Researchers have now found a way to use the computational abilities of AI in deciding the combination drug therapy for treating tuberculosis or TB. Mycobacterium tuberculosis is the causing bacteria of tuberculosis. Treating tuberculosis requires combination therapy because of the characteristics of TB bacteria. It differs in characteristics based on where or in which cells it is present. Also, Mycobacterium tuberculosis tends to produce resistance to the medication. Hence, based on the patient’s state and the stage of tuberculosis, the physician decides on a combination of drugs to administer to the patient. Tuberculosis is a serious disease. According to the data published by WHO, tuberculosis affected around 10 million people in 2020 alone. This included 3.3 million females, 5.6 million males, and 1.1 million pediatric patients. Tuberculosis also caused a death of around 1.5 million in 2020 alone. The region-wise data published by WHO suggests that the African region, Eastern Mediterranean Region, and South-East Asia Region have the highest incidence of tuberculosis. People who are suffering from HIV have the highest chances of getting ill because of tuberculosis. The regional incidence of tuberculosis reflects this fact, as the affected regions also have a higher incidence of HIV. The African region and South-East Asia Region also have the highest mortality rates when compared to the other regions. Deciding which combination of drugs to use for several drugs is difficult. For example, if there are 20 available drugs for treating tuberculosis and you have to use a combination of three or four drugs, this alone gives you over 6,000 various drug combinations. Deciding on a single combination that gives maximum output is nearly impossible. This is why, TB treatment involves a Directly observed therapy or DOT, where the patient is administered a drug combination and monitored to understand how well the administered combination is performing. Hence, researchers from Tufts University have found a way of doing this evaluation much more efficiently using AI. The researchers used the available data from earlier studies which included two-drug combinations of 12 different drugs. Researchers used this data to form a set of rules which must be followed when deciding on a three to four-drug combination therapy. When a two-drug combination is tested for its efficacy, it produces much more efficient results than directly testing out the three or four-drug combination therapy. As per the researchers from Tufts University, based on the assigned set of rules, one pair can be significantly tested with the other pair of the drug for their ability to treat tuberculosis. If the combination does not work, the pair can be then easily swapped with the next pair, and so on. As these operations are being performed using an AI, it has a much higher efficacy and success rate than the conventional methods of deciding on combination therapy for tuberculosis. Tufts University researchers said that the method they have developed is much more streamlined and accurate and uses a much less number of combinations to test out higher numbers of drugs. Hopefully, the new method can improve the available treatments for tuberculosis and help in reducing the mortality rates for TB.