Impact of High‐sensitivity Troponin Testing on Operational Characteristics of an Urban Emergency Department

Emergency department (ED) and hospital overcrowding is a major issue impacting health care institutions all over the country. Approximately 6% of all ED visits and approximately 27% of all admissions are related to chest pain.1 According to some studies, utilization of the newer high-sensitivity troponin (hs-TnT) assays is an effective way to improve early identification of acute coronary syndrome (ACS).2 However, to our knowledge, there are no studies evaluating the impact of hs-TnT testing on hospital operations. We hypothesize that implementation of the new hs-TnT with accelerated diagnostic pathway (ADP) will result in decreased admission rates compared to the use of older fourth-generation troponin T (TnT). Secondary outcomes were ED length of stay (LOS) and laboratory use. University of Texas Southwestern’s Clements University Hospital in Dallas, Texas, recently replaced a TnT and a point-of-care troponin I (TnI) assays with a new hs-TnT assay from Roche Diagnostics Elacsys. Prior to this change, patients evaluated for ACS were tested using TnT or TnI at the providers’ discretion. On the day the hs-TnT assay was implemented, both TnT and TnI were simultaneously discontinued. This resulted in three separate, significant changes in evaluation of ACS: the near instantaneous point-of-care testing was discontinued requiring all assays to be processed by a central laboratory, new cutoff limits required a change in interpretation (normal limits changed from <0.01 ng/mL for TnT and TnI to <12 ng/L for hs-TnT),3 and evaluation was standardized to follow a specific ADP4 instead of relying on an individual provider’s discretion. This ADP used both hs-TnT results and time of onset of symptoms to direct subsequent testing and disposition. The implementation of ADP was automated within a clinical decision support system using a deterministic finite state machine computational model that offered nursing and physician guidance on when next assay was required and ultimately the ADP had completed. To evaluate the impact of this transition, we examined a convenience sample of all patients who presented to the ED between June 1, 2018, and June 1, 2019 who underwent troponin testing as part of their ED evaluation. During the first part of the study between June 1, 2018, and October 22, 2018, patients requiring troponin testing were evaluated with a combination of TnT and TnI. Patients presenting after this time period were evaluated only with the new hs-TnT plus the ADP that resulted in each patient being classified as “ruled out” or “abnormal.”4 Providers were then encouraged to use the HEART score for risk stratification5 to help make the final disposition. Standard ED operational metrics were used to assess the implementation of the hs-TnT in association with an ADP on the operational characteristics. The aims of this before-and-after study are to compare the impact of the new hs-TnT assay and ADP on the hospital admission rates. We also examined its impact on ED LOS and phlebotomy/laboratory utilization. ED LOS is defined as the time between patient arriving to an ED treatment room and the time of disposition order; this metric reduces the impact of confounders such as ED overcrowding and ED boarding on this analysis. Continuous variables if normally distributed were depicted with mean (±SD) and compared with Welch’s t-test; otherwise, they were depicted with median (IQR) and compared with the Wilcoxon rank-sum test. Categorical data were depicted with total count and percentages and compared using two-sample test of proportions. The changes to ED operational characteristics are summarized in Table 1. During both study periods, approximately 27% of ED patients were evaluated with a troponin assay, and both groups had similar distributions of chief complaints. The admission rate for patients who did not undergo troponin testing was similar during TnT/TnI and hs-TnT periods (25.0% vs 28.3%, p = 0.08). On the other hand the admission rate for all patients evaluated with a troponin assay decreased from 54.3% for TnT/TnI to 51.4% for hs-TnT (p < 0.01), but resulted in a 7-minute increase in LOS (95% CI = 3 to 11 minutes, p < 0.01) and 0.76 increase in troponin assay utilization per patient encounter (p < 0.01). In the subgroup of patients with the chief complaint of chest pain, implementation of hs-TnT and ADP had even greater effect on admission rate and LOS—admissions decreased from 42.6% to 31.5% (p < 0.01) and the median LOS increased by 22 minutes (95% CI = 12 to 28 minutes, p < 0.01). A decrease in admission rates was not observed in patients who presented with chief complaints other than chest pain (Table 1). TnT/TnI (Jun 2018–Oct 2018) hs-TnT (Oct 2018–Jun 2019) Difference (95%CI) This study is one of the first to describe the impact of implementation of ADP using fifth-generation troponin on the ED throughput in the United States. The study demonstrates a significant reduction in patient admissions following hs-TnT and ADP implementation. This reduction is not only a reflection of improved sensitivity compared to prior fourth-generation troponin assays, but it also suggests that the ADP guides providers toward recommended management based on the final risk strata of their patients. Previously utilized risk stratification tools such as the HEART score5 or EDACS6 have been plagued by significant variation in testing strategies and disposition following troponin testing. However, in this study, the standardization of patient evaluation within the ADP helped minimize this variability and potentially contributed to the decrease in admission rates. The decrease in admission rates comes at a price of increased ED LOS and increased number of troponin assay tests. While the 22-minute increase in LOS and 0.7 increase in tests per patient were statistically significant, these numbers need to be placed in an appropriate perspective. In a subgroup with chest pain, for every nine patients tested, one less required an admission. This suggests that every obviated admission costs an additional 198 minutes in ED room occupancy and six extra tests. For some hospitals this might be a cheap price to free up an inpatient bed, while for others with small EDs it might be too high. This cost–benefit analysis needs to be made on a per system basis. Because multiple interventions occurred in parallel, change to troponin assay and initiation of the ADP, it is possible that the increased LOS is a result of an algorithm that encourages repeat testing. In addition, the new hs-TnT assay and ADP were implemented within a new electronic health records–based clinical decision support system that simplified interpretation of this diagnostic algorithm and offered the ability to alert providers when the ADP had completed and automatically discontinue any unnecessary troponin orders. This automation may directly impact LOS and limit the generalizability of our results. This study was conducted at a quaternary hospital with patients who likely demanded higher number of troponin tests within the ADP. Health systems with lesser acuity might require fewer troponin tests and therefore have shorter ED LOS and per patient testing while still benefiting from the decreased admission rate. Implementation of hs-TnT–based ACS evaluation utilizing a novel ADP was associated with decreased hospital admissions for chest pain patients, but at a cost of longer ED room times and more per patient testing. It also helped to standardize a highly variable chest pain evaluations and admission strategies for high-risk patients. With increased emphasis on decreasing costs and improving outcomes, implementation of the high-sensitivity troponin assay with ADP may offer improved risk stratification and resource utilization.