多量表聯合多層次干預在大學生心理危機預防中的作用

Abstract： health problems potential psychological crises affect the healthy growth learning performance colege students. Effective suitable prevention psychological crises among college students is a continuous challenge university managers face. To explore a method preventing psychological crises among colege students，we measured 38 661 students by using SCL-90 （symptom check list-9O） screened out 5 79O students with positive results. Then，we measured 33 188 students by using PHQ-9（patient health questionnaire-9） screened out 6O3 students with suicidal ideation or behavior；we interviewed 392 students by using GAQ（growth adversity questionnaire）. The number students who had positive results at both phases is 155.As a result，we obtained a data set（ N=76 ）by integrating the students who tested positive on the PHQ-9 （i.e. total score ?20 ） with those who completed the PHQ-9 GAQ. In addition，we obtained a data set （ N （20 =50 ）by excluding the cases in which the GAQ score is O. With regard to QCA（qualitative comparative analysis） results，the data set（ N=76 ）exhibits 5 constellations solutions with a coverage rate greater than O.7， the first eight indicators the PHQ-9 constitute the explanatory variables in the combined solutions. About the data set （ N=50 ），the combined solutions are extremely complicated the explanatory variables encompass indicators from both the PHQ-9 GAQ.Allthese mean that the multi-scale could more comprehensively reflect mental health states college students，thus enhance the accuracy effectiveness the corresponding hierarchical intervention， finally provide support for preventing psychological crises in universities.

Keywords：multi-scale；hierarchical intervention；preventing psychological crises

CLCnumber：O213.9;G40 Document code：A

Frequent occurrence psychological crisis events among college students has emerged as a significant issue in university management. Psychological crises among college students arise when they are confronted withunexpected events which can’t be copied with their personal coping styles or resources[1-2]． The “ Report on National Health Development in （2021-2022）”pointed out that the detection rate severe depressionis 4.94% among college students[3].Depressive symptoms are a significant factorcontributing to suicidal ideation or behavior among college students.Other research indicates that theratio college students contemplating suicide was on an upward trend[4]. According to the WHO’s Suicide worldwide in 2O19 Global Health Estimates， suicide has become the fourth leading cause death among individuals 15 to 29 years old ^?5? —a demographic range that includes the college student population.

Many strategies for preventing college students’ psychological crises are based on the measurement result a single scale assessing their mental statesat a given point in time. However，the accuracy effectiveness such strategies need further consideration when using a single scale to assess college students’ mental states.Thisis because there are dynamic combinations multiple factors influencing college students’mental states.

To address this issue，we propose a method to enhance the accuracy effctiveness preventing psychological crises among college students. Specifically， we employ multiple scales to measure mental states college students at different phases， utilize QCA to examine the concurrent causal relationships that lead to psychological crises among college students， soasto carry out hierarchical intervention in university management.

1 Materials Methods

1.1 Materials

1.1.1 Symptom Check List-90

Thesymptom check list-90（SCL-90）is a scale developed by Derogatis to measure an individual’s mental states over the previous week[6]. The SCL-90 scale consists a total 9O items has nine subscales. The factors measured by the SCL-9O are somatization，obsessive-compulsive symptoms，interpersonal sensitivity，depression，anxiety，hostility， phobic anxiety，paranoid ideation， psychoticism， in addition to other items that do not fall into any these specific categories. The SCL-9O adopts a fivepointscoringsystem（i.e.，1 indicates“notatall”， 2 indicates“a littlebit”，3indicates“moderately”， 4 indicates“quite a bit”， 5indicates“extremely\"）. Based on the criteria set bythe scale developers for interpreting themeasurement results， we set a cutf value as total score =160[6].

1.1.2 PatientHealthQuestionnaire-9

The patient health questionnaire-9（PHQ-9）is a scale developed by Spitzer that is based on the depression section the Diagnostic Statistical Manual Disorders-Fourth Edition[7]. PHQ-9 measures an individual’s depressive symptoms over the past two weeks. The PHQ-9 consists nine items，covering the dimensions depressed mood， loss interest，sleep disorders，fatigue，changes in appetite，feelings worthlessness or excessive guilt， difficultyconcentrating，psychomotoragitationor retardation， suicidal thoughts. The PHQ-9 adopts a four-point scoring system（i.e.O indicates“not at all”，1 indicates“several days”，2 indicates“more than half the days”， 3 indicates“nearly every day”）.

Based on the total score the respondent’s measurements， the PHQ-9 categorizes depressive symptomsinto five severitylevels（i.e.，thescore （2 0-4 indicates“no depressive symptoms”，5-9 scores indicates“mild depression”，1O-14 scores indicates“moderate depression”，15-19 scores indicates“moderately severe depression”，20.-27 scores indicates“severe depression”）.We set a total score ?20 as the cutf score in this study.

Because QCA processes data based ondichotomous variables，we code each item the PHQ-9 as follows ： responses O or1 asO， responses 2or3 as1.

1.1.3 Growth Adversity Questionnaire

The growth adversity questionnaire （GAQ） is a tool designed to measure the potentially traumatic events that an individual has experienced during their childhood. In our study，GAQ comprises 14 items employs a dichotomous scoring system，where selecting“yes”for each item assignsa score 1 selecting“no”assigns a score O. The first 1O items GAQ are derived from the Adverse Childhood Experiences Questionnaire in this book“The Deepest Well\"[8]. These items measure various dimensions adverse childhood experiences， such as emotional abuse，physical abuse，sexual abuse，domestic violence，livingwith someonewho hasasubstance abuse problem，livingwith someone who has a mental illness or is suicidal，livingwith someone who has beenin prison，loss family functioning，emotional neglect， physicalneglect.Furthermore，we added four interview items （i.e. peer bullying[9] residing in a dangerous neighborhood， witnessing community violence， facing financial difficulties within the family[10-11]）.

1.2 Methods

1.2.1 ConfirmatoryFactor Analysis

Confirmatory factor analysis（CFA）is a common application that involves analyzing sample data to verifywhether the relationship between a factor its corresponding measurement items conforms to the theoretical relationship designed by researchers[12-13]. To examine the applicability the SCL-9O in measuring the mental health among college students，we used RStudio to test factor indicators（i.e. loading，weight， averagevariance extracted（AVE）， communality）.

1.2.2 QualitativeComparativeAnalysis

The qualitative comparative analysis （QCA）isa method to study multiple concurrent causal relationships. Charles C. Ragin his colleagues first introduced

QCA in 1987^[14] . QCA posits that the emergence outcome Y isaresult combinations multiple conditions （ A*BY ）ordifferent combinations multiple conditions （ A*B+C*DY） ，rather than asingle condition. In certain specific scenarios，the presenceorabsence aparticularcondition can also lead to outcome Y . In this study，we adopted QCA to analyze the various dimensions depressive symptoms adverse childhood experiences involved in PHQ-9 GAQ，exploring the predictive explanatory power different combinations causes for the emergence psychological crises among college students.

1.2.3 VariableAssignment

Weset“suicidal or self-harming ideationor behavior”as the outcome variable Y ，corresponding to the ninth item PHQ-9，which is“thoughts death or self-harm”[7]． And we set 22 causal variables.Amongthem，thefirst14causalvariables， labeled X₁ to X₁₄ in sequence，are derived from the 14 items GAQ. The remaining 8 causal variables， labeled X₁₅ to X₂₂ in sequence，are derived from the first eight items PHQ-9. Tab.1 displays the variable assignments their respective dimensional descriptions for the GAQ PHQ-9.

1.3 Subject Process

Inthis study，we selected college students from 2021 to 2023 in a university Guizhou Province （including undergraduates postgraduates） as the research subjects.We divided the research into two phases（Fig.1）.

Phase 1： During the first academic year，we surveyed the mental states 38 661 college students using SCL-90 identified 5 790 students with positive test results. We used CFA to examinewhether SCL-9O is applicable for measuring the mental states college students.

Phase2：During the second or third academic year，we surveyed the depression level 33 188 college students using PHQ-9， identified 603 students with positive test results，indicating the presence suicidal ideation or behavior（those who chose2or3fortheninthitem）.Then，weconducted interviews using the GAQ to assess various dimensions adverse childhood experiences among 392 college students. This group included 198 studentswho completedthe PHQ-9，aswellas194studentsthat had not completed the PHQ-9 but were identified by tutors counsellors based on their information.We integrated the data college students who had positive PHQ-9 results participated in GAQ interviews to further explore the multiple concurrent causal factors leading to suicidal or self-harming ideation or behavior among college students. At the last，we obtained two data sets（ N=76 N=50 ）， afterexcludingthosewhoscored OonGAQ. Subsequently，we utilized QCA to analyze both groups.

2 Results

2.1 CFAResults SCL-90

After CFA process，we obtained the factor loading，weight，AVE， communality the SCL90.The AVE value eachfactorin SCL-9O iswithin therangeO.25to O.75 which indicatesthat the AVE values are acceptable [15]

It is worth noting that some loading values for the itemscorresponding to each factor SCL-9O are less thanO.7.We speculatethatthismay be due to the following reasons：first， the SCL-9O has a large number items overall， each factor corresponds to more than 5 items；second，SCL-9O measured 38 661 college students， representing a large sample size；third，the quality the responses varied among the students.Given the large sample size scale， wethink that loadingvalues thisstudyare acceptable.

2.2 Results Multi-scale Measurements

Phase1：We used the SCL-9Oto measure the mental states 38 661 college students identified 5 790 college students with positive results.

Phase 2：We used the PHQ-9 to measure the depression level 33 188 college students identified 6O3 students with positive results.We used the GAQto interview the adverse childhood experiences 392 college students identified 5O students with scores greater than O.

Based on the results multi-scale measurements interviews，we found 155 college students who showed positive results at both phases（Fig.2）.

During the phase 2，we identified 76 college students who had positive PHQ-9 result participated in the GAQ interview；this number was reducedto5Oafterweexcluded these students whose GAQ’s total score is O（Fig.3）.

2.3 QCA Results

Forthe data sets（ N=76 N=50 ）atthe second phase，we employed Tosmana 1.61 to conduct csQCA fsQCA obtained the corresponding truthtables composite solutions.The truth tables for both csQCA fsQCA can be found at the website provided in Appendix A.Tablesat the end the paper.

2.3.1 Results （ N=76 ）

Afterconducting csQCA，we obtained the csQCA truthtable（ N=76 ）.We found that the cause variables in the most concentrated group cases are entirely focused on the indicators corresponding to the first eight items PHQ-9（ X₁₅ to X₂₂ ），with a very lowdegree correlation with the cause variables corresponding to the 14 items GAQ.

The csQCA truth table （ N=76 ）revealsthat the characteristics the cases are relatively dispersed. To further improve the similarity among the studied cases，we conductedadimensionality reduction process on the causal variables. After selecting the combination causal variables （X₁₅，X₁₆，X₁₇，X₁₈，X₁₇，X₁₇，X₁₇，X₁₇，X₁₇，X₁₇，X₁₇，X₁₇，X₁₇，X₁₇，X₁₇，X₁₇，X₁₇，X₁₇，X₁₈，X₁₇，X₁₇，X₁₇，X₁₇，X₁₇，X₁₈，X₁₇，X₁₇，X₁₇，X₁₇，X₁₈，X₁₇，X₁₇，X₁₈） X₁₉ ， X₂₀ ， X₂₁ ， X₂₂ ）that showed the highest concentration cases，we proceeded with fsQCA obtained the fsQCA truth table （ N=76 ）．The results indicate that the cases college students experiencing suicidal ideation orbehavior tend to be most frequent when all the corresponding indicators for the first eight items the PHQ-9 are positive.

Whenwe set the consistency threshold to1，the data set reveals the existence five valid fsQCA solutions with a coverage greater than O.7（Tab.2）. The results show that when we set the consistency threshold to ¹ ，we got a total five valid combinations solutions ：

This means that when the consistency threshold is set to ¹ ，predicting the occurrence suicidal or self-harming ideation or behavior among college students，five constellations causalvariables achieve explanatory coverages 71.05% ， 80.20% ， 82.89% ， 71.05% 71.05% .Among these，the combination causal variables with the highest explanatorycoverage is“loss interest * depressed mood * sleep disorder feeling fatigue* changes in appetite * negative self-evaluation”.

It isworth noting that this combination comprises three measurement dimensions the PHQ-9 （i.e.， cognitive behavioral，emotional， somatization manifestations）.

2.3.2 Results （ ?（N=50 ））

When we usedthedataset （ N= 76 ）for analysis，we obtained only the combinedsolutions regarding the causal variables the PHQ-9.To explore the relationship between the causal variables the GAQ suicidal or self-harming ideation orbehavior among college students， we conducted csQCA using the data set （ N=50 ）， got the csQCA truth table（ N=50 ）．Based on the csQCA truthtable （ N=50 ），we selected 14 causal variables （ X₁ ， X₄ ， X₇ ， X₁₂ ， X₁₃ ， X₁₄ ， X₁₅ ， X₁₆ ， X₁₇ ， X₁₈ ， X₁₉ ， X₂₀ ， X₂₁ ， X₂₂ ）corresponding to the most concentrated cases. After retaining these 14 causal variables，we obtained another csQCA truth table（ N= 50）after filtering the causal variables further selected 12 causal variables （X₄，X₇，X₁₂，X₁₄，X₁₅， （20 X₁₆ ， X₁₇ ， X₁₈ ， X₁₉ ， X₂₀ ， X₂₁ ， X₂₂ ） for fsQCA.

Based on the selected 12positive causal variables，weobtained the fsQCAtruthtable（ N= 50） the solution combinations displayed in Tab.3.

Theresults in Tab.3 indicate that all the combinationscausalvariablesarefarbelowO.7O.We speculate thatafter addingthepositivecausal variables the GAQ，the predictive power the scales’measurements for suicidal or self-harming ideation or behavior among college students is weak. In other words，the causal combinations in Tab.3 suggestthat therelationship between adverse childhood experiences suicidal or self-harming ideation or behavior among college students is highly intricate， asmallnumbercombinationscausalvariables can't fully explain this causal relationship.

3 Discussion

This study suggests several applications. First， tutorscould give comprehensive consideration these complex causes （e.g. depressive symptoms childhood experiences） in daily student management[9，16]. Second，counsellors could assess students’ mental resilience based on their depressive symptoms or childhood experiences，so as to provide personalized support when they encounter unexpected events which can't be copied with their personal coping styles or resources. Finally， university managers could pay attention to the dynamic changes college students’ mental states the underlying cause mechanisms， hence building the dynamic prevention system with digital tools artificial intelligence technology[17].

Appendix A.Tables

Thesupplementarytablesin thisarticle can be found online at https：//cstr.cn/31253.11.scienced b. 19613.

References：

[1]CAPLAN G. The principles preventive psychiatry[M].NewYork：BasicBook，1964：26-38.

[2]樊富珉.“非典”危機反應與危機心理干預[J]．清華大學學報（哲學社會科學版），2003，18（4）：32-37.

[3］傅小蘭，張侃，陳雪峰，等．中國國民心理健康發展報告（2021—2022）[M]．北京：社會科學文獻出版社，2023：76-79.

[4]HUANGSQ，WANGDF，ZHAOJB，etal.Changes insuicidal ideation related influential factors in collegestudentsduring the COVID-19 lockdown in [J].Psychiatryresearch，2022，314（8）：40-49.

[5]World Health Organization.Suicide worldwidein 2019：global health estimates[EB/OL].（2019-06-16）[2024-07-20].https：//iris.who.int/bitstream/hle/10665/341728/9789240026643-eng. pdf？ sequence = 1.

[6]DEROGATISLR，RICKELSK，ROCKAF.TheSCL-90 theMMPI：astepin thevalidation anewself-reportscale[J].BritishJournal Psychiatry，1976，128（3）：280-289.

[7]KROENKEK，SPITZER RL，WILLIAMSJB.ThePHQ-9：validity a brief depression severity measure[J].Journal General Internal Medicine，2OO1，16（9）：606-613.

[8]HARRIS N B. The deepest well：healing the long-termeffects childhood adversity[M].Hangzhou：ZhejiangEducation PublishingHouse，2020：225-228.

[9］曾欣然，汪玥，丁俊浩，等．班級欺凌規范與欺凌行為：群體害怕與同輩壓力的中介作用[J]．心理學報，2019，51（8）：935-944.

[10]程剛，張大均．家庭社會經濟地位對大學新生抑郁情緒的影響：有調節的中介模型[J].心理與行為研究，2018，16（2） ： 247-252.

[11]金童林，烏云特娜，張璐，等．社會性創傷對大學生自殺意念的影響：心理痛苦和樂觀的作用[J].心理科學，2024，47（4）：1008-1015.

[12]MARSHWH，MORINJA，PARKERDP，etal.Exploratory structural equationmodeling：an integrationthebest featuresexploratory confirmatory factoranalysis[J].Annual Review Clinical Psychology，2014，10（1） ： 85-110.

[13]LLORET-SEGURAS，FERRERES-TRAVERA，HERNANDEZ-BAEZA A，et al.Exploratoryitem factoranalysis：a practical guide revised updated[J].AnnalsPsychology，2014，30（3）：1151-1169.

[14]RAGIN C C，RIHOUX B.Qualitative comparative analysisrelated techniques：configurational comparative methods[M].Beijing： Machine Press，2O17：1-104.

[15]FORNELL C，LARCKER D F. Evaluating structuralequationmodelswithunobservablevariables measure-menterror[J].JournalMarketingResearch，1981，18（1）：39-50.

[16]陳春平，何子耕，譚子妍，等．睡眠障礙與大學生自殺意念的關系：一個有調節的中介模型[J]．中國臨床心理學雜志，2023，31（5）：1112-1116.

[17]符虔，趙海騰，趙小青，等．大數據視角下數字社區用戶群體人格畫像［J]．貴州大學學報（自然科學版），2023，40（6）：62-68.

（責任編輯：周曉南）

摘要：大學生的心理健康問題與潛在心理危機影響其健康成長和學習表現。如何有效、恰當地開展大學生心理危機預防工作一直是高校管理者面臨的挑戰。為探索一種大學生心理危機預防的方法，首先，采用癥狀自評量表（SCL-90）對38 661名大學生進行測量，篩選出5790名陽性結果的大學生。其次，使用9條目患者健康問卷（PHQ-9）對33188名大學生進行測量，篩選出603名有自殺意念或行為的大學生;使用成長逆境問卷（GAQ）訪談了392名大學生。兩階段測量結果均呈陽性的人數為155。通過整合PHQ-9陽性結果大學生（總分 ?20 ））且同時完成 PHQ?9 測量與GAQ訪談的大學生，獲得了一個數據集（ N=76 ）；通過排除GAQ總分為0的個案，獲得了另一個數據集（ ）。定性比較分析的結果顯示，數據集（ N=76 ）存在5組覆蓋率大于0.7的組合解，且組合解的原因變量均由PHQ-9的前8項指標構成；數據集（ N=50 ）所有解的組合情況非常復雜，且解的原因變量包含了 PHQ-9 和GAQ的指標。結果表明，多量表能夠更全面地反映出大學生的心理健康狀態，進而提高對應的多層次干預的精準性與有效性，并為高校的心理危機預防工作提供重要支持。

關鍵詞：多量表；多層次干預；心理危機預防