Identifying Aortic Aneurysm Phenotype Cohorts Using Electronic Health Records
John Phair, MD1, Matthew Carnevale2, Evan Lipsitz, MD1, Jeffrey E. Indes, MD1.
1Montefiore Medical Center, Bronx, NY, USA, 2Albert Einstein College of Medicine, Bronx, NY, USA.
Introduction: Electronic medical records (EMRs) represent an underused data source that can potentially identify patients with abdominal aortic aneurysms (AAA). The utilization of EMRs as a screening tool for the recognition of comorbidities in AAA patients may aid in the identification of these patients, while also revealing unknown disease correlations.
Methods: Our institutuional EPIC EMR was queried utilizing Clinical Looking Glass (CLG), a patented software used to collect and evaluate EMR data. Medicare part B AAA screening was used as the basis for inclusion criteria. Search criteria identified all men aged 65-75 who presented to the hospital from 1/1/2016-1/1/2017. Demographics, included smoking history and family history of AAA. Outcomes included prior diagnosis of AAA, prior AAA ultrasound or abdominal CT, and prior EVAR or open AAA repair.
Results: A total of 7931 men aged 65-75 years were identified comprised of 4749(59.88%) inpatients and 3182(40.12%) ED patients. Patient demographics included: 1429(18.02%) white race, 3933(49.59%) Medicare beneficiaries, 4868(61.37%) diabetics, 329(4.14%) patients with CAD and 2680(33.79%) patients with hypertension. There were 2984(62.83%) inpatients and 1393(43.77%) ED patients with EMR diagnosed smoking history, 131(2.76%) inpatients and 43(1.35%) ED patients with documented family history of cardiovascular disease. There were 204(4.30%) inpatients and 60(1.89%) ED patients with a personal history of AAA, 18(0.38%) inpatients and 5(0.16%) ED patients had prior open AAA repair or EVAR. Finally, 173(3.64%) inpatients and 69(2.17%) ED patients had obtained an aorto-iliac duplex ultrasound while 285(6.0%) inpatients and 58(1.82%) ED patients obtained an aorto-iliac CT scan. After excluding all inpatients with no smoking history and no family history and those with prior aortic imaging, there was 2683 inpatients (88 with prior history of AAA, 6 of which had AAA repair). Similar analysis revealed 1306 ED patients, (61 with prior history of AAA, 5 of which had AAA repair). This left 2677 inpatients and 1301 ED patients eligible for AA screening.
Conclusions: The best method of recruitment for screening of those most at risk for AAA in the US remains to be determined. EMR data from clinical settings presents a potential source to identify patients at risk for AAA. Its universality, as well as, the real world nature of EMR’s make them an ideal system for screening and pre-emptive diagnosis of patients with vascular disease.
Back to 2019 ePosters