinsp;15.62, df\u0026thinsp;=\u0026thinsp;9234). Also, a social withdrawal component (PC2) was identified as a significant protective factor for all groups(chronic: \u003cem\u003eb\u003c/em\u003e = -0.70; 95% CI [-0.90-0.49], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003et\u003c/em\u003e = -6.57, df\u0026thinsp;=\u0026thinsp;9234, mild chronic: \u003cem\u003eb\u003c/em\u003e = -0.51; 95% CI [-0.62-0.40], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003et\u003c/em\u003e = -8.88, df\u0026thinsp;=\u0026thinsp;9234, early response: \u003cem\u003eb\u003c/em\u003e = -0.43; 95% CI [-0.56-0.30], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003et\u003c/em\u003e = -6.39, df\u0026thinsp;=\u0026thinsp;9234). An alcohol-related problem component (PC3) was identified as a significant protective factor for the chronic group(\u003cem\u003eb\u003c/em\u003e = -0.30; 95% CI [-0.49-0.10], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.01, \u003cem\u003et\u003c/em\u003e = -3.02, df\u0026thinsp;=\u0026thinsp;9234) and the mild chronic group (\u003cem\u003eb\u003c/em\u003e = -0.23; 95% CI [-0.34-0.13], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003et\u003c/em\u003e = -4.30, df\u0026thinsp;=\u0026thinsp;9234). Depression/anxiety component (PC4) was also identified as a significant risk factor for the early response group (\u003cem\u003eb\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.23; 95% CI [0.10\u0026ndash;0.35], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;.053, \u003cem\u003et\u003c/em\u003e\u0026thinsp;=\u0026thinsp;3.55, df\u0026thinsp;=\u0026thinsp;9234). Young age and unmarried status also showed a tendency as a risk factor for the chronic group, but no significant difference was identified.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eBaseline PTSS estimation\u003c/h2\u003e \u003cp\u003eIn order to provide an intuitive interpretation of PTSS prognosis from the baseline, we estimated \u0026ldquo;baseline PTSS'', which would have arisen even if the pandemic had not occurred. We also tested the estimation validity of baseline PTSS by using leave-one-time-point-out cross-validation analysis. The mean adjusted-\u003cem\u003eR\u003c/em\u003e\u003csup\u003e2\u003c/sup\u003e score of baseline PTSS estimation model showed 0.525 at training and 0.436 at testing. The leave-one-time-point-out cross-validation analysis showed higher than 0.5 adjusted-\u003cem\u003eR\u003c/em\u003e\u003csup\u003e2\u003c/sup\u003e score for all Tx conditions at training and higher than 0.4 adjusted-\u003cem\u003eR\u003c/em\u003e\u003csup\u003e2\u003c/sup\u003e score for all Tx conditions at test (see details in Supplementally).\u003c/p\u003e \u003cp\u003eThe baseline PTSS score of the chronic group (31.8; 95% CI [29.5\u0026ndash;34.1]) was significantly higher than those of the mild chronic (22.4; 95% CI [21.3\u0026ndash;23.6]), early-response (22.8; 95% CI [22.0-23.6]) and resilient groups (10.0; 95% CI [9.7\u0026ndash;10.4]), and scores of mild chronic and early-response groups were significantly higher than that of the resilient group (one-way ANOVA, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003eF\u003c/em\u003e\u003csub\u003e3,3811\u003c/sub\u003e = 512.9, \u003cem\u003eη\u003c/em\u003e\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.29; post-hoc Tukey test chronic versus resilient, mild chronic and early-response \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001; resilient versus mild chronic, early response \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001). No significant difference was found between mean scores of the mild chronic and early-response groups.\u003c/p\u003e \u003c/div\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eIn this study, we conducted an online longitudinal survey including COVID19-related PTSS over a period of 1.5 years since before the pandemic. Our LGMM trajectory analysis identified four PTSS trajectory types: resilient, chronic, mild chronic, and early response. We then demonstrated how these prognoses are modulated by psychiatric conditions prior to the pandemic. For demonstration purposes, we further estimated \u0026ldquo;baseline\u0026rdquo; PTSS before the pandemic, which we assume to correspond to the PTSS score that would have arisen during the time of the pandemic if this psychiatric exacerbation had not occurred. Generally, higher baseline PTSS was associated with higher PTSS during the pandemic, with more pronounced exacerbation in PTSS.\u003c/p\u003e \u003cp\u003eSome psychiatric conditions before the pandemic were identified as risk or protective factors for specific trajectory patterns, while others were common to all trajectories. This fact may lead to two possible strategies for prevention and treatment of stress responses, namely, those tailored to an individual trajectory type and those common to all types. The depression/anxiety component was identified as a risk factor specific to the early response group. It may be feasible to assume that PTSS in this type reflects temporal depression and anxiety which resemble a normal stress response in healthy individuals. As with adjustment disorders and acute stress disorders, symptomatic psychotherapy and pharmacotherapy may be more effective than curative pharmacotherapy, such as SSRIs. It may also be effective to allay individuals\u0026rsquo; anxieties through psychoeducation that promotes an awareness that the response is common, normal response to stressful or traumatic events.\u003csup\u003e27,28\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eIn contrast to a specific risk factor, the social withdrawal/general psychiatric burden component was identified as a common protective/risk factor for all trajectories. While most previous studies underlined the adverse effects of quarantine policies on mental health through inhibition of social communication,\u003csup\u003e29,30\u003c/sup\u003e effects of such policies may have varied according to individual social propensity. While individuals with high sociability may have experienced greatly exacerbated status during the quarantine, those with low sociability may have improved their psychological conditions by not having to participate in unwanted social activities in which they were forced to participate before the pandemic. Quarantine policies, thus, should be flexibly set according to individual propensities for social activities. After the COVID-19 pandemic, we should also shift to a lifestyle that accepts diverse social communication styles rather than simply returning to the lifestyle before the pandemic. The general psychiatric burden component was also identified as a common risk factor. This finding aligns with established observations indicating the vulnerability of psychiatric patients to stress.\u003csup\u003e7,31\u003c/sup\u003e From this perspective, those with high general psychiatric burden components may have experienced a more pronounced exacerbation of their psychological states. On the other hand, from the perspective that many psychiatric disorders arise from a single common factor,\u003csup\u003e32\u003c/sup\u003e stress-induced psychiatric symptoms may be attributable to emotional malaise that existed before the event. Individuals may have interpreted their malaise that existed before the pandemic rather than that was induced by the pandemic. In such a scenario, the individual\u0026rsquo;s holistic problem, rather than the effect of the trigger event, should be resolved to treat PTSS. To identify which perspective better explains the mechanism of PTSS development, it is necessary to separate psychiatric problems existing prior to the event from those exacerbated during the event.\u003c/p\u003e \u003cp\u003eAs a proof-of-concept, we estimated \u0026ldquo;baseline\u0026ldquo; PTSS. This was calculated using pre-pandemic psychiatric conditions based on the correspondence between PTSS and psychiatric conditions during the pandemic. Here, we assume that baseline PTSS corresponds to symptoms attributed to psychiatric conditions prior to traumatic events. Leave-one-time-out regression suggested that PTSS scores could be estimated from scores of other psychiatric disorders at other time points (mean adjusted-R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.411). The ANOVA conducted on the mean baseline PTSS score for trajectory types revealed three levels: high for the chronic group, low for the resilient group, and intermediate for the others. This result suggests a general trend in which higher baseline PTSS levels correspond to higher levels of PTSS during the pandemic. Under our assumptions, our results suggest mechanisms by which stress symptoms evolved, especially in the context of long-lasting stressful events like the COVID-19 pandemic. Higher baseline PTSS levels contribute to an elevation in PTSS during the initial phase of the pandemic. The increased PTSS, serving as the new baseline, further intensified PTSS during the middle phase of the pandemic. This cyclic change of exacerbation continued throughout the pandemic. If this hypothesis is confirmed, it is crucial to make concerted efforts to reduce levels of \u0026ldquo;baseline\u0026rdquo; PTSS in the early phase of the pandemic. While precise interpretation of baseline PTSS requires further discussion and clarification, our proposal represents an initial step toward addressing the intricate relationship between stress vulnerability and exacerbations triggered by stress.\u003c/p\u003e \u003cp\u003eThere are two limitations in our study. First, while most results were consistent with previous studies, the delayed-onset trajectory was not identified in our study. This may be explained by the difference in observation period, interval, and the event type.\u003csup\u003e12\u003c/sup\u003e Note that delayed-onset PTSD would have developed after our observation period (24 months).\u003csup\u003e18\u003c/sup\u003e Second, it is also important to verify whether the results of our study can be generalized to other types of traumatic experience; therefore, it is necessary to apply the same analysis to other cases.\u003c/p\u003e \u003cp\u003eIn this study, we identified four types of stress response with different modifying factors. In addition, we also found that these four trajectories were partly predictable, even before the onset of the traumatic event. These findings may help to develop personalized-care according to the psychiatric states underlying their pathogenesis.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAuthors Contributions\u003c/strong\u003e: FN, TK, and TC had full access to all of the data in the study and take responsibility for the integrity of the work.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eConcept and design\u003c/em\u003e: FN, TK, TO, TC.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eData acquisition\u003c/em\u003e: TO, TH, NK, MH, MM, TC.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eStatistical analyses\u003c/em\u003e: FN, TK, TO, TC.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eInterpretation of data\u003c/em\u003e: FN, TK, TO, AC, TC.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eDrafting the first version of the manuscript\u003c/em\u003e: FN, TK, AC, TC.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eRevised and approved the final version of the manuscript\u003c/em\u003e: All authors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflicts of Interest\u003c/strong\u003e: NK, TH, MH, and YM are employees of KDDI Research Inc. NK, TH, MH, and YM were involved in the data acquisition and revised the manuscript. For the remaining authors, there were no conflicts of interest.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding/ Support\u003c/strong\u003e: This research was supported by a KDDI collaborative research contract. This work was also supported by AMED under Grant Number JP20dm0307008.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRole of Funder/Sponsor\u003c/strong\u003e: The sponsor, KDDI, reviewed the design, the collection, and the analysis. The funder had no role in the conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDisclaimers\u003c/strong\u003e: All information and materials are original.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Sharing\u003c/strong\u003e: The main summary statistical data that support the findings of this study are available in the Supplementary Data. Owing to company cohort data sharing restrictions, individual-level data cannot be publicly posted. However, data are available from the authors upon request and with permission of KDDI Corporation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAdditional Contributions\u003c/strong\u003e: We thank Rumi Yorizawa (The Department of Decoded Neurofeedback, Computational Neuroscience Laboratories, Advanced Telecommunications Research Institute International, Kyoto), and Miho Nagata (The Department of Decoded Neurofeedback, Computational Neuroscience Laboratories, Advanced Telecommunications Research Institute International, Kyoto) for data collection and organization.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eBarzilay R, Moore TM, Greenberg DM, DiDomenico GE, Brown LA, White LK, Gur RC, Gur RE. Resilience, COVID-19-related stress, anxiety and depression during the pandemic in a large population enriched for healthcare providers. Transl Psychiatry. 2020;10(1):291. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41398-020-00982-4\u003c/span\u003e\u003cspan address=\"10.1038/s41398-020-00982-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 32820171; PMCID: PMC7439246.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBr\u0026uuml;lhart M, Klotzb\u0026uuml;cher V, Lalive R, Reich SK. Mental health concerns during the COVID-19 pandemic as revealed by helpline calls. Nature. 2021;600(7887):121\u0026ndash;126. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41586-021-04099-6\u003c/span\u003e\u003cspan address=\"10.1038/s41586-021-04099-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Epub 2021 Nov 17. PMID: 34789873; PMCID: PMC9973557.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePierce M, McManus S, Hope H, Hotopf M, Ford T, Hatch SL, John A, Kontopantelis E, Webb RT, Wessely S, Abel KM. Mental health responses to the COVID-19 pandemic: a latent class trajectory analysis using longitudinal UK data. Lancet Psychiatry. 2021;8(7):610\u0026ndash;619. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/S2215-0366(21)00151-6\u003c/span\u003e\u003cspan address=\"10.1016/S2215-0366(21)00151-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Epub 2021 May 6. PMID: 33965057; PMCID: PMC9764381.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOka T, Hamamura T, Miyake Y, Kobayashi N, Honjo M, Kawato M, Kubo T, Chiba T. Prevalence and risk factors of internet gaming disorder and problematic internet use before and during the COVID-19 pandemic: A large online survey of Japanese adults. J Psychiatr Res. 2021;142:218\u0026ndash;225. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jpsychires.2021.07.054\u003c/span\u003e\u003cspan address=\"10.1016/j.jpsychires.2021.07.054\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Epub 2021 Aug 4. Erratum in: J Psychiatr Res. 2022;148:275\u0026ndash;276. PMID: 34385071.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOka T, Kubo T, Kobayashi N, Nakai F, Miyake Y, Hamamura T, Honjo M, Toda H, Boku S, Kanazawa T, Nagamine M, Cortese A, Takebayashi M, Kawato M, Chiba T. Multiple time measurements of multidimensional psychiatric states from immediately before the COVID-19 pandemic to one year later: a longitudinal online survey of the Japanese population. Transl Psychiatry. 2021;11(1):573. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41398-021-01696-x\u003c/span\u003e\u003cspan address=\"10.1038/s41398-021-01696-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 34759293; PMCID: PMC8581018.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChiba T, Ide K, Murakami M, Kobayashi N, Oka T, Nakai F, Yorizawa R, Miyake Y, Hamamura T, Honjo M, Toda H, Kanazawa T, Boku S, Kubo T, Hishimoto A, Kawato M, Cortese A. Event-related PTSD symptoms as a high-risk factor for suicide\u0026mdash;Longitudinal observational study. Nature Mental Health. 2023;1(12):1013\u0026ndash;22..\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHarris TB, Carlisle LL, Sargent J, Primm AB. Trauma and diverse child populations. Child Adolesc Psychiatr Clin N Am. 2010;19(4):869\u0026thinsp;\u0026ndash;\u0026thinsp;87. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.chc.2010.07.007\u003c/span\u003e\u003cspan address=\"10.1016/j.chc.2010.07.007\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 21056351.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBrown GW, Bifulco A, Harris T, Bridge L. Life stress, chronic subclinical symptoms and vulnerability to clinical depression. J Affect Disord. 1986 Jul-Aug;11(1):1\u0026ndash;19. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/0165-0327(86)90054-6\u003c/span\u003e\u003cspan address=\"10.1016/0165-0327(86)90054-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 2944923.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJudkins JL, Moore BA, Collette TL, Hale WJ, Peterson AL, Morissette SB. Incidence Rates of Posttraumatic Stress Disorder Over a 17-Year Period in Active Duty Military Service Members. J Trauma Stress. 2020;33(6):994\u0026ndash;1006. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/jts.22558\u003c/span\u003e\u003cspan address=\"10.1002/jts.22558\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Epub 2020 Jun 29. PMID: 32598575.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNagamine M, Giltay EJ, Shigemura J, van der Wee NJ, Yamamoto T, Takahashi Y, Saito T, Tanichi M, Koga M, Toda H, Shimizu K, Yoshino A, Vermetten E. Assessment of Factors Associated With Long-term Posttraumatic Stress Symptoms Among 56 388 First Responders After the 2011 Great East Japan Earthquake. JAMA Netw Open. 2020;3(9):e2018339. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1001/jamanetworkopen.2020.18339\u003c/span\u003e\u003cspan address=\"10.1001/jamanetworkopen.2020.18339\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 32990742; PMCID: PMC7525349.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAndersen SB, Karstoft KI, Bertelsen M, Madsen T. Latent trajectories of trauma symptoms and resilience: the 3-year longitudinal prospective USPER study of Danish veterans deployed in Afghanistan. J Clin Psychiatry. 2014;75(9):1001-8. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.4088/JCP.13m08914\u003c/span\u003e\u003cspan address=\"10.4088/JCP.13m08914\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 25188895.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBonanno GA, Mancini AD. Beyond resilience and PTSD: Mapping the heterogeneity of responses to potential trauma. Psychological trauma: Theory, research, practice, and policy. 2012;4(1):74. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1037/a0017829\u003c/span\u003e\u003cspan address=\"10.1037/a0017829\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBonanno GA, Mancini AD, Horton JL, Powell TM, Leardmann CA, Boyko EJ, Wells TS, Hooper TI, Gackstetter GD, Smith TC; Millennium Cohort Study Team. Trajectories of trauma symptoms and resilience in deployed U.S. military service members: prospective cohort study. Br J Psychiatry. 2012;200(4):317\u0026thinsp;\u0026ndash;\u0026thinsp;23. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1192/bjp.bp.111.096552\u003c/span\u003e\u003cspan address=\"10.1192/bjp.bp.111.096552\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Epub 2012 Feb 23. PMID: 22361018.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDickstein BD, Suvak M, Litz BT, Adler AB. Heterogeneity in the course of posttraumatic stress disorder: trajectories of symptomatology. J Trauma Stress. 2010;23(3):331-9. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/jts.20523\u003c/span\u003e\u003cspan address=\"10.1002/jts.20523\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 20564365.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOrcutt HK, Erickson DJ, Wolfe J. The course of PTSD symptoms among Gulf War veterans: a growth mixture modeling approach. J Trauma Stress. 2004;17(3):195\u0026ndash;202. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1023/B:JOTS.0000029262.42865.c2\u003c/span\u003e\u003cspan address=\"10.1023/B:JOTS.0000029262.42865.c2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 15253091.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKarstoft KI, Armour C, Elklit A, Solomon Z. The role of locus of control and coping style in predicting longitudinal PTSD-trajectories after combat exposure. J Anxiety Disord. 2015;32:89\u0026ndash;94. doi: 10.1016/j.janxdis.2015.03.007. Epub 2015 Mar 28. PMID: 25890288.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSaito T, van der Does FHS, Nagamine M, van der Wee NJ, Shigemura J, Yamamoto T, Takahashi Y, Koga M, Toda H, Yoshino A, Vermetten E, Giltay EJ. Risk and resilience in trajectories of post-traumatic stress symptoms among first responders after the 2011 Great East Japan Earthquake: 7-year prospective cohort study. Br J Psychiatry. 2022 Feb 22:1\u0026ndash;8. doi: 10.1192/bjp.2022.2. Epub ahead of print. PMID: 35191369.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAndrews B, Brewin CR, Philpott R, Stewart L. Delayed-onset posttraumatic stress disorder: a systematic review of the evidence. Am J Psychiatry. 2007;164(9):1319-26. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1176/appi.ajp.2007.06091491\u003c/span\u003e\u003cspan address=\"10.1176/appi.ajp.2007.06091491\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 17728415.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDohrenwend BP, Turner JB, Turse NA, Adams BG, Koenen KC, Marshall R. The psychological risks of Vietnam for U.S. veterans: a revisit with new data and methods. Science. 2006;313(5789):979\u0026ndash;82. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1126/science.1128944\u003c/span\u003e\u003cspan address=\"10.1126/science.1128944\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 16917066; PMCID: PMC1584215.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePan L, Xu Q, Kuang X, Zhang X, Fang F, Gui L, Li M, Tefsen B, Zha L, Liu H. Prevalence and factors associated with post-traumatic stress disorder in healthcare workers exposed to COVID-19 in Wuhan, China: a cross-sectional survey. BMC Psychiatry. 2021;21(1):572. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12888-021-03589-1\u003c/span\u003e\u003cspan address=\"10.1186/s12888-021-03589-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 34781901; PMCID: PMC8591965.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBomyea J, Risbrough V, Lang AJ. A consideration of select pre-trauma factors as key vulnerabilities in PTSD. Clin Psychol Rev. 2012;32(7):630\u0026ndash;41. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cpr.2012.06.008\u003c/span\u003e\u003cspan address=\"10.1016/j.cpr.2012.06.008\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Epub 2012 Jul 20. PMID: 22917742; PMCID: PMC3444146.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBonanno GA, Chen S, Galatzer-Levy IR. Resilience to potential trauma and adversity through regulatory flexibility. Nature Reviews Psychology. 2023 Sep 19:1\u0026ndash;3.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSteenkamp MM, Dickstein BD, Salters-Pedneault K, Hofmann SG, Litz BT. Trajectories of PTSD symptoms following sexual assault: is resilience the modal outcome? J Trauma Stress. 2012;25(4):469\u0026ndash;74. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/jts.21718\u003c/span\u003e\u003cspan address=\"10.1002/jts.21718\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Epub 2012 Jul 13. PMID: 22807251.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003edeRoon-Cassini TA, Mancini AD, Rusch MD, Bonanno GA. Psychopathology and resilience following traumatic injury: a latent growth mixture model analysis. Rehabil Psychol. 2010;55(1):1\u0026ndash;11. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1037/a0018601\u003c/span\u003e\u003cspan address=\"10.1037/a0018601\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 20175629.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMurphy D, Smith KV. Treatment Efficacy for Veterans With Posttraumatic Stress Disorder: Latent Class Trajectories of Treatment Response and Their Predictors. J Trauma Stress. 2018;31(5):753\u0026ndash;763. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/jts.22333\u003c/span\u003e\u003cspan address=\"10.1002/jts.22333\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Epub 2018 Oct 19. PMID: 30338573; PMCID: PMC6456014.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCohen JA, Mannarino AP. Trauma-focused Cognitive Behavior Therapy for Traumatized Children and Families. Child Adolesc Psychiatr Clin N Am. 2015;24(3):557\u0026ndash;70. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.chc.2015.02.005\u003c/span\u003e\u003cspan address=\"10.1016/j.chc.2015.02.005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Epub 2015 Apr 22. PMID: 26092739; PMCID: PMC4476061.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBersani FS, Delle Chiaie R. The end method: Normalization. Empathy, Normalization and De-escalation: Management of the Agitated Patient in Emergency and Critical Situations. 2021:57\u0026ndash;64.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evan der Velden PG, Hyland P, Contino C, von Gaudecker HM, Muffels R, Das M. Anxiety and depression symptoms, the recovery from symptoms, and loneliness before and after the COVID-19 outbreak among the general population: Findings from a Dutch population-based longitudinal study. PLoS One. 2021;16(1):e0245057. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1371/journal.pone.0245057\u003c/span\u003e\u003cspan address=\"10.1371/journal.pone.0245057\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 33411843; PMCID: PMC7790276.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBrunoni AR, Suen PJ, Bacchi PS, Razza LB, Klein I, Dos Santos LA, de Souza Santos I, Valiengo LD, Gallucci-Neto J, Moreno ML, Pinto BS. Prevalence and risk factors of psychiatric symptoms and diagnoses before and during the COVID-19 pandemic: findings from the ELSA-Brasil COVID-19 mental health cohort. Psychological medicine. 2023;53(2):446\u0026ndash;57.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGoh C, Agius M. The stress-vulnerability model how does stress impact on mental illness at the level of the brain and what are the consequences? Psychiatr Danub. 2010;22(2):198\u0026ndash;202. PMID: 20562747.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCaspi A, Moffitt TE. All for One and One for All: Mental Disorders in One Dimension. Am J Psychiatry. 2018;175(9):831\u0026ndash;844. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1176/appi.ajp.2018.17121383\u003c/span\u003e\u003cspan address=\"10.1176/appi.ajp.2018.17121383\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Epub 2018 Apr 6. PMID: 29621902; PMCID: PMC6120790.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChiba T, Oka T, Hamamura T, Kobayashi N, Honjo M, Miyake Y, Kubo T, Toda H, Kanazawa T, Boku S, Hishimoto A. PTSD symptoms related to COVID-19 as a high risk factor for suicide-Key to prevention. medRxiv. 2020 Dec 18:2020\u0026ndash;12. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1101/2020.12.15.20246819\u003c/span\u003e\u003cspan address=\"10.1101/2020.12.15.20246819\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWeiss DS. The impact of event scale: revised. Cross-cultural assessment of psychological trauma and PTSD. 2007:219\u0026thinsp;\u0026ndash;\u0026thinsp;38.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRadloff LS. The CES-D scale: A self-report depression scale for research in the general population. Applied psychological measurement. 1977;1(3):385\u0026ndash;401.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFoa EB, Kozak MJ, Salkovskis PM, Coles ME, Amir N. The validation of a new obsessive\u0026ndash;compulsive disorder scale: The Obsessive\u0026ndash;Compulsive Inventory. Psychological assessment. 1998;10(3):206.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMeerkerk GJ, Van Den Eijnden RJ, Vermulst AA, Garretsen HF. The compulsive internet use scale (CIUS): some psychometric properties. Cyberpsychology \u0026amp; behavior. 2009;12(1):1\u0026ndash;6.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKessler RC, Adler L, Ames M, Demler O, Faraone S, Hiripi EV, Howes MJ, Jin R, Secnik K, Spencer T, Ustun TB. The World Health Organization Adult ADHD Self-Report Scale (ASRS): a short screening scale for use in the general population. Psychological medicine. 2005;35(2):245\u0026ndash;56.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBaron-Cohen S, Wheelwright S, Skinner R, Martin J, Clubley E. The autism-spectrum quotient (AQ): Evidence from asperger syndrome/high-functioning autism, malesand females, scientists and mathematicians. Journal of autism and developmental disorders. 2001;31:5\u0026ndash;17.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBaker SL, Heinrichs N, Kim HJ, Hofmann SG. The Liebowitz social anxiety scale as a self-report instrument: a preliminary psychometric analysis. Behaviour research and therapy. 2002;40(6):701\u0026ndash;15.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSpielberger CD. State-Trait Anxiety Inventory for Adults. 1983. DOI:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1037/t06496-000\u003c/span\u003e\u003cspan address=\"10.1037/t06496-000\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSaunders JB, Aasland OG, Babor TF, De la Fuente JR, Grant M. Development of the alcohol use disorders identification test (AUDIT): WHO collaborative project on early detection of persons with harmful alcohol consumption-II. Addiction. 1993;88(6):791\u0026ndash;804.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMuth\u0026eacute;n B. Latent variable analysis. The Sage handbook of quantitative methodology for the social sciences. 2004;345(368):106-9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMuth\u0026eacute;n B, Muth\u0026eacute;n LK. Integrating person-centered and variable-centered analyses: growth mixture modeling with latent trajectory classes. Alcohol Clin Exp Res. 2000;24(6):882\u0026ndash;91. PMID: 10888079.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eProust-Lima C, Philipps V, Diakite A, Liquet B, Proust MC. Package \u0026lsquo;lcmm\u0026rsquo;. CRAN R. 2023 Feb 20.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWardenaar K. Latent Class Growth Analysis and Growth Mixture Modeling using R: A tutorial for two R-packages and a comparison with Mplus.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNylund KL, Asparouhov T, Muth\u0026eacute;n BO. Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural equation modeling: A multidisciplinary Journal. 2007;14(4):535\u0026ndash;69. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1080/10705510701575396\u003c/span\u003e\u003cspan address=\"10.1080/10705510701575396\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWold S, Esbensen K, Geladi P. Principal component analysis. Chemometrics and intelligent laboratory systems. 1987;2(1\u0026ndash;3):37\u0026ndash;52. doi: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/0169-7439(87)80084-9\u003c/span\u003e\u003cspan address=\"10.1016/0169-7439(87)80084-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVapnik V, Golowich S, Smola A. Support vector method for function approximation, regression estimation and signal processing. Advances in neural information processing systems. 1996;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBishop CM, Nasrabadi NM. Pattern recognition and machine learning. New York: springer; 2006 Aug 17.\u003c/span\u003e\u003c/li\u003e \u003c/ol\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cspan\u003eProcedures and outcomes\u003c/span\u003e\u003c/p\u003e\n\u003cp\u003e\u003cspan\u003eThe online survey was conducted by Macromill Inc.(Japan) as a large, longitudinal survey to investigate mental health of the general Japanese adult population (see our previous studies\u003csup\u003e4,5,33\u003c/sup\u003e for details). The original panel survey was conducted in December 2019 (T0) before identification of the first COVID-19 case in Japan (January 2020). In response to the outbreak of COVID-19, follow-up surveys of T0 participants were conducted in August 2020 (T1), December 2020 (T2), April 2021 (T3), August 2021 (T4), and December 2021 (T5). Invitations for the original survey were sent to 5955 individuals, of which 478 were excluded due to inconsistencies or contradictions in their answers. An additional 481 individuals were also excluded because of unreliable answers. They answered using only the maximum or minimum answers in questionnaires including reverse items. The original and follow-up research designs were approved by the Ethics Committees of the Advanced Telecommunications Research Institute International (ATR) (approval No. 21\u0026ndash;195 for the original study \u0026amp; 21\u0026ndash;749 for the follow-up study). All participants read a full explanation of the study and gave informed consent before each survey.\u003c/span\u003e\u003c/p\u003e\n\u003cp\u003e\u003cspan\u003eQuestionnaires were constructed from questions for psychiatric, demographic and COVID-19 related items. Psychiatric questions included 10 types of validated questionnaires for psychiatric disorders; PTSS (IES-R\u003csup\u003e34\u003c/sup\u003e), major depressive disorder (CES-D\u003csup\u003e35\u003c/sup\u003e), obsessive-compulsive disorder (OCI\u003csup\u003e36\u003c/sup\u003e), internet-related problems (CIUS\u003csup\u003e37\u003c/sup\u003e), attention-deficit/hyperactivity disorder (ADHD) (ASRS\u003csup\u003e38\u003c/sup\u003e), autistic spectrum disorder (AQ\u003csup\u003e39\u003c/sup\u003e), social anxiety (LSAS-fear/avoid\u003csup\u003e40\u003c/sup\u003e), general anxiety (STAI-Y-state\u003csup\u003e41\u003c/sup\u003e), and alcohol-related problems (AUDIT\u003csup\u003e42\u003c/sup\u003e). Collected demographic data included sex (women and men), age, job status (self-employed, employed, unemployed and other), marital state and household income per year (lowest; less than four million yen, 2nd; four\u0026thinsp;~\u0026thinsp;six million yen, 3rd; six\u0026thinsp;~\u0026thinsp;eight million yen, 4th; eight\u0026thinsp;~\u0026thinsp;ten million yen, Highest; more than ten million yen, and Missing) and they were analyzed as pre-COVID demographics. The temporal change of PTSS score was tested by one-way ANOVA with post-hoc Tukey-HSD test.\u003c/span\u003e\u003c/p\u003e\n\u003cp\u003e\u003cspan\u003eTrajectory analysis using the Latent Growth Mixture Model\u003c/span\u003e\u003c/p\u003e\n\u003cp\u003e\u003cspan\u003eIn this study, we assumed that the global heterogeneity of PTSS trajectories can be explained by a set of homogeneous trajectories. Consistent with previous studies, the latent growth mixture model\u003csup\u003e43,44\u003c/sup\u003e (LGMM) was used to identify latent class trajectories of PTSS measured by IES-R from T1 to T5, during the COVID-19 pandemic. The \u0026ldquo;\u003cem\u003elcmm\u003c/em\u003e\u0026rdquo; package\u003csup\u003e45\u003c/sup\u003e in R version. 4.1.0 was used to identify the latent class.\u003csup\u003e46\u003c/sup\u003e The classification was performed under all parameter conditions (6\u0026times;3: the number of latent classes (1 to 6 class) \u0026times; the model function (linear, quadratic and exponential)). The optimal condition was determined according to the indicators. Commensurate with recommendations in the previous study\u003csup\u003e47\u003c/sup\u003e, we relied on three types of indicators to determine the best model for clustering; AIC (Akaike information criterion), BIC (Bayesian information criterion), sample-size-adjusted BIC. A grid search approach with 100 iterations was used to estimate optimal values of model parameters and 100 repetitions were applied to achieve stable results. After repetition, the optimal parameter was determined using the maximum likelihood method. After estimating optimal model parameters, the membership probability was calculated for each participant to assign a class label. Each trajectory was also analyzed by unpaired t-test to confirm the temporal change (multi-comparison was adjusted using the Bonferroni method).\u003c/span\u003e\u003c/p\u003e\n\u003cp\u003e\u003cspan\u003eMultinomial logistic regression analysis\u003c/span\u003e\u003c/p\u003e\n\u003cp\u003e\u003cspan\u003eWe used a multinomial logistic regression model to identify risk/protective factors for each trajectory class. Prior to apply the model, the nine psychiatric scores of all participants were compressed into four dimensional data using principal component analysis (PCA) on T0 data.\u003csup\u003e5,48\u003c/sup\u003e Then, according to the estimated loading of the top four PCs, orthogonal transformation was applied and scores of each participant at each time point were converted into four-dimensional orthogonal scores.Then, logistic regression was performed using class assignment as a dependent variable and compressed psychiatric scores at T0 and other demographics (sex, age, income, employment status, and marital status) as independent variables. The relative risk b from the resilient group for each independent variable was estimated for each trajectory group, and p-values were adjusted by Bonferroni correction. We fitted the model using the \u003cem\u003eStatistical Machine Learning Toolbox\u003c/em\u003e in \u003cem\u003ematlab\u003c/em\u003e.\u003c/span\u003e\u003c/p\u003e\n\u003cp\u003e\u003cspan\u003eBaseline PTSS estimation\u003c/span\u003e\u003c/p\u003e\n\u003cp\u003e\u003cspan\u003eWe estimated \u0026ldquo;baseline PTSS\u0026quot; by using multiple psychiatric scores collected prior to COVID-19 in multiple ridge regression analysis.\u003csup\u003e49,50\u003c/sup\u003e First, we analyzed the linear association of PTSS score with nine other psychiatric scores at T1 to T5 during the pandemic. All data from T1 to T5 were randomly split in half to create training and test sets. To obtain reliable results, this random sampling was repeated 100 times. For each data set, both training and test data were normalized based on the mean and variance of the training dataset. Training data were further divided into training data and validation data with a 10-fold cross validation approach to tune the regularization parameter. Specifically, the optimal penalty coefficient was searched from 0 to 1 for every 0.001. In each cross-validation, regression coefficients and the penalty coefficient were determined based on minimization of the cost function that was estimated with the validation dataset. Successively, for each repetition, optimal parameters were selected based on minimization of the cost function among the folds of cross-validation. Each model was evaluated using the test dataset based on the adjusted-\u003cem\u003eR\u003c/em\u003e\u003csup\u003e2\u003c/sup\u003e metrics and the model for the baseline PTSS estimation was evaluated by the mean score of adjusted-\u003cem\u003eR\u003c/em\u003e\u003csup\u003e2\u003c/sup\u003e metrics for all models. Averages of the coefficients were used later in estimation of baseline PTSS.\u003c/span\u003e\u003c/p\u003e\n\u003cp\u003e\u003cspan\u003eSecond, we evaluated the applicability of the model across time points with a similar approach. Specifically, we performed leave-one-time-point-out cross validation. In this validation, the model was trained on scores at four time points, except fixed time point Tx (selected between T1 to T5), and the scores at Tx were used as test data.\u003c/span\u003e\u003c/p\u003e\n\u003cp\u003e\u003cspan\u003eFinally, based on the estimated linear association, baseline PTSS scores of each subject were calculated from nine other psychiatric scores at T0. The difference in the mean baseline PTSS score for each group was tested by one-way ANOVA, with post-hoc Tukey test.\u003c/span\u003e\u003c/p\u003e"},{"header":"Table 1","content":"\u003cp\u003eTable 1 is available in the Supplementary Files section.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-3865066/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3865066/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eRecent time-dependent analyses of stress-related disorders have identified heterogeneity of trajectories and their modifying factors. While psychiatric patients are vulnerable to stress events, it is unclear how psychiatric conditions in the general population modulate subsequent stress responses. Using our longitudinal online survey from before the COVID-19 pandemic to post-pandemic follow-ups (n\u0026thinsp;=\u0026thinsp;3815 Japanese adults), here we identified four trajectories of post-traumatic stress symptoms (PTSS) a latent growth mixture model; resilient, chronic, mild chronic, and early response. The depression/anxiety were identified as specific risk factors for the early-response trajectory. In contrast, general psychiatric burden and social withdrawal were identified as common risk/protective factors. Further, we estimated \u0026ldquo;baseline\u0026rdquo; PTSS to determine the predictability of the PTSS prognosis from pre-pandemic states. The chronic group showed significantly higher baseline PTSS scores than the mild-chronic and early-response groups, both of which were significantly higher than the resilient group. We concluded that prior psychiatric conditions significantly affect the PTSS trajectory. These results suggest that prior psychiatric conditions may be considered for the prevention and treatment of maladaptive stress responses.\u003c/p\u003e","manuscriptTitle":"Pre-event psychiatric states predict trajectories of post-traumatic stress symptoms during the COVID-19 pandemic","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-01-26 17:52:53","doi":"10.21203/rs.3.rs-3865066/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"993458e8-4ef1-469b-a49c-454b8c103307","owner":[],"postedDate":"January 26th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":28355131,"name":"Health sciences/Diseases/Trauma"},{"id":28355132,"name":"Health sciences/Diseases/Psychiatric disorders/Post-traumatic stress disorder"},{"id":28355133,"name":"Health sciences/Health care/Public health/Population screening"},{"id":28355134,"name":"Health sciences/Diseases/Psychiatric disorders/Depression"},{"id":28355135,"name":"Health sciences/Diseases/Psychiatric disorders/Anxiety"}],"tags":[],"updatedAt":"2024-05-10T18:45:25+00:00","versionOfRecord":[],"versionCreatedAt":"2024-01-26 17:52:53","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-3865066","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3865066","identity":"rs-3865066","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}