Effect of the Period From COVID-19 Symptom Onset to Confirmation on Disease Duration: Quantitative Analysis of Publicly Available Patient Data

Abstract

Background In general, early intervention in disease based on early diagnosis is considered to be very important for improving health outcomes. However, there is still insufficient evidence regarding how medical care that is based on the early diagnosis of confirmed cases can affect the outcome of COVID-19 treatment. Objective We aimed to investigate the effect of the duration from the onset of clinical symptoms to confirmation of COVID-19 on the duration from the onset of symptoms to the resolution of COVID-19 (release from quarantine). Methods For preliminary data collection, we performed data crawling to extract data from social networks, blogs, and official websites operated by local governments. We collected data from the 4002 confirmed cases in 33 cities reported up to May 31, 2020, for whom sex and age information could be verified. Subsequently, 2494 patients with unclear symptom onset dates and 1349 patients who had not been released or had no data about their release dates were excluded. Thus, 159 patients were finally included in this study. To investigate whether rapid confirmation reduces the prevalence period, we divided the duration from symptom onset to confirmation into quartiles of ≤1, ≤3, ≤6, and ≥7 days, respectively. We investigated the duration from symptom onset to release and that from confirmation to release according to these quartiles. Furthermore, we performed multiple regression analysis to investigate the effects of rapid confirmation after symptom onset on the treatment period, duration of prevalence, and duration until release from isolation. Results We performed multiple regression analysis to investigate the association between rapid confirmation after symptom onset and the total prevalence period (faster release from isolation). The time from symptom onset to confirmation showed a negative association with the time from confirmation to release (t1=−3.58; P<.001) and a positive association with the time from symptom onset to release (t1=5.86; P<.001); these associations were statistically significant. Conclusions The duration from COVID-19 symptom onset to confirmation date is an important variable for predicting disease prevalence, and these results support the hypothesis that a short duration of symptom onset to confirmation can reduce the time from symptom onset to release.