The Hindu : And now, artificial intelligence as a medical tool

Online edition of India's National Newspaper
Monday, Jun 09, 2003

Southern States

Southern States - Andhra Pradesh-Hyderabad

And now, artificial intelligence as a medical tool

By Aniket Alam

HYDERABAD June 8. In a few small rooms in a non-descript commercial building in Secunderabad, a group of scientists and entrepreneurs have been busy for the last three years developing an Artificial Intelligence tool which would provide doctors a `decision support system' in making diagnosis fool proof. A similar venture in the US took over 25 years, 70 million dollars in funding and nearly 30 Ph.Ds to complete it.

Impressed by the progress made so far, the Technology Information, Forecasting and Assessment Council (TIFAC) of the Department of Science and Technology (DST), Government of India, has tied up with Karishma Software to help them `scale up' the technology in collaboration with leading Government scientific institutions.

Sajid Mubashir, who heads the `Home Grown Technology' division of DST, terms the work done by Karishma `risky but exciting' and says that their involvement would be to enable this project to become commercially viable. He feels that it will be three more years before the product is on the market, but says that it has the potential of changing the way doctors work.

There are reasons behind Dr. Mubashir's confidence.

Every time a doctor is confronted with a sick person, he has to guess at the possible disease by working through the `symptoms' that the disease `expresses' in the patient. There are over 19,000 symptoms which have over 56,000 discrete `expressions' for the 620 internal medicine diseases known as of now.

In layman's language, this means that every sick person's symptoms can indicate many different diseases. For example, combined symptoms of fever, cough and body ache can indicate well over 100 diseases. It is a set of geographical, epidemological and social factors which, along with the famous "doctor's intuition,'' lead the physician to suspect one particular disease over another. This is further verified through tests. Failure to catch one crucial symptom or forgetting one rare disease probabilities can often have tragic consequences.

It is here that the new Artificial Intelligence tool being developed promises to help by taking out the guesswork involved.

The proposed software is based on a mathematical tool called the Bayesian Probabilistic Belief Networks developed over the last few decades at MIT, Stanford, UCLA and Berkeley in the USA. It has powered advanced AI equipment like the Mars Rover as well as simple things like the predictive `Help' of MS Word.

Bapi Raju Surampudi, of the Computer Science department of the University of Hyderabad, calls the Bayesian framework `elegant and robust' for tackling such problems and says that he was `impressed' by the work done by Dr. Mohan Rao and his team.

Dr. Bapi Raju lists out two main problems for developing a `decision support system' for medical diagnosis. One, the problems of `data mining' and two, adapting the mathematical tool into a practical software. This would require extensive medical knowledge combined with smart software skills.

If this is successful, the final product would not only list out all the known diseases which match the given set of symptoms, but also provide the percentage probability of its occurrence along with the reasons for that. It will tell the doctor not only what the probable diseases are, but will also explain why one disease has 28 per cent probability of occurrence and another only 2 per cent!

A.M. Mohan Rao, a US returned medical doctor, who heads the team at Karishma Software, says that they have followed the International Classification of Diseases and all data regarding diseases and their symptoms are taken from standard medical textbooks. It is now necessary to include geographical, climatic and epidemological nuances to its diagnosis software, he adds, as disease patterns and probabilities differ with region, season and social factors.

V. Giridharan, Managing Director, Karishma Software, who has been nurturing the project from its inception, says that this software's USP would be its ability to understand `natural language' and be compatible with any regular PC on the market. It would provide `back up' to doctors, help standardise medical practice as well as help medical education.

Printer friendly page
Send this article to Friends by E-Mail

Southern States

Stories in this Section

Two Himachal women with forged visas held

And now, artificial intelligence as a medical tool

Archakas threaten stir over demands

Maintain communal amity, says CPI

Panel for training Cong. workers