Promises and pitfalls of electronic health record analysis

Routinely collected electronic health care records (EHRs) are increasingly used for research; with their use comes the opportunity for large scale, high quality studies that can address questions not easily answered by randomised clinical trials or classical cohort studies involving bespoke data collection. However, the use of electronic health records generate challenges in terms of ensuring methodological rigour, which may be greater when studying complex chronic diseases such as diabetes. This review aims to describe the promises and potential of EHRs in the context of diabetes research, and outline key areas of caution with examples. In particular we consider the difficulties in identifying and classifying diabetes patients, in distinguishing between prevalent and incident cases, and in dealing with the complexities of diabetes progression and treatment; we discuss the dangers of introducing time-related biases; and we describe problems of inconsistent data recording, missing data, and confounding. Throughout we provide practical recommendations for good practice in conducting EHR studies and interpreting their results.RF and NC are funded by a Diabetes UK/British Heart foundation award (no. 15/0005250). RM is supported by a Sir Henry Wellcome Postdoctoral Fellowship from the Wellcome Trust (WT/ 201375/Z/16/Z). SVE is supported by a Sir George Alberti Training Fellowship (17/0005588). KB holds a Sir Henry Dale fellowship jointly funded by the Wellcome Trust and the Royal Society (107731/Z/15/Z). LS is supported by a Wellcome Trust Senior Research Fellowship in Clinical Science (098504/Z/12/Z