Background: Unnecessary hospitalizations and testing for syncope remain common (1), and physician attitudes continue to favor overuse for syncope (2). Structured protocols, specialized units, and decision support algorithms may reduce health service use, but lack of data precludes the ability to advocate for use of a specific decision support algorithm (3).
Purpose: Our primary aim for this quality improvement project is to reduce unnecessary transthoracic echocardiograms (TTE) among patients diagnosed with true syncope by developing and implementing a diagnostic timeout that triangulates several methods of changing providers’ behavior. In accordance with guideline recommendations, necessity is being defined by whether a cardiac etiology is suspected by the admitting provider. Secondary outcome measures are the short-term rates of critical interventions and adverse outcomes at 7 and 30 days. Secondary process measures include other diagnostic testing rates, such as intracranial imaging, laboratory testing, continuous cardiac monitoring, and stress testing.
Description: This project is occurring in the medical short stay unit of a large, tertiary care academic hospital. Patients are being consecutively screened and included by admitting unit providers who then implement the timeout, which consists of three steps: 1) patient selection; 2) differentiating between true and not true syncope; and 3) assigning a preliminary diagnosis. In Oct 2018, our first cycle began with a run-in period that utilized the patient selection portion of the timeout in order to develop a culture of appropriate subject selection and establish baseline rates of TTE. During this cycle, 56% (60/108) of patients admitted with suspected syncope had true syncope, and 18% (21/108) underwent TTE. In Jul 2019, our second cycle began with providers using the full timeout and is ongoing. The assignment of a preliminary diagnosis triggers subsequent decisions that: 1) establish rates of diagnostic confidence; 2) elicit accountable justification from providers about their decisions; 3) provide peer comparisons for TTE utilization rates; and 4) provide reassurance data for both providers and subjects. Preliminary diagnosis selections are being compared to final diagnoses to establish rates of diagnostic accuracy and calibration.
Conclusions: By incorporating several methods designed to change provider behavior into a diagnostic timeout, we expect to reduce resource overuse among patients presenting with true syncope to a medical short stay unit. Compared to current practice, we also expect this timeout to neither adversely affect short-term morbidity and mortality nor other diagnostic testing rates. By establishing this standard work, we expect to identify deviations from best practice and use scientific problem solving skills to improve our efficiency and effectiveness in an iterative fashion.