兒童虐待事件中社工風險評估與成案決策之相關性探究──以桃園縣家暴中心為例

Abstract

近幾年兒童虐待的通報與盛行率持續增加，相較於其它保護案件，兒童虐待議題亦不容小覷。為了有效預防及阻絕兒虐再次發生，第一線社工的評估與決策顯得十分重要。2011年監委曾彈劾一件兒虐案中失職的社工，該事件也開始反思社工一連串評估與決策過程，以及影響社工誤判的可能因素。回顧文獻後，發現許多研究已提出有信、效度的風險評估模型，以及可能影響社工決策之生態因素，但相關研究在目前台灣仍缺乏。有鑑於此，本研究希望了解社工在兒童虐待風險評估與成案決策的現況；進而探究社工兒虐成案決策，與風險評估、社工背景與外部環境之關係，以及評估和決策不一致的原因；最後，對社工決策兒童虐待風險的實務面及政策面提出建議。 本研究採取混合式研究，結合量化與質性研究方法。量化部分以「100年度兒少保護案件成案調查評估報告」為次級資料，以分層隨機抽樣的方法，抽取不同決策類別的個案，使用SPSS統計軟體進行多元邏輯迴歸分析，了解在控制情況下，社工的風險評估與成案決策之關係。得出結果後，再以質性訪談桃園縣家暴中心第一線兒保社工，了解評估與決策不一致原因，以及其他與決策相關之因素。 本研究共抽取607件有效個案，與訪談四位兒保社工。研究結果發現79.6%個案社工整體評估為低危機，72 %個案決策不成案，成案（家庭維繫或家外安置）的比例並不高（28%）。其中兒童的性別與族群，和社工成案決策無關。在以階層邏輯迴歸控制其他變項的情況下，發現「整體危機評估」最能解釋社工成案決策，其次為「壓力與危機（經濟）」、「受施虐者威脅程度」、「受生活照顧狀況」「對兒少受虐態度（期待）」、「受傷部位」、「受傷頻率」等評估指標。代表整體危機評估若為高風險、家庭中的經濟壓力越大、受施虐者威脅風險高、受生活照顧狀況越糟、照顧者對兒少期待越不合理、受傷頻率越高，以及受傷部位在頭臉部，相較於不成案決策，社工決策成案（家庭維繫或家外安置）之可能性也越大。 在質性結果中，發現社工使用風險評估但與最後決策不一致的理由，包含人為操作的技術層面與量表構面問題。人為操作面主要是社工因為時間壓力或對量表態度消極，因此會採取先決策後評估方式。此外，因為網絡分工結果，改由其他機構社工成案，因此即使評估高危機但仍會決策不成案。在量表構面部分，則發現目前使用的兒童少年受虐待暨被疏忽危機診斷表，當中有幾個問題：包括單一指標的決策原則、無法反應實務上關鍵的評估指標、量表項目意義不明、評估指標的危機分類有異議、未納入保護因子考慮等。另外，除了風險因子外，社工的年資背景、外部網絡與壓力團體、機構內部環境與督導制度，以及受訓情況等，也都與社工決策密切相關。 依據研究結果，本研究提出實務、工具及政策上三方面的建議。在實務上，應重視成案決策之預測因子，提供督導討論或新進社工之訓練，並且提供密集且足夠的風險工具使用之訓練，增加社工評估與決策之準確性。再者，應納入實務經驗，定期檢討評估與決策工具。在評估工具上，建議內容描述應統一、具體，並針對社工的評估重點，區分主要評估與次要參考評估因素。在政策上，應重新檢討通報分級制度並增加社工人力，加強宣導風險評估工具的使用，同時增進網絡間交流與溝通平台，以了解兒保相關人員的評估與決策原則。最後，建議檢討〈兒少法〉的調查時限，增加兒保社工評估與決策的彈性。

The incidence of reported child abuse is fast growing in the past few years. To identify a child abuse case, a reliable and valid risk assessment is pertinent for a child protection worker to make appropriate treatment decision. In 2011, according to the Control Yuan, a child protection worker, who once made inconsistent decision from risk assessment of possible harm after a child going home, was impeached for a child death incidence due to that worker’s negligence. The extant literature in Taiwan gave no valid information of how a risk assessment model could be effective to identify a child abuse case. This study aims to test predicting power of a proposed risk assessment model and examine how child protection workers would make inconsistent decision from their risk assessment. A data set provided by Tao-Yuan County Child Protection Team was used to do analysis. A particular form of risk assessment tool was completed by every child protection worker after investigation during 2011. A total of 2,142 cases reported were investigated in 2011, but only 2003 cases were risk assessed in forms. Due to statistical efficiency, a total of 607 cases were stratified randomly sampled from the sampled population. And a qualitative approach was used to collect data from four child protection workers regarding inconsistent decisions they made in risk assessment procedure. Results showed that almost eighty percent of the sampled cases assessed were rated as children with the lower risk and 86.3% were not substantiated for further treatment. In 20.4% of the sampled cases rated as children of higher risks, 61.2% were substantiated as child abuse cases for further treatment services, but the others did not receive treatment. In prediction, the indicators of global risk level, wounds of the body, frequency of the abuse acts, care and living conditions, degree of threat from the abuser, attitude and expectations towards the child, and economic stress were found to have prediction power on child abuse case substantiation. Since some cases with higher level assessed were not substantiated, four first line workers were in-depth interviewed to explain the inconsistency. Two sets of reasons were given regarding the technical issues related to assessing process (decision before filling the form of risk assessment) and the difficulties related to using the risk assessment tool (some concern for the validity). Implications about developing effective risk assessment models and promoting precision between investigation and decision-making are included.