Betterhumans > News > Artificial Intelligence Could Reverse HIV Treatment Failure. Imagine being able to mathematically model an HIV-infected patient's "viral indicators" long-term. According to the above article, a group headed by Julio Montaner at the University of British Columbia has developed a neural network to create a dynamic model (using artificial intelligence) that predicts a person's projected viral load and matches the numbers with a particular drug regimen. This neural network, according to the group, was able to predict with 87% accuracy (within the narrowest time frame of 8 weeks prior to the treatment change) which drug regimen could have been used to reduce viral load. Of course, the study was performed using data from a treatment cohort and the neural network was not at all implicated in decision making for drug therapy. However, if the model holds true, perhaps it can be used in future decision-making processes to evaluate drug therapy changes. Imagine that, in the future, you sit in the doctor's office and s/he punches in your numbers and the computer spits out what drugs you should be taking and when you should be taking them. I'm sure it's probably already done to a certain degree in the clinic.
Posted by johnvu at June 18, 2003 12:32 AM