1. What concerns might a physician have about using a decision support system such as Isa-bel or Watson to make a medical diagnosis? How might those concerns be alleviated?

2. Is it possible that in a decade this type of technology could be easily accessible by laypeople who could then perform self-diagnosis, thus helping to reduce the cost of medical care?

3. Does the use of decision support systems to support healthcare decisions seem like an effective way to reduce healthcare costs? Why or why not?

Diagnosis errors (including missed, wrong, or delayed diagnosis) are a frequent and serious problem in the healthcare industry. It is estimated that such errors result in death or permanent injury for up to 160,000 U.S. patients each year. In a recent Johns Hopkins University study examining malpractice claims, researchers found that claim payments for diagnostic errors added up to $38.8 billion over the time period 1986 to 2010. Failure to fully diagnose a patient’s condition puts the patient at risk of suffering a recurrence of the problem—such as incurring further damage from another accident caused by, for example, an undiagnosed brain injury. Misdiagnosis of a patient’s condition can lead to costly, painful, potentially harmful, and inappropriate treatments. A delay in the diagnosis of a patient can allow an otherwise reversible condition to advance to the point that it is no longer treatable. Over the past decade, several decision support systems to aid in healthcare diagnosis have been developed, including DiagnosisPro® , DXPlain® , First Consult^©, PEPID, and Isabel^©. A decision support system is an interactive computer application that aids in decision making by gathering data from a wide range of sources and presenting that data in a way that aids in decision making. Isabel, one of the more advanced healthcare decision support systems, is a Web-based system developed in the United Kingdom. Isabel uses key facts from the patient’s history, physical exam, and laboratory findings to identify the most likely diagnosis based on pattern matches in the system’s database. The system can interface with electronic medical records systems to obtain patient data, or the data can be entered manually. Each diagnosis is linked to information in commonly used medical reference sources such as The 5 Minute Clinical Consult, Oxford Textbook of Medicine,and Medline—the U.S. National Laboratory of Medicine’s online bibliographic database. Isabel can also suggest bioterrorism agents that might be responsible for a patient’s symptoms, as well as identify drugs or drug combinations  that might be the cause. The cost of using Isabel ranges from a few thousand dollars for a family practice to as much as $400,000 for a health system. United Hospital, a large hospital in St. Paul, Minnesota, recently implemented the Isabel system to help physicians investigate and diagnose patient cases. The system will integrate directly with the hospital’s electronic medical record system and physicians will be able to access Isabel from mobile devices. On another front, medical researchers at Memorial Sloan-Kettering Cancer Center in New York are busy feeding data from medical textbooks and journals into IBM’s Watson super-computer to create a world-class healthcare diagnostic tool. Watson is the same supercomputer that gained recognition in 2011 for beating the world’s best players on the TV game show Jeopardy!. Watson is now being programmed to understand plain language so that it can absorb data about a patient’s symptoms and medical history, form a diagnosis, and suggest an appropriate course of treatment. When presented with a set of symptoms, Watson will be able to provide several diagnoses, ranked in order of its confidence. One incentive hospitals have to adopt such systems is concern that a failure to adopt new technology could subject the hospital to liability in cases where it could be shown that adoption of the technology would not have been overly costly and could have prevented patient injury.