醫學論文摘要主要為結構式摘要(structured abstract) ，它由加拿大mc master大學臨牀流行病學和生物統計學教授haynes博士[3]於1990年首先提出。同年，被美國《內科學紀事》(ann intern med)率先採用[4]。據統計，目前世界上60%以上的生物醫學期刊採用結構式摘要，我國大多數醫學期刊都採用了4段式結構式摘要，即目的、方法、結果和結論[ 5，6 ]。
1.1 目的(objective)：文獻[7]指出，摘要的“目的”不能簡單地重複文章題名;應該提供研究背景，這也是國際醫學期刊編輯委員會的要求[8]。在摘要的“目的”中應簡要説明提出問題的緣由，表明研究的國內外研究背景，特別是一些國內或世界首次報道的研究，闡明這些內容就更為重要。例如，《豚鼠聽力衰退模型的aflp分子標記篩選》(screening of aflp markers on the hearing-loss model of guinea pig)，其英文摘要的“目的”部分原譯為: to screen amplified fragment length polymorphism (aflp) markers by whole genome scanning on the hearing-loss model of guinea pig and to reveal the molecular mechanism of presbycusis. 除了重複題名外，尚缺乏研究背景的闡明。宜修改為: the long fragment deletion of mitochondria dna has important relation to hearing decline. but the effect of factors including karyogene should not be neglected. this study was designed to screen amplified fragment length polymorphism (aflp) markers by whole genome scanning on the hearing-loss model of guinea pig and to reveal the molecular mechanism of presbycusis. (國外研究顯示了線粒體dna的大片段缺失與聽力衰退具有重要關係，但是，不能排除核基因等因素對聽力衰退的作用。該研究獲取了擴增片段長度多態性(aflp)標記，以揭示老年性耳聾的分子發病機制。)修改後讀者能清晰地瞭解到該項工作的研究背景，其研究目的亦非常明瞭。
1.2 方法(methods)：簡要説明研究課題的基本設計，使用材料和方法，如何分組對照，研究範圍及精確程度，數據是如何取得的，經何種統計學方法處理。例如，原摘要中未進行分組介紹，修改後方法較詳實，methods: the vascular dementia (vd) rat model was established by permanent bilateral occlusion of both common carotid arteries. the rats were divided into 4 groups randomly as control， vd model (0.9% sodium chloride)， edaravone (3 mg/kg per day) and duxil groups (gastric perfusion， 20 mg/kg per day) after place navigation training. the treatment was from the 7th day after operation and lasted for 20 d. at the 1st and 2nd month the heads of rats were cut down for brain to measure the activity of cox and the change of genetic expression.
1.3 結果(results)：簡要列出研究的主要結果和數據及置信值、統計學顯著性檢驗的確切值。只有列出了有意義的數據，讓讀者有個客觀的比較，研究結果才讓人可信，才更具説服力[8]。例如，results 3 and 7 d after mcao， the neurological deficiency score in fns group (2.25±0.46 and 1.0±0.82 respectively) had no significant difference from those (2.33±0.82 and 1.43±0.79) at same time point in mcao group. it was decreased to 0.50±0.55 at 14 d after mcao in fns group ，which was significantly different from 1.25±0.71 in mcao group (p<0.05). the level of of gap-43 mrna expression in fns group at 12 h and 3 d posterior of mcao (0.70±0.23 and 1.44±0.62，respectively) had no significant difference from those (0.57±0.25 and 1.18±0.42) in mcao group(p>0.05). at 7 d and 14 d time point， they reached the height of 3.03±1.45 and 1.82±0.9 respectively in fns group， significantly higher than 1.55±0.72 and 0.92±0.52 in mcao group(p<0.05)。原文未提供具體的有統計學意義的數據，修改後更具有説服力。
1.4 結論(conclusion)：簡要説明經驗證、論證取得的正確觀點其理論價值或應用價值，是否可推薦或推廣等。例如，conclusion eeg doesn’t change necessarily with aging. the elevation of slow wave power may be related to the mmse decline and mci. 修改後宜對其應用價值加以補充為：it would help improve the diagnostic value of recognition function decline and provide a practical and feasible method for its prevention.