Transcript 請點選此處進行下載
由研究設計到文章發表 國立成功大學醫學院公衛所 陳國東 教授 引言 科學的特性 •公開的(Public) •客觀的(Objective) •可重複產生的(Reproducible) •預測的(Predictive) •累積的(Cumulative) •系統性的(Systematic) 這些都得依靠文字記載來呈現 科學性文章的型態 1. 觀察性文章 病例報導 病例對照研究 世代研究 2. 理論性文章 探討物理、生物或社會 法則的文章 3. 實驗性文章 臨床試驗研究 4. 方法學文章 5. 回顧性文章 陳述性回顧文章 系統性回顧文章 6. 短文 7. 編者註 8. 與主編通訊 如何產生研究計畫？ • 一、研究計畫才有取研究經費 二、有計畫、經費才有好的研究 團隊 三、有計畫、經費、好的研究團 隊才有好成果 研究問題 •它是研究者想解決的疑問， 也就是你的研究的目的 •那麼，你的研究想陳述什麼 問題？ 靈感來至何處？ •文獻 •對於新的觀念及技術保持警覺 參加會議 (i.e. SARS in Vietnam 啟示) •抱好奇，懷疑的態度 BCG真的有效嗎 •保持想像力，觀察力，意志力 人為何吃三餐 •選擇導師 提出的問題最好縮小到 可作為研究的主題 問題: 人是否應該吃魚? vs 台灣人多久吃ㄧ次魚? 吃魚有好處嗎? vs 吃魚會降低罹患心血管疾病的風險嗎? 好的研究問題應該有其作用， 不應讓讀者感覺so what? 好的研究問題應具備以下條件 • Feasible Adequate number of subjects Affordable in time and money Manageable in scope • Interesting Getting the answer intrigues the investigator and her friends • Novel Confirms, refutes or extends previous findings Provides new findings • Ethical Institutional review board will • Relevant To scientific knowledge To clinical and health policy To future research approve 不同角度有不同的問題 比方說對於腸病毒71型感染 臨床醫師可能的疑問是，受到腸病毒 71型感染後，什麼轉機造成感染者死 亡? 流行病學家可能要問，1998年EV 71 為什麼造成大流行? 研究法 • 觀察研究法 1.世代研究 2.橫斷研究 3.病例對照研究 • 臨床實驗研究法 1.隨機 2.非隨機 描述性研究 ↓ 分析性研究 ↓ 臨床實驗研究 問題：你會由何研究法開始？ 1999年6月14日宜蘭靖廬通報，有12 位大陸客因下肢水腫、手足無力、食 慾不振、呼吸困難而住院。其中有2 位雖經加強治療，仍在住院24小時內 不治死亡。 (Chen KT et al. Public Health Reports 2003;118:59-64) 首先進行描述性研究 人 時 地 選擇研究對象 問題：你的病例定義？ 1996年8月間，有6位在台灣某電子工 廠工作的菲勞，因皮疹、發燒、肝功 能異常而住院。其中有一人死亡，另 外5人經住院治療後痊癒。 (Chen KT et al. Am J Public Health 2003;93:489-492) 如果要探討菲勞罹患SJS/TEN的危 險因子 病例定義:1996年2月至1997年元月 間，因SJS/TEN而住院的菲勞 對照定義:同期間來台，與病例同 一國籍、性別、年齡差距在3歲 以內，同一工廠工作但不同寢室 的工作人員 如果要探討口服避孕藥造成靜脈栓 塞的危險性 case: between March 1999 and September 2004，consecutive patients aged 18-50 years with first episode of deep venous thrombosis or pulmonary embolism in Netherlands (BMJ: 2009;339:b2921) Control: Partners of patient aged 18-50 years were invited as controls Exclusion: postmenopausal, pregnant, or within 4 weeks postpartum at the time of the thrombotic event and women using hormonal contraception other that oral contraceptives Clinical versus community population 1.選擇醫院病人為研究對象要考 慮選樣偏差的問題，比方醫院 知名度高，較多嚴重病人 2.以郵寄問卷訪問受訪者也同樣 要考慮選樣偏差的問題，比如 願意受訪的是較健康的人 測量 Measurement describe phenomena in terms that can be analyzes statistically 測量尺度 (scale) • 連續變項 • 類別變項 Common Types of Measurements Types of Measurement Examples Medical history Diagnoses, medications, operations, symptoms, physical findings Psychosocial factors Depression, family history Anthropometric Height, weight, body composition Biochemical measures Serum cholesterol, plasma fibrinogen Genetic/molecular tests Single neucleotide polymorphisms, human leukocyte antigen type Imaging Bone density, coronary calcium Electromechanical Arrhythmia, congenital heart disease 可以用類別變項 Table. Risk of SJS/TEN occurrence by anthelmintic drug use among foreign laborers Drug use Cases (n=46) Controls (n=92) Odds (95%CI) Pyrantel pamoate only Yes No 0 46 44 48 No estimable Mebendazole only Yes No 6 40 19 73 0.5 (0.1-1.6) Metronidazole only Yes No 5 41 6 86 1.8 (0.3-8.6) 35 11 23 69 9.5 (3.9-23.9)** Mebendazole + metronidazole Yes No Chi-square test, * p <0.05, **p< 0.001 (Chen KT, et al. Am J Public Health 2003;93:489-492) 可以是連續變項 Table. Risk of occurrence by the exposure dosage No. Case Dosage of metronidazole† 500mg-1000 mg 4 No. Control Odds ratioa (95% CI) 14 1.0 (referent) 1500 mg 4 4 3.5 (0.4-30.9) >= 2000 mg 27 5 18.9 (3.6-112.8)** 28 3 4 11 1 11 Dosage of mebendazole‡ 500-1000 mg 1500 mg >= 2000 mg a Chi-square test, * p <0.05, **p< 0.001. (Chen KT, et al. Am J Public Health 2003;93:489-492) 1.0 (referent) 1.2 (0.1-32.9) 0.1 (0.03-0.6)* 以多個item測量 Scales and Scores to measure abstract variable: e.g. quality Using multiple items to assess a concept may have other advantages over single question or several questions asked in different ways that cannot combined Internal consistent of a scale can be tested statistically using measure such as Cronbach’s alpha Likert scales: to quantify attitude, behaviors, and domains of health-related quality of life. Table. Risk factors for drug abuse by stepwise logistic regression analysis Odds ratio 95% CI Sex (male/female) 0.29 0.10-0.79* Alcohol drinking (yes/no) 3.00 1.26-7.20* Single/married 12.18 2.04-72.67* Depression score (increased by 1) 1.27 1.18-1.36** Variables Depression score was measured using Self-rating Depression Scale (SDS). *p<0.05, **p<0.001 (YU et al. J Addiction Nursing 2005;16:195-198) 如何撰寫研究計畫 步驟 一、 先決定向誰申請 二、 組成研究小組 三、 根據提供經費單位之規定撰寫 四、 建立時間表及定期開會 五、 找到前輩的研究計畫書 六、 根據要點撰寫 七、 回顧、前測、修正 研究計畫書的主要成分 writing and Funding a Research Proposal 一、 開頭 1.題目：descriptive and concise 2. 摘要：concise summary of protocol 二、 行政部分 1.預算 2. 研究人員資歷 3. 現有資源、設備 三、 研究目的 The specific aims are statements of the research question and plan using concise format that specific in concrete terms the desired outcome Purpose (目的) Goal (目標) Objective (預期成果) Purpose The purpose of this public health initiative is to make the vaccine against human papillomavirus available to the Anytown community, especially to girls and women between the ages of 11 and 26. Goals To accomplish this purpose, we have set 2 goals: 1) To prepare an educational brochure targeted to teenage girls, their parents, and young women that describes cervical cancer and its risk factors and the reasons for getting vaccinated. 2) To distribute these brochures to schools, physicians, pharmacists, and libraries in the community. Objectives for Goal #1 We want the brochure to educate readers about cervical cancer and the vaccine and to leave a lasting impression of the reasons for getting vaccinated. Therefore, this goal has 2 objectives. Object 1: Immediately after reading the brochure, readers will be able to answer correctly 80% of the questions on cervical cancer and the vaccine. Object 2: Two weeks after reading the brochure, readers will be able to remember at least one reason for getting vaccinated. • 研究問題 探討胸腔手術病人，術後使用繃帶對 於傷口復原有多大效果? • 無效假說(null hypothesis) 接受胸腔手術的病人，術後使用繃帶 處理並不影響傷口復原時間。 • 替代假說(alternative hypothesis) 接受胸腔手術的病人，術後使用繃帶 處理會縮短傷口復原的時間。 example 1. To recruit 400 healthy men, 40-59 years old into a randomized blinded trial of the effects of a testosterone patch 2. To test the hypothesis that compared with men assigned to recruit a placebo patch, those assigned to receive the testosterone patch will have a less bone loss 四、背景及重要意義 describing the background in the field under study 五、研究方法 The subjects, measurements, pretest, data management, quality control, statistical methods, timetable, references 好研究計畫的特徵 一、好的研究問題 二、好的研究方法 三、好的研究團隊 四、計畫書敘述清楚完整 計畫書易犯錯誤 • • • • • 計畫目的不符合提供經費機構的宗旨 研究計畫構想或研究目的不夠完善 研究計畫在技術性方面的瑕疵 研究計畫不夠說服力 經費編列不夠完善 尋找研究經費 It takes more than just a good idea to be competitive in the grants arena–especially in today’s tight funding market (by Reif-Lehrer Liane) 計畫成功的要件 • • • • • • 多與提供經費機構負責人聯繫。 計畫之撰寫要符合綱要(guideline)。 計畫撰寫詳實(well written)。 與審查委員說明作者有能力完成計畫。 具體、實際、不落入空想。 經費實在。 如何撰寫論文 • • • • • 使用有效的用詞 論文訴求的對象? 論文探討什麼目的? 不僅提供讀者新的資訊，也協助其作決定 將你的論文主軸濃縮成一句話，作為你的 論文題目 列出你想呈現的表圖 標題種類 • 宣示性標題(declarative title): 內容包括標題關鍵語以及論文主要結論 例如: Lack of fiber cell induction stops normal growth of rat lenses in Organ culture • 詢問式標題(interrogative title): 例如: Measurement of colonic polyps by radiologists and endoscopists: Who is the most accurate? • 含有副標題 例如: Low-Air-Loss Beds vs. Foam Mattress for Treating Pressure Ulcers: A Cost-Benefit Analysis. 問題：你認為哪個題目較吸引讀者 Outcomes of Heart Transplantation: The Tinytown Community Hospital Experience. vs. Outcomes of Heart Transplantation: A Review of 250 Cases with 3-Year Followup. Hepatocyte growth factor prevents renal fibrosis in chronic renal disease vs. Effects of hepatocyte growth factor on DNA synthesis of tubular epithelial cells in a mouse model of chronic renal disease Epidemiology of measles in Taiwan vs. Absence of endemic measles transmission in Taiwan 前言 1. 2. 3. 4. 背景(Background statement) 問題(Problem statement) 應對行動(Activity statement) 研究作法(Forecasting statement) Examples [Part 1: Background statement] In patients with atherosclerotic vascular disease, aspirin is widely recommended to prevent myocardial infarction, graft occlusion after coronary artery bypass surgery, and stroke. [Part 2: Problem statement] However, aspirin is also associated with prolonged bleeding. Patients are often asked to stop taking aspirin for several days before undergoing bronchoscopy, to reduce the presumed risk of bleeding. The effectiveness of this practice has never been tested, but it dose mean that patients must, for a short while, stop taking a medication with proven benefits, and it can also delay the planned bronchoscopy if aspirin use is not stopped soon enough. [Part 3: Activity statement] Thus, we sought to determine whether aspirin really dose increase the risk of bleeding after bronchoscopy. [Part 4: Forecasting statement] In this article, we describe a prospective trial of 138 consecutive patients undergoing bronchoscopy in which we compared the number and severity of bleeding events in those taking aspirin with those who were not. 研究方法 目的: •讓其他學者可重複類似研究 •讓讀者瞭解本研究所使用的方法已很恰當 實驗研究法要提供的資訊 •本研究是否經倫理委員會(IRB)或動 物照顧使用委員會(IACUC)審查通過? •研究樣本如何選取? •有哪些可能錯誤、干擾因子、偏誤?是 否已作處理? •使用哪些材料?誰提供? • 避免使用突兀、不合邏輯的方法。 例如: We used Colin’s method, with modifications. •可使用副標題使讀者更容易瞭解研究 方法。 測量 • 要敘述測量的環境或實驗室條件? • 測量的單位? • 測量的指標,生物標幟,endpoint? 例如:根據Beck Depression Inventory指 數高於17以上定義為憂鬱。 •測量的精密度?信度?效度? 統計 • 統計方法 • 統計的閥值 • 統計的軟體 Figure 1. Case fatality rates of Severe Acute Respiratory by age and sex in Taiwan, February-July, 2003 Case-fatality rate (%) 70 60 Male Female 50 40 30 20 10 0 0-9yrs 10-19yrs 20-29 yrs 30-39 yrs 40-49 yrs 50-59 yrs 60-69 yrs 70-79 yrs >80 yrs Age group Source:www.medscape.com/features/slideshow 結果 • 避免太多結論 • 在每一節不需加註摘要 • 臨床研究首先呈現的常是人口學基本 資料 討論 • • • • • • • 以一、二句將本研究結果作一摘要整理 對於本研究結果作解釋 與已有的文獻作比較 本研究結果作推論的可行性 本研究結果的意涵 本研究的研究限制 結論 撰寫題目練習 • List the most important words associated with your research (e.g., variable names, processes, reagents, outcomes, interventions, experimental conditions) skin testing humoral immunity streptococcal vaccination ankylosing spondylitis HLA B27 cell-mediated immunity • Arrange the terms in a logical order or orders: skin testing humoral immunity cell-mediated immunity ankylosing spondylitis streptococcal vaccination HLA B27 • Connect the terms with phrases, conjunctions, or transitions. Do NOT worry about how long the resulting text becomes: skin testing to determine humoral immunity and cellmediated immunity responses to streptococcal vaccination in patients with HLA B27 and ankylosing spondylitis • Begin to form text into a title by rearranging or removing unnecessary words, Immune responses in HLA B27-positive patients and ankylosing spondylitis: responses to streptococcal vaccination • Continue to refine the words and their order until you arrive at a final title: Immune Responses to Streptococcal Vaccination in HLA B27-Positive Patients with Ankylosing Spondylitis 謝 謝