JMIR Publications

ABSTRACT

Background:

Syncope is a prevalent and recurrent condition. Because of concerns that patients presenting with syncope are at risk for an impending catastrophic event, overuse of testing and admission are common. Though clinical decision support (CDS) tool is an effective strategy, CDS tools in syncope are lacking and limited.

Objective:

Develop and demonstrate the viability of a clinical decision support for syncope diagnosis and prognosis.

Methods:

With input from clinicians representing diverse health systems, we developed the MISSION Syncope OptimalCare Pathway based on 2017 Syncope ACC/AHA Guidelines. This protocol provides flexible decision-making and tailoring of treatment to patient’s clinical presentation, medical history and needs while emphasizing the care continuum. We designed a mobile App for the implementation of this protocol. Following human-centered design principles and applying research evidence, development of the App was a multi-step process including: 1) ideation; 2) assessment variable decision; 3) mathematical model development for differential diagnosis; 4) clinical risk tool selection; 5) recommendation decisions; and 6) technical buildout of the application. We used baseline prevalence of different etiologies of syncope, pre-test odds ratio (OR) and likelihood ratios (LR), for each indicator in available literature to predict each etiology. Multiple binary regression models were used to calculate post-test ORs. Clinicians guided refinement of the model where LRs were lacking or clinical indicators warranted assessment.

Results:

The App product is consistent with guideline recommendations. It walks through clinical assessment in a concise manner that is consistent with evidence-based practice and provides recommendations for diagnosis and prognosis based on user input. Though a ranked differential diagnosis is displayed, the user can pick their most likely differential. The same set of questions also determines a Canadian Syncope Score, a validated clinical scoring system to identify those at higher risk for cardiovascular events after a syncopal episode.

Conclusions:

Our App demonstrates the viability of using evidenced-based literature in developing a CDS tool, and the importance of applying clinical experience to fill gaps in research. It is essential for a successful App to be deliberate in pursuing a practical clinical model instead of striving for a perfect mathematical model. Our experience can be applied to similar CDS tool development, especially in clinical scenarios without definitive research.