Drug development is an expensive, time-consuming, and often unsuccessful endeavor — developing new drugs can often take more than 10 years, cost more than $2 billion, and fail nearly 90% of the time. And these numbers have not been improving, which is especially worrisome for people suffering from life-threatening diseases and patients with chronic conditions for which there are no good therapies or treatments. Researchers hope artificial intelligence (AI) can help speed up the process, create greater efficiency, and make the process less expensive. Take a look at four ways AI could do that.
AI can find new and targeted drugs more quickly. The discovery of new drugs is about predictions — that is, how well a drug may work for a specific disease and how individual persons who have the disease might benefit. AI can help fine-tune these predictions. AI might also help anticipate how new drugs, especially new drug structures, might function. And since drugs are often used in combination — most people take more than one medicine — learning how a new drug and existing drugs might interact is critical for researchers to see if it is worthwhile to pursue various phases of clinical trials.
AI can use the power of computing to tap into large databases. AI can look at associations and patterns more effectively than human researchers to find new indications for existing drugs. For example, the drug Viagra was being studied for hypertension and heart disease when researchers noticed men who were taking it were having erections. In another example, researchers noticed that the drug Latisse, studied to treat high eye pressure and glaucoma, also caused eye lashes to grow and become thicker. AI could help find those patterns sooner so that researchers don’t have to rely on chance.
AI can improve clinical trial design, recruitment, and participation. Drug studies have large dropout rates and often don’t enroll enough women and minorities. Some trials never get started or completed because researchers lack enough participants. AI can help identify the reasons for dropout or lack of enrollment. That information could help refine a clinical trial’s process and perhaps require fewer participants. This is especially important for rare disease studies that are limited by a small patient population available for well-designed trials.
AI could provide earlier warnings of serious side effects. Every drug has risks and benefits associated with taking them. Sometimes serious side effects are not discovered until a drug has been on the market for a few years. In addition, the FDA sometimes requires additional safety studies after a drug is approved, which costs money. AI might help find these potential safety problems earlier, prevent harm to patients, and save time and money.
Finally, know that AI will not replace people with computers and robots, but it may change the type of work humans do. We always will need scientists, researchers heading up clinical trials, regulators, pharmacology experts, and patients to help develop the safest and most effective drugs.