Data Sharing of Continuous Glucose Monitoring Data: The Need for a New Paradigm?

Data sharing has the potential to advance scientific progress. When scientists have access to comprehensive data material, they can investigate new research questions, validate findings, increase transparency, and conduct large-scale analyses across multiple trials—increasing the potential of producing robust and generalizable findings.^1,2 For researchers sharing data, this could lead to an increased academic acknowledgment. Despite the obvious benefits, support from politicians, funding, and many scientific publishers, academic scientists infrequently make their research data available. ³

In diabetes research, continuous glucose monitoring systems (CGM) has over the two last decades proven to be a paradigm shift in both management of diabetes and understanding of the disease. CGM offers patients with diabetes the opportunity to improve their glucose control by monitoring glucose levels continuously. This aids in finding nonoptimal glucose patterns and variations that would be difficult to detect using the traditional self-measurements of blood glucose (SMBG). Studies have shown how CGM could be an essential part of an intelligent control system, detecting hypoglycemia or guiding the administration of insulin. ⁴ But in general, many studies don’t have access to a large data material of patients using CGM under different conditions. ⁵ Therefore, most of this novel research is applied on few homogeneous samples, which limits the generalizability and applicability into diabetes management. ⁶

Numerous clinical studies including CGM as an important intervention or measure of outcome have been conducted over the years; still, very few data have been made available for the purpose of reuse. Nevertheless, some initiatives in the direction of research transparency and data sharing have been made, which include data on patients with diabetes using CGM—such as the open D1NAMO dataset, the OhioT1DM dataset, t1dexchange, and the Novo Nordisk data science initiative. There is still a need for high-quality datasets including CGM to be made available to researchers worldwide. Especially large heterogeneous CGM datasets could be used for studies as a benchmark, which would make it easier to compare different approaches between studies.