From the 2018 HVPAA National Conference
Remus Popa (University of California Riverside), Sara Patrawala (University of Rochester Medical Center), Leslie Oyama (University of California San Diego)
Evidence-based medicine is inconsistently used in the evaluation of syncope, leading to unnecessary testing and increased healthcare costs.
To study the effect of a multifaceted intervention, including a best practice algorithm, on the adherence to evidence-based recommendations.
A multidisciplinary group (Cardiology, Neurology, Emergency Medicine, Hospital Medicine) developed an evidence-based syncope algorithm. Included was a risk stratification tool used to divide the patients in low-risk (LR) and high-risk (HR) groups for adverse events. Educational (ED nursing education on orthostatic vital signs or OVS, physician education on syncope guidelines) and electronic medical record (EMR) interventions (preselected ECG and OVS orders, note template to prompt collection of pertinent history, link to syncope algorithm on intranet) were implemented. Emergency and Internal Medicine physicians independently reviewed charts of syncope patients seen in the ED April – June 2011 (Pre) and 2013 (Post). Clinical data and tests performed or recommended were analyzed for their appropriateness per protocol. Adverse outcomes (ICU admission, blood transfusions, cardiac catheterization, pacemaker/AICD placement, ED readmission or death within 30 days) were tracked. Fisher exact test run with STATA was used to determine statistical significance.
There were 138 cases in the Pre and 175 in the Post intervention group. In the Pre group, 77 patients were LR and 61 were HR compared to 112 LR and 63 HR in the Post group. There was a significant improvement in obtaining OVS (p<0.001) and documentation of history pertinent to risk stratification (p<0.05) for all patients, as well as an overall reduction in excessive testing. In the HR group, a significant reduction in unnecessary testing (Troponin, Telemetry, Echocardiogram, head CT) post intervention was observed. There was a significant reduction in LR telemetry utilization and a trend towards a reduction in head CTs. Reductions in more advanced imaging (Stress testing, MRI, EEG, etc.) were not seen. LR patients had low rate of adverse events (limited to a few ED revaluations for benign problems), suggesting it is safe to utilize such an algorithm for risk stratification in patients with syncope of unclear etiology.
The most significant gains in adherence to an evidence-based syncope protocol came with interventions that were hard wired into the EMR. There was a higher variability and deviation from evidence-based protocols in testing that required physician judgment without EMR prompts (i.e. troponin, Head CT, echocardiogram). This is similar to conclusions made by other investigators that physician education alone is not enough to spur adherence to evidence-based recommendations. We hypothesize that practice guidelines embedded in the EMR workflow would lead to a decrease in test overutilization and inappropriate admissions. An EMR order set is both feasible and practical to achieve. A larger study size is needed to detect differences in the use of cardiac testing in LR populations as well as more advanced imaging in both the LR and HR groups due to the overall infrequent use of these tests.
Implications for the Patient
Our intervention reduces variation from the standard of care as well as decreases diagnostic errors and overtesting. Both the emotional and financial burden of over testing is important and often overlooked. Having a standardized approach to certain clinical presentations will have a positive impact on the overall patient experience.