JMIR Publications

ABSTRACT

Background:

Massive amount of covid-19 related data is generated everyday by Twitter users. Self-reports of covid-19 symptoms on Twitter can reveal a great deal about the disease and its prevalence in the community. In particular, self-reports can be used as a valuable resource to learn more about the common symptoms and whether their order of appearance differs among different groups in the community. With sufficient available data, this has the potential of developing a covid-19 risk- assessment system that is tailored toward specific group of people.

Objective:

The aim of this study is to identify the most common symptoms reported by covid-19 patients in the Arabic language and order the symptoms appearance based on the collected data.

Methods:

We search the Arabic content of Twitter for personal reports of covid-19 symptoms from March 1st to May 27th, 2020. We identify 463 Arabic users who tweeted testing positive for covid-19 and extract the symptoms they publicly associate with covid-19. Furthermore, we ask them directly through personal messages to opt in and rank the appearance of the first three symptoms they experienced right before (or after) diagnosed with covid-19. Finally, we track their Twitter timeline to identify additional symptoms that were mentioned within +-5 days from the day of tweeting having covid-19. In summary, a list of 270 covid-19 reports were collected and symptoms were (at least partially) ranked from early to late.

Results:

The collected reports contained roughly 900 symptoms originated from 74% (n=201) male and 26% (n=69) female Twitter users. The majority (82%) of the tracked users were living in Saudi Arabia (46%) and Kuwait (36%). Furthermore, 13% (n=36) of the collected reports were asymptomatic. Out of the users with symptoms (n=234), 66% (n=180) provided a chronological order of appearance for at least three symptoms. Fever 59% (n=139), Headache 43% (n=101), and Anosmia 39% (n=91) were found to be the top three symptoms mentioned by the reports. They count also for the top-3 common first symptoms in a way that 28% (n=65) said their covid journey started with a Fever, 15% (n=34) with a Headache and 12% (n=28) with Anosmia. Out of the Saudi symptomatic reported cases (n=110), the most common three symptoms were Fever 59% (n=65), Anosmia 42% (n=46), and Headache 38% (n=42).

Conclusions:

This study demonstrates that Twitter is a valuable resource to analyze and identify COVID-19 early symptoms within the Arabic content of Twitter. It also suggests the possibility of developing a real-time covid-19 risk estimator based on the users’ tweets.