Reconciling data actionability and accountability in global health research: The case of SARS-CoV-2.

Abstract

The requirements for actionability and accountability in data infrastructures are often viewed as incompatible, creating a trade-off where enhancing one diminishes the other. Through a comparative analysis of two data infrastructures used to share genomic data about the SARS-CoV-2 virus, we argue that making data actionable for knowledge development involves a commitment to ensuring that the data in question are representative of the phenomena being studied and accountable to data subjects and users. This in turn presupposes that: (1) enough data are contributed by a wide and diverse set of relevant sources; (2) mechanisms of feedback and inclusion are set up to ensure that data contributors can participate in data governance and interpretation, thereby helping to adequately contextualise data; and (3) accountability extends to the ways in which data infrastructures are run, financed and positioned vis-à-vis the communities they are meant to serve. Such a model of data sharing can only work on the understanding that data do not need to be easily accessible to be actionable; rather, actionability depends on the responsiveness and accountability of data infrastructures, and the efforts invested in ensuring open communication among contributors.