To evaluate the effectiveness of electronic clinical decision support on the diagnostic yield of CT pulmonary angiography for suspected pulmonary embolism (PE). Electronic decision support tools for suspected PE integrating PE Rule-out Criteria, Revised Geneva Score, and age-adjusted d-dimer prior to imaging, was implemented in the electronic health record of two health care systems hospitals each with >80,000 patients/year, 1 Epic site and 1 Cerner site. CT pulmonary angiography diagnostic yield and overall CTPA rates were evaluated and compared to matched controls. 95,382 patients were evaluated, with 53% female and mean age of 39. Prior to ePE tool availability 1.8% of all patients underwent CTPA compared to 1.9% with use of the ePE tool. CTPA yield was 11.3% at the Epic site and 11.2% at the Cerner site before and 14.6% and 15.6% after the ePE tool availability, respectively, representing an improvement by 3.3% (95% CI= 2.16-4.32, p<0.001) at the Epic site and at the Cerner site 4.4% (95% CI= 3.06-5.24, p<0.001). CTPA ordering rates were not significantly changed with ePE tool implementation, but CTPA yield improved with use of an electronic decision support for PE diagnosis.

Clinical decision support (CDS) for the diagnosis of acute pulmonary embolism (PE) has been recommended by multiple professional societies, yet its use remains uncommon. We sought to compare the CTPA positive rate at 4 hospitals where electronic CDS is optional, 1 hospital where eCDS was mandatory, and 1 hospital where eCDS was not available. We performed a retrospective cohort study at 6 hospital emergency departments (EDs), 3 academic level 1 trauma centers and 3 large suburban community hospitals with combined visits of 286,096 from January-December 2016. Included patients were age >17, with suspected PE determined by the treating emergency physician. Patients were excluded if age <18, or pregnant at the time of enrollment. eCDS consisted of PE Rule-out Criteria (PERC), Revised Geneva Score (RGS) or Wells score followed by age-adjusted d-dimer when appropriate and was implemented in 5 hospitals electronic medical record; 4 where the use of eCDS was optional and 1 where it was mandatory prior to ordering CTPA. We compared the relative risk and percent relative risk of having a positive CTPA using combined eCDS utilization data from the mandatory use hospital plus the cases where eCDS was used at the 4 optional use hospitals, to the hospital where eCDS was not available and cases where eCDS was not utilized at the optional hospitals. At the 5 hospitals with eCDS available 1147 CTPA’s were performed using the eCDS tools with 184 PE diagnoses for a CTPA positive rate of 16% (95% CI: 14-18%). At the hospitals where eCDS was not used or not available, 4560 CTPA were performed with 302 PE’s diagnosed for a CTPA positive rate of 6.6% (95% CI: 5.9-7.4%). With eCDS use, relative risk of having a positive CTPA is 2.42 (95% CI: 2.04-2.88; p<0.001), compared to no use of eCDS. Percent relative risk shows CTPA’s are 142% more likely to be positive with eCDS utilization. eCDS utilization is associated with a higher CTPA positive rate.Tabled 1CTPA positive rate with eCDSCTPAN=PE Diagnoses N=% Positive (95% CI)eCDS114718416% (14.1-18.3)No eCDS45603026.6% (5.9-7.4) Open table in a new tab

The Intermountain Risk Score (IMRS) has proven to predict mortality in multiple patient groups. We sought to determine if it was predictive of mortality in patients with PE diagnosed in the emergency department. A retrospective chart review was performed on 3500 patients that had a computed tomography pulmonary angiogram (CTPA) performed for the workup of PE in two urban hospital EDs with a combined annual census of ∼120,000 patients between January 2009 and December 2010. We tracked recurrent (venothromboembolic events) VTE at 90-days and death at 90-days and 5-years. A total of 343 patients were enrolled following diagnosis of PE by CTPA in the ED between January 2009 and December 2010. 52% of patients were female and mean age was 55.41. At 90 days, women have mortality rates of 0.0% (0 deaths of n=98), 7.1% (4 deaths of n=56) and 15.0% (3 deaths of n=20) for low-, moderate-, and high-risk IMRS, respectively (p-trend<0.001). At 90 days, men have 0.0% (0 of n=84), 7.4% (4 deaths of n=54), and 11.8% (2 deaths of n=17) for low-, moderate-, and high-risk IMRS, respectively (p-trend=0.005). For recurrent VTE at 90 days the results were 4.2%, 5.6%, and 5.9% for females (p-trend=0.68) and 1.3%, 7.8%, and 5.9% for males (p-trend=0.12) for low-, moderate-, and high-risk IMRS. Mortality for the 5-year follow-up period produced similar results (0%, 11.5%, and 38.7% death for females, p<0.001; and 6.8%, 21.7%, 75.0% death for males, p<0.001). The IMRS is highly predictive of both 90-day and 5-year all cause mortality in PE patients.View Large Image Figure ViewerDownload Hi-res image Download (PPT)

The Intermountain Risk Score (IMRS) is an easily computerized risk assessment tool based on results of common lab tests, age, and sex that has been shown to predict mortality related to pulmonary embolism (PE) and other conditions. We sought to determine if it could better predict PE mortality in patients across all classifications of PE Severity Index (PESI) in order to expand the population that may be eligible for outpatient management. A retrospective chart review was performed on 3500 patients that had a computed tomography pulmonary angiogram (CTPA) performed for the workup of PE in two urban hospital EDs with a combined annual census of ∼120,000 patients between January 2009 and December 2010. We then calculated PESI scores on all patients diagnosed with PE and tracked mortality at 90-days and 5-years. Of the 3500 patients with CTPA, 343 were diagnosed with PE (9.8%). 52% of patients were female and mean age was 55.4. Patients were then grouped into PESI classifications; classes 1-2 combined (n=179), class 3 (n=69), and classes 4-5 combined (n=81) and IMRS performed within each of the three groups to identify mortality over 5 years. The p-trend across the three IMRS categories in PESI 1-2 is p=0.68, for PESI 3 the p-trend is p=0.002, and for PESI 4-5 the p-trend is p=0.05. There was no significant difference in mortality between PESI 1-2 and IMRS. The patients in the low-risk IMRS category had a 0.0% mortality rate at 90-days (p<0.001) and low 5-year mortality across all PESI classes. IMRS predicts mortality equally well in the "low-risk" PESI classes (1-2), but stratifies patient mortality in the higher PESI classes (3-5) that might identify an additional patient population suited for outpatient management of PE.View Large Image Figure ViewerDownload Hi-res image Download (PPT)

Outpatient treatment for low-risk pulmonary embolism (PE) is recommended by many professional societies; however, in the US this practice has not yet been adopted as standard of care due to lack of US-based prospective studies proving safety. We performed a prospective, non-inferiority multicenter trial comparing outpatient treatment to historic controls, starting in January 2013 to present. Consecutive patients with PE determined to be low-risk by PESI score <86 were screened for enrollment at 5 hospitals in Utah, US. Patients who qualified for the study were discharged directly from the emergency department or following short observation stay (<23 hours) and were treated with enoxaparin/warfarin, rivaroxaban, or apixaban per patient preference. The primary outcome was a composite of 90-day mortality, recurrent symptomatic VTE or major bleeding. This trial was registered with ClinicalTrials.gov, and IRB approved. To date we have enrolled 193/200 total patients, with 90-day follow-up completed on 178 patients; 55% of patients are women. There have been 3 composite outcome events; no mortality events, 1 recurrent VTE event, 2 major bleeding events; 22 patients (11.4%) were discharged on warfarin/enoxaparin, 21 patients (10.9%) on apixaban, 150 patients (77.7%) on rivaroxaban. Carefully selected patients with low-risk PE can be safely treated as an outpatient.