Background: COVID-19 is an infectious disease characterized by various clinical presentations. Knowledge of possible symptoms and their distribution allows for the early identification of infected patients. Objective: To determine the distribution pattern of COVID-19 symptoms as well as possible unreported symptoms, we created an app-based self-reporting tool. Methods: The COVID-19 Symptom Tracker is an app-based daily self-reporting tool. Between April 8 and May 15, 2020, a total of 22,327 individuals installed this app on their mobile device. An initial questionnaire asked for demographic information (age, gender, postal code) and past medical history comprising relevant chronic diseases. The participants were reminded daily to report whether they were experiencing any symptoms and if they had been tested for SARS-CoV-2 infection. Participants who sought health care services were asked additional questions regarding diagnostics and treatment. Participation was open to all adults (≥18 years). The study was completely anonymous. Results: In total, 11,829 (52.98%) participants completed the symptom questionnaire at least once. Of these, 291 (2.46%) participants stated that they had undergone an RT-PCR (reverse transcription-polymerase chain reaction) test for SARS-CoV-2; 65 (0.55%) reported a positive test result and 226 (1.91%) a negative one. The mean number of reported symptoms among untested participants was 0.81 (SD 1.85). Participants with a positive test result had, on average, 5.63 symptoms (SD 2.82). The most significant risk factors were diabetes (odds ratio [OR] 8.95, 95% CI 3.30-22.37) and chronic heart disease (OR 2.85, 95% CI 1.43-5.69). We identified chills, fever, loss of smell, nausea and vomiting, and shortness of breath as the top five strongest predictors for a COVID-19 infection. The odds ratio for loss of smell was 3.13 (95% CI 1.76-5.58). Nausea and vomiting (OR 2.84, 95% CI 1.61-5.00) had been reported as an uncommon symptom previously; however, our data suggest a significant predictive value. Conclusions: Self-reported symptom tracking helps to identify novel symptoms of COVID-19 and to estimate the predictive value of certain symptoms. This aids in the development of reliable screening tools. Clinical screening with a high pretest probability allows for the rapid identification of infections and the cost-effective use of testing resources. Based on our results, we suggest that loss of smell and taste be considered cardinal symptoms; we also stress that diabetes is a risk factor for a highly symptomatic course of COVID-19 infection.

This is the abstract only. Read the full article on the JMIR site. JMIR is the leading open access journal for eHealth and healthcare in the Internet age.

中文翻译：

---

基于App的自我报告COVID-19症状跟踪：问卷数据分析。

背景：COVID-19是一种具有多种临床表现特征的传染病。了解可能出现的症状及其分布，可以及早发现受感染的患者。目的：为了确定COVID-19症状以及可能未报告的症状的分布模式，我们创建了一个基于应用程序的自我报告工具。方法：COVID-19症状跟踪器是基于应用程序的每日自我报告工具。在2020年4月8日至5月15日之间，共有22,327个人在其移动设备上安装了此应用。初始问卷要求提供人口统计学信息（年龄，性别，邮政编码）和包括相关慢性病的既往病史。每天都提醒参与者报告他们是否有任何症状以及是否已接受SARS-CoV-2感染检测。向寻求医疗保健服务的参与者询问了有关诊断和治疗的其他问题。所有成年人（≥18岁）均可参加。这项研究是完全匿名的。结果：总共11,829名（52.98％）参与者至少完成一次症状问卷调查。其中，291（2.46％）名参与者表示，他们已经接受过SARS-CoV-2的RT-PCR（逆转录-聚合酶链反应）测试；65个（0.55％）的测试结果为阳性，而226（1.91％）的测试结果为阴性。未经测试的参与者中报告的症状平均数为0.81（SD 1.85）。测试结果呈阳性的参与者平均有5.63个症状（标准差2.82）。最重要的危险因素是糖尿病（几率[OR] 8.95，95％CI 3.30-22.37）和慢性心脏病（OR 2.85，95％CI 1.43-5.69）。我们发现了寒战，发烧，嗅觉丧失，恶心和呕吐以及呼吸急促是COVID-19感染的前五项最强预测指标。气味损失的比值比是3.13（95％CI 1.76-5.58）。先前已报道恶心和呕吐（OR 2.84，95％CI 1.61-5.00）是一种罕见的症状。但是，我们的数据表明具有重要的预测价值。结论：自我报告的症状追踪有助于识别COVID-19的新症状，并估计某些症状的预测价值。这有助于开发可靠的筛选工具。具有较高的预检概率的临床筛查可以快速识别感染并以经济有效的方式使用检测资源。根据我们的结果，我们建议将嗅觉和味觉的丧失视为主要症状。

这仅仅是抽象的。阅读JMIR网站上的全文。JMIR是互联网时代电子健康和医疗保健领域领先的开放获取期刊。