Evaluating the side effects of Covid-۱۹ vaccines with artificial intelligence methods

Background and aims: During and after the Covid-۱۹ vaccination, various side effects of vaccinesmade up a global debate. In addition, realizing the side effects of vaccines as soon as possiblecan lead to accurate monitoring of the vaccination process during pandemics. A precise andquick way to collect side effect reports of these vaccines is through artificial intelligence (AI)techniques. We conducted a systematic search to evaluate the side effects noticed through theartificial intelligence aid.Method: We searched for the related articles in three databases; PubMed, Scopus, and Web ofScience with the keywords related to Covid-۱۹ vaccines, side effects, and artificial intelligenceterminology. Out of ۲۰۳ articles for screening, we selected ۱۴ articles after applying exclusioncriteria which consisted of congress, non-English, lacking full-text articles, etc.Results: According to the results, the most used data for constructing AI methods source is theVaccine Adverse Event Reporting System (VAERS), but data sources such as Twitter, Kaggle, andelectronic health records (EHR) were also frequently employed. Most of the AI models are basedon CNN (Convolutional Neural Network), regression networks, and neural networks. Text miningis the most frequently assisted method for data extraction. According to our results, death and allergicsymptoms are the basis of side effects that people paid more attention to them. We revealedthat in all reports women to complain more about side effects. We found that headache is the mostcomplained side effect of Covid-۱۹ vaccines in overall articles, and fatigue is the second mostcomplained side effect in overall results. According to our study, we concluded that in social mediasuch as Twitter, more severe side effects such as thrombosis and anaphylaxis are mentioned,which causes a discrepancy with the results of other data sources. Despite the reports mentioningnot observing anaphylaxis after the Moderna vaccine injection, studies based on AI reveal that itis a vital side effect of that brand, which shows the ability of AI to check more information aboutthe disease and can be used properly in pandemics where previous information is not available.Conclusion: It seems that there is a problem in investigating the effect of gender on the side effectsthrough social-media data, and in our screened articles, there was no gender separation inthe data that used social media. We recommend that an AI approach should be developed to solvethis problem in future health data monitoring. We suggest that during future pandemics modelinga single platform for extracting patients’ data makes the results more accurate.

کلیدواژه ها:

Artificial Intelligence ، Covid-۱۹ Vaccine ، side effects ، text mining