Rxivist logo

Transparency Assessment of COVID-19 Models

By Mohammad S. Jalali, Catherine DiGennaro, Devi Sridhar

Posted 21 Jul 2020
medRxiv DOI: 10.1101/2020.07.18.20156851

As the COVID-19 pandemic has caused major societal unrest, modelers have worked to project future trends of COVID-19 and predict upcoming challenges and impacts of policy action. These models, alone or in aggregate, are influential for decision-makers at every level. Therefore, the method and documentation of COVID-19 models must be highly transparent to ensure that projections and consequential policies put forth have sound epistemological grounds. We evaluated 29 COVID-19 models receiving high attention levels within the scientific community and/or informing government responses. We evaluated these models against 27 transparency criteria. We found high levels of transparency in model documentation aspects such as reporting uncertainty analysis; however, about half of the models do not share code and a quarter do not report equations. These discrepancies underscore the need for transparency and reproducibility to be at the forefront of researchers' priorities, especially during a global health crisis when stakes are critically high.

Download data

  • Downloaded 255 times
  • Download rankings, all-time:
    • Site-wide: 108,958
    • In infectious diseases: 4,581
  • Year to date:
    • Site-wide: 120,250
  • Since beginning of last month:
    • Site-wide: None

Altmetric data

Downloads over time

Distribution of downloads per paper, site-wide