Using Machine Learning to Predict Cardiovascular Disease

By Aarushi Bellani

Cardiovascular diseases are the leading cause of death globally today (WHO) and more research and resources are being devoted to creating methods to predict and prevent them. Apart from diagnosing cardiovascular diseases accurately, assessing the impact of different risk factors to predict susceptible individuals is being explored. This has opened up an avenue for artificial intelligence to do the same.

An increasing influence of artificial intelligence has been seen in the field of medicine in recent times, especially through the use of machine learning algorithms to predict disease. Machine learning (ML) algorithms essentially use a given input of factors and associated consequences to form patterns which are used to predict consequences for a completely new set of factors. This concept is used to create models built on thousands of patient data entries which are used to understand disease susceptibility in data collected from other individuals.

A study conducted on over 400,000 biobank samples from the United Kingdom showed that an ML based model was able to predict cardiovascular disease with a greater accuracy than traditional models that are based on conventional risk factors (Alaa et al., 2019). The model was devised using AutoPrognosis which is an algorithmic tool that selects and tunes an ML modeling pipeline automatically and includes factors that are not usually considered in existing risk prediction models, such as the individuals’ usual walking pace. Along with introducing more risk factors, the model was also able to enhance predicting cardiovascular disease in relevant sub-populations such as those with a history of diabetes. 

Research in this field has shown that an ensemble of ML models might also be an approach to diagnosing these diseases early and precisely. As the American population is greatly affected by diabetes and cardiovascular disease, data from the National Health and Nutrition Examination Survey (NHANES) was used to determine risk factors that were used as variables in the various ML models that were tested (Dinh et al., 2019). A weighted ensemble model was then created combining all the other singular ones, which was found to be more robust and accurate. 

From the many types, coronary artery disease (CAD) is one of the most common cardiovascular diseases and if diagnosed early on can provide opportunity for appropriate treatment to prevent mortality. In a study conducted by Abdar et al. data mining methods such as Support Vector Machines (SVMs) and Neural Networks (NNs) coupled with machine learning techniques were used to create a cost and time effective, precise diagnosis method for CAD (Abdar et al., 2019). 

Since traditional models of risk prediction have worked well, using ML to enhance accuracy has only benefited the field of diagnosis. Most mortalities related to cardiovascular diseases occur in developing countries and so mechanisms such as these will help combat this issue well. Further research proposes developing a universal prediction system would be taking it one step further and would provide a standardized tool for physicians to use (Maini et al., 2019). 

References:

Cardiovascular Diseases . Available from: https://www.who.int/health-topics/cardiovascular-diseases/#tab=tab_1 [Accessed 20th February 2021].

Abdar, M., Książek, W., Acharya, U. R., Tan, R., Makarenkov, V. & Pławiak, P. (2019) A new machine learning technique for an accurate diagnosis of coronary artery disease. Computer Methods and Programs in Biomedicine. 179 104992. Available from: doi: https://doi.org/10.1016/j.cmpb.2019.104992.

Alaa, A. M., Bolton, T., Di Angelantonio, E., Rudd, J. H. F. & van der Schaar, M. (2019) Cardiovascular disease risk prediction using automated machine learning: A prospective study of 423,604 UK Biobank participants. Plos One. 14 (5), e0213653. Available from: https://doi.org/10.1371/journal.pone.0213653.

Dinh, A., Miertschin, S., Young, A. & Mohanty, S. D. (2019) A data-driven approach to predicting diabetes and cardiovascular disease with machine learning. BMC Medical Informatics and Decision Making. 19 (1), 211. Available from: https://doi.org/10.1186/s12911-019-0918-5

Maini, E., Venkateswarlu, B. & Gupta, A. (2019) Applying Machine Learning Algorithms to Develop a Universal Cardiovascular Disease Prediction System. In: Hemanth, J., Fernando, X., Lafata, P. & Baig, Z. (eds.) International Conference on Intelligent Data Communication Technologies and Internet of Things (ICICI) 20. 18, Cham, Springer International Publishing. pp.627-632.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s