Can a Computer Substitute Your Doctor? IBM's Watson Almost Can

Impact

Doctors in the United States are responsible for keeping up-to-date with the newest innovations in health care. However, it is nearly impossible to do this as new treatments are constantly being created and best practice guidelines are constantly being revised. The New England Journal of Medicine found that only 54.9% of patients receive the recommended care and the Archives of Internal Medicine found that an increase in years of practice leads to a decrease in quality of care. This is where IBM's Watson comes into play. Watson is a computer designed to think like a doctor and will soon be used across the country in hospitals, family practices, and insurance companies. Watson, programmed by complex computer algorithms, is able to tell doctors which preventative care measures to suggest, which diagnosis is most likely to be correct for the given symptoms, and which treatment options to consider given the diagnosis and patient history.

Watson was first developed by a team within IBM who wanted to create a computer that could win a game of Jeopardy! By perfecting a question and answer technology, Watson was successful at winning in 2011 and the team took on bigger challenges. They began feeding Watson medical information and perfected his question and answer technology so that Watson could answer more complicated questions, sometimes even offering multiple solutions. Doctors struggle to remember preventative care measures, stay up-to-date with evidence-based care, and they often forget details of certain disease and ailments they haven't seen for a while. Watson has none of these problems. For instance, remembering whether a women is supposed to start getting pap smears at age 20 or 21 and if its every year or every other year depending on her health history becomes a lot to remember when you have 20 different preventative care tests to think about with recommendations changing every few years. Watson has no problem remembering this information and changing as the new recommendations come out. Along the same lines, knowing what disease to suspect and how urgently to treat it is often considered an art, based on taking into consideration what the patient looks like, what they are telling you, their history and their family history, but using real probabilities for research based evidence is often better. This is because doctors skew their practice of medicine based on things other than science, like how sincere they felt the patient was when speaking about pain, what the last journal article they read was on, a case that they saw last week with similar symptoms. A computer, does not have this problem, but simply uses hard science to filter out the most likely diagnosis.

Watson will also soon be helping insurance companies. Insurance companies approve payments based on whether they believe patients are getting the most reasonable cost effective treatment. A team of doctors and nurses decide whether or not proper care was given and then decide if the insurance company will pay for it based on a person's plan. This process takes weeks and leaves many patients waiting before they can receive proper care. Watson would take seconds to decide this, and use statistical methods to immediately decide reimbursement. The terrible process of patients waiting for their procedures approved and the expensive cost of hiring doctors and nurses to review cases would be gone.

Watson is the future of our health care both in treating and administration. Time and time again we have seen that the roles of doctors have become overly complicated, and although doctors are intelligent and well trained, they make mistakes. While Watson cannot replace doctors all together, Watson can reduce error by presenting doctors with the most reasonable approaches to patient care and allow the doctor to make the final decisions. Watson can have the same impact on the insurance end of health care by taking over the role of the doctor and quickly determine if patients are receiving proper care.