STANFORD, Calif.-- Patients' advocates, healthcare leaders and policymakers are increasingly advocating the use of computer-based systems that promise to improve patient care by giving physicians on-the-spot medical information and treatment recommendations based on the latest scientific evidence. While several such systems have been developed at U.S. hospitals and clinics, studies evaluating some of the systems aimed at chronic disease have found that physicians used them only a small fraction of the time -- typically less than 5 percent. This may be because some of the systems were cumbersome, required extra work (such as requiring clinicians to enter data twice), or were simply viewed by physicians as not very helpful.

At Stanford Medical Center and the VA Palo Alto Health Care System, however, a system designed by physicians and medical informatics developers to improve hypertension care appears to have succeeded where others have struggled. According to a study published in the September/October issue of the Journal of the American Medical Informatics Association, physicians who took part in a randomized controlled test of the system, known as ATHENA, used it with relatively high frequency and found it helpful.

The ATHENA "decision-support" system was rolled out to 91 physicians and other clinicians at nine VA clinics in northern California and Durham, N.C., as part of a clinical trial comparing it with a simple reminder to the clinicians about patients' hypertension. Over the 15-month test period, the clinicians interacted with ATHENA for 63 percent of their eligible patients. For example, clinicians entered new blood pressure results into the system and obtained updated treatment recommendations for 43 percent of the patients. Overall, the clinicians interacted with ATHENA at 54 percent of the patient visits.

Compared with reports of other computer-based systems for managing chronic disease, "these numbers are quite high; it's very encouraging," said Mary Goldstein, MD, MS, a core faculty member at Stanford's Center for Primary Care and Outcomes Research (PCOR), an associate professor of medicine at the VA Palo Alto, and lead investigator for the ATHENA project. "We've shown that this type of system is feasible, it provides helpful information aimed at improving quality of care, and doctors will actually use it." ATHENA's impact on patients' blood pressure results is being evaluated.

While ATHENA is easy to use and requires little extra work for physicians, Goldstein explained, it is also highly sophisticated, relying on a complex set of rules and tailoring its recommendations to each individual patient. ATHENA's recommendations and other information are contained in a pop-up window that appears on the physician's computer screen when he clicks on a patient's electronic file at each office visit. The system pulls information from the patient's electronic medical record, analyzes the information in accordance with nationally recognized guidelines for hypertension management, and generates treatment recommendations based on the patient's latest blood pressure results, lab-test results, medical history, allergies and response to previous treatments.

ATHENA's recommendations could include increasing the dose of a medication, prescribing a new medication, or ordering a lab test. It also includes information on counseling patients about diet and exercise. Since ATHENA is connected to a clinic's pharmacy system, it indicates when a patient has not filled a prescription. It also generates easy-to-interpret graphs, such as one that shows a patient's blood pressures over time, with the antihypertensive medications prescribed for him during the same time frame, so the physician can see how the patient has responded to different drugs.

Significantly, ATHENA provides brief explanations of why a particular therapy is being recommended, along with links to supporting material from medical journals and best-practice guidelines. "Doctors are more willing to accept recommendations if they're backed up by the scientific literature," Goldstein explained. Another advantage: ATHENA is not tied to any particular electronic medical record system, but can potentially be integrated with a variety of EMR systems.

Based on these pluses and ATHENA's promising pilot-test results, the system has attracted interest from several non-VA clinics and hospitals, though it is not being offered commercially. Plans call for expanding ATHENA to incorporate other diseases, beginning with diabetes.

"There are all sorts of exciting applications for this down the road," Goldstein said. "In the future, I believe systems like ATHENA will become the standard of care, as the technology improves and more healthcare providers convert to electronic medical records."

ATHENA's successful implementation, Goldstein believes, is due largely to the project's collaborative approach, involving individuals from several fields including medical informatics developers, clinicians and healthcare administrators. The team worked closely with physicians, including hypertension experts, to ensure that the final product would be user-friendly and that its recommendations would be accurate and useful.

When it came time to roll out the system at the VA clinics, the project team identified a physician champion at each clinic who would organize training sessions for the participating doctors, promote ATHENA among their colleagues, and field questions or concerns. This aspect was crucial in getting physicians to use the system, said Eugene Oddone, MD, director of the Center for Health Services Research at the Durham VA Medical Center. "You definitely need a champion on-site who is respected, enthusiastic and willing to troubleshoot problems," said Oddone, who assumed this role at the VA Durham clinic. "Otherwise physicians tend to be skeptical."

Oddone said he and his colleagues who used ATHENA found it helpful, and have continued using it after the test period. Compared with other healthcare information systems, he said, ATHENA is more sophisticated and more useful. "It doesn't just tell you, 'Your patient's blood pressure is out of control.' It says, 'Here's what you can do about it.' And it shows you the supporting evidence." In his own practice, Oddone has found the system particularly helpful for patients with multiple health conditions who take several medications. For such patients, he said, "when I see that their blood pressure is still high, I always consult ATHENA, because it helps me make decisions based on a more complex set of rules than I could remember in my head."

Oddone said ATHENA has prompted him to manage his patients' hypertension more aggressively. "Doctors sometimes get complacent; when we've got the patient's systolic BP down to 145 we figure that's OK," he said, while the recommended systolic blood pressure is less than 140, and for diabetics it is less than 130. "ATHENA is that voice in the background, reminding you that you need to do more."

Co-principal investigator for the ATHENA project is Brian B. Hoffman, MD, professor of medicine at Harvard Medical School and chief of medicine at the VA Boston-West Roxbury. Michael Shlipak, MD, MPH, assistant professor of medicine at UCSF and the San Francisco VA, is a co-investigator. Other key collaborators include Robert Coleman, MS Pharm, and Susana Martins, MD, at the VA Palo Alto Health Care System; and Mark Musen, MD, PhD, Albert Chan, MD, and Samson Tu, MS, at Stanford Medical Informatics. The project was funded by grants from the Department of Veterans Affairs and the National Library of Medicine.

ATHENA's successful implementation was discussed in articles in two healthcare industry magazines: the December 2004 issue of Healthcare IT News, and the January 2005 issue of HealthLeaders.