SciOps as Key Research Infrastructure: A decade of data to inform next steps for big science

Large, multi-institutional research programs now account for a substantial share of NIH-funded science, and Congress has called for greater accountability, transparency and coordination at the agency. Yet the operational demands of these programs fall mainly on principal investigators (PIs), who are selected and trained for scientific expertise. We examine the impacts of this mismatch using three sources of data. First, 232 participants in a large NIH-funded consortium answered an open-ended question about problems with collaboration and communication. Of 390 coded quotations, 337 named problems, and these mapped to all ten characteristics of effective teams described by NIH officers a decade earlier, most often to “effective leadership and management skills.” Second, we reviewed the PI eligibility policies of the 17 institutions that received the most NIH direct costs from 2014 to 2023. None requires demonstrated leadership, team facilitation or project-management competence. Third, we found that only 13 contact PIs each received more than $200 million in NIH direct costs over the same period; in some cases, their awards accounted for up to 85% of their institution’s NIH direct costs. If any of these PIs lacked operational or leadership expertise, it could have an outsized impact on their projects and on the people in their charge. To remedy these persistent problems, we propose SciOps: a small team of operational, financial, legal and communications experts with decision-making authority over operations, working alongside scientific leaders on large research programs. We further propose that NIH run a pre-registered pilot of SciOps with appropriate outcome measures. Such a pilot would show whether dedicated operational expertise improves how large-scale science runs.