What is the clinical evidence from this study?

Study design: Observational. Population: Tetanus (n=110). Intervention: 2D-CNN-Transformer network vs. Baseline machine learning methods (e.g., 1D-CNN, 2D-CNN, Random Forest). Primary outcome: Classification of tetanus severity (mild vs severe).

What does this research mean for the field?

A novel hybrid CNN-Transformer model using ECG spectrograms outperforms state-of-the-art methods in automatically classifying tetanus severity, although a Random Forest model with manually selected features can achieve comparable performance. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

November 3, 2022Open Access

Improving Classification of Tetanus Severity for Patients in Low-Middle Income Countries Wearing ECG Sensors by Using a CNN-Transformer Network

Key Result

The proposed 2D-CNN-Transformer model outperformed state-of-the-art methods in classifying tetanus severity from ECG spectrograms, achieving an F1 score of 0.82 and an accuracy of 0.88.

Study Design

Type

Observational (n=110)

Multicenter

Structured PICO

Does a 2D-CNN-Transformer hybrid model improve the classification of tetanus severity in ICU patients using wearable ECG sensors compared to traditional CNNs and machine learning methods?

Population

110 tetanus patients ≥16 years old admitted to the ICU at the Hospital for Tropical Diseases, Ho Chi Minh City, Vietnam.

Intervention

2D-CNN-Transformer hybrid model analyzing 2D logarithmic spectrograms from 60-second ECG signals acquired via low-cost wearable sensors (ePatch).

Comparator

Baseline machine learning and deep learning methods including 1D-CNN, 2D-CNN, 2D-CNN + Dual Attention, 2D-CNN + Channel-wise Attention, and Random Forest.

Outcome

Classification of tetanus severity (mild vs. severe) based on Ablett score.

A novel CNN-Transformer hybrid model using wearable ECG sensor data can accurately classify tetanus severity, potentially aiding clinical decision-making and resource allocation in low- and middle-income countries.

Limitations

Small tetanus dataset size
Classification was made using ECG data recorded on both day 1 and day 5, rather than predicting day 5 severity from day 1 data alone
Potential confounding if patients suffer from pre-existing heart diseases
Use of ventilators in severe tetanus patients may influence heart rate variability

Abstract

Tetanus is a life-threatening infectious disease, which is still common in low- and middle-income countries, including in Vietnam. This disease is characterized by muscle spasm and in severe cases is complicated by autonomic dysfunction. Ideally continuous vital sign monitoring using bedside monitors allows the prompt detection of the onset of autonomic nervous system dysfunction or avoiding rapid deterioration. Detection can be improved using heart rate variability analysis from ECG signals. Recently, characteristic ECG and heart rate variability features have been shown to be of value in classifying tetanus severity. However, conventional manual analysis of ECG is time-consuming. The traditional convolutional neural network (CNN) has limitations in extracting the global context information, due to its fixed-sized kernel filters. In this work, we propose a novel hybrid CNN-Transformer model to automatically classify tetanus severity using tetanus monitoring from low-cost wearable sensors. This model can capture the local features from the CNN and the global features from the Transformer. The time series imaging - spectrogram - is transformed from one-dimensional ECG signal and input to the proposed model. The CNN-Transformer model outperforms state-of-the-art methods in tetanus classification, achieves results with a F1 score of 0.82±0.03, precision of 0.94±0.03, recall of 0.73±0.07, specificity of 0.97±0.02, accuracy of 0.88±0.01 and AUC of 0.85±0.03. In addition, we found that Random Forest with enough manually selected features can be comparable with the proposed CNN-Transformer model.

Ask AI

Helpful

Bookmark

View Full Paper