Back

The impact of early scientific literature in response to COVID-19: a scientometric perspective

Golinelli, D.; Nuzzolese, A. G.; Boetto, E.; Rallo, F.; Greco, M.; Toscano, F.; Fantini, M. P.

2020-04-18 health policy
10.1101/2020.04.15.20066183 medRxiv
Show abstract

BackgroundIn the early phases of a new pandemic, identifying the most relevant evidence and quantifying which studies are shared the most can help researchers and policy makers. The aim of this study is to describe and quantify the impact of early scientific production in response to COVID-19 pandemic. MethodsThe study consisted of: 1) review of the scientific literature produced in the first 30 days since the first COVID-19 paper was published; 2) analysis of papers metrics with the construction of a "Computed-Impact-Score" (CIS) that represents a unifying score over heterogeneous bibliometric indicators. In this study we use metrics and alternative metrics collected into five separate categories. On top of those categories we compute the CIS. Highest CIS papers are further analyzed. Results239 papers have been included in the study. The mean of citations, mentions and social media interactions resulted in 1.63, 10 and 1250, respectively. The paper with highest CIS resulted "Clinical features of patients[...]" by Chaolin Huang et al., which rated first also in citations and mentions. This is the first paper describing patients affected by the new disease and reporting data that are clearly of great interest to both the scientific community and the general population. ConclusionsThe early response of scientific literature during an epidemic does not follow a pre-established pattern. Being able to monitor how communications spread from the scientific world toward the general population using both traditional and alternative metric measures is essential, especially in the early stages of a pandemic.

Matching journals

The top 3 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.