自殺防治熱線年輕來電者立即自殺風險狀態相關因素之配對病例對照研究

Abstract

背景：在臺灣，自殺為年輕族群之主要死因之一。自殺防治專線為常見之自殺預防策略，惟過去研究較少運用其通話資料，深入分析年輕族群之自殺相關風險。本研究以全國自殺防治專線之年輕來電者為研究對象，分析其社會人口學特徵、語言情緒傾向及自殺風險，並探討與僅具自殺意念者相較下，上述因素與來電者是否屬於具立即自殺風險者之關聯性。 方法：本研究納入2019年7月1日至2020年5月21日期間撥打全國自殺防治專線小於30歲之來電者。研究對象須具有近期自殺意念，並進一步區分為具立即自殺風險者與僅具自殺意念者。其中，具立即自殺風險者定義為近期每週或每日具有自殺意念，且同時出現自殺準備行為或正在進行自殺行為。研究採1:1配對設計，依性別及年齡（±1歲）與僅具近期自殺意念之對照來電者進行配對。使用由熱線志工編碼之通話資料進行描述性統計分析，以呈現研究對象之社會人口學特徵、心理健康狀況及通話特性之分布。此外，透過文本情感分析（sentiment analysis）及自殺風險進行編碼，以探討來電者情緒狀態及自殺風險於通話過程中之變化。最後，採用條件式羅吉斯迴歸分析（conditional logistic regression）評估各變項與自殺風險狀態之關聯性。 結果：由熱線志工編碼之社會人口學特徵、心理健康特徵及文本情感分析之分數等資料分析結果，在具立即自殺風險狀態者（病例組）與僅具自殺意念者（對照組）之間並無顯著差異。相較於對照組，病例組於通話歷程中呈現較高之整體自殺風險分數與自殺意念分數，並於通話開始時具有較高之自我傷害意念分數，且於通話結束時具有較高之自殺意圖程度。在條件式羅吉斯迴歸分析中，通話結束時之自殺意圖以及通話歷程中之自殺意念皆與較高立即自殺風險狀態之勝算比相關。同時，與他人同住與較低立即自殺風險狀態之勝算比相關。整體而言，病例組與對照組在通話進行過程中皆呈現語言情緒趨於正向、且自殺風險分數逐漸下降之趨勢，惟兩組於通話開始至結束之變化幅度上並未發現顯著差異。 結論：自殺防治專線之通話資料，為理解其使用族群之自殺風險與相關因子提供重要之研究素材。對臺灣之年輕族群而言，本研究結果顯示自殺防治專線為具效果之自殺預防措施。後續研究需進一步釐清對話過程中情緒與自殺風險改善之機轉。此外，本研究期能協助自殺防治專線工作人員與志工更早期辨識具較高自殺風險之來電者，並作為未來訓練設計之參考。

Background: In Taiwan, suicide continues to be a leading cause of death for children, adolescents, and young adults. Suicide crisis hotlines are a common strategy for suicide prevention, yet few studies have utilized their data to analyze adolescent and youth suicide. This study analyzed the socio-demographic characteristics, linguistic sentiment, and suicide risk of young callers to the National Suicide Prevention Hotline to assess their association with being a caller who is in an active state of suicidality versus callers who had only suicidal ideation. Methods: Callers younger than 30 from July 1, 2019 through May 21, 2020 with a recent history of suicide ideation and in an active state of suicidality, defined as having both weekly or daily suicidal thoughts and either preparing for or having an ongoing suicide attempt, was matched 1:1 by sex and ±1 year of age with control callers who only had a recent history of suicide. Descriptive statistics were used to describe the distribution of socio-demographic, mental health status, and call characteristics using data coded by the hotline volunteer. Sentiment analysis and suicide risk were coded to investigate changes in caller emotion and suicide risk over the course of the call. Conditional logistic regression analysis was used to assess associations with suicide state. Results: Demographic and mental health characteristics coded by the hotline volunteers and sentiment analysis scores were not different between callers who were in an active state of suicidality (i.e., cases) compared to controls. Cases were more likely to have higher overall suicide risk scores and suicidal ideation scores throughout the call, higher self-harm ideation scores at the beginning of calls, and higher suicidal intent at the end of calls. Suicidal intent at the end of the call and suicidal ideation throughout the call were found to be associated with higher odds of being in an active state of suicidality, while living with others was found to be associated with lower odds of active suicidality. Overall, there was a general trend toward more positive sentiment and lower suicide risk scores as calls progressed for both cases and controls, with no difference found in the level of changes from the start to the end of the calls. Conclusion: Hotlines provide abundant data to understand suicide risk and the associated factors among the user population. For young people in Taiwan, our data suggest that hotlines are an effective prevention measure. Further research is needed to better understand the mechanisms underlying the improvement of sentiment and suicide risk during the hotline conversation. In addition, this research will hopefully help hotline workers and volunteers to better identify more severe callers and inform their training.