(B-037) Quantifying Risk in Patients Diagnosed with Pulmonary Embolism Using Electronic Health Record Data with Truveta

Background: Physicians classify patients who present with a Pulmonary Embolism (PE) using a prognostic tool called the simplified Pulmonary Embolism Severity Index (sPESI). This risk score estimates the risk of 30-day mortality for patients with PE and is used to help determine the appropriateness of outpatient therapy.

Objectives: To assess the feasibility and capacity of Truveta Data to create a new PE risk score that incorporates sPESI criteria and other components available in EHR data.

Methods: Truveta is a growing collective of 30 US health systems which includes more than 120 million de-identified patients. This retrospective study attributed PE patients to one of three risk classifications: highest risk, high risk, and low risk. Patients were characterized in accordance with their vital signs, laboratory results, and medical history.

Patients were classified as highest risk if procedure or diagnostic codes for any of the following were present within 1 day of their first PE diagnosis: shock, ventilator support, cardiac arrest, or need for CPR. Patients whose sPESI score was 0, had normal troponins and no vasopressor or inotrope prescription within 1 day of the PE were classified as low risk. Patients who did not fall into low or highest risk categories were classified as high risk. Both advanced therapies (AT) and medication only therapies (MOT) that occurred within 7 days of the PE diagnoses were quantified for each group. Inpatient mortality within 30 days of the encounter and the mean hospital length of stay (LOS) were also calculated.

Results: A total of 276,178 patients whose first PE diagnosis was associated with an inpatient or emergency room encounter type were identified. The population had slightly more females (50.6%) than males and was predominantly white (73.3%). Mean age and IQR were 63.3 (52.4 - 76.0).

Most patients fell into the high risk category (high risk: 63.2%, low risk: 26.7%, highest risk: 10.1%). Highest risk patients saw the greatest rate of AT (highest risk: 17.8%, high risk: 8.8%, low risk: 3.0%; X2 (2) = 8180.5, p < .001) and lowest rates of MOT (highest risk: 67.8%, high risk: 81.4%, low risk: 71.5%; X2(2) = 4543.5, p < 0.001) . Highest risk patients had the highest 30-day inpatient mortality rate (highest risk: 29.9%, high risk: 5.2%, low risk: 1.3%; X2(2) = 25709, p < .001) and the longest LOS (highest risk: 11.9 days, SD: 19.8; high risk: 7.0 Days, SD: 23.0; low risk: 5.7 days, SD: 8.8).

Conclusions: For patients with PE, Truveta’s novel data set using the components of the sPESI score has potential to classify highest risk patients, identify those most at risk of inpatient mortality, and inform on approach to treatment.