Byte solution for the heart

even when computers are increasingly being used as major diagnostic tools, the role of physicians has not diminished. Computers are especially helpful in cases when complex analytical procedures have to be performed with small samples. But physicians are still required to diagnose, analyse results in conjunction with clinical symptoms and suggest suitable treatment to patients.

Now, researchers have devised for the first time a computer program that would assist doctors in interpreting the electrocardiograms (ecg) of patients. It has been found to work better than cardiologists in the diagnosis of heart attacks (Circulation , Vol 96, No 1798-802).

Lars Edenhardt and Bo Heden at the Department of Clinical Physiology and Cardiology, University Hospital, Lund, Sweden, have used the computer program with neural networks. It is a form of artificial intelligence that mimics human learning. The system can recognise electro-cardiographic patterns of ischaemic heart disease by exposing the machine to thousands of ecg tracings with matching case histories.

During trials, the computer performs as many ecg s in a short time as several cardiologists do in their life time. Then the crucial test was conducted to know who can do a better job -- the human or the machine. During 1990-95, when nearly 11,500 ecgs were screened, the machine performed better than the human. Out of the total ecg s, 1,120 belonged to patients with cardiac problems and 10,452 from normal people. The computer outscored cardiologists by nearly 10.5 per cent in correctly diagnosing myocardial infarction.

Correct diagnosis is especially difficult for cardiac problems when doctors are under considerable pressure. About 25 per cent of patients with acute myocardial infarction (ami) are either misjudged or overlooked and sent home in us hospitals. About 80 per cent of people admitted to coronary care units for suspected ami are discharged without having the diagnosis confirmed. According to doctors, even the conditions in India are not up to expectations.

Though the performance of the computer was far better than physicians in identifying subtle indications of ischaemia, there were a few strong points for doctors also. The doctors were found to be better in recognising clear cut changes of acute vascular blockages. The researchers conclude that the computer and artificial intelligence would have a prominent role in the diagnosis of heart disease. Especially when nearly 25 per cent of ecgs are misinterpreted by doctors.

But despite the computer's skills in interpreting ecgs, it can only assist cardiologists in taking rational decisions. Edenhardt says ecg reading is merely one part of the diagnostic exercise. There are other tests that are required to diagnose heart disease. It is especially important to know the case history. Only a doctor can get such informations by talking to patients about symptoms and past medical history.

Subscribe to Daily Newsletter :

SUPPORT US

We are a voice to you; you have been a support to us. Together we build journalism that is independent, credible and fearless. You can further help us by making a donation. This will mean a lot for our ability to bring you news, perspectives and analysis from the ground so that we can make change together.