Development of a deep learning model for prediction of cardiovascular disease
DOI:
https://doi.org/10.15282/daam.v6i1.12496Keywords:
Deep learning, Cardiovascular disease, Long short-term memory, Convolutional neural networks, Predictive modellingAbstract
15.1% of medically certified deaths in 2023 were due to ischemic heart disease (IHD), according to Department of Statistics Malaysia (DOSM) statistics on causes of death in Malaysia published in October 2024. Despite the slight decline, IHD remains a significant health concern in Malaysia, especially among males and individuals aged 41–59 years, where it accounted for 19.8% of deaths in that age group. Regular checks are one approach to preventing heart disease in its early stages; however, they can be expensive and time-consuming. With the advancement of technology, people can now conveniently check their blood pressure, heart rate, and electrocardiogram (ECG) using smartwatches. However, since some people lead busy lives and occasionally forget to track or monitor their health through the applications, monitoring alone is insufficient. The primary goal of this research was to develop a deep learning model for predicting cardiovascular disease (CVD) using data from smartwatches, which offer non-invasive and real-time health monitoring capabilities. The research employs two deep learning techniques: Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks. ECG and heart rate data were collected from 20 volunteers using local smartwatches supplemented with a publicly available dataset from Kaggle. Data pre-processing involved denoising ECG signals and normalising heart rate readings to ensure accuracy and reliability. The models were evaluated using precision, recall, F1-score, and accuracy metrics, achieving over 99 per cent across all measures. While both models demonstrated high predictive power, the LSTM model outperformed the CNN in computational efficiency, completing model training in 31 minutes compared to 87 minutes for the CNN. The study highlights the potential for wearable devices for real-time CVD monitoring and early diagnosis. Future work will explore the inclusion of additional data sources and advanced modelling involving ensemble techniques.
References
[1] Department of Statistics Malaysia (DOSM). Statistics on Causes of Death, Malaysia, 2024. Retrieved from https://www.dosm.gov.my/portal-main/release-content/statistics-on-causes-of-death-malaysia-2024.
[2] Grundy SM, Pasternak R, Greenland P, Smith S, Fuster V. Assessment of cardiovascular risk by use of multiple-risk-factor assessment equations: a statement for healthcare professionals from the American Heart Association and the American College of Cardiology. Journal of the American College of Cardiology. 1999;34(4):1348-59.
[3] Park K, Hwang SY. 10-year risk for cardiovascular disease among male workers in small-sized industries. Journal of Cardiovascular Nursing. 2015;30(3):267-73.
[4] Ahsan MM, Luna SA, Siddique Z. Machine-learning-based disease diagnosis: A comprehensive review. In Healthcare 2022 Mar 15 (Vol. 10, No. 3, p. 541). MDPI.
[5] Ahmad I, Alqurashi F. Early cancer detection using deep learning and medical imaging: A survey. Critical Reviews in Oncology/Hematology. 2024;204:104528.
[6] Parthiban G, Srivatsa SK. Applying machine learning methods in diagnosing heart disease for diabetic patients. International Journal of Applied Information Systems. 2012;3(7):25-30.
[7] Taye MM. Understanding of machine learning with deep learning: architectures, workflow, applications and future directions. Computers. 2023;12(5):91.
[8] Hochreiter S, Schmidhuber J. Long short-term memory. Neural computation. 1997;9(8):1735-80.
[9] Kim MJ. Building a cardiovascular disease prediction model for smartwatch users using machine learning: Based on the Korea national health and nutrition examination survey. Biosensors. 2021;11(7):228.
[10] Fawagreh K, Gaber MM, Elyan E. Random forests: from early developments to recent advancements. Systems Science & Control Engineering: An Open Access Journal. 2014;2(1):602-9.
[11] Lee HY, Shin J, Kim GH, Park S, Ihm SH, Kim HC, Kim KI, Kim JH, Lee JH, Park JM, Pyun WB. 2018 Korean Society of Hypertension Guidelines for the management of hypertension: part II-diagnosis and treatment of hypertension. Clinical hypertension. 2019;25(1):20.
[12] Martin A, Fitzsimons C, Jepson R, Saunders DH, van der Ploeg HP, Teixeira PJ, Gray CM, Mutrie N. Interventions with potential to reduce sedentary time in adults: systematic review and meta-analysis. British journal of sports medicine. 2015;49(16):1056-63.
[13] Rens N, Gandhi N, Mak J, Paul J, Bent D, Liu S, Savage D, Nielsen-Bowles H, Triggs D, Ata G, Talgo J. Activity data from wearables as an indicator of functional capacity in patients with cardiovascular disease. Plos one. 2021;16(3):e0247834.
[14] Kachuee M, Fazeli S, Sarrafzadeh M. Ecg heartbeat classification: A deep transferable representation. In 2018 IEEE international conference on healthcare informatics (ICHI) 2018 (pp. 443-444). IEEE.
[15] Albawi S, Mohammed TA, Al-Zawi S. Understanding of a convolutional neural network. 2017 International Conference on Engineering and Technology (ICET), Antalya, Turkey, 2017:1-6.
[16] Sowmya S, Jose D. Contemplate on ECG signals and classification of arrhythmia signals using CNN-LSTM deep learning model. Measurement: Sensors. 2022;24:100558.
[17] Mehmood A, Iqbal M, Mehmood Z, Irtaza A, Nawaz M, Nazir T, Masood M. Prediction of heart disease using deep convolutional neural networks. Arabian Journal for Science and Engineering. 2021;46(4):3409-22.
[18] Moulton S. Long-term health monitoring using wearable technology: Promises and pitfalls. Journal of Health Informatics. 2019:25(1), 28-32.
[19] Raghupathi W, Raghupathi V. Big data analytics in healthcare: promise and potential. Health information science and systems. 2014;2(1):3.
Downloads
Published
Issue
Section
License
Copyright (c) 2025 The Author(s)
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
