I'm not sure how ECG relates, how would ECG predict heart disease risk? I don't see any evidence of that, it's not even mentioned in the article you linked, which is odd considering the whole approach of clustering is gathering as many risk factors and relevant test results as are available.