Machine learning from electronic health records in primary care identifies missed people with heart failure
What is the aim?
A Belgian study aimed to address the issue of over- and underdiagnosis in primary care by assessing the misclassification of heart failure and testing the performance of a machine learning algorithm using electronic health records.
What has been achieved?
The algorithm improved the identification of people with heart failure. First, it was trained using data from electronic health records. It then identified people with heart failure using routine clinical data, such as heart failure risk factors, signs, symptoms and medications. The algorithm showed that almost half of people with a registered heart failure diagnosis did not have the syndrome, and more than two thirds of people with heart failure did not have a registered diagnosis.
References
Raat W, Smeets M, Henrard S, et al. 2022. Machine learning optimization of an electronic health record audit for heart failure in primary care. ESC Heart Fail 9(1): 39-47