In our study, the AAA screening rate among all eligible patients significantly improved after initiation of a Web-based point-of-care clinical decision support system (GDMS) in our primary care practice. The system's utilization in face-to-face encounters led to increased completion of screening after a clinic visit by identifying eligible patients and alerting their providers, who then engaged patients in the discussion, regardless of the reason for the visit. This was evidenced by a fivefold improvement in screening rates for eligible patients in 2008, when compared to the same period in 2007. This makes sense given that there are many tests that a patient might have for unrelated reasons that can also visualize the abdominal aorta; if the patient already had any of these tests, they do not need an ultrasound test for AAA screening. These tests include abdominal ultrasound, abdominal or pelvic computed tomography, magnetic resonance angiography of the abdominal aorta and computed tomography colonography.
Taking time out of a 15-minute primary care consultation to check the patient's EMR for a record of such a test is not possible. This is particularly true when the physician is seeing the patient for an acute condition (e.g. back pain, headache) and has a limited time to address it. The GDMS and our pre-visit work processes are designed to cue the provider if the patient has already had a test that visualized the abdominal aorta, saving us the time of searching for that information. Even though our study was active for only 3 months, five patients with AAA were identified, two of whom benefited from elective repair that may have saved their lives.
However, not all patients identified by the decision support as eligible for screening received it. In fact, 81.7% of eligible patients did not receive the screening. Although the system helped providers by identifying patients who needed preventive services and tests for chronic conditions, the limited time that providers have for visits might have been insufficient to address all the patients' needs. This highlights the importance of patient health records for enabling patients to be proactive in seeking the care they need. If our systems allowed patients to access their own version of GDMS before the visit (i.e. Web-based patient portal), they might have been proactive in asking for this service.
We have not investigated the possible reasons for lack of screening of eligible patients; however, we hypothesize that, because the GDMS alerts occur during all patient visits concerning any number of recommended tests or treatments unrelated to AAA, the providers might not have had time to address all the recommendations and therefore left them for a future visit. Recent literature also suggests that provider reminders have to be appropriately inserted into their workflow to have the desired effect [17
]. A recent study showed that providers might ignore over 90% of alerts because of alert fatigue [18
This finding underscores the importance of having standardized processes in place to support the clinical decision support system for services that are due. Primary care physicians have many tasks to perform during a 15- to 20-minute visit, and having other members of the health care team take responsibility for helping the physicians deliver the necessary services can lead to better results by getting the orders for the services ‘teed up’ for them [8
With the upcoming changes spurred by health care reform and an increasing focus on quality and value, the emphasis on delivery of evidence-based preventive services and management of chronic conditions will place increasing demands on primary care providers. The decision support systems that help improve quality will play an increasingly important role. Without these systems, providers will not be able to provide high-value care to their patients, as they will need to determine the need for screening through a time-consuming and unreliable manual process.
As US efforts to improve health care quality and cap costs shift to HIT, carefully designed clinical decision support systems can optimize care delivery in primary care. As we move towards achieving cost-effective care and reducing unnecessary laboratory tests and radiologic imaging studies, it is important to note that our clinical decision support model enabled us to identify up to 31.92% of patients who had prior imaging of their aorta from an unrelated examination. This information helped us eliminate the need for an additional ultrasound examination, saving time and expense for patients and unnecessary cost for the organization and third-party payers.
As with any practice-based quality study, our study of the effect of decision support technology implementation had its limitations: (1) Our control population was retrospective from the prior year. We would have preferred a randomized controlled trial, but that was not feasible, because our goal was to provide improved care to all of our patients for all of their preventive services and chronic conditions. (2) We did not ask the primary care providers why the remaining eligible patients did not get the AAA screening as they were identified as needing. (3) We did not explore the relationship between the visit type and the likelihood of the patient getting the necessary screening.