Aortic Dissection Detection risk score prevalence in emergency department patients.

Background: Aortic Dissection (AoD) is a high morbidity diagnosis based on imaging but dependent on clinical suspicion. There are no commonly used clinical prediction rules. The Aortic Dissection Detection (ADD) risk score was proposed (Rogers 2011) based on: 1) past medical history, 2) pain description, and 3) physical exam. Objective: Evaluate the degree to which specific components of the ADD risk score are documented in patients with suspected AoD in the emergency department (ED) setting. Methods: Single-site retrospective medical record review of all ED patients evaluated for suspected AoD. We electronically searched for all chest computed tomographic angiography (CTA) orders then manually excluded those for trauma or pulmonary embolism and examined for MD documentation of suspicion for AoD and results. Excluded: Known previous history of AoD. We examined the medical record and noted whether each of the individual components of the ADD risk score were documented via a standardized data collection instrument. For each component, we recorded if not documented, documented as negative, or documented as positive. Results: 114 cases of suspected AoD over 6 months. Prevalence of AoD was 4.4% (95% CI 1.9%-9.9%). At least one component of all three domains was documented in 27% of charts. Documentation was common for physical exam and pain features, where 100% and 96% of charts noted the presence or absence of at least one corresponding variable, respectively. Past medical history features were documented in only 27%. Individual variable documentation: Marfan 1%, family history 10%, Ao valve disease 9%, recent Ao manipulation 11%, thoracic Ao aneurysm 7%, abrupt pain onset 28%, severe pain 94%, ripping or tearing 43%, pulse deficit or BP difference 42%, focal neuro deficit 65%, murmur 79%, hypotension or shock 96%. Taking all 12 possible variables of the ADD risk score into account, the mean number of variables documented as present or absent overall was only 4.8. Conclusion: While classically-taught predictor variables of AoD such as severity, onset and nature of pain, and pulse deficits were commonly documented in the chart, the past medical history risk factors were poorly documented. This risk score may have utility for future research and/or clinical care, but either clinician education or electronic record decision support should be explored.