Introduction
COVID-19 is a novel virus and the course of disease in non-hospitalized patients has not been well characterized. Most of the literature early in the pandemic and to date has focused on the hospital course; there is little information on outpatient disease progression1, 2, 3. This was especially true when the pandemic spread to the United States in January 2020. To guide care for patients at home and prevent unnecessary entry into an overwhelmed health care system, we sought to describe the outpatient course of COVID-19. At the time of this review, our outpatient primary care providers advised patients based on anecdotal evidence gleaned from patients actively ill with COVID-19. This experience-based advice included that patients with COVID-19 at home may experience a relapsing remitting disease course and that a relapse sometimes is severe enough to require hospitalization. To better inform our clinical practice, we used publicly available data to investigate if our guidance reflected what was being observed elsewhere. We systematically analyzed data in the form of first-person accounts from people who reported having COVID-19 while at home published online (e.g., on social media platforms). These data sources provided a rich description of COVID-19 symptomatology and progression and were easily searchable via Google, allowing us to quickly confirm our guidance.
Methods
We searched Google, Twitter, and YouTube to identify adults (≥18 years) who chronicled their COVID-19 symptoms at home on an online platform. We first searched these domains using broad search terms (SARS, SARS-CoV, SARS-CoV2, COVID-19, COVID, coronavirus, and corona), then used the find function within positive searches for more specific terms (positive, test, home, hospital, hospitalized, emergency, ambulance) and, finally, read accounts positive for one or more terms for inclusion criteria. When searching on Twitter, the above terms were preceded with a hashtag. Narrator’s accounts were included if they had self-reported lab test positive for SARS-CoV2, diagnosis of COVID-19 by a doctor based on clinical picture, or described close contact with a patient who had lab-confirmed COVID-19. Narrators were excluded if they did not describe their symptoms in a chronological order. Inclusion/exclusion criteria were screened using the online narration and were not confirmed in any additional way. The search began March 25 and was ended March 30, 2020, when continued web-searching with the above terms failed to reveal new accounts meeting criteria. Data collection included age, sex, self-reported diagnostic criteria, self-reported comorbidities, length of disease, symptom type and timing, timing of symptom remissions/relapses (defined as self-reported symptom improvement followed by worsening within 1 week), hospitalization timing, and discharge timing. Data were analyzed using descriptive data software on Microsoft Excel 2020.
Results
We reviewed 55 first-person accounts, of which 32 met criteria—of these, 23 were found via Google search, 7 YouTube, and 2 social media (Facebook or Twitter). Mean age of the narrators was 43 (95%CI 36-50) years old, 43% were male, and 9% described comorbid conditions. The majority (84%) self-reported that they had lab-confirmed COVID-19. Eighteen cases (56%) described a relapsing remitting course with a mean day of remission of 5 (95%CI: 4.3-5.7) and mean day of relapse of 6 (95%CI:5.2-6.8) Two people with a relapsing and remitting course had a second relapse; one recovered on day 11 with relapse day 12 and the other recovered day 16 with relapse day 17. Of those who relapsed, 7 (39%) noted worsening respiratory symptoms. Eight (25%) of patients were admitted to the hospital and 3 of them were admitted after a relapse. Mean admission day was 5.1 (95%CI: 2.6-7.6) and was earlier for those admitted without a relapse than those admitted after a relapse.
Fatigue and cough were the most common initial symptoms. At the time of the review, we also identified many novel symptoms that had not been widely recognized and reported in the peer-reviewed literature. The most common novel symptoms identified were feeling like there was a “weight on [their] chest,” dizziness, and ocular complaints (eye pain/soreness, burning, and sensitivity to light). At the time of this project done in March 2020, the majority of narrators (75%) had reported symptoms not frequently described in the peer-reviewed hospital focused literature during their disease course.
Discussion
In this analysis of 32 first-person accounts of the COVID-19 symptom course, we found that the majority of narrators had a relapsing remitting course and the majority of them recovered at home without the need for hospitalization. Almost 40% of narrators who relapsed had worsening of their respiratory symptoms and almost half were hospitalized after their relapse. Our findings informed our guidance for patients and healthcare providers. As a result, we counseled patients about the possibility of relapse around day six, after a day of improvement, and that if patients have worsening respiratory symptoms, they should call their physician. While initial symptoms of the narrators were similar to those documented in hospitalized patients, a significant portion of them described novel symptoms that had not been widely described at the time of our review, including feeling like a “weight on [their] chest,” dizziness and eye symptoms.1, 3, 4 This was significant because patients with these symptoms may not suspect they are infected and could spread COVID-19. Our study was limited by the small number of self-reported COVID-19 infections and the inability to confirm self-reported information in a reliable way. It was also limited by the manual extraction of information as it relied on clinician ability to identify articles via search engines. Further research is needed to better understand the course of COVID-19 symptoms, especially in the home setting, and the long-term sequelae of COVID-19 in patients managed outside of the hospital.
This study exemplifies the role of open online platforms in delivering pertinent clinical information in the setting of a global pandemic, particularly when little is known about the infectious agent or disease course. Generalist physicians play a uniquely important role in pandemic management but are often left relying on early studies focused on diseases in the hospital setting. This creative approach to gathering outpatient-specific disease information is specifically suited to informing the generalist practice and should be considered in any future situations like that of COVID-19.