A machine learning algorithm can detect diabetes and prediabetes from ECG features, according to a study published in BMJ Innovations. Anoop R. Kulkarni, PhD, and colleagues combined noninvasive ECG with machine learning to detect diabetes and prediabetes using data from 1,262 individuals and 10,461 time-aligned heartbeats recorded digitally. A classifier that used the signal-processed ECG as input predicted membership in no diabetes, prediabetes, or T2D classes. Prevalence rates of T2D and prediabetes were about 30% and 14%, respectively. Training was quick and smooth, with convergence achieved within 40 epochs. The algorithm predicted the classes in the independent test set with 97.1% precision, 96.2% recall, 96.8% accuracy, and 96.6% F1 score. “In theory, our study provides a relatively inexpensive, noninvasive, and accurate alternative that can be used as a gatekeeper to effectively detect diabetes and prediabetes early in its course,” Dr. Kulkarni and colleagues wrote.