臨床試驗同意書可讀性與受試者對於同意書理解程度之探討

Abstract

背景：隨著臨床試驗的環境改變，全世界的臨床試驗有越來越多的趨勢，目前已經有266,161 個臨床試驗案於美國50州和203國家進行中，其中台灣進行的臨床試驗登錄資料高達5241筆(National Institutes of Health, 2018)。各法規與倫理規範均強調了知情同意應有的程序、內容與其重要性，但臨床試驗同意書內容具有改善的空間。本研究目的在於探討臨床試驗受試者在知情同意過程的理解程度與重要影響因素，並且探索癌症臨床試驗受試者、研究主持人和研究護理師的實際經驗，並且測試癌症臨床試驗同意書之可讀性。 方法：本研究採三角交叉法(Triangulation)，併用質性與量性研究方法(between-method triangulation)，透過不同方法與面向來探討知情同意的現況。採用量性方法橫斷面調查方式探討受試者於知情同意過程中的經驗與理解程度，以質性深度訪談方式探討研究主持人、研究護理師、病人於臨床試驗同意書解釋過程中的經驗，以內容分析法從人體試驗委員會審查資料中分析同意書常見問題，再以量性研究方式測試範本同意書之可讀性。 結果：在375位受試者的知情同意過程中，同意書是受試者重要資訊，主要同意書訊息來源為研究護理師(51.7％)、研究助理(17.9％)、醫師(17.9％)；資訊提供者的態度親切友善(84.8％)；且以受試者慣用且熟悉的語言解說(97.3％)。受試者理解程度各項次得分於3.81~4.17分(滿分為5分)，對於｢研究的風險｣(3.81分)與｢研究的好處｣(3.98分)、｢我擁有足夠的時間去閱讀受試者同意書｣(3.99分)自覺分數較低。以探索性因子分析(Exploratory factor analysis, EFA)來評估量表的因素結構分析，進行Kaiser-Meyer-Olkin(KMO)測試和Bartlett球形測試，KMO值為0.937，Bartlett球形測試結果顯著(X2= 4185.925，df = 190，p <0.0001)；並且發現有三個因素提供了最有意義的因素模式，因素一為「對研究案的理解(Understanding of the research)」，因子負荷量從0.581到0.759，變異量為26.763％，Cronbach’s α為0.925。因素二為「信任和信心(Trust and confidence)」，因子負荷量從0.359到0.658，變異量為16.817％，Cronbach’s α為0.865。因素三為「懷疑和不確定性(Doubt and uncertainty)」，因子負荷量從0.531到0.707，變異量為9.74％，Cronbach’s α為0.713。整體的變異量為52.954%，Cronbach’s α為0.917。在9位受訪者接受深度訪談，在進行文本編碼與分析後，發展出四個主題，包含｢準備｣、｢溝通｣、｢猶豫｣、｢決策｣等四個階段，以及十二個次主題。探討2015-2017年間，於中部某醫學中心人體試驗委員會接受審查且資料齊全的29份癌症臨床試驗同意書，發現平均頁數為24.7頁，平均字數為15643.5字，常見問題為｢試驗方法、程序與相關檢驗｣佔42.3%；再從中抽樣進行同意書可讀性分析，得到54筆受測資料，平均答對率為91.9537分(滿分100分)；難字率平均為0.35172分；進行分析比較後，發現第三期藥品查驗登記案同意書，可讀性較低；且對於具備｢生物醫療科技背景｣者，則癌症臨床試驗同意書對其之可讀性較佳。 結論：本研究建立本土性的實證資料，找出影響同意書理解程度以及影響同意書可讀性之重要因素，可供臨床實務者改善與教育訓練之參考。

Background: With the environmental changes in clinical trials, there are more and more clinical trials in the world. There are currently 266,161 clinical trial cases in 50 states in the United States and 203 countries. The clinical trial registration number from Taiwan is up to 5241 (National Institutes of Health, 2018). Various laws, regulations and ethical norms emphasize the informed consent process, consent form content and importance of informed consent, but there is room for improvement. The purpose of this study was to explore the level of understanding and key influencing factors in the informed consent process, to explore the practical experience of cancer clinical trial subjects, principal investigator and research nurses, and to investigate the readability of consent form. Methods: This study uses between-method triangulation with qualitative and quantitative research. The cross-sectional survey was used to explore the subjects' experience and understanding in the process of informed consent. The qualitative interview was used to explore the experience of the clinical trial investigators, research nurses and patients in explaining the clinical trial consent. Content analysis method was to use to analyze the consent formed from the institutional review board, and the quantitative research methods to test the readability of consent forms. Results: We recruited 375 participants who had signed an informed consent of clinical trials to join the cross-sectional survey, the provider of informed consent form information were research nurse (51.7%), Research assistant(17.9%), and Doctor(17.9%). The demeanor of the information provider was Friendly (84.8%), and used the familiar language of the subjects to explain the study (97.3%). The understanding level scores were between 3.81~4.17(out of 5). The item of “I feel I know the risks of the study” (score is 3.81), “I feel I know the benefits of the study” (score is 3.98), “I had enough time to read the consent document” (score is 3.99) were lower than others. In Exploratory factor analysis (EFA), three major themes were found in the scale: Factor 1: “Understanding of the research” , factor loading is from 0.581 to 0.759, variance is 26.763%, Cronbach’s α is 0.925. Factor 2: “Trust and confidence” , factor loading is from 0.359 to 0.658, variance is 16.817%, Cronbach’s α is 0.865. Factor 3: “Doubt and uncertainty”, factor loading is from 0.531 to 0.707, variance is 9.74%, Cronbach’s αis 0.713. The total variance is 52.954%, with Cronbach’s α of 0.917. The Kaiser–Meyer–Olkin value was 0.937, and indicating excellent sampling adequacy and relatively compact patterns of correlation. Bartlett’s test of sphericity was significant( x2= 4185.925, df = 190, p <0.0001). Nine interviewees received in-depth interviews. After text encoding and analysis, we developed four themes, including 'preparation', 'communication', 'hesitation' and 'decision making' and twelve themes. We explored 29 cancer clinical trial consent forms were submitted for review in the institutional review board of a medical center in Central Taiwan during 2015-2017. The average number of consent form pages was 24.7 pages with an average of 15643.5 words. The common questions were 'Investigation Methods, Procedures and Related Testing '(42.3%). In the readability analysis of consent form study, we obtained 54 test data, the average correct rate was 91.9537 points (out of 100 points); difficult words average rate of 0.35172 points. We found that the readability of the consent forms for phase III new drug clinical trial were lower than others and having better readability to those with 'biomedical science and technology background'. Conclusion: This study establishes local empirical data to find out the important factors that affect the understanding of consent and the readability of consent, which can be used as a reference for clinical practitioners to improve and education.