Structured query language tool to identify untreated individuals with abdominal aortic aneurysms

Abstract

Objective

Despite well-established guidelines for the diagnosis and surveillance of abdominal aortic aneurysms (AAAs), many patients are still not discovered until they have already sustained a rupture. It is estimated that as many as 50% of individuals with AAAs are undiagnosed until they present with a life-threatening rupture and hemorrhagic shock. Structured query language (SQL) is a tool that can scan existing electronic medical record (EMR) systems to identify patients with specific characteristics or morbidities. Here we evaluate the use of SQL to determine whether it can identify untreated patients with AAA disease and other vascular comorbidities.

Methods

An SQL disease discovery query code (Medtronic) was developed based on the clinical criteria recommendations of the United States Preventative Services Task Force for AAA screening (male sex; age, 65-75 years; and former/current smoking history). The code was then integrated with 18 International Classification of Diseases, 10th Revision (ICD-10) diagnostic codes for AAA and AAA-related diseases. Over 130 ICD-10 or Current Procedural Terminology (CPT) codes for previous aortic repair were overlaid as an exclusion criterion. The query was applied to 1 year of patient data (June 2021 to June 2022) at our single-center, university-affiliated, regional referral medical center. Manual chart review was performed on identified patients to confirm the incidence of AAA disease (aortic diameter ≥3.0 cm), as well as the incidence of other vascular conditions such as hypertension, coronary artery disease (CAD), peripheral arterial disease (PAD), and/or carotid artery stenosis (CAS).

Results

The SQL code identified a total of 457 patients (449 male, 8 female). Of these patients, 167 (36.5%) had confirmed AAAs with an average diameter of 4.2 cm (range, 3.0-9.7 cm), with a positive predictive value of 28.2%. Prior AAA repair, with either endovascular or open repair, was observed in 39 patients with AAAs (23.4%). Among the remaining untreated 128 patients, six (5 male, 1 female) met the traditional size criteria for repair, and time between last radiological assessment and last clinical follow-up was 4.8 ± 8.8 months. Interestingly, only 71 patients (55.5%) were evaluated by a vascular surgeon. Additionally, we observed that the SQL code identified concomitant PAD in 31 patients (18.5%), and CAS in 17 patients (10.1%). Among patients with treated and untreated AAAs, 26.8% had either PAD or CAS.

Conclusions

An SQL tool can be incorporated in modern healthcare systems to facilitate identification of patients with untreated AAA disease and other vascular comorbidities. Such tools can enhance prompt disease recognition, referral to vascular surgery specialists, and early implementation of appropriate surveillance and/or treatment algorithms.