Reassessing the Feasibility of PPG-Based Non-Invasive Blood Glucose Level Estimation
2026-08-03 • Human-Computer Interaction
Human-Computer InteractionMachine Learning
AI summaryⓘ
The authors studied how well methods estimate blood sugar levels using a light-based sensor (PPG) without needing blood samples. They found that many current tests overstate how good these methods are because they mix up data from the same people when training and testing. When they tried stricter tests separating data by participants, the models performed poorly and often only matched simple average guessing. Interestingly, standard clinical accuracy metrics made the models look better than they really were. The authors conclude that better testing is needed before trusting these models for real-world use.
blood glucose levelphotoplethysmographymachine learning evaluationdata leakagetrain-test splitR-squaredClarke Error Gridwearable health monitoringparticipant-aware splittingclinical validation
Authors
Supraja Ramesh, Markus Neufeld, Michael Küttner, Tobias Röddiger, Michael Beigl
Abstract
Non-invasive blood glucose level (BGL) estimation from photoplethysmography (PPG) holds great promise for wearable health monitoring, but results across studies are hard to compare due to inconsistent datasets, data leakage, and non-standardized evaluation metrics. We present the first reproducible, extensible evaluation pipeline and use it to reassess five representative PPG-based BGL methods on published datasets under three increasingly strict data-split protocols: random window-level, participant-aware, and leave-some-participants-out (LSPO). Models appeared competitive under random splitting but collapsed under participant-aware and LSPO evaluation, with nearly all yielding near-zero or negative R$^2$ values comparable to a mean-prediction baseline. Critically, across every model and split, over 90% of predictions fell within clinically acceptable zones (Clarke Error Grid A+B), including the baseline. This reveals a fundamental disconnect: clinical zone metrics systematically conceal model failure in this domain. Our findings demonstrate that random train-test splits substantially overestimate the generalization of PPG-based BGL models due to sample-level data leakage, and that robust ML evaluation must precede clinical validation to meaningfully assess real-world utility.