BACKGROUND A number of prior studies have demonstrated that research participants with limited English proficiency in the United States are routinely excluded from clinical trial participation. Systematic exclusion through study eligibility criteria that require trial participants to be able to speak, read, and/or understand English affects access to clinical trials and scientific generalizability. We sought to establish the frequency with which English language proficiency is required and, conversely, when non-English languages are affirmatively accommodated in US interventional clinical trials for adult populations. METHODS AND FINDINGS We used the advanced search function on ClinicalTrials.gov specifying interventional studies for adults with at least 1 site in the US. In addition, we used these search criteria to find studies with an available posted protocol. A computer program was written to search for evidence of English or Spanish language requirements, or the posted protocol, when available, was manually read for these language requirements. Of the 14,367 clinical trials registered on ClinicalTrials.gov between 1 January 2019 and 1 December 2020 that met baseline search criteria, 18.98% (95% CI 18.34%-19.62%; n = 2,727) required the ability to read, speak, and/or understand English, and 2.71% (95% CI 2.45%-2.98%; n = 390) specifically mentioned accommodation of translation to another language. The remaining trials in this analysis and the following sub-analyses did not mention English language requirements or accommodation of languages other than English. Of 2,585 federally funded clinical trials, 28.86% (95% CI 27.11%-30.61%; n = 746) required English language proficiency and 4.68% (95% CI 3.87%-5.50%; n = 121) specified accommodation of other languages; of the 5,286 industry-funded trials, 5.30% (95% CI 4.69%-5.90%; n = 280) required English and 0.49% (95% CI 0.30%-0.69%; n = 26) accommodated other languages. Trials related to infectious disease were less likely to specify an English requirement than all registered trials (10.07% versus 18.98%; relative risk [RR] = 0.53; 95% CI 0.44-0.64; p < 0.001). Trials related to COVID-19 were also less likely to specify an English requirement than all registered trials (8.18% versus 18.98%; RR = 0.43; 95% CI 0.33-0.56; p < 0.001). Trials with a posted protocol (n = 366) were more likely than all registered clinical trials to specify an English requirement (36.89% versus 18.98%; RR = 1.94, 95% CI 1.69-2.23; p < 0.001). A separate analysis of studies with posted protocols in 4 therapeutic areas (depression, diabetes, breast cancer, and prostate cancer) demonstrated that clinical trials related to depression were the most likely to require English (52.24%; 95% CI 40.28%-64.20%). One limitation of this study is that the computer program only searched for the terms "English" and "Spanish" and may have missed evidence of other language accommodations. Another limitation is that we did not differentiate between requirements to read English, speak English, understand English, and be a native English speaker; we grouped these requirements together in the category of English language requirements. CONCLUSIONS A meaningful percentage of US interventional clinical trials for adults exclude individuals who cannot read, speak, and/or understand English, or are not native English speakers. To advance more inclusive and generalizable research, funders, sponsors, institutions, investigators, institutional review boards, and others should prioritize translating study materials and eliminate language requirements unless justified either scientifically or ethically.

中文翻译：

---

评估临床试验资格标准中英语语言要求的频率：使用 ClinicalTrials.gov 的系统分析。

背景 许多先前的研究表明，在美国英语能力有限的研究参与者通常被排除在临床试验参与之外。通过要求试验参与者能够说、阅读和/或理解英语的研究资格标准进行系统排除会影响临床试验的可及性和科学的普遍性。我们试图确定要求英语语言能力的频率，相反，美国成人干预性临床试验中肯定地适应非英语语言的频率。方法和结果 我们使用了 ClinicalTrials.gov 上的高级搜索功能，指定针对在美国至少有 1 个站点的成人的介入研究。此外，我们使用这些搜索标准来查找具有可用发布协议的研究。编写了一个计算机程序来搜索英语或西班牙语语言要求的证据，或者在可用时手动阅读已发布的协议以了解这些语言要求。在 2019 年 1 月 1 日至 2020 年 12 月 1 日期间在 ClinicalTrials.gov 注册的 14,367 项符合基线搜索标准的临床试验中，18.98%（95% CI 18.34%-19.62%；n = 2,727）需要阅读、说话和/或或懂英语，2.71% (95% CI 2.45%-2.98%; n = 390) 特别提到了翻译成另一种语言的适应。此分析中的其余试验和以下子分析未提及英语语言要求或英语以外的语言适应。在联邦资助的 2,585 项临床试验中，28.86% (95% CI 27. 11%-30.61%；n = 746) 要求英语语言能力和 4.68% (95% CI 3.87%-5.50%; n = 121) 指定其他语言的适应；在 5,286 项由行业资助的试验中，5.30%（95% CI 4.69%-5.90%；n = 280）需要英语，0.49%（95% CI 0.30%-0.69%；n = 26）需要其他语言。与传染病相关的试验比所有注册试验都不太可能指定英语要求（10.07% 对 18.98%；相对风险 [RR] = 0.53；95% CI 0.44-0.64；p < 0.001）。与所有注册试验相比，与 COVID-19 相关的试验也不太可能指定英语要求（8.18% 对 18.98%；RR = 0.43；95% CI 0.33-0.56；p < 0.001）。发布协议的试验（n = 366）比所有注册的临床试验更有可能指定英语要求（36.89% vs 18.98%；RR = 1.94，95% CI 1.69-2.23；p < 0.001)。对 4 个治疗领域（抑郁症、糖尿病、乳腺癌和前列腺癌）中已发布协议的研究进行的单独分析表明，与抑郁症相关的临床试验最有可能需要英语（52.24%；95% CI 40.28%-64.20% ）。这项研究的一个限制是计算机程序只搜索术语“英语”和“西班牙语”，可能遗漏了其他语言适应的证据。另一个限制是我们没有区分阅读英语、说英语、理解英语和母语为英语的要求；我们将这些要求归为英语语言要求的类别。结论 美国成人介入临床试验中，有相当比例的人排除了无法阅读、说话、和/或懂英语，或者母语不是英语。为了推进更具包容性和可推广性的研究，资助者、赞助商、机构、研究人员、机构审查委员会和其他人应优先考虑翻译研究材料并取消语言要求，除非在科学或道德上有正当理由。