JMIR Publications

ABSTRACT

Background:

Hyperkalemia is a critical condition, especially in the intensive care unit. So far, there have been no accurate and noninvasive methods for recognizing hyperkalemia events on ambulatory electrocardiogram monitors.

Objective:

This study aimed to improve the accuracy of hyperkalemia predictions from ambulatory electrocardiogram (ECG) monitors using a personalized transfer learning method; this would be done by training a generic model and refining it with personal data.

Methods:

This retrospective cohort study used open-source data from the Waveform Database Matched Subset of the Medical Information Mart from Intensive Care III (MIMIC-III). We included patients with multiple serum potassium test results and matched ECG data from MIMIC-III database. A one-dimensional, convolutional neural network-based deep-learning model was first developed to predict hyperkalemia in a generic population. Once the model achieved a state-of-art performance, it was utilized in an active transfer learning process to perform patient-adaptive heartbeat classification tasks.

Results:

The results show that by acquiring data from each new patient, the personalized model can improve the accuracy of hyperkalemia detection significantly, from an average of 0.604 ± 0.211 to 0.980 ± 0.078, when compared with the generic model. Moreover, the Area under the Receiver Operating Characteristic Curve level improved from 0.729 ± 0.240 to 0.945 ± 0.094.

Conclusions:

By utilizing the deep transfer learning method, we were able to build a clinical standard model for hyperkalemia detection using ambulatory ECG monitors. These findings could potentially be extended to applications that continuously monitor one's ECG for early alerts of hyperkalemia and help avoid unnecessary blood tests. Clinical Trial: This study was approved by the Institutional Review Board of Chang Gung Medical Foundation (number: 202001217B0, date of approval: 21/07/2020).