AI Prescribes Better Treatment than Doctors


As America stares down the needle of an empty syringe called Healthcare, we are realizing with more tangible worry everyday that something must be done quickly to solve our medical woes. President Obama has offered solutions such as the Patient Protection and Affordable Care Act, but we are all well aware now of the cracks in the already unsteady foundation of Obamacare. It doesn’t seem like any viable solution is available, so maybe the answer is to get creative, or even artificial.

In the spirit of bringing the singularity even closer to fruition, a study has found that artificial intelligence may be the answer to the issue at the forefront of political and social debate. Artificial intelligence (AI), aka the intelligence that will one day rule over us with the sweet aroma of logic and rationality (I choose to welcome our robotic overlords), has a history of impressing its human creators, even if one of the most powerful AIs in the world, IBM’s Watson, has a potty mouth.

Don’t let the word ‘artificial’ fool you: AI is smart, so smart that it has recently been shown to outperform doctors at their very job description; prescribing proper treatments. Using a combination of AI designs, namely Markov Decision Processes and Dynamic Decision Networks, researchers Casey Bennett and Kris Hauser from the Indiana University School of Informatics and Computing have revealed that AI can consistently prescribe better treatment than doctors and dramatically save on healthcare costs.

The study considered 6700 patients, and randomly chose 500 from the group.  After comparing the results of AI doctors and human doctors, the researchers found that the AI improved patient outcomes by close to 50%, while reducing overall required healthcare costs from an average of $497 to $189, a difference of more than 50%.  We are talking about a 50% improvement in care, and a 50% reduction in costs; these are revolutionary results.

The specific AI that the researchers used in their study was able to think like a doctor.

By using a new framework that employs sequential decision-making, the previous single-decision research can be expanded into models that simulate numerous alternative treatment paths out into the future; maintain beliefs about patient health status over time even when measurements are unavailable or uncertain; and continually plan/re-plan as new information becomes available.

watson MD

The AI can think exactly like a doctor, the difference being that it can do it faster, and can review a patient’s entire history in the blink of an eye. It can also instantaneously factor in new information and compare that information with known and potentially unknown variables. The AI is faster, smarter, has a better memory, costs less, and doesn’t ever yell at its wife.  Moreover, while doctors must spend decades in school specializing in a specific field, AI is a highly trained doctor from birth, and does not require specialization to function optimally in all fields. Keep in mind, if new, relevant data becomes available, all it takes is a momentary upload and the AI has already integrated the new information into every aspect of its being.

Most medical decision made by doctors are based on individual, experience based-approaches, including using intuition. The researchers suggest that in the majority of cases, modeling, rather than case-by-case decision making, is a better solution in every way. The researchers not that:

Modeling lets us see more possibilities out to a further point, which is something that is hard for a doctor to do. They just don’t have all of that information available to them.


AI has a wealth of resources and computation speed at its disposal.  The researchers believe that

using the growing availability of electronic health records, health information exchanges, large public biomedical databases and machine learning algorithms…the approach could serve as the basis for personalized treatment through integration of diverse, large-scale data passed along to clinicians at the time of decision-making for each patient.

Keep in mind that the researchers are not insinuating a total removal of humans from medical professions, rather

even with the development of new AI techniques that can approximate or even surpass human decision-making performance, we believe that the most effective long-term path could be combining artificial intelligence with human clinicians. Let humans do what they do well, and let machines do what they do well. In the end, we may maximize the potential of both.

We are talking about the seamless and lightning fast integration of all medical knowledge and inquiry around the globe.  Instead of being treated by a single doctor, wouldn’t you rather be treated by the collective knowledge and understanding of every doctor that has ever existed?

If you are interested in additional reading regarding AI being used in healthcare, IBM has done extensive research into the matter using Watson.  Using AI like Watson to improve healthcare is becoming an exponentially growing potential.


Wondergressive: ObamaCare and the 49-Employee Company

Wondergressive: The Singularity is Nigh Upon Us

Fortune: Teaching IBM’s Watson the Meaning of OMG

Markov Decision Processes

Dynamic Decision Networks

Indiana University: Can Computers Save Healthcare? 

IBM: Watson and Healthcare