中文關鍵字 生命態度量表；青少年憂鬱；自殺行為

• 系統編號 RG9811-0373

• 計畫中文名稱 憂鬱及自殺青少年檢測及處遇模式之建立與評值

• 計畫英文名稱 --

• 主管機關 行政院衛生署 • 計畫編號 NHRI-EX98-9821PC

• 執行機構 台北醫學大學護理學系

• 本期期間 9801 ~ 9812

• 報告頁數 30 頁 • 使用語言 中文

• 研究人員 張秀如；顏淑華；詹曇璇；；； Chang, Hsiu-Ju(Shi；Yen, Shu-Hua；Chan, Tanm-Hsuan；；；

• 中文關鍵字 生命態度量表；青少年憂鬱；自殺行為；；；；；

• 英文關鍵字 ；；；；；；；

• 中文摘要

（1） 中文摘要 目的 生命態度量表(Life Attitude Schedule, LAS)用於評估青少年自殺危險性及威脅生命的行為。本研究目的為翻譯並發展中文版 生命態度量表(Chinese Version of the Life Attitudes Schedule , LAS-C)，並檢測其信效度。 研究背景 青少年自殺行為的評估應包含保護性因素及危 險因素，生命態度量表可用於評估青少年及成人早期自殺危險性及威脅生命的行為。 研究方法 研究設計：本研究為縱貫性及前瞻性研究設計，

以社區青少年為樣本，探討中文版生命態度量表的信效度。量表翻譯係採用 Brislin (1970)所建議的步驟。量表翻譯已取得加拿大 Multi-Health Systems (MHS)出版公司授權同意。 研究對象：以方便取樣從台北市高中名單中，選取五所同意進行研究的學校，共收集 1492 位研究樣本完整填寫研究 相關問卷，回收率為 97.2%。於三週後，從 1492 位研究樣本中選取 135 位研究樣本進行後測，以檢測量表之再測信度。此外，以兒童憂鬱量表 (Children’s Depression Inventory, CDI)得分 19 分以上(含 19 分)以評估社區中具憂鬱高風險青少年。Kovacs (1982, 1981)建議以 19 分為切點分數評估 及確認具臨床診斷意義之憂鬱症狀。本研究樣本中發現，若以 19 分為切點分數，約 47.4%的青少年具憂鬱症狀的高危險性。 生命態度量表翻譯、

回覆翻譯及文化對等性之評估：首先由一位熟悉中文及英文，且其專業領域與本研究主題相關的專業翻譯者將英文版生命態度量表翻譯成中文版 生命態度量表，再邀請另一位沒有讀過英文版生命態度量表的翻譯專家將中文版本回覆翻譯成英文版本。再由研究者組成審核小組比對及檢視二 種英文版本的差異，若文意表達有差異，則反覆進行上述過程，直到翻譯的中文版與原始版本達一致性為止。為評估文化對等性的問題，本研究 邀請七位兒童青少年精神心理衛生領域且精通中文及英文的專家學者擔任專家效度的評量，其中包含二位兒童及青少年精神科醫師、一位護理教 師(精通兒童及青少年精神科護理及心理計量研究)、三位精神科臨床心理師及一位精神科臨床社工師。七位專家分別就內容適切性及語意清晰性，

以 4 分計分法評量並提供修改意見。之後計算各題的平均分數，以評估個題目包括量表的適切性。依 Waltz, Strickland, & Lenz（1991）其所得結 果大於 0.80 表示具有良好效度指標，結果發現最終的中文版生命態度量表專家效度指數為 0.99。 研究者將最終版的中文版生命態度量表施測於

36 位高中學生，多數學生反應容易理解及閱讀，填寫量表時間約 10 分鐘。 研究工具：研究工具包括兒童憂鬱量表(Children’s Depression Inventory, CDI)，貝克無望感量表(Beck Hopelessness Scale, BHS),認知三元素量表(Cognitive Triad Inventory for Children, CTI-C),正向及負向自殺意念量表 (Positive and Negative Suicidal Ideation Scale, PANSI),及自殺危險程度量表 (Scale of Suicidal Risk)。 資料分析：本研究運用 SPSS/PC for Windows version 15.0 及 Amos version 7.0 進行統計分析。信度分析採用 Cronbach’s α 係數及再測信度，以組內相關係數(Interclass correlation coefficient, ICC) 來檢定再測信度，信度係數高於 0.7 是可接受範圍。採用驗證性因素分析(Confirmatory Factor Analysis ,CFA)檢視量表的理論性因素結構以建立建 構效度。以獨立樣本 t 檢定比較社區憂鬱高危險群及低危險群在中文版生命態度量表的得分是否有差異。以專家效度指標(Content Validity Index, CVI)建立內容效度。以中文版生命態度量表之總分與貝克絕望量表、負向自殺意念(PANSI-NSI)之總分是否具有高度正相關檢測聚歛效度

(Convergent validity)。以中文版生命態度量表之總分與兒童認知三元素量表(CTI-C)、正向意念(PANSI-PI)之總分是否具有高度負相關檢測分歧效度 (Divergent validity)。以中文版生命態度量表與兒童憂鬱量表(CDI)、自殺危險程度量表是否具有相關檢測同時效度(Concurrent validity)。 研究結果 基本人口學特質：1492 名樣本中，241 人就讀公立高中(16.2%)、731 人就讀私立高中(49.0%)、520 人就讀高職(34.9%)，平均年齡為 16.24(SD = .97)，

男性佔 42.8，女性佔 57.2%。多數學生來自完整家庭(父母結婚且同住) (n = 1153, 77.4%)。 描述性統計及研究變項間相關 主要研究變項間皆呈統 計上的顯著相關，相關係數範圍為.25 至.94。 信度 依中文版生命態度量表及各次量表呈現 Cronbach α 值如下：四大內容次量表，包含死亡關連 (death-related; DR) .78、健康關連(health-related; HR) .70、傷害關連(injury-related; IR) .75、自我的關連(self-related; SR) .85。 三種行為項目次量表 則包含行動(Actions) .75、想法(Thoughts) .79 和感覺(Feelings) .81。兩種複合次量表包含生理層面(Physical).82 及心理層面(Psychological).89，兩種 價值次量表包含正向.86 及負向.86，總量表則為.91。以 Cronbach’s α-if-item-deleted 分析也發現沒有任何一題刪除後，總量表或各次量表之 Cronbach’s α 能有更大改善空間，以上得分顯示中文版生命態度量表具適當之內在一致性。 題項-題項之相關 總量表及各次量表之題項與題項之 間的相關係數皆低於.80，顯示中文版生命態度量表沒有所謂共線性的問題。 再測信度 於第一次施測後三週，自樣本中選取 135 名學生再次填寫 中文版生命態度量表，各次量表及總量表之組內相關係數如下：四大內容次量表：死亡關連(death-related; DR) .68、健康關連(health-related; HR) .67、

傷害關連(injury-related; IR) .75、自我的關連(self-related; SR) .70。三種行為項目次量表：(Actions) .72、想法(Thoughts) 1.0 和感覺(Feelings) .68。

兩種複合次量表包含生理層面(Physical).76 及心理層面(Psychological).72，兩種價值次量表包含正向.75 及負向.68，總量表則為.75。以上組內相關 係數皆趨近或大於.70，且皆達到統計上的顯著差異(p < .001)，顯示本量表具良好之再測信度。 效度 建構效度 以聚斂及分岐效度呈現本量表與 相關概念間的關係，在聚斂效度方面，結果顯示中文版生命態度量表(LAS-C)與中文版正向及負向自殺意念量表中的負向意念次量表(PANSI-NSI) 具正向且顯著的相關(r = .54)，與貝克無望感量表(BHS)亦呈現正向且顯著的相關(r = .67)。在分歧效度方面，結果顯示中文版生命態度量表(LAS-C) 與中文版正向及負向自殺意念量表中的正向意念次量表(PANSI-PI)具負向且顯著的相關(r = -.57)，與兒童認知三元素量表(CTI-C)亦呈現負向且顯著 的相關(r = -.72)。以上所有相關係數皆達到統計上的顯著相關(p < .001)。以獨立樣本 t 檢定發現中文版生命態度量表總量表及各次量表得分，在社 區高憂鬱風險及低憂鬱風險的青少年樣本中具顯著差異(p < .001)，以上結果支持本量表具良好的建構效度。 本研究進一步以驗證性因素分析檢視 建構效度，採用 Lewinsohn 等人(2004)所建構之理論模式。將 96 題題項重新組合成 24 個新的變項，以確保題項之分佈較為平均。模型之適配指 數如下：GFI=.93, AGFI=.90, NFI=.89, NNFI=.89, CFI=.91, RMSEA=.06, , p<.001. GFI, NFI, AGFI, RMSEA, CFI, 及 NNFI 的適配指數皆接近可接受 的切點分數，結果支持 本研究所採用以四大內容向度、三種行為項目及二種價值取向所見建構的理論模式。 同時效度 在同時效度方面，結果顯 示中文版生命態度量表(LAS-C)與中文版兒童憂鬱量表(CDI)具正向且顯著的相關(r = .67)，與自殺危險性量表呈現正向且顯著的相關(r = .56)。 結

論 中文版生命態度量表具良好之信度及效度，可用以評量青少年的自殺危險性及威脅生命的行為，未來研究宜進一步與本研究結果比較以探討研 究結果的推論性。

• 英文摘要

1-(2)-1 （2）英文摘要 AIM This paper is a report of a study conducted to develop a Chinese version of the Life Attitudes Schedule (LAS-C), a risk assessment for suicidal and life-threatening behaviors, and evaluate its psychometric properties. BACKGROUND Comprehensive assessment of suicidal behavior in youths should incorporate both protective factors and risk factors. The LAS was developed to identify adolescents and young adults who are at increased risk of suicidal and/or life-threatening behaviors. METHODS Design：A longitudinal and prospective design was used to evaluate the reliability and validity of the LAS-C with a community-based sample of adolescents.We used a procedure proposed by Brislin (1970) to guide the translation of instruments. Permission for the translation was obtained from the publisher Multi-Health Systems (MHS). Participants：A convenient sampling was used to select five schools in Taipei city. A sample of 1492 normal community-based adolescents completed the questionnaires. The response rate was 97.2%. The instruments were administered to a convenience sample of 135 adolescents and were then re-administered to them three week later to evaluate test-retest reliability. Furthermore, the CDI was used to identify a high-depression-risk community group. Kovacs (1982, 1981) suggested that a score of ≥19 can be used to identify a clinical level of depression. Using the cutoff score of 19, 47.4% of adolescents were found to have a clinical level of depressive symptoms, and these students were categorized as at high risk for depression. Translation, back-translation, and equivalence assessment of the LAS-C 1-(2)-2 The English version of the LAS was initially translated into Chinese by one of the professional translator, who is bilingual in Chinese and English and experienced in the study topic. A second qualified and blinded bilingual expert back-translated the Chinese version of the LAS, creating a new English version. A monolingual reviewer on the research team then compared the original and back-translated versions of the LAS. To guarantee cultural

equivalence, a refinement group of 7 bilingual experts in child and adolescent mental health was formed, including two child and adolescent psychiatrists, one nursing faculty who’s field is inchild and adolescent psychiatric nursing and psychometric property testing, three clinical mental health psychologists, and one mental health social worker. These experts were asked to rate the relevance of the content independently using a content validity index (CVI). They were asked to rate each item on the LAS-C based on relevance and semantic equivalence using the following 4-point Likert scale: 1 = not relevant (not appropriate), 2 = somewhat relevant (somewhat appropriate), 3 = relevant (quite appropriate), 4 = very relevant (very appropriate). The CVI was then computed based on the percentage of total items rated by the experts as either 3 or 4. According to Norwood (2000), a CVI rating exceeding .80 can be considered to show good content validity. All items in the LAS-C were rated by the panel experts as having a CVI greater than .90. For the final version, the CVIs for relevance and semantic equivalence both reached .99. The LAS-C was then pilot-tested with a small sample of 36 high school students recruited from one high school. Most of the adolescents confirmed that all items on the LAS-C were easy to read and understand, and the form took approximately 10 minutes to complete. Instruments Instruments used included the Children’s Depression Inventory(CDI), Beck Hopelessness Scale (BHS), Cognitive Triad Inventory for Children 1-(2)-3 (CTI-C), Positive and Negative Suicidal Ideation Scale (PANSI), and Scale of Suicidal Risk. Data analysis SPSS/PC forWindows version 15.0 and Amos version 7.0 software were used for the statistical analyses. Several methods were used to evaluate the LAS’s reliability and validity.Reliability was evaluated with Cronbach α’s coefficients and test-retest reliability. Test-retest reliability was measured with the intraclass correlation coefficient (ICC). An ICC of ≥ .70 was considered acceptable (Scientific Advisory Committee of the Medical Outcomes Trust 2002).

Confirmatory factor analysis (CFA) was used to confirm the theoretical factor structures to support the evidence of construct validity. An

independent-sample t-test was used to compare scores on the subscales between the community high depression-risk and low depression-risk groups. Content validity was evaluated through 7 experts’ comments on the translated LAS. Convergent validity was demonstrated by a positive relationship between the LAS and the PANSI-NSI and BHS. Divergent validity was demonstrated by a negative relationship between the LAS and the CTI-C and PANSI-PI.

Concurrent validity was evidenced by a significant relationship between the LAS and the CDI and Scale of Suicidal Risk. RESULTS Demographics The sample of 1492 adolescents included 241 public high-school students (16.2%), 731 private high-school students (49.0%), and 520 (34.9%) occupational high-school students with a mean age of 16.24 (SD = .97) years. Gender was represented equally, with 42.8% boys and 57.2% girls. The majority of the adolescents were from intact families (n = 1153, 77.4%). Descriptive statistics and intercorrelations among the study variables Correlations were used to examine relationships among the variables. 1-(2)-4 Statistically significant correlations were present among the major study variables. The correlation coefficients ranged from .25 to .94. All correlations were significant at the p < .001 level. RELIABILITY ESTIMATE Internal Consistency The LAS

encompass 4 Content scales (Death-Related 【DR】, Health-Related 【HR】, Injury-Related 【IR】, and Self-Related【SR】), 3 Modality Type scales (Actions, Feelings, and Thoughts), 2 Composite Scales (Physical and Psychological), 2 Valence Scales (Negative and Positive), and a Total Score. The Cronbach’s α coefficients for the DR, HR, IR, and SR were were .78 and .70, .75, and .85, respectively. The Cronbach’s α coefficients for the Action, Feelings, and

Thoughts were .75 .81, and 79, respectively. The Cronbach’s α coefficients for the Physical and Psychological were .82 and 89, respectively. The Cronbach’s α coefficients for the Positive and Negative were .86 and 86, respectively. The Cronbach’s α coefficients for the total scores on the LAS were .91. All the results indicated adequate internal consistency based on the suggested criterion level for a coefficient’s α of .70 or above (Nunnally & Bernstein 1994). The Cronbach’s α-if-item-deleted showed that no items were unreliable. Inter-Item Correlations Inter-item correlations were all below .80, indicating lack of multicollinearity (Tabachnick & Fidell 2007). Test-retest reliability The LAS-C was re-administered at a three-week period for the same sample. ICC coefficients for the DR, HR, IR, and SR were .68 and .67, .75, and .70, respectively. ICC coefficients for the Action, Feelings, and Thoughts were .72, 1.0 and .68, respectively. ICC coefficients for the Physical and Psychological 1-(2)-5 were .76 and .72, respectively. ICC coefficients for the Positive and Negative were .75 and 68, respectively. ICC coefficients for the total scores on the LAS were .75. All the ICC coefficients were close to or greater than the accepted value .70, indicating adequate test-retest reliability. All correlations were statistically significant at the p < .001 level. VALIDITY ESTIMATE Construct Validities Convergent and divergent validities were demonstrated by relationships between the subscales and the conceptually-related constructs.

Evidence of convergent validity was demonstrated by statistically significant and positive correlations between total scores on the LAS-C and the

PANSI-NSI (r = .54) and BHS (r = .67). Evidence of divergent validity was demonstrated by statistically significant and negative correlations between the total scores on the LAS-C and the PANSI-PI (r = −.57) and CTI-C (r = −.72). Contrasting group comparisons were also used to support the construct validity of the subscales. The independent-sample t-test revealed that the total scores on the LAS and its subscales in the community high-depression-risk sample were statistically significantly higher than those in the community low-depression-risk sample (p < .001). High-depression-risk community sample demonstrated higher scores on the whole scale of the LAS-C and its subscales. Construct validity was examined further using CFA.We replicated the theoretical model suggested by Lewinsohn et al., (2004).We parceled the 96 LAS-C items into 24 composite items in order to form better item distributions

for the CFA. The fit of this model was found to be good: GFI=.93, AGFI=.90, NFI=.89, NNFI=.89, CFI=.91, RMSEA=.06,2(243)1231.9, p<.001. The goodness-of-fit indices for the GFI, NFI, AGFI, RMSEA, CFI, and NNFI were all close to the cutoff values for an acceptable fit. This model provides support 1-(2)-6 for the validity of the 4 Content and 3 Modality and 2 Valence Scales. Concurrent Validity Concurrent validity was evidenced by a

significant relationship between the LAS-C and the CDI (r = .67) and Scale of Suicidal Risk (r = .56). All correlations were statistically significant at the p

< .001 level (Table 2). CONCLUSION The LAS-C may be a reliable and valid instrument to measure the risk of suicidal and life threatening behaviors in adolescents, subject to further research to test the generalizability of the present findings.