Determining the effectivenes of medications based on patient reviews collected on medical social media

This article highlights solution of the problem of determining the medications’ effectiveness
based on sentiment analysis of patient reviews collected in the medical segment of social
media. Public opinion about media subjects (physicians, nurses, clinics, pharmaceutical
companies, etc.) can be determined based on the information collected in medical social
media. One of the most discussed topics in medical social media is related to medications
(drugs), their effectiveness, and determining public opinion based on collected user comments
is one of the current issues. To analyze patient reviews about drugs, the Kaggle platform
drugsComTest_raw.csv medical database is used, lexicon-based sentiment analysis, statistical
methods and machine learning algorithms are applied. The problem is solved in stages on the
patient-disease, disease-drug and patient-drug segments, the issues of which diseases are most
often used for drugs, and which drugs are most often used for each disease are resolved.
Based on the integration of the results obtained from the problem solutions, a mechanism for
forming public opinion on the effectiveness of drugs is developed. The proposed approach
takes into account not only positive but also negative opinions when determining public
opinion about the effectiveness of drugs. Such results can be used to support appropriate
decision-making in the healthcare sector, specifically in pharmaceutical companies.