運用人為因素分析及分類工具(HFACS)於異常事件根本原因分析之探討

背景：醫療是複雜的系統，醫療不良事件的發生很多可以歸咎於人為因素(Human Factors)及系統失效等原因，這幾年在錯誤原因的探討中，漸漸突顯人為因素帶給醫療照護產業的影響。目前國內對於醫療不良事件調查中之人為因素辨識與分類，缺乏相關工具應用，對於過去各機構調查與分析的執行狀況更是不得而知，能否有效掌握根本原因進行改善，令人存疑。因此本研究欲探討台灣醫院目前內部進行不良事件根因調查分析現況，與中譯人為因素分析及分類架構工具，進一步實證此工具應用於重大不良事件根因回溯性分析之結果。研究方法：本研究採橫斷性調查，以全國地區教學醫院以上負責病人安全醫療不良事件通報及進行根因調查之單位同仁作為研究對象，以自擬之問卷進行大規模調查，探討目前國內醫院內部對於醫療不良事件之根因調查分析現況，並以雙變量分析及多元線性迴歸進行假說驗證。此外，將Wiegmann與Shapell學者2017年發展之醫療版人為因素分析及分類工具中文化，並補充微分類碼，進行工具信效度檢測。以某醫學中心自2006至2017年間發生之重大醫療不良事件，運用中譯之HFACS工具進行回溯性根因調查分析。研究結果：共122家醫院及負責病安通報之窗口、590名根本原因分析調查員接受調查，問卷回收率分別為93.1%及90.1%。在醫院內部通報機制方面，97.5%醫院設有專責單位及相當資歷之專責人員處理醫療不良事件通報，89.3%為自主通報，通報管道多元，各家通報的件數分佈不一，91%的醫院設有根本原因分析小組，可見醫院推行異常事件通報及分析調查，在國內已相當成熟。本研究調查員有18.6%未接受過訓練，在接受過訓練者中仍有35.8%覺得不足，且有49.1%的調查員僅一半甚至更少的信心以現行的調查工具能有效挖掘人為因素與掌握根本原因，反映出現行調查工具以及國內醫院對於調查員培訓的不足。高達94.2%調查員表示在進行事件分析調查有障礙，障礙前三位為「事後回溯困難，有記憶偏差」、「訪談或調查事件相關人員排斥」及「調查人員經驗不足」。醫院進行不良事件調查人員之「個人經驗」，不論參與調查年資、參與調查件數、是否接受過根因調查訓練、參與根因調查訓練總時數、自覺訓練是否足夠以及人因掌握信心程度等因素，皆與調查不良事件時之「人為因素整體考量程度」、「顯性失效考量程度」呈現顯著影響。對「隱性失效考量程度」呈現顯著影響的有參與調查件數、參與調查年資、參與根因調查訓練總時數、人因掌握信心程度等。有接受根因訓練及自覺訓練非常足夠的調查員，人因整體、顯性及隱性失效考量程度皆顯著優於無接受訓練者，接受訓練時數16小時以上的調查員對人因整體、顯性及隱性失效考量成績皆顯著較高，接受調查訓練的時數越多，整體考量成績越高。調查件數達30件以上的調查員，人因整體、顯性及隱性考量程度皆顯著較高。在「醫院病安通報與調查分析制度」方面，調查員之「人為因素掌握程度50%」者，對人為因素考量程度顯著低於「人為因素掌握程度75~100%」者，「能杜絕事件再發生25%及50%」者，對人因考量程度顯著低於「能杜絕事件再發生75%」者。在調查人員之「基本資料」方面，呈現女性調查員對於人因整體、顯性及隱性失效考量掌握度顯著較男性調查員高(p<0.05)，而「服務醫院權屬別方面，顯示(準)醫學中心進行不良事件調查時之人為因素整體考量程度及顯性失效考量程度顯著高於區域及地區醫院。本研究完成醫療版HFACS工具中文化，包含HFACS架構四個層次26個次項目、根本原因分析快速參考指引、訪談指引及199項微分類碼，經過翻譯及回覆翻譯，確定與原意相符，並於第二層次不安全行為之前置條件之團隊因素補充微分類碼項目，新增「病人因素」及「照顧者因素」二個項目。信效度檢測結果專家效度CVI為1.0，分析員信度分析Cronbach’s α結果介於0.78~0.99，Cohen’s Kappa Coefficient介於0.5~1.0。以HFACS回溯123件調查分析案件的結果，案件中人為因素佔了100%，可被預防的比例97.6%。人為因素編碼的結果，共2081個微分類貢獻因子，HFACS四個層面人為因素比例分佈，與本研究問卷調查結果，調查員於進行不良事件人為因素根因考量時的趨勢一致，皆呈現顯性失效(第一層不安全的行為及第二層不安全的前置條件)高於隱性失效(第三層監督及第四層組織影響)。此外，分析實際案例中有40例與病家因素相關，證實補充二項微分類碼的必要性。本研究應用HFACS工具重新分析編碼結果多挖掘了61%的微分類貢獻因子，且以Wilcoxon Sign Rank檢定結果，有19個項目呈現前後二組編碼結果有顯著差異，證實使用HFACS可更完整挖掘人為因素，有系統的全面識別導致錯誤的顯性及隱性失效，客觀的量化人為因素錯誤的性質與及影響，並據以發展改善措施證實此工具於醫療領域的適用性，能作為醫院進行不良事件根本原因分析調查時之輔助工具，應推廣使用。結論：研究結果顯示過去皆聚焦在顯性失效，忽略隱性失效帶來的影響，調查員的個人經驗為重要的因素，個人的專業背景以及是否接受訓練、訓練時數長短、年資累積與參與調查件數的經驗等皆會影響調查時對人為因素掌握度。此外，機構應積極排除根因調查時遇到的障礙，運用HFACS訪談指引，提升訪談技巧，培養全員認知，有助於調查員調查事件的進行。本研究人員在經過HFACS培訓後，實際應用HFACS工具於重大不良事件分析，能協助調查員系統性、全面性地的檢視事件相關人為因素，找出潛藏的貢獻因子，重視監督與組織層級的影響。病人安全推動十餘年，隨著各項病人安全政策的推動，可見重大異常事件減少的成效，然而隨著科技日新月異，錯誤的型態轉型，呼籲醫療機構即將面臨資訊科技帶來病安風險的挑戰，醫院應積極培養人員對人為因素的認識以及導入人因工程以發展有效之對策，以維護病人安全。

關鍵字

人為因素分析及分類工具；人為因素；根本原因分析；醫療不良事件；病人安全

並列摘要

Background: Many medical adverse events can attribute to human factors and system failures. At present, there is a lack of relevant tools for the identification and classification of human factors in medical adverse event investigations. It is still doubtful whether the implementation status of the root cause analysis (RCA) of various institutions in the past can effectively control the root cause for improvement. Therefore, this study intends to explore the current status of RCA in Taiwan's hospitals for adverse events, as well as to translate and validate the Human Factor Analysis and Classification System (HFACS) tool into Chinese language. We aim to provide further empirical results of this tool applying to retrospective investigation of the RCA reports. Method: This research adopts a cross-sectional survey, recruiting colleagues from institutions responsible for medical adverse event notification and root cause investigation above the teaching hospitals in the country as the research object, and conducts a large-scale survey with a self-developed questionnaire to discuss the current internal the root cause of medical adverse events is investigated and analyzed. We use the validated HFACS tool developed by Drs Wiegmann and Shapell in 2016.Based on the major medical adverse events that occurred in a medical center from 2006 to 2017, the Chinese translation of HFACS was used for retrospective RCA reports. Results: A total of 122 responsible units for hospital patient safety, and 590 RCA investigators were surveyed. The questionnaire response rate reached 93.1% and 90.1%. 97.5% of the hospitals have dedicated units and qualified personnel to handle adverse event report, 89.3% are independent report, multiple report channels, and the distribution of the number of reporting varies. 91% of hospitals have RCA team. It shows that use of RCA in adverse event investigation has become a common practice in Taiwan.18.6% of the investigators in this study have not been trained, and 35.8% of the trainees still feel that they are insufficient, and 49.1% of the investigators have only half or less confident that the current survey tools can effectively uncover human factors and grasp the root cause, reflect the emergence of investigative tools and the lack of training of investigators in domestic hospitals. As many as 94.2% of investigators said that there were obstacles in the event analysis and investigation. The top three obstacles were "difficult retrospective afterwards, with memory deviation", "rejection of personnel related to interviews or investigation incidents" and "experience of investigators"," lack of experience".The "personal experience" of the RCA investigators from hospital, regardless of factors such as the "years of participation" in the survey, the number of surveys participated, whether they have received RCA training, the total hours in RCA training, and the degree of confidence in human factors , all have significant impacts on the "overall consideration of human factors" and "active failure consideration" when investigating adverse events. For investigators with sufficient RCA and conscious training, the "overall , active and latent failure"considerations of human factors are significantly better than those without training. Regarding the "hospital adverse report and RCA policy", the investigator's "human factor mastery degree is 50%", the human factor consideration level is significantly lower than the "human factor mastery degree 75~100%", "can prevent the incident recurrence of 25% and 50%" is significantly lower than that of "can prevent 75% of recurrence". In terms of "service hospital ownership", it is shown that the overall consideration of human factors and the active failure of the medical center when conducting adverse event investigations is significantly higher than that of regional and regional hospitals.This study completed translation and validation of the "HFACS in healthcare" Chinese version, including 26 categories at four levels of the HFACS framework, RCA quickly reference, interview guide and 199 nanocodes. Complete dual translation, and add 2 nanocodes (patient factor and caregiver factor) to the unsafe act precondition. The CVI is 1.0, the Cronbach’s α results are between 0.78 and 0.99, and Cohen’s Kappa Coefficient is between 0.5 and 1. Based on the results of HFACS reviewing 123 RCA cases, human factors accounted for 100% of the cases, and the preventable proportion was 97.6%.A total of 2081 contribution factors, and the proportional distribution of human factors at the four levels of HFACS are consistent with the results of the questionnaire survey of this study. The active failure are higher than latent failure. In addition, the analysis of 40 cases are related to the factors of patients and caregiver, confirming the necessity of supplementing the two nanocodes. In this study, the HFACS was used to re-analyze the coding results, and 61% of the contribution factors were mined. Based on the results, 19 categories showed significant differences between the two sets of coding results before and after, confirming that HFACS can be more complete excavate human factors, systematically and comprehensively identify the explicit and implicit failures of errors. Conclusion: The results of the study show the personal experience of the investigator is an important factor, the individual’s professional background and whether he has received training, the experience of the number of surveys will affect the mastery of human factors during the survey.In addition, the organization should actively eliminate the obstacles encountered during the RCA investigation. After being trained by HFACS, and applied HFACS tools in the analysis of RCA, which can assist the investigator to systematically and comprehensively examine the human factors related to the event, and pay attention to the impact of supervision and organization level. The hospital should actively cultivate personnel's understanding of human factors and introduce human factors engineering to develop effective countermeasures to maintain patient safety. Key Word: HFACS, human factor, root cause analysis, adverse event, patient safety.