Decision support systems prove vital to healthcare
Doctors on TV might get the diagnosis wrong a few times to fill up the broadcast hour, but in real life, things go much better if the doctor is right the first time. Information systems that bring medical and patient information to physicians to support diagnostic and treatment decisions help improve patient safety and manage costs.
Isabel is a diagnosis decision support and knowledge mobilizing system from Isabel Healthcare. The company was co-founded by Jason Maude after his daughter Isabel nearly died from secondary infections that followed a bout of chicken pox in 1999. Initially misdiagnosed at a local hospital in England, the correct diagnosis of toxic shock syndrome and necrotizing fasciitis, the "flesh-eating" bacteria, was made by Dr. Joseph Britto at St. Mary's Hospital in London. Britto and the Maude family established a foundation to help prevent such misdiagnoses in the future, and Britto led the development and validation of the Isabel system as a commercial product.
"Biomedical knowledge is increasing exponentially," says Britto. "It is simply not possible for doctors to carry in memory all the knowledge they need." Given a patient's symptoms, signs and results of investigations, Isabel is designed to help the clinician consider all the likely diagnoses and to make the latest medical information readily available at the point of care.
The first step in developing the Isabel decision support system was to create a taxonomy of more than 10,000 diseases and 4,000 drugs. Content for the system is stored on a server and consists of the latest medical textbooks, journals and other articles, which are indexed and searched by Autonomy's natural language processing software.
"When the healthcare professional enters the clinical features of the patient, Isabel instantly provides the user with a checklist of likely diagnoses," says Britto, "and with diagnosis-specific knowledge from medical textbooks and journals on tests to be performed, treatment options, lessons learned and recent advances."
Numerous hospitals and physicians throughout the United States use Isabel to suggest diagnoses. "Isabel does not provide clinical probability of a diagnosis in a given patient," emphasizes Britto. "Its great value is in presenting likely options, so that physicians are able to consider all the possibilities before coming to a conclusion."
At an annual subscription starting from $500 for an individual physician or $180 per hospital bed per year, Isabel is within range of many providers. In contrast to the clinical diagnosis decision support expert systems of several decades ago, users do not have to respond to a long series of questions about the patient. The short turnaround time benefits doctors and patients alike, providing a quick path to improved diagnoses, thereby enhancing patient safety and quality of care.
Another way to provide decision support to physicians is by analyzing data about clinical outcomes. The New York City Health and Hospitals Corp. (HHC) aggregates data from its electronic medical records (EMRs) to provide valuable insights as to the use and effectiveness of various procedures and medications. HHC oversees all the public medical facilities in New York City, including 11 hospitals, four long-term care facilities and more than 80 community-based clinics and programs. It serves 1.3 million individuals per year, 60 percent of whom are on Medicaid.
HHC was an early adopter of EMRs, implementing Misys CPR from Misys about 15 years ago. Misys CPR contains data from multiple departments and presents a comprehensive view for each patient.
"The use of an EMR lets any of the patient's providers see data that relates to allergies or other potentially harmful conditions," says Louis Capponi, chief medical informatics officer at HHC. More recently, HHC began implementing the Misys Data Warehouse, which is populated with data from the transactional system.
The clinical data is aggregated and analyzed to determine the degree to which HHC is meeting its goals for healthcare delivery. "One of our most exciting projects relates to patients with chronic illnesses," Capponi says. "We have a registry of about 48,000 diabetics, and we are now able to easily identify our sickest patients, make sure they are getting appropriate care and monitor their progress." Blood sugar levels, blood pressure and other clinical measurements are among the indicators used. The aggregated data for patients in HHC's registries is provided to clinical leaders as well as to care teams so that they can evaluate population outcomes. Care teams vary in the way they utilize the registry; some interacting with it at the point of care, and others producing reports at the department level and forwarding relevant information to the physicians. HHC's experience in analyzing