An early Bayesian Network Meta-analysis of Coronavirus Diseases 2019 (COVID-19) clinical trials

In this paper, we conducted a Bayesian Network Meta-analysis of the latest COVID-19 clinical trials including 5 studies on 591 patients receiving 4 different agents of Arbidol, Favipiravir, lopinavir- ritonavir, and Hydroxychloroquine and standard treatment protocol. We ranked the best agent based on patient improvement using Markov-Monte-Carlo-Chain. Hydroxychloroquine showed the best efficiency following the Favipiravir, Arbidol, lopinavir-ritonavir, and standard regimen in the first week of treatment. In the second week, with excluding Hydroxychloroquine arm (as some reporting studies hadnchr('39')t addressed its efficacy in the second week), Favipiravir was the best treatment following by lopinavir-ritonavir, standard care, and Arbidol. As we saw a huge change in the ranking of the drugs by evaluating outcomes in the second week of treatment, we think that COVID-19 randomized clinical trials should be performed based on a standard study protocol worldwide, that could help  policy makers to make a decision on the treatment protocol.

Keywords:

Network Meta-analysis , Favipiravir , SARS-CoV-2 , COVID-19 , lopinavir-ritonavir.