Association of D-Dimer Screening With Emergency Department Length of Stay

Although guidelines recommend a D-dimer-first approach to the diagnostic evaluation for pulmonary embolism (PE) in most patients, literature suggests physicians may avoid D-dimer due to concerns about potential delays in care. We are aware of no studies on the association of D-dimer screening with ED patient throughput. We hypothesized that a D-dimer-first strategy was associated with a faster time to ED disposition than a strategy using only computed tomography (CT).

We performed a retrospective cohort study of adult patients (18-79 years) with D-dimer testing or CT pulmonary angiography at 3 EDs between 2017-23. Our primary outcome was the time between a patient’s first PE test order and the time of their ED disposition order, in minutes. We excluded patients with hypotension on arrival, a chief complaint of PE, and those who died in the ED, eloped, left against medical advice, were transferred, were initially evaluated as a trauma alert, or had D-dimer ordered after CT. Covariates included demographics (age, sex, race, language, insurance); initial vital signs; arrival characteristics (time; ED volume); comorbidities; chief complaint; year; hospital identifier; and other diagnostic tests (influenza and Covid tests; pregnancy tests; chest x-ray). To address confounding by indication, we applied inverse probability weighting using the Stata stteffect ipw command with a Weibull survival distribution to assess the marginal treatment effect of D-dimer on time to ED disposition. Estimates included the average treatment effect (ATE) and the average treatment effect on the treated (ATT), with robust standard errors to account for uncertainty in treatment and outcomes models.

The sample included 13,701 patients, including 7,983 who had D-dimer screening and 5,718 with CT only. The D-dimer group was younger (50.4 vs. 58.5 years), had lower rates of chronic heart disease (17.8% vs. 25.7%) and cancer (13.3% vs. 31.5%); lower rates of respiratory viral testing (29.7% vs. 44.2%); and a lower hospitalization rate (28.4% vs. 57.4%). D-dimer screening was associated with a shorter time to ED disposition, with an ATE of -10.4 minutes (95% confidence interval [CI] [-14.7, -6.1], p < 0.001) and an ATT of -14.3 minutes (95% CI [-19.1, -9.4], p < 0.001).

A D-dimer-first approach to evaluation of PE was not associated with a delay in disposition