0 
QHHL 
How do doctors predict cardiovascular disease? Well, that blood work that goes with the annual wellness monitoring check-up has a lot to do with it. The blood work, along with a check of blood pressure and medical history, are used to determine the risk of cardiovascular disease according to the guidelines published by the American College of Cardiology and the American Heart Association (ACC/AHA). Doctors traditionally use these guidelines when completing health risk assessments of cardiovascular disease for their patients. The primary factors included in the guidelines are age, gender, total cholesterol level, high-density lipoprotein cholesterol, systolic blood pressure, use of blood pressure medication, diabetes, and current smoking.
How accurate are these guidelines in predicting cardiovascular disease? Well, even when these guidelines are used, it can be difficult for doctors to predict the development of active cardiovascular disease. However, a new study shows that artificial intelligence (AI) can help to improve the accuracy of predicting the development of cardiovascular disease. The results of this study could change the way doctors test risk factors in the future.
What the study says
This study, published in PLOS ONE in April 2017, examined the ability of AI to predict cardiovascular events based on data from the electronic medical records of 378,256 patients in the United Kingdom. Four machine-learning algorithms were used in the study. The algorithms first “trained themselves” by processing 78% of the records and building their own internal guidelines for predicting cardiovascular events. The algorithms then tested the guidelines against the remaining medical records by identifying cardiovascular events in the records that occurred between 2005 and 2015. The results of these tests were compared to the predictive results of the traditional ACC/AHA guidelines on the same set of records. A statistic called Area Under the Curve, or AUC, was used to compare the different sets of guidelines. An AUC score of 1.0 signifies 100% accuracy.
The results of the study showed that all four AI algorithms were significantly better at predicting cardiovascular events than the ACC/AHA guidelines used by doctors. The ACC/AHA guidelines achieved an AUC score of 0.728, while the four AI algorithms’ AUC scores ranged from 0.745 to 0.764. The best algorithm predicted 7.6% more events and 1.6% fewer false alarms than the ACC/AHA guidelines. Out of the sample of about 83,000 records, this percentage would have amounted to 355 additional patients whose lives might have been saved, according to researchers.
The AI algorithms identified 22 additional data points not included in the ACC/AHA guidelines, including ethnicity, arthritis, and kidney disease among others. Several of the strongest predictors of cardiovascular disease identified by the algorithms, including severe mental illness and taking oral corticosteroids, are not listed in the ACC/AHA guidelines. Diabetes, which is included in the ACC/AHA guidelines, was not considered a strong predictor of cardiovascular disease by any of the algorithms.
What this means for doctors and patients
The results of this study suggest that a much larger set of factors could be used by doctors to help predict the possibility of cardiovascular disease and that AI could be used to help analyze these factors. Using this larger set of factors would improve the accuracy of prediction and would help patients realize when they are truly at risk of developing cardiovascular disease. Patients who can see this risk based on the larger number of risk factors would be more likely to take steps to prevent cardiovascular disease, including lifestyle changes and regular health and wellness tracking.
An additional potential benefit to patients is that using AI to analyze risk factors for cardiovascular disease could eventually be adapted for health screening at home. Since AI is computer based, it would not be difficult for a person to enter data into an AI program to analyze risk for cardiovascular disease themselves. Equipment and questionnaires already included in online health measurement systems like HomeLab by Quantihealth could be used to collect and upload this data to AI programs. A person who already self-checks health using a system like HomeLab could easily generate his or her own risk of developing cardiovascular disease without having to visit a doctor. This system could be used to help people know when to change lifestyle based on their levels of risk and if those changes are effective over time. It could also be used to store data for doctors in case a cardiovascular event or the symptoms that precede an event occur. This type of system could change the way people monitor health at home for hidden disease.
Using a home health testing system like HomeLab by Quantihealth is a crucial step toward being ready for this kind of powerful health risk assessment. For more information about HomeLab by Quantihealth, visit www.thequantihealth.com.
Tags: health risk assessments, wellness monitoring, health and wellness tracking, health screening at home, hidden disease.
