目录 · 跳到任一小节
科研入门 · 中英对照精读 · 04/05

ARRIVE 指南 2.0:用于报告动物研究的更新版指南

Percie du Sert et al. 2020 · PLoS Biology | 英文原题 The ARRIVE guidelines 2.0: Updated guidelines for reporting animal research
下划线 = 中文容易"变味"或丢信息的词句:实线带编号 = 40 条注释卡(点击跳转);虚线 = 另有 0 处同样有损耗、未单列。译文之后附完整英文原文

作者:Nathalie Percie du Sert 等(保留原名,不音译)

出处:PLoS 生物学(PLoS Biology)2020;18(7):e3000410

DOI:10.1371/journal.pbio.3000410

PMCID:PMC7360023

译自英文原题:The ARRIVE guidelines 2.0: Updated guidelines for reporting animal research

摘要

可重复的科学要求透明的报告。ARRIVE 指南(Animal Research: Reporting of In Vivo Experiments,动物研究:体内实验报告规范)最初制定于 2010 年,目的是改进动物研究的报告。该指南由一份核查清单构成,列出的是描述体内实验(in vivo experiments)的论文中应当包含的信息,以便他人能够充分审视这项工作、评估其方法学严谨性,并重现其方法与结果。尽管多年来资助机构和期刊的认可程度相当高,但对指南的遵守情况并不一致,动物研究论文报告质量方面预期的改善也尚未实现。在此,我们介绍 ARRIVE 2.0。我们对指南作了更新,并重新组织了其中的信息,以便于在实践中使用。我们采用德尔菲法对指南条目作优先级排序,并把它们分为两组:构成最低要求的"ARRIVE 十大要点"(ARRIVE Essential 10),以及描述研究背景的"推荐条目集"(Recommended Set)。这一划分支持分步实施,从而促进动物研究报告的改进。这有助于期刊编辑和审稿人核查稿件中最重要的条目是否已得到报告。我们还编写了配套的《说明与阐释》(Explanation and Elaboration,E&E)文件,其作用在于:(1)解释指南中每一条目背后的理由;(2)澄清关键概念;(3)提供示例说明。我们希望通过这些改动,帮助确保研究者、审稿人和期刊编辑更有能力提升科学过程的严谨性与透明度,进而提升可重复性。


各作者在本文投稿时的任职单位,见 S1 标注作者署名(S1 Annotated Byline)。

术语对照表

译者附注:本表不在原文之内,是为便于对照与日后查规范而增设的辅助附录。下表所列中译即为本译文全篇统一采用的定名。

英文原名 中译 一句话说明
ARRIVE (Animal Research: Reporting of In Vivo Experiments) 动物研究:体内实验报告规范 面向活体动物实验论文的国际报告清单,本文即其 2.0 版
ARRIVE Essential 10 ARRIVE 十大要点(基本十条) 论文必须包含的最低限度信息,共 10 条
Recommended Set 推荐条目集 交代研究背景、伦理与数据等信息的 11 个条目
E&E document (Explanation and Elaboration) 《说明与阐释》文件 逐条解释理由、澄清概念、给出示例的配套文件
study design 研究设计 比较了哪些组、实验单位是什么
sample size 样本量 每组实验单位的准确数目及其确定依据
inclusion and exclusion criteria 纳入与排除标准 事前规定哪些动物、哪些数据进入分析
randomisation 随机化 用随机序列把实验单位分配到各组
blinding 盲法 谁不知道分组,须分阶段交代
outcome measures 结局指标 被测量、被分析的指标
primary outcome measure 主要结局指标 用于确定样本量的那个结局指标
statistical methods 统计方法 每项分析所用的方法与软件
experimental animals 实验动物 物种、品系、性别、年龄等特征
experimental procedures 实验操作 做了什么、何时、何地、为何做
results 结果 描述统计、变异度量与效应量
abstract 摘要 目的、物种、方法、主要发现与结论
background 背景 研究的理由与实验路径
objectives 目的 研究问题与待检验的假设
ethical statement 伦理声明 伦理审查机构名称与批准编号
housing and husbandry 饲养与管护 笼具环境与日常饲养管理
animal care and monitoring 动物照护与监测 减痛措施、不良事件与监测安排
humane endpoints 人道终点 预先设定的、提前终止实验的指征
interpretation / scientific implications 解释/科学意义 结合当前理论解释结果
generalisability / translation 可推广性/转化 结果能否外推到其他物种或人类
protocol registration 方案注册 研究前登记研究问题与分析计划
data access 数据获取 数据是否公开、在哪里可得
declaration of interests 利益声明 利益冲突与资助来源
harm–benefit analysis 伤害—受益分析 动物所受伤害与研究社会收益之间的权衡
3Rs 3R 原则 replacement(替代)、reduction(减少)、refinement(优化)
Delphi exercise 德尔菲法 多轮匿名征询专家意见以形成共识
preregistration 预注册 在采集数据前公开研究计划
FAIR Data Principles FAIR 数据原则 可发现、可访问、可互操作、可重用
MDAR 材料、设计、分析与报告最低标准 生命科学领域的最低报告标准框架与清单
experimental unit 实验单位 统计分析中独立的最小单位(一只动物、一窝或一笼)
confounder 混杂因素 与分组和结局均相关、可扭曲比较的因素
effect size 效应量 组间差异大小的标准化度量
NIH (National Institutes of Health) 美国国立卫生研究院 美国主要的生物医学研究资助机构

为什么良好的报告很重要

近年来,科学家、资助机构、研究使用者和政策制定者都提出了对研究结果可重复性的担忧 [1, 2]。导致可重复性不佳的因素包括:研究设计与分析存在缺陷,试剂及其他生物材料的变异性与验证不足,方法与结果报告不充分,以及数据获取存在障碍 [3]。生物科学界已推出一系列举措来应对这一问题——从开放获取与开放实践,使研究的各个方面都能被审视 [4, 5],到研究预注册,把关注点转向方法的稳健而非结果的新颖 [6, 7],以及用于改进实验设计与统计分析的资源 [8–10]。

对研究方法和发现的透明报告是可重复性的必要组成部分。缺少这一点,研究的方法学严谨性就无法得到充分审视,结果的可靠性无法评估,他人也无法重复或在此基础上继续推进。尽管针对临床前研究和临床研究已制定了专门的报告指南,但证据表明,科学论文常常缺少关键信息,改进的空间仍然相当大 [11–18]。动物研究就是一个很好的例证:报告质量差会影响治疗手段的研发,而无法重现的结果可能催生出一整个研究领域,或引发临床研究,使患者接受很可能无效的干预 [2, 19, 20]。

为改进动物研究的报告,动物研究:体内实验报告规范(Animal Research: Reporting of In Vivo Experiments,ARRIVE)指南于 2010 年发表。该指南由一份核查清单构成,列出的是任何报告体内实验的稿件都应当包含的条目,以确保描述全面而透明 [21–30]。它们适用于使用活动物物种的任何研究领域,尤其适用于描述在实验室或其他正式试验环境中所做的比较研究。该指南在更广阔的语境中也同样适用,例如观察性研究、在野外开展的研究,以及使用动物组织的情形。发表后的 10 年间,ARRIVE 指南已获得生命科学领域一千多种期刊的认可。认可通常包括在给作者和审稿人的指引中倡导使用该指南。然而,尽管有如此程度的支持,近期研究显示,在抽样的多数论文中,ARRIVE 指南所规定的关键信息仍然缺失。这些信息包括随机化的细节(仅在 30%–40% 的论文中报告)、盲法的细节(仅在约 20% 的论文中报告)、样本量依据(在不到 10% 的论文中报告),以及动物特征(所有基本特征均报告的论文不到 10%)[11, 31, 32]。

证据表明,有两个主要因素限制了 ARRIVE 指南的影响力。第一个是编辑部和期刊工作人员在多大程度上积极参与执行报告标准。发表于 PLOS ONE 的一项随机对照试验说明了这一点,该试验旨在检验在稿件投稿流程中要求提交一份填写完整的 ARRIVE 清单所产生的作用。这一单一的编辑干预未包含期刊工作人员的进一步核查,未能改善已发表论文中的信息披露 [33]。相比之下,其他使用更简短清单(主要聚焦于实验设计)且编辑跟进更多的研究显示,论文所载信息的性质与详细程度均有显著改善 [34–36]。目前,指南要求期刊和编辑投入的资源水平很可能使全部条目的落实难以实现。

第二个问题是,研究者以及其他对研究过程完整性负有责任的个人和组织,对报告不完整的后果认识不足。有一些证据表明,对 ARRIVE 的了解与采用更严谨的实验设计标准相关 [37];然而,研究者往往并不熟悉研究发表之中、以及某些结果乃至整个领域的可靠性之中存在的范围大得多的系统性偏倚[33, 38–40]。这种认识上的欠缺,会影响实验如何设计、经费申请如何撰写、实验室中动物如何使用与数据如何记录,也影响作者如何撰写稿件、期刊工作人员、编辑和审稿人如何评估稿件。

涉及动物的实验获得批准,通常基于伤害—受益分析(harm–benefit analysis),即权衡对所用动物造成的伤害与该研究对社会的益处。如果研究报告得不够详细,即便研究开展得十分严谨,其益处也可能无法实现,伤害—受益分析以及公众对研究的信任都会受到损害 [41]。作为一个共同体,我们必须做得更好,确保在使用动物之处,研究既设计良好、分析得当,又得到透明的报告。在此,我们介绍修订后的 ARRIVE 指南,即 ARRIVE 2.0。其中的信息已经更新、扩展并重新组织,以便于指南的使用,帮助确保研究者、编辑和审稿人——以及其他相关期刊工作人员——更有能力提升动物研究的严谨性与可重复性。


▲ 本部分翻译难点(对照学习用)

〔注1〕
原文
transparent reporting
直译
透明的报告
失味点
中文"透明"日常多指"公开、不隐瞒",英文 transparent 在此兼指"信息足够完整,使他人能够逐步追溯、核查每一步",重点在"可核查"而非"不遮掩"
学习提示
读到 transparent reporting,要问"哪些信息可被第三方逐步核查",而不是"作者态度是否坦诚"。
〔注2〕
原文
in vivo experiments
直译
在活体内的实验
失味点
in vivo / in vitro / ex vivo / in situ 是拉丁短语,中文常笼统译作"体内实验",但"体内"在中文里也可指"身体内部"这一空间概念,与"活体动物"的方法学含义不同
学习提示
ARRIVE 只覆盖 in vivo(活体动物)研究;细胞实验属 in vitro,不在其适用范围,但指南也提到"使用动物组织"的情形亦相关。
〔注3〕
原文
scrutinise
直译
审视/细看
失味点
中文"审视"偏中性、偏静态;scrutinise 含"挑剔地、逐条地检查"的对抗性意味,是同行评议与批判性阅读的核心动作
学习提示
这个词常与 rigour、bias 同现,出现在"读者应当能……"的句子里,提示你读文献时该带着"找茬"的姿态。
〔注4〕
原文
methodological rigour
直译
方法学严谨性
失味点
rigour 在英文中兼有"严格、苛刻、不留松动"之义,中文"严谨"只保留了"细致"一层,力度被削弱
学习提示
这是方法学审稿的高频词,与 reproducibility 构成因果链:rigour → reliability → reproducibility。
〔注5〕
原文
endorsement
直译
认可/赞同
失味点
endorsement 指机构或期刊"正式、公开地表示支持并纳入自身政策",带制度约束意味;中文"认可"可指私下同意,力度差一档
学习提示
本文的关键矛盾正在这里——"被一千多家期刊 endorsement"与"遵守率极低"并存,说明"表态支持"≠"执行"。
〔注6〕
原文
adherence to the guidelines
直译
对指南的遵守
失味点
adherence 兼有"黏附、坚持"之义,隐含"持续依从";中文"遵守"像一次性动作,且医学语境中"依从性"多指患者遵医嘱,易造成语境串味
学习提示
评价报告质量的研究里 adherence/compliance 常互换出现,都指"实际做到"而非"声称支持"。
〔注7〕
原文
Delphi exercise
直译
德尔菲练习
失味点
exercise 在此不是"练习/运动",而是"一套按既定程序实施的方法学活动";直译会让人误以为是培训或演练
学习提示
Delphi 是共识形成方法(多轮匿名问卷+反馈+排序),本文用它把 21 个条目排成两档,属于"专家共识"类证据,不是实验证据。
〔注8〕
原文
research users
直译
研究使用者
失味点
中文没有自然对应说法;research users 指"使用研究结论的人"——临床医生、政策制定者、其他领域研究者、产业界,而非"被研究的对象"(那是 research participants/subjects)
学习提示
这个群体在转化医学语境中极重要,本文反复强调"研究使用者能否判断该给结果多大信任"。
〔注9〕
原文
should be included
直译
应当包含
失味点
指南类文本中 should 表示"强烈建议、不照做须有理由",must 表示"强制";中文"应当"与"必须"在口语中常被混用,而本文的力度分层恰恰是理解全文的钥匙
学习提示
看到 should 记作"推荐级(有回旋余地)",看到 must 记作"要求级(无回旋余地)"。全文用 must 仅寥寥数处(如"我们必须做得更好"),其余多为 should/can——这个比例本身就是信息。
〔注10〕
原文
randomisation
直译
随机化
失味点
中文"随机"的日常含义接近"随便、碰运气",而 randomisation 是严格的统计学操作——每个实验单位被分配到某组的概率已知,且不受研究者意志影响
学习提示
方法学审稿必查项。读到 randomisation 要追问三件事:"序列怎么生成的?谁生成的?分配时是否被隐藏(allocation concealment)?"
〔注11〕
原文
blinding
直译
盲法
失味点
本文统一用 blinding,未用 masking(二者同义,但不同学科习惯不同,久读文献者会注意到这一用词选择);中文"盲法"又易与"盲审""盲测"混淆
学习提示
盲法必须分阶段说清——分配时、实施时、结局评估时、数据分析时,各是谁"盲"。只说"实施了盲法"在 ARRIVE 下是不合格报告。
〔注12〕
原文
sample size justification
直译
样本量辩护/正当化
失味点
justification 是"给出理由以证明其合理",含论证义务;中文"依据"只表示"有个出处",丢掉了"你必须说服审稿人"这一层
学习提示
无 justification 的样本量是动物实验最常见的伦理与方法学双重缺陷——既浪费动物,也做不出可重复的结论。
〔注13〕
原文
the extent to which editorial and journal staff are actively involved
直译
编辑部和期刊工作人员积极参与的程度
失味点
英文用 the extent to which 把"程度"提为主语,句子重心在"程度高低";中文若照搬显得冗赘,若简化为"是否参与"则把连续的程度问题压成了是非题
学习提示
the extent to which 是英文社科与指南类文本的高频框架,译时须保留"程度"这一连续量。
〔注14〕
原文
enforcing reporting standards
直译
执行报告标准
失味点
enforce 含"强制、以权力迫使服从"之义,中文"执行"可指自己照做,也可能是被动配合,主语的责任被抹平了
学习提示
本文核心论点之一是"reporting 不是作者一个人的事",enforce 的主语是编辑部和期刊,这正是责任分配的落点。
〔注15〕
原文
It is likely that the level of resource required ... prohibits the implementation
直译
很可能,所需的资源水平阻碍了实施
失味点
这是作者对原因的推测(hedged claim),不是实测结论;中文若译成"是因为资源不足",就把推测说成了定论
学习提示
学术写作中 likely / may / suggest 是"证据强度的刻度",读文献时要分清"作者证明了"与"作者认为很可能"。
〔注16〕
原文
systemic bias
直译
系统性偏倚
失味点
systemic(系统层面的)与 systematic(有系统的、按步骤的)只差两个字母,含义不同——前者指"整个科研生态系统结构性的问题",后者指"方法上按部就班"
学习提示
这是高频误读点。本文语境是"发表偏倚、领域性不可靠",属 systemic;写成"系统偏倚"尚可,写成"系统评价式的偏倚"就跑偏了。
〔注17〕
原文
harm–benefit analysis
直译
伤害—受益分析
失味点
中文文献更常见"风险—受益分析"(risk–benefit),但 harm 指"已实际造成或必然造成的伤害",比 risk(可能发生的风险)更重;换成"风险"会淡化动物伦理审查的分量
学习提示
动物伦理审查的核心文书就是 harm–benefit analysis,这是你进课题组后写伦理申请要用的第一个概念。
〔注18〕
原文
we must do better
直译
我们必须做得更好
失味点
全文少见的 must,语气是"共同体层面的义务宣告",兼有呼吁与自责;中文"必须"能传达强制,但容易丢掉"我们(包括作者自己)也在其中"的自指意味
学习提示
注意主语是 we 而不是 researchers——作者把自己也算进被批评的对象,这是学术共同体写作的典型姿态。

介绍 ARRIVE 2.0

在 ARRIVE 2.0 中,我们提升了指南的清晰度,对条目作了优先级排序,增加了新的信息,并编写了配套的《说明与阐释》(E&E)文件,为每一条目提供背景与理由 [42](亦可见 https://www.arriveguidelines.org)。新增的内容包括:纳入与排除标准,这是数据处理的一个关键方面,可防止对数据进行临时性(ad hoc)排除 [43];方案注册,这是一种新近出现的方法,可促进科学严谨性,鼓励研究者在采集任何数据之前仔细考虑实验设计与分析方案 [44];以及数据获取,与 FAIR 数据原则 [45](Findable, Accessible, Interoperable, and Reusable,即可发现、可访问、可互操作、可重用)相一致。S1 表概述了这些变化。

与原指南最显著的差异在于,条目被划分为两个有优先次序的组,如表 1 和表 2 所示。各组内部条目之间不分先后。第一组是"ARRIVE 十大要点"(ARRIVE Essential 10),它描述的是稿件中应当包含的基本最低限度信息,因为缺少这些信息,审稿人和读者就无法有信心地评估所呈现结果的可靠性。它包括研究设计、样本量、减少主观偏倚的措施、结局指标、统计方法、动物、实验操作以及结果等方面的细节。第二组称为"推荐条目集"(Recommended Set),为所描述的研究补充背景信息。它包括伦理声明、利益声明、方案注册和数据获取,以及关于方法学的更详细信息,如动物的饲养、管护、照护与监测。关于摘要、背景、目的、解释和可推广性的条目,也说明了稿件中更偏叙述性的部分应当包含什么内容。

表 1. ARRIVE 十大要点(Essential 10)

ARRIVE 十大要点
研究设计 1 对每一项实验,简要说明研究设计的细节,包括:a. 所比较的各组,包括对照组。若未设对照组,应说明理由。b. 实验单位(例如单只动物、一窝动物或一笼动物)。
样本量 2 a. 指明分配到各组的实验单位的准确数目,以及每项实验中的总数。同时说明所用动物的总数。b. 解释样本量是如何确定的。如事先做过样本量计算,请提供细节。
纳入与排除标准 3 a. 描述实验过程中用于纳入和排除动物(或实验单位)的任何标准,以及分析过程中用于纳入和排除数据点的任何标准。说明这些标准是否事先(a priori)确立。若未设定标准,请明确说明。b. 对每个实验组,报告任何未纳入分析的动物、实验单位或数据点,并解释原因。若无排除,请说明。c. 对每项分析,报告每个实验组中 n 的准确数值。
随机化 4 a. 说明是否使用随机化将实验单位分配到对照组和处理组。若使用了随机化,请提供生成随机化序列所用的方法。b. 描述用于尽量减少潜在混杂因素(例如处理与测量的顺序,或动物/笼具的位置)的策略。若未控制混杂因素,请明确说明。
盲法 5 描述在实验的不同阶段(分配阶段、实验实施阶段、结局评估阶段和数据分析阶段),哪些人知道分组情况。
结局指标 6 a. 清晰界定所有评估的结局指标(例如细胞死亡、分子标志物或行为学改变)。b. 对于检验假设的研究,请指明主要结局指标,即用于确定样本量的那个结局指标。
统计方法 7 a. 提供每项分析所用统计方法的细节,包括所使用的软件。b. 描述用于评估数据是否满足该统计方法前提假设的任何方法,以及在不满足假设时做了何种处理。
实验动物 8 a. 提供与物种相适应的所用动物细节,包括物种、品系与亚系、性别、年龄或发育阶段,以及(如相关)体重。b. 提供关于动物来源、健康/免疫状态、遗传修饰状态、基因型以及既往任何操作的进一步相关信息。
实验操作 9 对每个实验组(包括对照组),以足以让他人重现的详细程度描述操作流程,包括:a. 做了什么、如何做的、使用了什么。b. 何时做、多久做一次。c. 在哪里做(包括任何适应期的细节)。d. 为什么做(提供操作的理由)。
结果 10 对所做的每项实验(包括独立重复实验),报告:a. 每个实验组的汇总/描述性统计量,并在适用时给出变异程度的度量(例如均数与 SD,或中位数与范围)。b. 如适用,给出效应量及其置信区间。

表 2. ARRIVE 推荐条目集(Recommended Set)

推荐条目集
摘要 11 对研究目的、动物物种、品系与性别、关键方法、主要发现和研究结论给出准确的概述。
背景 12 a. 纳入足以理解该研究理由与背景的科学背景资料,并解释实验路径。b. 解释所用的动物物种与模型如何回应上述科学目的,并在适当之处说明其与人类生物学的相关性。
目的 13 清晰描述研究问题、研究目的,并在适当之处描述所检验的具体假设。
伦理声明 14 提供批准本研究中动物使用的伦理审查委员会或同等机构的名称,以及任何相关的许可证号或方案编号(如适用)。若未申请或未获得伦理批准,请提供理由。
饲养与管护 15 提供饲养与管护条件的细节,包括任何环境丰富化措施。
动物照护与监测 16 a. 描述实验方案中为减轻疼痛、痛苦和应激所采取的任何干预或措施。b. 报告任何预期或非预期的不良事件。c. 描述为本研究设定的人道终点、所监测的体征以及监测频率。若研究未设人道终点,请说明。
解释/科学意义 17 a. 解释结果,并考虑研究目的与假设、当前理论以及文献中其他相关研究。b. 评述研究的局限性,包括潜在的偏倚来源、动物模型的局限,以及与结果相关的不精确性。
可推广性/转化 18 评述本研究的结果是否、以及如何可能推广到其他物种或实验条件,包括(在适当之处)其与人类生物学的相关性。
方案注册 19 提供说明,指出是否在研究开始之前制定了方案(包括研究问题、关键设计特征和分析计划),以及该方案是否注册、注册于何处。
数据获取 20 提供说明,描述研究数据是否可得以及可在何处获得。
利益声明 21 a. 声明任何潜在的利益冲突,包括财务性和非财务性的。若不存在,应予说明。b. 列出所有资助来源(包括资助编号)以及资助方在研究设计、分析和报告中所起的作用。

修订该指南是一项广泛而协作性的工作,科学界的意见被审慎地纳入整个过程。ARRIVE 指南的修订由一个国际工作组——即本文作者——承担,其专长涵盖整个生命科学界,包括资助机构、期刊编辑、统计学家、方法学家以及来自学术界和产业界的研究者。我们与外部利益相关方开展了一项德尔菲法 [46],以最大化专业领域和地理位置的多样性;来自 19 个国家的专家对每一条目提供反馈、提出新条目,并按各条目对评估研究结果可靠性的相对重要性进行排序。这一排序使指南条目被划分为两个优先级组。德尔菲专家组的构成特征以及完整的方法与结果,见支撑信息 S1 Delphi 和 S1 Data。修订后的指南与 E&E 在 BioRxiv 上发表之后,还与正在撰写体内研究稿件的作者一起进行了实地试用,以确保这些文件能被目标使用者充分理解并确实有用。该研究见支撑信息 S1 Road Testing 和 S2 Data。

虽然遵循 ARRIVE 2.0 全部 21 个条目来报告动物研究代表最佳实践,但将条目划分为两组的目的在于,通过允许先聚焦于最关键的问题,促进动物研究报告的改进。这更有助于期刊工作人员、编辑和审稿人核查稿件中各条目是否得到了充分报告。第一步应当是确保依从ARRIVE 十大要点这一最低要求。随后可以逐步、并结合具体编辑政策加入推荐条目集中的条目,直至所有条目在所有稿件中都得到常规报告。ARRIVE 2.0 与近年来发表的其他指南完全兼容且互为补充。通过提供一套专门针对体内研究描述而定制的完整建议,它帮助报告动物实验的作者遵循美国国立卫生研究院(National Institutes of Health,NIH)标准 [43] 以及最低标准框架与清单(材料、设计、分析与报告,Materials, Design, Analysis and Reporting,MDAR [47])。修订后的指南也与许多期刊的政策相一致,并将帮助作者满足以下方面的信息要求:研究的伦理审查 [48, 49]、数据呈现与获取 [50–52]、统计方法 [51, 52] 以及利益冲突 [53, 54]。

尽管指南的撰写考虑的是研究者和期刊编辑政策,但需要强调:透明报告的责任不应当仅由研究者独自承担。迄今为止,资助机构、学术机构和出版方对 ARRIVE 的认可,对提高认识起到了关键作用;它们如今在建设能力、以及倡导改进报告实践所需的行为改变方面,可发挥关键作用。这包括将 ARRIVE 2.0 嵌入适当的培训、工作流程和流程之中,以支持承担不同角色的研究者。虽然该指南的首要关注点是动物研究的报告,但 ARRIVE 在研究过程的更早阶段还有其他用途,包括在体内实验的规划与设计阶段。例如,在经费申请或伦理审查申请中要求按指南描述研究设计,可以确保在研究周期之初就考虑到尽量减少实验偏倚的措施 [55]。

结论

如果动物研究要有助于扩充知识库,并为未来的研究、政策和临床实践提供依据,透明报告显然是必不可少的。ARRIVE 2.0 优先强调与研究可靠性有关信息的报告。这使研究使用者能够评估应当给予这些结果多大的权重,并与此同时促进在体内实验的规划与实施中采用严谨的方法学 [37],从而提高结果可靠、并最终可重复的可能性。

ARRIVE 2.0 的意图不是取代各期刊各自的要求,而是促进各期刊采取统一的做法,以确保所有稿件都包含评价该研究所必需的关键信息。期刊通常共享一个共同目标,即提升其所发表研究的方法学严谨性与可重复性,但不同期刊强调的信息有所不同 [56–58]。在此,我们提出一份关于信息优先次序的专家共识。这将为作者提供明确指引,便于稿件在期刊之间转投,并加快报告标准的改进。

将研究和出版界的努力集中于 ARRIVE 十大要点,为高效评估报告质量、以及评估旨在改进动物实验报告的干预措施与政策的效果,提供了一种可掌控的途径。它为开发用于评估报告情况的操作化清单提供了一个起点,并最终导向构建能够快速检测缺失信息的自动化或半自动化人工智能工具 [59]。

改进报告是一项协作性的事业,需要生物医学研究界的一致努力才能确保最大影响。我们欢迎与该领域的其他团队合作,也欢迎对 ARRIVE 2.0 及其实施策略提出反馈。

支撑信息

译者注:原文该节为支撑信息(Supporting information)文件目录。正文中按编号引用的支撑文件包括 S1 标注作者署名(S1 Annotated Byline)、S1 表(S1 Table)、S1 Delphi、S1 Data、S1 Road Testing 与 S2 Data。清洗后的英文文本未保留该目录的条目著录,故此处仅保留节标题,不另作补写。

参考文献

注:参考文献为著录信息,按学术惯例保留英文原文,见文末附录。


▲ 本部分翻译难点(对照学习用)

〔注19〕
原文
ad hoc exclusion
直译
为此目的而设的排除
失味点
ad hoc 是拉丁语,意为"为眼前这一具体情形临时凑出来的、无事先规则的",带明确的贬义;中文若译"临时"尚可,译"特殊"或"个别"就把作者对数据挑选的批评语气洗掉了
学习提示
动物实验中最常见的作弊式操作就是"看着数据不顺眼就剔掉几只",纳入/排除标准必须是 a priori 的,这正是本条存在的原因。
〔注20〕
原文
encourages researchers to carefully consider
直译
鼓励研究者仔细考虑
失味点
指南文字里 encourage 属于"软性引导",没有强制力;中文"鼓励"虽可对应,但读者容易当成客套话而略过
学习提示
在本族指南中,must > should > encourage/can 构成三级强度阶梯,读条目时先给它定级,再判断"不这么做会不会被拒稿"。
〔注21〕
原文
the most significant departure from the original guidelines
直译
与原指南最显著的偏离
失味点
departure 含"离开原处、另走一路"的意象,暗含"这是有意的改弦更张";译作"差异"太平,"背离"又偏贬义
学习提示
departure from 在文献中常用于描述"与既有范式的分野",读到时要注意作者是在标榜创新还是在承认偏离。
〔注22〕
原文
control groups
直译
控制组
失味点
control 在英文中兼有"对照"(实验设计)与"控制"(操纵、抑制)两义;中文若一律译"控制组",会把"作为比较基准的组"误解为"被控制的组"
学习提示
实验设计语境下一律译"对照":control group 对照组、vehicle control 溶剂对照、sham control 假手术对照。看到 control 先判断它在句中是名词(组)还是动词(控制)。
〔注23〕
原文
experimental unit
直译
实验单位
失味点
中文"实验单位"容易被误读为"一个实验室"或"一次实验";其统计学含义是"独立接受处理、并提供独立数据点的最小实体"——可能是一只动物、一窝、一笼,甚至一块脑片
学习提示
这是初学者最常踩的坑:若一窝 10 只小鼠同笼、只按"窝"给药,则 n=1 窝而不是 n=10 只,样本量会被虚假放大(伪重复,pseudoreplication)。
〔注24〕
原文
established a priori
直译
事先确立的
失味点
a priori 是拉丁语,哲学语境中意为"先于经验的",此处指"在看到数据之前就定好",与统计学中的"事前"(pre-specified)同义;中文若只写"事先",读者不易意识到这是防止事后挑数据的关键
学习提示
与 a priori 相对的是 post hoc(事后)。凡是"事后才决定"的分析,都必须在论文里标注为探索性分析。
〔注25〕
原文
outcome measures
直译
结果测量
失味点
outcome 在临床与实验研究中是术语,指"被用来判断干预效果的指标",不指"实验做完得到的那些结果"(那是 results);中文"结局指标"较准,但"结局"一词在日常中文里带有"最终下场"的色彩,容易让初学者觉得别扭
学习提示
区分四件套——outcome measure(指标)/result(结果数值)/endpoint(终点,指判定的时间点或阈值)/finding(发现)。本文第 6 条要求"清晰界定所有 outcome measures",即先定义尺子,再谈读数。
〔注26〕
原文
assumptions of the statistical approach
直译
统计方法的各种假定
失味点
assumption 在统计中译"假设/假定",但中文"假设"常与 hypothesis(待检验的假设)混淆,二者是两回事
学习提示
hypothesis 是被检验的对象(如"药物降低脑水肿");assumption 是方法成立的前提(如"数据服从正态分布、方差齐")。本文第 7 条 b 要求交代"前提是否满足、不满足时怎么办"(如改用非参数检验)。
〔注27〕
原文
acclimatisation periods
直译
适应期(acclimatisation)
失味点
该词指动物运抵实验室后、正式实验前"适应新环境、稳定生理状态"的那段时间;中文常译"适应期",但"适应"也可指动物对药物产生耐受,语义会打架
学习提示
这是动物实验里极容易被省略、又极影响结果的细节——应激激素水平未回落就开刀,组间差异可能全来自运输应激。
〔注28〕
原文
effect size with a confidence interval
直译
带置信区间的效应大小
失味点
effect size 是标准化后的"差异有多大",与 p 值回答的问题完全不同;中文"效应量"是行内定译,但"量"字易让人以为只是个绝对数值
学习提示
p 值只说"有没有",效应量说"有多大",置信区间说"估得准不准"。只报 p 值不报效应量,是本文批评的报告缺陷之一。
〔注29〕
原文
housing and husbandry
直译
居住与畜牧
失味点
husbandry 在实验动物科学中指"日常饲养管理"(喂食、换笼、清洁、群养方式),与农业"畜牧业"无关;中文若译"饲养"会与 feeding 混同
学习提示
housing(住在哪、几只一笼)与 husbandry(怎么养)共同决定动物的基线状态,是 3R 中 refinement 的主要着力点。
〔注30〕
原文
environmental enrichment
直译
环境丰富
失味点
这是实验动物福利的专门术语,指在笼内添加筑巢材料、玩具、跑轮等,以提供行为表达机会;中文"环境丰富化"是定译,但"丰富"听上去像"装饰",容易低估其对脑结构与行为学结果的实质影响
学习提示
你做脑出血/神经行为学实验时,enrichment 会显著改变基线行为学表现,是组间可比性的隐藏变量。
〔注31〕
原文
pain, suffering, and distress
直译
疼痛、受苦与痛苦
失味点
这是三个并列的法定概念,不是同义反复:pain 是生理性疼痛,suffering 是长期的身心煎熬,distress 是应激性痛苦(恐惧、焦虑、不适);中文"痛苦"一个词会把三者压平
学习提示
伦理审查表里必须分别对应处置措施,把 distress 译成"痛苦"会让你在写申请时漏掉"应激"这一类(如孤养、强光、束缚)。
〔注32〕
原文
adverse events
直译
不利事件
失味点
adverse event 是"实验中发生的、对动物不利的任何意外情况"(死亡、感染、体重骤降、癫痫发作等),不等于"不良反应"(那是副作用);且须区分 expected(预期)与 unexpected(非预期)
学习提示
这是临床研究质量管理(GCP)借到动物实验里的词,报告它不是为了追责,而是为了让别人能判断你的数据是否被异常事件污染。
〔注33〕
原文
humane endpoints
直译
人道的终点
失味点
humane endpoint 指"预先设定的、达到即提前终止动物生命的判定标准",目的是不让动物承受超出批准的痛苦;中文"人道终点"字面像"体面地结束",容易读成修辞而非操作标准
学习提示
它不是"实验终点"(endpoint of experiment),而是"提前终止点"。写方案时必须给出可观测指标(如体重下降 >20%、无法自主进食)与监测频率。
〔注34〕
原文
Generalisability/translation
直译
可普遍化/翻译
失味点
translation 在此不是"翻译语言",而是"转化医学"意义上的转化——把动物结果转用到人;中文读者极易在此处读错
学习提示
这是全文最典型的"一词多义只能选一支"。同理,translational research 译"转化研究",绝不译"翻译研究"。
〔注35〕
原文
While reporting animal research in adherence to all 21 items ... represents best practice, the classification ... is intended to facilitate
直译
虽然……代表最佳实践,这一分类意在促进……
失味点
while 在此是让步连词(虽然),不是时间连词(当……时);中文若漏译这一转折,句子的逻辑会变成"遵守 21 条是为了分类",因果彻底颠倒
学习提示
英文长句中 while / whereas / albeit 是逻辑路标,漏掉一个,整句意思可能反向。读指南类文本时,先找这些路标再读内容。
〔注36〕
原文
represents best practice
直译
代表最好的做法
失味点
best practice 是质量治理术语,指"本领域公认应达到的做法水准",带规范性;中文"最佳实践"已通用,但常被误解为"最高标准、只有少数人能达标"
学习提示
本文特意区分了"best practice(全部 21 条)"与"minimum requirement(十大要点)"——前者是理想,后者是及格线。
〔注37〕
原文
ensure compliance with the ARRIVE Essential 10
直译
确保对十大要点的顺从
失味点
compliance 在法规与质量语境中指"符合规定",中文"依从"在医学里多指患者听医嘱,"遵守"又太日常;此处实际含义是"做到、达标"
学习提示
compliance/adherence/conformity 三词在评审意见中高频出现,都指向"你的稿子有没有真的照做"。
〔注38〕
原文
assess how much weight to ascribe to the findings
直译
评估该给这些发现分配多少重量
失味点
ascribe weight to 是"赋予证据分量"的比喻,中文若直译"分配重量"完全不通;译"权重"又会被统计学读者误认为加权系数
学习提示
这是证据分级(GRADE 等)的核心动作——读者要根据报告完整度决定"这条结论能信几分"。报告缺项越多,可赋予的分量越低。
〔注39〕
原文
is not to supersede individual journal requirements
直译
不是要取代各期刊的要求
失味点
supersede 指"以更优者取代旧的",含"新版覆盖旧版"之意;中文"取代"能达意,但丢掉了"层级上更高、因而覆盖"的意味
学习提示
作者在此做政治性表态——ARRIVE 是"共识"而非"上级规定",这解释了为什么它没有强制力,也解释了全文为何反复呼吁而非命令。
〔注40〕
原文
promote a harmonised approach across journals
直译
在各期刊间推广一种被和谐化的做法
失味点
harmonised 指"把彼此不一致的东西调成一致",是标准化的专门动词;中文"统一"可,但"统一的做法"容易被读成"一家说了算"
学习提示
harmonisation(协调一致)是国际监管与报告规范领域的常用词,与 standardisation(标准化)近义但更强调"多主体协商后趋同"。

附录 · 英文原文

以下为未经改动的英文原文(由 Europe PMC 全文 XML 抽取),用于与上方译文对照。 参考文献保留英文著录格式,未作翻译。

The ARRIVE guidelines 2.0: Updated guidelines for reporting animal research

Authors: Percie du Sert Nathalie; Hurst Viki; Ahluwalia Amrita; Alam Sabina; Avey Marc T.; Baker Monya; Browne William J.; Clark Alejandra; Cuthill Innes C.; Dirnagl Ulrich; Emerson Michael; Garner Paul; Holgate Stephen T.; Howells David W.; Karp Natasha A.; Lazic Stanley E.; Lidster Katie; MacCallum Catriona J.; Macleod Malcolm; Pearl Esther J.; Petersen Ole H.; Rawle Frances; Reynolds Penny; Rooney Kieron; Sena Emily S.; Silberberg Shai D.; Steckler Thomas; Würbel Hanno; Boutron Isabelle

Journal: PLoS Biology 2020;18(7):e3000410 DOI: 10.1371/journal.pbio.3000410 PMCID: PMC7360023

Abstract

Reproducible science requires transparent reporting. The ARRIVE guidelines (Animal Research: Reporting of In Vivo Experiments) were originally developed in 2010 to improve the reporting of animal research. They consist of a checklist of information to include in publications describing in vivo experiments to enable others to scrutinise the work adequately, evaluate its methodological rigour, and reproduce the methods and results. Despite considerable levels of endorsement by funders and journals over the years, adherence to the guidelines has been inconsistent, and the anticipated improvements in the quality of reporting in animal research publications have not been achieved. Here, we introduce ARRIVE 2.0. The guidelines have been updated and information reorganised to facilitate their use in practice. We used a Delphi exercise to prioritise and divide the items of the guidelines into 2 sets, the “ARRIVE Essential 10,” which constitutes the minimum requirement, and the “Recommended Set,” which describes the research context. This division facilitates improved reporting of animal research by supporting a stepwise approach to implementation. This helps journal editors and reviewers verify that the most important items are being reported in manuscripts. We have also developed the accompanying Explanation and Elaboration (E&E) document, which serves (1) to explain the rationale behind each item in the guidelines, (2) to clarify key concepts, and (3) to provide illustrative examples. We aim, through these changes, to help ensure that researchers, reviewers, and journal editors are better equipped to improve the rigour and transparency of the scientific process and thus reproducibility.


See S1 Annotated Byline for individual authors’ positions at the time this article was submitted.

Why good reporting is important

In recent years, concerns about the reproducibility of research findings have been raised by scientists, funders, research users, and policy makers [1, 2]. Factors that contribute to poor reproducibility include flawed study design and analysis, variability and inadequate validation of reagents and other biological materials, insufficient reporting of methodology and results, and barriers to accessing data [3]. The bioscience community has introduced a range of initiatives to address the problem, from open access and open practices to enable the scrutiny of all aspects of the research [4, 5] through to study preregistration to shift the focus towards robust methods rather than the novelty of the results [6, 7], as well as resources to improve experimental design and statistical analysis [8–10].

Transparent reporting of research methods and findings is an essential component of reproducibility. Without this, the methodological rigour of the studies cannot be adequately scrutinised, the reliability of the findings cannot be assessed, and the work cannot be repeated or built upon by others. Despite the development of specific reporting guidelines for preclinical and clinical research, evidence suggests that scientific publications often lack key information and that there continues to be considerable scope for improvement [11–18]. Animal research is a good case in point, where poor reporting impacts on the development of therapeutics and irreproducible findings can spawn an entire field of research, or trigger clinical studies, subjecting patients to interventions unlikely to be effective [2, 19, 20].

In an attempt to improve the reporting of animal research, the Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines were published in 2010. The guidelines consist of a checklist of the items that should be included in any manuscript that reports in vivo experiments, to ensure a comprehensive and transparent description [21–30]. They apply to any area of research using live animal species and are especially pertinent to describe comparative research in the laboratory or other formal test setting. The guidelines are also relevant in a wider context, for example, for observational research, studies conducted in the field, and where animal tissues are used. In the 10 years since publication, the ARRIVE guidelines have been endorsed by more than a thousand journals from across the life sciences. Endorsement typically includes advocating their use in guidance to authors and reviewers. However, despite this level of support, recent studies have shown that important information as set out in the ARRIVE guidelines is still missing from most publications sampled. This includes details on randomisation (reported in only 30%–40% of publications), blinding (reported in only approximately 20% of publications), sample size justification (reported in less than 10% of publications), and animal characteristics (all basic characteristics reported in less than 10% of publications) [11, 31, 32].

Evidence suggests that 2 main factors limit the impact of the ARRIVE guidelines. The first is the extent to which editorial and journal staff are actively involved in enforcing reporting standards. This is illustrated by a randomised controlled trial at PLOS ONE, designed to test the effect of requesting a completed ARRIVE checklist in the manuscript submission process. This single editorial intervention, which did not include further verification from journal staff, failed to improve the disclosure of information in published papers [33]. In contrast, other studies using shorter checklists (primarily focused on experimental design) with more editorial follow-up have shown a marked improvement in the nature and detail of the information included in publications [34–36]. It is likely that the level of resource required from journals and editors currently prohibits the implementation of all the items of the ARRIVE guidelines.

The second issue is that researchers and other individuals and organisations responsible for the integrity of the research process are not sufficiently aware of the consequences of incomplete reporting. There is some evidence that awareness of ARRIVE is linked to the use of more rigorous experimental design standards [37]; however, researchers are often unfamiliar with the much larger systemic bias in the publication of research and in the reliability of certain findings and even of entire fields [33, 38–40]. This lack of understanding affects how experiments are designed and grant proposals prepared, how animals are used and data recorded in the laboratory, and how manuscripts are written by authors or assessed by journal staff, editors, and reviewers.

Approval for experiments involving animals is generally based on a harm–benefit analysis, weighing the harms to the animals involved against the benefits of the research to society. If the research is not reported in enough detail, even when conducted rigorously, the benefits may not be realised, and the harm–benefit analysis and public trust in the research are undermined [41]. As a community, we must do better to ensure that, where animals are used, the research is both well designed and analysed as well as transparently reported. Here, we introduce the revised ARRIVE guidelines, referred to as ARRIVE 2.0. The information included has been updated, extended, and reorganised to facilitate the use of the guidelines, helping to ensure that researchers, editors, and reviewers—as well as other relevant journal staff—are better equipped to improve the rigour and reproducibility of animal research.

Introducing ARRIVE 2.0

In ARRIVE 2.0, we have improved the clarity of the guidelines, prioritised the items, added new information, and generated the accompanying Explanation and Elaboration (E&E) document to provide context and rationale for each item [42] (also available at https://www.arriveguidelines.org). New additions comprise inclusion and exclusion criteria, which are a key aspect of data handling and prevent the ad hoc exclusion of data [43]; protocol registration, a recently emerged approach that promotes scientific rigour and encourages researchers to carefully consider the experimental design and analysis plan before any data are collected [44]; and data access, in line with the FAIR Data Principles [45] (Findable, Accessible, Interoperable, and Reusable). S1 Table summarises the changes.

The most significant departure from the original guidelines is the classification of items into 2 prioritised groups, as shown in Tables 1 and 2. There is no ranking of the items within each group. The first group is the “ARRIVE Essential 10,” which describes information that is the basic minimum to include in a manuscript, as without this information, reviewers and readers cannot confidently assess the reliability of the findings presented. It includes details on the study design, the sample size, measures to reduce subjective bias, outcome measures, statistical methods, the animals, experimental procedures, and results. The second group, referred to as the “Recommended Set,” adds context to the study described. This includes the ethical statement, declaration of interest, protocol registration, and data access, as well as more detailed information on the methodology such as animal housing, husbandry, care, and monitoring. Items on the abstract, background, objectives, interpretation, and generalisability also describe what to include in the more narrative parts of a manuscript.

Table 1 ARRIVE Essential 10.

ARRIVE Essential 10
Study design 1 For each experiment, provide brief details of study design including:a. The groups being compared, including control groups. If no control group has been used, the rationale should be stated.b. The experimental unit (e.g., a single animal, litter, or cage of animals).
Sample size 2 a. Specify the exact number of experimental units allocated to each group, and the total number in each experiment. Also indicate the total number of animals used.b. Explain how the sample size was decided. Provide details of any a priori sample size calculation, if done.
Inclusion and exclusion criteria 3 a. Describe any criteria used for including and excluding animals (or experimental units) during the experiment, and data points during the analysis. Specify if these criteria were established a priori. If no criteria were set, state this explicitly.b. For each experimental group, report any animals, experimental units, or data points not included in the analysis and explain why. If there were no exclusions, state so.c. For each analysis, report the exact value of n in each experimental group.
Randomisation 4 a. State whether randomisation was used to allocate experimental units to control and treatment groups. If done, provide the method used to generate the randomisation sequence.b. Describe the strategy used to minimise potential confounders such as the order of treatments and measurements, or animal/cage location. If confounders were not controlled, state this explicitly.
Blinding 5 Describe who was aware of the group allocation at the different stages of the experiment (during the allocation, the conduct of the experiment, the outcome assessment, and the data analysis).
Outcome measures 6 a. Clearly define all outcome measures assessed (e.g., cell death, molecular markers, or behavioural changes).b. For hypothesis-testing studies, specify the primary outcome measure, i.e., the outcome measure that was used to determine the sample size.
Statistical methods 7 a. Provide details of the statistical methods used for each analysis, including software used.b. Describe any methods used to assess whether the data met the assumptions of the statistical approach, and what was done if the assumptions were not met.
Experimental animals 8 a. Provide species-appropriate details of the animals used, including species, strain and substrain, sex, age or developmental stage, and, if relevant, weight.b. Provide further relevant information on the provenance of animals, health/immune status, genetic modification status, genotype, and any previous procedures.
Experimental procedures 9 For each experimental group, including controls, describe the procedures in enough detail to allow others to replicate them, including:a. What was done, how it was done, and what was used.b. When and how often.c. Where (including detail of any acclimatisation periods).d. Why (provide rationale for procedures).
Results 10 For each experiment conducted, including independent replications, report:a. Summary/descriptive statistics for each experimental group, with a measure of variability where applicable (e.g., mean and SD, or median and range).b. If applicable, the effect size with a confidence interval.

Table 2 ARRIVE Recommended Set.

Recommended Set
Abstract 11 Provide an accurate summary of the research objectives, animal species, strain and sex, key methods, principal findings, and study conclusions.
Background 12 a. Include sufficient scientific background to understand the rationale and context for the study, and explain the experimental approach.b. Explain how the animal species and model used address the scientific objectives and, where appropriate, the relevance to human biology.
Objectives 13 Clearly describe the research question, research objectives and, where appropriate, specific hypotheses being tested.
Ethical statement 14 Provide the name of the ethical review committee or equivalent that has approved the use of animals in this study, and any relevant licence or protocol numbers (if applicable). If ethical approval was not sought or granted, provide a justification.
Housing and husbandry 15 Provide details of housing and husbandry conditions, including any environmental enrichment.
Animal care and monitoring 16 a. Describe any interventions or steps taken in the experimental protocols to reduce pain, suffering, and distress.b. Report any expected or unexpected adverse events.c. Describe the humane endpoints established for the study, the signs that were monitored, and the frequency of monitoring. If the study did not have humane endpoints, state this.
Interpretation/scientific implications 17 a. Interpret the results, taking into account the study objectives and hypotheses, current theory, and other relevant studies in the literature.b. Comment on the study limitations, including potential sources of bias, limitations of the animal model, and imprecision associated with the results.
Generalisability/translation 18 Comment on whether, and how, the findings of this study are likely to generalise to other species or experimental conditions, including any relevance to human biology (where appropriate).
Protocol registration 19 Provide a statement indicating whether a protocol (including the research question, key design features, and analysis plan) was prepared before the study, and if and where this protocol was registered.
Data access 20 Provide a statement describing if and where study data are available.
Declaration of interests 21 a. Declare any potential conflicts of interest, including financial and nonfinancial. If none exist, this should be stated.b. List all funding sources (including grant identifier) and the role of the funder(s) in the design, analysis, and reporting of the study.

Revising the guidelines has been an extensive and collaborative effort, with input from the scientific community carefully built into the process. The revision of the ARRIVE guidelines has been undertaken by an international working group—the authors of this publication—with expertise from across the life sciences community, including funders, journal editors, statisticians, methodologists, and researchers from academia and industry. We used a Delphi exercise [46] with external stakeholders to maximise diversity in fields of expertise and geographical location, with experts from 19 countries providing feedback on each item, suggesting new items, and ranking items according to their relative importance for assessing the reliability of research findings. This ranking resulted in the prioritisation of the items of the guidelines into the 2 sets. Demographics of the Delphi panel and full methods and results are presented in Supporting Information S1 Delphi and S1 Data. Following their publication on BioRxiv, the revised guidelines and the E&E were also road tested with researchers preparing manuscripts describing in vivo studies, to ensure that these documents were well understood and useful to the intended users. This study is presented in Supporting Information S1 Road Testing and S2 Data.

While reporting animal research in adherence to all 21 items of ARRIVE 2.0 represents best practice, the classification of the items into 2 groups is intended to facilitate the improved reporting of animal research by allowing an initial focus on the most critical issues. This better allows journal staff, editors, and reviewers to verify that the items have been adequately reported in manuscripts. The first step should be to ensure compliance with the ARRIVE Essential 10 as a minimum requirement. Items from the Recommended Set can then be added over time and in line with specific editorial policies until all the items are routinely reported in all manuscripts. ARRIVE 2.0 are fully compatible with and complementary to other guidelines that have been published in recent years. By providing a comprehensive set of recommendations that are specifically tailored to the description of in vivo research, they help authors reporting animal experiments adhere to the National Institutes of Health (NIH) standards [43] and the minimum standards framework and checklist (Materials, Design, Analysis and Reporting [MDAR] [47]). The revised guidelines are also in line with many journals’ policies and will assist authors in complying with information requirements on the ethical review of the research [48, 49], data presentation and access [50–52], statistical methods [51, 52], and conflicts of interest [53, 54].

Although the guidelines are written with researchers and journal editorial policies in mind, it is important to stress that researchers alone should not have to carry the responsibility for transparent reporting. Funders, institutions, and publishers’ endorsement of ARRIVE has been instrumental in raising awareness to date; they now have a key role to play in building capacity and championing the behavioural changes required to improve reporting practices. This includes embedding ARRIVE 2.0 in appropriate training, workflows, and processes to support researchers in their different roles. While the primary focus of the guidelines has been on the reporting of animal studies, ARRIVE also has other applications earlier in the research process, including in the planning and design of in vivo experiments. For example, requesting a description of the study design in line with the guidelines in funding or ethical review applications ensures that steps to minimise experimental bias are considered at the beginning of the research cycle [55].

Conclusion

Transparent reporting is clearly essential if animal studies are to add to the knowledge base and inform future research, policy, and clinical practice. ARRIVE 2.0 prioritises the reporting of information related to study reliability. This enables research users to assess how much weight to ascribe to the findings and, in parallel, promotes the use of rigorous methodology in the planning and conduct of in vivo experiments [37], thus increasing the likelihood that the findings are reliable and, ultimately, reproducible.

The intention of ARRIVE 2.0 is not to supersede individual journal requirements but to promote a harmonised approach across journals to ensure that all manuscripts contain the essential information needed to appraise the research. Journals usually share a common objective of improving the methodological rigour and reproducibility of the research they publish, but different journals emphasise different pieces of information [56–58]. Here, we propose an expert consensus on information to prioritise. This will provide clarity for authors, facilitate transfer of manuscripts between journals, and accelerate an improvement of reporting standards.

Concentrating the efforts of the research and publishing communities on the ARRIVE Essential 10 items provides a manageable approach to evaluate reporting quality efficiently and assess the effect of interventions and policies designed to improve the reporting of animal experiments. It provides a starting point for the development of operationalised checklists to assess reporting, ultimately leading to the build of automated or semi-automated artificial intelligence tools that can detect missing information rapidly [59].

Improving reporting is a collaborative endeavour, and concerted effort from the biomedical research community is required to ensure maximum impact. We welcome collaboration with other groups operating in this area, as well as feedback on ARRIVE 2.0 and our implementation strategy.

Supporting information


References

  1. 1GoodmanSN, FanelliD, IoannidisJPA. What does research reproducibility mean? Science translational medicine. 2016;8(341):341ps12 10.1126/scitranslmed.aaf5027 27252173
  2. 2BegleyCG, IoannidisJP. Reproducibility in science: improving the standard for basic and preclinical research. Circ Res. 2015;116(1):116–26. Epub 2015/01/02. 10.1161/CIRCRESAHA.114.303819 .25552691
  3. 3FreedmanLP, VenugopalanG, WismanR. Reproducibility2020: Progress and priorities. F1000Research. 2017;6:604 Epub 2017/06/18. 10.12688/f1000research.11334.1 28620458PMC5461896
  4. 4KidwellMC, LazarevicLB, BaranskiE, HardwickeTE, PiechowskiS, FalkenbergLS, et al Badges to acknowledge open practices: a simple, low-cost, effective method for increasing transparency. PLoS Biol. 2016;14(5):e1002456 Epub 2016/05/14. 10.1371/journal.pbio.1002456 27171007PMC4865119
  5. 5ElseH. Radical open-access plan could spell end to journal subscriptions. Nature. 2018;561(7721):17–8. Epub 2018/09/06. 10.1038/d41586-018-06178-7 .30181639
  6. 6NosekBA, EbersoleCR, DeHavenAC, MellorDT. The preregistration revolution. Proc Natl Acad Sci U S A. 2018;115(11):2600–6. Epub 2018/03/14. 10.1073/pnas.1708274114 29531091PMC5856500
  7. 7ChambersCD, ForstmannB, PruszynskiJA. Registered reports at the European Journal of Neuroscience: consolidating and extending peer-reviewed study pre-registration. Eur J Neurosci. 2017;45(5):627–8. Epub 2016/12/28. 10.1111/ejn.13519 .28027598
  8. 8BateST, ClarkRA. The design and statistical analysis of animal experiments. Cambridge, United Kingdom: Cambridge University Press; 2014 310 p.
  9. 9Percie du SertN, BamseyI, BateST, BerdoyM, ClarkRA, CuthillI, et al The Experimental Design Assistant. PLoS Biol. 2017;15(9):e2003779 Epub 2017/09/29. 10.1371/journal.pbio.2003779 28957312PMC5634641
  10. 10LazicSE. Experimental design for laboratory biologists: maximising information and improving reproducibility. Cambridge: Cambridge University Press; 2016.
  11. 11MacleodMR, Lawson McLeanA, KyriakopoulouA, SerghiouS, de WildeA, SherrattN, et al Risk of bias in reports of in vivo research: a focus for improvement. PLoS Biol. 2015;13(10):e1002273 Epub 2015/10/16. 10.1371/journal.pbio.1002273 26460723PMC4603955
  12. 12MacleodMR, FisherM, O'CollinsV, SenaES, DirnaglU, BathPM, et al Good laboratory practice: preventing introduction of bias at the bench. Stroke. 2009;40(3):e50–2. Epub 2008/08/16. STROKEAHA.108.525386 [pii] 10.1161/STROKEAHA.108.525386 .18703798
  13. 13RiceAS, Cimino-BrownD, EisenachJC, KontinenVK, Lacroix-FralishML, MachinI, et al Animal models and the prediction of efficacy in clinical trials of analgesic drugs: a critical appraisal and call for uniform reporting standards. Pain. 2009;139(2):243–7. Epub 2008/09/26. S0304-3959(08)00508-3 [pii] 10.1016/j.pain.2008.08.017 .18814968
  14. 14McCanceI. Assessment of statistical procedures used in papers in the Australian Veterinary Journal. Aust Vet J. 1995;72(9):322–8. Epub 1995/09/01. 10.1111/j.1751-0813.1995.tb07534.x .8585846
  15. 15HackamDG, RedelmeierDA. Translation of research evidence from animals to humans. JAMA. 2006;296(14):1731–2. Epub 2006/10/13. 10.1001/jama.296.14.1731 .17032985
  16. 16KilkennyC, ParsonsN, KadyszewskiE, FestingMF, CuthillIC, FryD, et al Survey of the quality of experimental design, statistical analysis and reporting of research using animals. PLoS ONE. 2009;4(11):e7824 Epub 2009/12/04. 10.1371/journal.pone.0007824 .19956596PMC2779358
  17. 17van der WorpHB, HowellsDW, SenaES, PorrittMJ, RewellS, O'CollinsV, et al Can animal models of disease reliably inform human studies? PLoS Med. 2010;7(3):e1000245 Epub 2010/04/03. 10.1371/journal.pmed.1000245 .20361020PMC2846855
  18. 18GlasziouP, AltmanDG, BossuytP, BoutronI, ClarkeM, JuliousS, et al Reducing waste from incomplete or unusable reports of biomedical research. Lancet. 2014;383(9913):267–76. Epub 2014/01/15. 10.1016/S0140-6736(13)62228-X .24411647
  19. 19BegleyCG, EllisLM. Drug development: raise standards for preclinical cancer research. Nature. 2012;483(7391):531–3. Epub 2012/03/31. 10.1038/483531a .22460880
  20. 20ScottS, KranzJE, ColeJ, LincecumJM, ThompsonK, KellyN, et al Design, power, and interpretation of studies in the standard murine model of ALS. Amyotroph Lateral Scler. 2008;9(1):4–15. Epub 2008/02/15. 789666722 [pii] 10.1080/17482960701856300 .18273714
  21. 21KilkennyC, AltmanDG. Improving bioscience research reporting: ARRIVE-ing at a solution. Lab Anim. 2010;44(4):377–8. Epub 2010/07/28. la.2010.0010021 [pii] 10.1258/la.2010.0010021 .20660161
  22. 22KilkennyC, BrowneW, CuthillIC, EmersonM, AltmanDG. Animal research: reporting in vivo experiments: the ARRIVE guidelines. J Gene Med. 2010;12(7):561–3. Epub 2010/07/08. 10.1002/jgm.1473 .20607692
  23. 23KilkennyC, BrowneWJ, CuthillIC, EmersonM, AltmanDG. Improving bioscience research reporting: the ARRIVE guidelines for reporting animal research. PLoS Biol. 2010;8(6):e1000412 Epub 2010/07/09. 10.1371/journal.pbio.1000412 20613859PMC2893951
  24. 24KilkennyC, BrowneWJ, CuthillIC, EmersonM, AltmanDG. Improving bioscience research reporting: the ARRIVE guidelines for reporting animal research. Journal of Pharmacology & Pharmacotherapeutics. 2010;1(2):94–9. 10.4103/0976-500X.72351 21350617PMC3043335
  25. 25KilkennyC, BrowneWJ, CuthillIC, EmersonM, AltmanDG. Animal Research: Reporting In Vivo Experiments: the ARRIVE guidelines. J Physiol. 2010;588(Pt 14):2519–21. Epub 2010/07/17. 588/14/2519 [pii] 10.1113/jphysiol.2010.192278 20634180PMC2916981
  26. 26KilkennyC, BrowneWJ, CuthillIC, EmersonM, AltmanDG. Animal Research: Reporting In Vivo Experiments: the ARRIVE guidelines. Exp Physiol. 2010;95(8):842–4. Epub 2010/07/09. 95/8/842 [pii] 10.1113/expphysiol.2010.053793 .20610776
  27. 27McGrathJC, DrummondGB, McLachlanEM, KilkennyC, WainwrightCL. Guidelines for reporting experiments involving animals: the ARRIVE guidelines. Br J Pharmacol. 2010;160(7):1573–6. Epub 2010/07/24. BPH873 [pii] 10.1111/j.1476-5381.2010.00873.x 20649560PMC2936829
  28. 28KilkennyC, BrowneW, CuthillIC, EmersonM, AltmanDG. Animal Research: Reporting In Vivo Experiments-the ARRIVE Guidelines. J Cereb Blood Flow Metab. 2011 Epub 2011/01/06. jcbfm2010220 [pii] 10.1038/jcbfm.2010.220 .21206507PMC3070981
  29. 29KilkennyC, BrowneWJ, CuthillIC, EmersonM, AltmanDG. Improving bioscience research reporting: the ARRIVE guidelines for reporting animal research. Vet Clin Pathol. 2012;41(1):27–31. Epub 2012/03/07. 10.1111/j.1939-165X.2012.00418.x .22390425
  30. 30KilkennyC, BrowneWJ, CuthillIC, EmersonM, AltmanDG. Improving bioscience research reporting: the ARRIVE guidelines for reporting animal research. Osteoarthritis Cartilage. 2012;20(4):256–60. 10.1016/j.joca.2012.02.010 22424462
  31. 31AveyMT, MoherD, SullivanKJ, FergussonD, GriffinG, GrimshawJM, et al The Devil Is in the Details: Incomplete Reporting in Preclinical Animal Research. PLoS ONE. 2016;11(11):e0166733 Epub 2016/11/18. 10.1371/journal.pone.0166733 27855228PMC5113978
  32. 32LeungV, Rousseau-BlassF, BeauchampG, PangDSJ. ARRIVE has not ARRIVEd: Support for the ARRIVE (Animal Research: Reporting of in vivo Experiments) guidelines does not improve the reporting quality of papers in animal welfare, analgesia or anesthesia. PLoS ONE. 2018;13(5):e0197882 Epub 2018/05/26. 10.1371/journal.pone.0197882 29795636PMC5967836
  33. 33HairK, MacleodMR, SenaES, SenaES, HairK, MacleodMR, et al A randomised controlled trial of an intervention to improve compliance with the ARRIVE guidelines (IICARus). Research Integrity and Peer Review. 2019;4(1):12 10.1186/s41073-019-0069-3 31205756PMC6560728
  34. 34The NPQIP Collaborative group. Did a change in Nature journals’ editorial policy for life sciences research improve reporting? BMJ Open Science. 2019;3(1):e000035 10.1136/bmjos-2017-000035PMC864760835047682
  35. 35HanS, OlonisakinTF, PribisJP, ZupeticJ, YoonJH, HolleranKM, et al A checklist is associated with increased quality of reporting preclinical biomedical research: a systematic review. PLoS ONE. 2017;12(9):e0183591 Epub 2017/09/14. 10.1371/journal.pone.0183591 28902887PMC5597130
  36. 36RamirezFD, MotazedianP, JungRG, Di SantoP, MacDonaldZD, MorelandR, et al Methodological rigor in preclinical cardiovascular studies: targets to enhance reproducibility and promote research translation. Circ Res. 2017;120(12):1916–26. Epub 2017/04/05. 10.1161/CIRCRESAHA.117.310628 28373349PMC5466021
  37. 37ReichlinTS, VogtL, WurbelH. The Researchers' View of Scientific Rigor-Survey on the Conduct and Reporting of In Vivo Research. PLoS ONE. 2016;11(12):e0165999 Epub 2016/12/03. 10.1371/journal.pone.0165999 27911901PMC5135049
  38. 38HurstV, Percie du SertN. The ARRIVE guidelines survey. Open Science Framework. 2017 10.17605/OSF.IO/G8T5Q
  39. 39FraserH, ParkerT, NakagawaS, BarnettA, FidlerF. Questionable research practices in ecology and evolution. PLoS ONE. 2018;13(7):e0200303 Epub 2018/07/17. 10.1371/journal.pone.0200303 30011289PMC6047784
  40. 40The Academy of Medical Sciences. Reproducibility and reliability of biomedical research: improving research practice 2015. Available from: https://acmedsci.ac.uk/policy/policy-projects/reproducibility-and-reliability-of-biomedical-research. [cited 2020 June 16].
  41. 41SenaES, CurrieGL. How our approaches to assessing benefits and harms can be improved. Animal Welfare. 2019;28(1):107–15. 10.7120/09627286.28.1.107 WOS:000455904200011.
  42. 42Percie du SertN, AhluwaliaA, AlamS, AveyMT, BakerM, BrowneWJ, et al Reporting animal research: Explanation and elaboration for the ARRIVE guidelines 2.0. PLoS Biol. 2020;18(7):e3000411 10.1371/journal.pbio.3000411PMC736002532663221
  43. 43LandisSC, AmaraSG, AsadullahK, AustinCP, BlumensteinR, BradleyEW, et al A call for transparent reporting to optimize the predictive value of preclinical research. Nature. 2012;490(7419):187–91. Epub 2012/10/13. 10.1038/nature11556 23060188PMC3511845
  44. 44KimmelmanJ, AndersonJA. Should preclinical studies be registered? Nature biotechnology. 2012;30(6):488–9. Epub 2012/06/09. 10.1038/nbt.2261 .22678379PMC4516408
  45. 45WilkinsonMD, DumontierM, AalbersbergIJ, AppletonG, AxtonM, BaakA, et al The FAIR Guiding Principles for scientific data management and stewardship. Scientific Data. 2016;3:160018 10.1038/sdata.2016.18 26978244PMC4792175
  46. 46MoherD, SchulzKF, SimeraI, AltmanDG. Guidance for developers of health research reporting guidelines. PLoS Med. 2010;7(2):e1000217 Epub 2010/02/20. 10.1371/journal.pmed.1000217 .20169112PMC2821895
  47. 47ChambersK, CollingsA., GrafC., KiermerV., MellorD.T., MacleodM., SwaminathanS., SweetD., VinsonV. Towards minimum reporting standards for life scientists. MetaArXiv. 2019 10.31222/osf.io/9sm4x
  48. 48RandsSA. Inclusion of policies on ethical standards in animal experiments in biomedical science journals. J Am Assoc Lab Anim Sci. 2011;50(6):901–3. Epub 2012/02/15. 22330784PMC3228928
  49. 49OsborneNJ, PayneD, NewmanML. Journal editorial policies, animal welfare, and the 3Rs. Am J Bioeth. 2009;9(12):55–9. Epub 2009/12/17. 10.1080/15265160903318343 .20013503
  50. 50VasilevskyNA, MinnierJ, HaendelMA, ChampieuxRE. Reproducible and reusable research: are journal data sharing policies meeting the mark? PeerJ. 2017;5 10.7717/peerj.3208 WOS:000400303800002. 28462024PMC5407277
  51. 51GiofreD, CummingG, FrescL, BoedkerI, TressoldiP. The influence of journal submission guidelines on authors' reporting of statistics and use of open research practices. PLoS ONE. 2017;12(4). 10.1371/journal.pone.0175583 WOS:000399874800038. 28414751PMC5393581
  52. 52MichelMC, MurphyTJ, MotulskyHJ. New author guidelines for displaying data and reporting data analysis and statistical methods in experimental biology. Mol Pharmacol. 2020;97(1):49–60. Epub 2019/12/29. 10.1124/mol.119.118927 .31882404
  53. 53Rowan-LeggA, WeijerC, GaoJ, FernandezC. A comparison of journal instructions regarding institutional review board approval and conflict-of-interest disclosure between 1995 and 2005. J Med Ethics. 2009;35(1):74–8. 10.1136/jme.2008.024299 WOS:000261929300018. 19103950
  54. 54AnckerJS, FlanaginA. A comparison of conflict of interest policies at peer-reviewed journals in different scientific disciplines. Sci Eng Ethics. 2007;13(2):147–57. Epub 2007/08/25. 10.1007/s11948-007-9011-z .17717729
  55. 55Updated RCUK guidance for funding applications involving animal research 2015. Available from: https://mrc.ukri.org/news/browse/updated-rcuk-guidance-for-funding-applications-involving-animal-research/. [cited 2020 June 16].
  56. 56PragerEM, ChambersKE, PlotkinJL, McArthurDL, BandrowskiAE, BansalN, et al Improving transparency and scientific rigor in academic publishing. J Neurosci Res. 2018 Epub 2018/12/07. 10.1002/jnr.24340 .30506706PMC12990824
  57. 57Enhancing reproducibility. Nat Methods. 2013;10(5):367 Epub 2013/06/14. 10.1038/nmeth.2471 .23762900
  58. 58CurtisMJ, AlexanderS, CirinoG, DochertyJR, GeorgeCH, GiembyczMA, et al Experimental design and analysis and their reporting II: updated and simplified guidance for authors and peer reviewers. Br J Pharmacol. 2018;175(7):987–93. Epub 2018/03/10. 10.1111/bph.14153 29520785PMC5843711
  59. 59HeavenD. AI peer reviewers unleashed to ease publishing grind. Nature. 2018;563(7733):609–10. Epub 2018/11/30. 10.1038/d41586-018-07245-9 .30482927