Real-World Databases in Diabetes: A Systematic Review.

AIM: To identify publications that describe RWD databases in the field of diabetes and map their key characteristics.

METHODS

Databases were searched (MEDLINE, SCOPUS, CINAHL from January 2018 to October 2025) to identify relevant publications describing RWD databases. Eligible studies were identified as RWD databases that contained more than 1000 adults and were followed up for one year or longer. Metadata including population, sample size, and outcomes were then extracted.

RESULTS

A total of 610 eligible RWD databases were identified from 1779 included articles. Databases were identified in 74 countries, with the majority collected in Europe (n = 193). Only 22 databases were identified in lower- and- middle income countries (as defined by the World Bank). Demographics, prescription, diagnosis, and biochemical information were collected frequently; however, mortality and technology related variables (CGM usage) were found in fewer datasets. Very few databases were open access however 205 mentioned having the possibility for data sharing.

CONCLUSION

We have produced an online dashboard to summarise the RWD databases available in the field of diabetes. We found many RWD databases globally that contain a range of variables related to individuals' health; however, lower- and middle-income nations are underrepresented within these data.