Decision support systems prove vital to healthcare

information contained in a transactional system will help as it begins participating in emerging regional health information organizations (RHIOs), according to Capponi. "Most of the RHIOs will be using federated systems with normalized data at the periphery," he says. In other words, data will remain in its original repository and will be accessed by tools that standardize it for analysis. Interoperability will be a key factor in the ability of RHIOs to provide medical information outside their own networks.

Misys Data Warehouse is built on the Oracle 10g platform and uses Oracle Discoverer for query, reporting and analysis. Because the warehouse is integrated with Misys CPR, configuring the system to select and analyze information contained in Misys CPR is straightforward. Other business intelligence products can also be used with the Misys Data Warehouse product.

"Use of the analytical capability is a good way for doctors to check out their hypotheses," says Gary Baluta, Misys product manager of enterprise services. "They may already have anecdotal evidence, but using a reporting solution allows them to validate their ideas and move to evidence-based decisions."

RHIOs

Regional health information organizations aim to improve patient health and safety and increase cost-effective use of health resources by sharing information among a coalition of providers, payers, employers and other stakeholders. The Department of Health and Human Services views RHIOs as a central component in developing a national health information network (NHIN).

In central Appalachia, the CareSpark RHIO was established to foster improved health outcomes and more effective use of resources. Covering 17 counties in three states, Care- Spark serves about 750,000 individuals. The region's population has a high rate of chronic conditions such as heart disease, lung disease and cancer. Some residents live in areas remote from medical facilities, and household income is low relative to other regions of the country. A high percentage of the area's residents are uninsured.

CareSpark was in the process of developing its strategic business plan when it came across the CareEngine System, a clinical decision support solution from ActiveHealth Management, a clinically-based, technology-driven health management services company.

"We wanted to find a mechanism for applying best practices," says Liesa Jo Jenkins, executive director of CareSpark, "but we did not think there was a product available that matched our goals. We planned to start with simpler steps such as providing a list of patient medications and then encourage the adoption of clinical best practices."

What appealed to CareSpark leaders was the opportunity to leapfrog some of the earlier steps and start combining point-of-care guidance with patient-specific data. After discussing the use of CareEngine with its stakeholders, CareSpark decided to move forward with implementation. One major employer in the region has already signed on for its 28,000 employees and dependents as of the first quarter of 2007. Jenkins expects the idea to catch on and draw additional participants as the value of the system is demonstrated.

ActiveHealth's founder, Dr. Lonny Reisman, saw a need for a product that would relate clinical data for patients to evidence-based standards on an ongoing basis. He also recognized the difficulty of keeping up with literature in the field. CareEngine is a decision support system designed to bring together clinical data with best practices.

"CareEngine is constantly accessing clinical data on our 13 million members, and updates its evaluation each time a patient has an encounter with the medical system," says Reisman. "As each piece of data is entered, the cumulative record is exposed to thousands of clinical rules relating to safety and evidence-based practice."

As the literature changes, new guidelines are entered into the system. Information is pushed to doctors automatically, but rather than issue general alerts about products or treatment, the system points to specific patients to which the issue applies and delivers the message to the patient's EMR. Physicians are then contacted with an alert called a Care Consideration that provides the rationale for the alert and the citation from which the alert was derived.

Pain-free SOA

The University of Texas M. D. Anderson Cancer Center is the largest specialized facility for cancer care and research in the United States. Its combined missions of patient care and clinical research posed particular challenges in identifying an appropriate EMR system. After searching unsuccessfully for a commercial EMR that was designed to accommodate data from both settings, M. D. Anderson decided to build its own system.

In 2001, M. D. Anderson deployed ClinicStation, a custom-built application written in Visual Basic. ClinicStation provided the required functionality, accessing information from more than 40 separate databases that included clinical care and research data. However, when Microsoft announced last year that it was no longer supporting Visual Basic, Dr. Lynn Vogel, VP and CIO at M. D. Anderson, began a review of options for re-architecting the system.

Choosing to stay with the Microsoft platform, Vogel determined that starting from scratch with .NET was more desirable than attempting to convert Visual Basic code into C#, the programming language used by .NET. Because his department had limited experience with .NET, Vogel sought out Avanade, a Microsoft applications developer.

"We were already moving toward a service-oriented architecture [SOA]," says Vogel, "because we wanted the flexibility for each department to retain its own data models, but also to view all the patient data through one interface." The first release of the new system is scheduled for March.

Personalized medicine is a major issue in healthcare today, and is particularly important in cancer care. "Certain kinds of genetic patterns may make an individual more susceptible to particular diseases," observes Vogel. "So a physician may want to test a patient's DNA, for decisions about both diagnosis and treatment."

Vogel predicts that within the next five years, such custom-tailored patient treatment will become standard in medical care. As a result, decisions about treatment will be based not on group data but on specific factors related to individual patients. Information systems will need to support a much greater degree of granularity in data, and will be expected to bring together a much more comprehensive set of information for each individual patient.

M. D. Anderson's broad range of activities makes ClinicStation a potential model for the healthcare industry as a whole, according to Kevin Kelley, regional CTO for Avanade. "ClinicStation covers every aspect of patient care, and it also has standard-based security, privacy and compliance built in," he says. "It was literally designed for change, which will let the system adapt easily as new 14 requirements emerge."