Abstract
Background
In 2021, 7 professional organizations jointly released guidelines for chest pain evaluation advocating for use of high-sensitivity troponin (hs-cTnT) and structured clinical pathways.
Objective
We explored the impact of hs-cTnT recommendations on emergency department (ED) care quality, downstream testing, hospitalization, and standardization of physician practice after automated reflex testing alongside an accelerated diagnostic protocol.
Methods
A pre- and postobservational study in 9 EDs introduced hs-cTnT in February 2022. Adults with complaints suggestive of acute coronary syndrome were included. Primary outcomes were odds of troponin testing, ED discharge, and chest pain center observation; secondary outcomes included proportion of acute myocardial infarction diagnosis, odds of downstream testing, and changes in physician practice variation. Multivariable logistic regression was used adjusting for patient demographics and secular trends during the baseline period.
Results
Of 68,301 visits in the 6-month preintervention period and 121,376 visits in the 12-month postintervention period, troponin was ordered 33,061 (48.4%) and 52,382 (43.2%), respectively. The intervention was associated with lower odds of troponin testing (odds ratio 0.79, 95% CI [0.76, 0.82]) and no difference in ED discharge (0.99 [0.95, 1.03]). There was increased odds of acute myocardial infarction diagnosis (1.17 [1.001, 1.36]), echocardiogram (1.08 [1.01, 1.15]), and cardiac catheterization (1.21 [1.07, 1.38]); decreased odds of cardiology consultation (0.88 [0.83, 0.93]); but no difference in stress testing (1.01 [0.87, 1.17]). Physicians with low preintervention discharge rates increased their discharge rates by 3.11% (2.9%, 3.3%) relative to peers.
Conclusion
hs-cTnT transition was associated with a decrease in troponin testing, varied effects on downstream cardiac testing, and a concurrent increase in diagnostic yield without appreciable impact upon ED disposition. Physicians with preintervention outlier practice moved toward department norms postintervention. Together, these findings suggest the intervention promoted diagnostic quality for ED patients with suspected acute coronary syndrome.
Keywords: care pathways, emergency medicine, high-sensitivity troponin, physician variation
The Bottom Line.
In this real-world study of over 189,000 emergency department visits, implementing high-sensitivity troponin with an accelerated diagnostic protocol reduced troponin testing and increased myocardial infarction diagnoses without affecting overall discharge rates. The intervention also standardized physician behavior, reducing variation in discharge and admission decisions. These findings highlight how integrating high-sensitivity troponin with clinical pathways can improve diagnostic efficiency and promote consistent decision-making in emergency care.
1. Introduction
1.1. Background
Chest pain is the chief complaint or presenting problem in over 8 million annual emergency department (ED) visits in the United States, making it one of the most common ED presentations.1 As coronary artery disease is a leading cause of death, rapid diagnosis and treatment are paramount.1, 2, 3 Although select patients will have electrocardiogram (ECG) suggestive of ST-elevation myocardial infarction and others will have clear alternative (ie, noncardiac) explanation of symptoms, many will require additional evaluation for acute coronary syndrome (ACS) using cardiac enzyme troponin (cTnT) assays. High-sensitivity cardiac troponin (hs-cTnT) assays have been used internationally since 2010, gained U.S Food and Drug Association approval in 2017, and increasingly are replacing older assays due to the ability to detect clinically important myocardial injury.4, 5, 6, 7 Although some prior work has explored the implementation of hs-cTnT in the ED, there are a variety of proposed ED testing strategies, concerns about downstream testing consequences, and lack of clarity around implementation techniques, all of which may contribute to slow hs-cTnT adoption across the United States.8, 9, 10, 11
1.2. Importance
In November 2021, 7 professional organizations released clinical practice guidelines for the evaluation and diagnosis of chest pain.2 These guidelines outline a variety of evidence-based best practices for chest pain evaluation including the use of hs-cTnT and practice consensus on stress testing to improve risk stratification. However, implementing guidelines has historically been delayed and variably successful.12, 13, 14 Real-world limitations to implementation can include limited effectiveness of educational or communicational interventions such as emails or presentations, as well as mixed evidence regarding the use of data-driven audit and feedback interventions in the ED setting given the variable time lag between delivery and clinical application of the information.15, 16, 17, 18 Furthermore, remembering guidelines can be challenging amidst frequent interruptions in an ED. Clinical decision support (CDS) tools are often implemented within the electronic health record (EHR) to promote best practice within the clinical workflow but have mixed outcomes.19, 20, 21 Furthermore, traditional CDS tools are not well suited to the complex, sequential decision required for interpreting hs-cTnT results.
1.3. Goals of This Investigation
Accordingly, we sought to evaluate the real-world transition from conventional troponin to hs-cTnT alongside the introduction of an associated EHR-integrated accelerated diagnostic protocol (ADP) and automated reflex testing on ED physician use of troponin testing, diagnostic yield, downstream cardiac testing, patient disposition, and diagnostic delay. As a secondary objective, we examined physician variation in practice before and after implementation.
2. Methods
2.1. Design and Setting
We conducted a retrospective pre- and postobservational study of adult ED patients with presenting complaints suggestive of ACS at 9 EDs in a single health system. The EDs have a combined annual volume of approximately 400,000 patient visits. The composition of the EDs include one academic, 6 community, and 2 free-standing EDs. The academic ED is the largest site and is a tertiary care center.
2.2. Selection of Subjects
Data were collected on ED patients, aged 18 years or older, presenting from July 1, 2021, to February 2, 2022 (preintervention period), and February 10, 2022, to February 9, 2023 (postintervention period), with a 1-week “washout” period surrounding the transition from Roche 4th generation troponin (herein termed “conventional troponin”) to Roche 5th generation assay (hs-cTnT). We extended the postintervention period to 12 months to mitigate biases common in quality improvement studies, such as regression to the mean and early adoption effects, ensuring a more stable and sustainable assessment of the intervention’s impact.22,23 Patients were included if they presented to the ED with chief complaints suggestive of ACS as determined by consensus among a panel of 7 ED physicians based on American Heart Association/American College of Cardiology (AHA/ACC) Guidelines, prior work, and expert opinion (see Table S1 for included chief complaints); primary analysis was limited to this population as troponin testing in this group likely represented evaluation for ACS rather than other reasons, such as risk stratification in gastrointestinal bleed or sepsis.2,24,25 This strategy worked well as the proportion of final diagnosis of gastrointestinal bleed or sepsis was similar in both periods (Table S7). Patients were excluded if they were transferred out of the ED, left without being seen, or left against medical advice as workup in these cases was considered incomplete; presented as a trauma or stroke activation, as troponin testing is included in EHR order sets for these activations and is less likely to represent evaluation for ACS; if they presented in cardiac arrest or required intubation, as their illness severity likely altered diagnostic reasoning and downstream testing; if they were pregnant, as this may have altered downstream testing; or if they presented with ST-segment elevation myocardial infarction, as this is primarily an electrocardiographic rather than laboratory diagnosis (Fig S1).
Data were obtained from the institutional data warehouse (Epic Systems) and included patient demographics, presenting complaint(s), ADP use, final diagnoses and dispositions, troponin orders, downstream cardiac workup (echocardiograms, stress testing, coronary catheterization, and cardiology consultations), and treating attending physician. The study was deemed exempt from IRB review under 45CFR46.104.
2.3. Intervention
On February 3, 2022, hs-cTnT (Roche) was introduced across the health system using an order panel (Fig S2) in conjunction with an ADP. Clinicians had the option of ordering automated reflex testing or stand-alone testing. For the reflex testing order, the hs-cTnT result value and delta thresholds were programmed into the EHR, which generated orders for second and, if needed, third hs-cTnT order at the programmed time interval to reduce variation and clinician burden in ordering subsequent tests. Hs-cTnT replaced the troponin options in order sets and standing orders to promote adoption of the new test. Collectively, these changes (hs-cTnT order panel, ADP, automated reflex testing) are described herein as the “intervention.”
With the exception of the tertiary ED, all EDs only had the option to utilize hs-cTnT after February 3, 2022. The tertiary ED was still able to utilize a conventional point-of-care (POC) troponin assay, but it was removed from order sets, standing nursing orders, and the preference list of common orders to discourage POC testing. Therefore, POC testing was only available if manually searched for and ordered. It was removed within 2 months of implementation of hs-cTnT, and there were no patients who only had POC testing ordered (12 total POC troponin ordered postimplementation). Before and after implementation, there were no differences in standing order triggers for nursing to place troponin orders and the order sets that had POC testing were swapped with hs-cTnT testing.
In conjunction with the hs-cTnT implementation, an ADP in the form of an EHR-integrated clinical pathway was introduced simultaneously across the health system. This CDS tool (AgileMD), described previously,26 was an interactive flow diagram within the EHR (Epic) that displays synthesized guidelines and local expert consensus and allows direct deployment of the recommended orders through the pathway for the patient whose chart was open. The ADP was developed to incorporate best practice from the AHA/ACC Guidelines and local health system cardiology recommendations using multidisciplinary input from ED, laboratory medicine, nursing, and cardiology using the pathway framework described previously.19,26 This ADP delineated hs-cTnT cutoffs and delta for myocardial injury based on a patient’s modified HEART score guiding clinicians to disposition and workup decisions (Figure S3A-C). The ADP is a passive tool that clinicians can access and use at their discretion. For the 1 month before the hs-cTnT and ADP implementation, educational sessions were held with all clinical groups including faculty physicians, resident physicians, non-physician practioners, and nurses to provide context and orientation to the guidelines, ADP, and answer questions. The ADP was not available outside of the EHR for use.
2.4. Measures/Outcomes
The primary outcomes included the proportion of ED encounters with at least one troponin ordered (troponin utilization rate), the proportion of patients discharged from the ED (discharge rate), and proportion of patients placed in the ED chest pain center (CPC) observation unit. The CPC is an ED managed cardiac observation unit that performs further cardiac testing resulting in hospital admission or discharge. To examine the more direct effects of ADP use, we also examined these outcomes among users of the ADP as tracked through EHR access data acknowledging this does not account for clinicians who have learned the ADP and do not feel the need to access it. Secondary outcomes included the proportion of patients tested with troponin for suspected ACS with a final ICD10 diagnosis code of acute myocardial infarction (AMI) (diagnostic yield) and rates of downstream testing and assessment including echocardiograms, cardiac stress testing, coronary catheterization, and cardiology consultations.
2.5. Data Analysis
The primary analysis examined the usage of hs-cTnT compared to the conventional troponin and POC troponin among patients with presentations potentially suggestive of ACS. The primary analysis was performed with logistic regression adjusting for age dichotomized as less than 65 or greater than or equal to 65, Charlson comorbidity index (binned as 0, 1, or 2 or more),27 sex (binary male/female), language, hs-cTnT implementation date, and secular trend during the baseline period. Secondary outcomes utilized the same covariates as the primary outcomes and calculated on the study population as well as the subset of patients who received troponin testing. Delayed ACS diagnosis was defined as a patient visit without an ACS diagnosis and a discharge disposition with a subsequent ED visit within 30 days with an ACS diagnosis.
As a secondary analysis, we examined whether individual physician behavior changed with respect to primary outcomes after hs-cTnT implementation. Downstream testing and yield were not included as part of this physician-level analysis as the ED physician is not solely making these decisions. Analysis was limited to ED attending physicians who saw at least 50 patients meeting inclusion criteria in the preintervention period and at least 50 patients in the postintervention period of the study. High and low discharge rate physicians were defined as one standard deviation (SD) from the mean physician performance.
The physician-level analysis utilized mixed effects logistic regression with random effect for the physician to account for physician-level variation as well as a fixed effect for years since medical school graduation. Data were analyzed at the visit level with the secular trend variable calculated by week. Data analysis was performed with R (version 4.2.2).
2.6. Sensitivity Analysis
Given the imperfect nature of chief complaints in identifying patients with concern for ACS, we performed a sensitivity analysis for the primary outcomes utilizing all chief complaints in the dataset.
3. Results
3.1. Study Population
Our study population consisted of 189,677 ED visits after exclusions (Fig S1). The mean patient age was 53.4 years (SD 20.4), and 110,682 (58.4%) patients were female. There were no significant differences in age or language before or after the intervention. Of the 121,376 ED visits after the intervention, 6804 (5.6%) had the ADP explicitly clicked on overall with 799 (12.6%) occurring in the first month of the intervention (Fig 1). Summary of study demographics are shown in Table 1.
Figure 1.
Electronic health record-integrated clinical pathway use over time. Line graph illustrating pathway encounter count (blue line) and percent of all patients with concern for acute coronary syndrome whom the pathway was used (orange line) between February 2022 and February 2023.
Table 1.
Summary study demographics and pathway usage among patients presenting a chief complaint suggestive of acute coronary syndrome.
| Overall | Conventional troponin (pre) | High-sensitivity troponin (post) | |
|---|---|---|---|
| Total | 189,677 | 68,301b | 121,376b |
| Mean age (SD) | 53.38 (20.46) | 53.45 (20.34) | 53.34 (20.52) |
| Age group (y) | |||
| <65 | 128,593 (67.8%) | 46,446 (68%) | 82,147 (67.68%) |
| ≥65 | 61,084 (32.2%) | 21,855 (32%) | 39,229 (32.32%) |
| Sex | |||
| Female | 110,682 (58.35%) | 39,624 (58.01%) | 71,058 (58.54%) |
| Male | 78,995 (41.65%) | 28,677 (41.99%) | 50,318 (41.46%) |
| Racea | |||
| White non-Hispanic | 102,399 (53.99%) | 37,090 (54.3%) | 65,309 (53.81%) |
| Black non-Hispanic | 38,775 (20.44%) | 14,013 (20.52%) | 24,762 (20.4%) |
| Hispanic | 39,967 (21.07%) | 14,190 (20.78%) | 25,777 (21.24%) |
| Other/unknown | 8536 (4.5%) | 3008 (4.4%) | 5528 (4.55%) |
| Charlson Comorbidity Index | |||
| 0 | 92,643 (48.84%) | 32,928 (48.21%) | 59,715 (49.2%) |
| 1 | 50,312 (26.53%) | 18,228 (26.69%) | 32,084 (26.43%) |
| 2+ | 46,722 (24.63%) | 17,145 (25.1%) | 29,577 (24.37%) |
| Preferred language | |||
| English | 170,294 (89.78%) | 61,365 (89.84%) | 108,931 (89.75%) |
| Non-English | 19,383 (10.22%) | 6938 (10.16%) | 12,445 (10.25%) |
| Clinical pathway | |||
| Did not use pathway | 182,827 (96.39%) | 68,257 (99.93%) | 114,572 (94.39%) |
| Used pathway | 6850 (3.61%) | 46 (0.07%)c | 6804 (5.61%) |
| Pathway use in month 1 | 799 (12.6%) | - | 799 (12.6%) |
First listed race was extracted and analyzed.
Proportion of total emergency department arrivals with chief complaints suggestive of acute coronary syndrome was similar in pre/post-period (33.1% pre vs 34.2% post).
Forty-six pathway uses in the preintervention period were related to testing in the electronic health record production environment but not applied to actual patient care.
3.2. Primary Outcomes
In the preintervention period, 33,061 (48.4%) of patients underwent at least one troponin test compared to 52,382 (43.2%) in the postintervention period. The proportion of patients discharged from the ED remained similar between the 2 periods, with 45,436 (66.5%) discharged preintervention and 82,414 (67.9%) discharged postintervention. The proportion of patients placed in chest pain observation decreased from 713 (1.0%) preintervention to 855 (0.7%) postintervention.
The intervention was associated with lower odds of troponin ordering (OR 0.79; 95% CI [0.76, 0.82]), with no difference in odds of discharge from the ED (0.99; [0.95, 1.03]) or being placed in the CPC (0.83; [0.67, 1.01]). The use of the ADP during an ED visit was associated with lower odds of ED discharge (0.84; [0.79, 0.89]) and higher odds of CPC (9.16; [7.82, 10.71]). Abbreviated results available in Table 2 with full regression results available in Table S2.
Table 2.
Multivariable analysis of association between intervention and primary outcomes for patients with a chief complaint consistent with acute coronary syndrome.
| Outcome | Conventional troponin (pre) N = 68,301 |
High-sensitivity troponin (post) N =121,376 |
ORa | 95% CI |
|---|---|---|---|---|
| Troponin orderedb | 33,061 (48.4%) | 52,382 (43.2%) | 0.79 | 0.76, 0.82 |
| Discharge disposition | 45,436 (66.5%) | 82,414 (67.9%) | 0.99 | 0.95, 1.03 |
| CPC dispositionc | 713 (1.0%) | 855 (0.7%) | 0.83 | 0.67, 1.00 |
All models adjusted for (reference range): age (<65), sex (male), Charlson comorbidity index (CCI = 0), language (English), time trend.
The proportion of patients with a positive COVID-19 in the troponin-tested population was similar in the pre- and postintervention periods (5.9% pre vs 5.1% post)
Chest pain center (CPC) was calculated on the proportion of nondischarged patients and only included the 3 EDs with access to this resource.
3.3. Secondary Outcomes
Among patients with chief complaints consistent with ACS, the intervention was associated with increased odds of downstream echocardiograms (1.08; [1.01, 1.15]) and cardiac catheterization (1.21; [1.07, 1.38]). There was a decrease in cardiology consultations (0.88; [0.83, 0.93]) and no difference in rates of stress testing (1.01; [0.87, 1.17]). Abbreviated results available in Table 3 with full regression results available in Table S3.
Table 3.
Multivariable analysis of association between intervention and secondary outcomes for all patients with a chief complaint consistent with acute coronary syndrome.
| Conventional troponin (pre) | High-sensitivity troponin (post) | ORa | 95% CI | |
|---|---|---|---|---|
| Echocardiogram | 6847 (10.02%) | 12,015 (9.90%) | 1.08 | 1.01, 1.15 |
| Stress testing | 1194 (1.75%) | 1451 (1.20%) | 1.01 | 0.87, 1.17 |
| Cardiology consult | 8880 (13.0%) | 13,171 (10.85%) | 0.88 | 0.83, 0.93 |
| Coronary catheterization | 1362 (1.99%) | 2271 (1.87%) | 1.21 | 1.07, 1.38 |
All models adjusted for (reference range) age (<65), sex (male), Charlson comorbidity index (CCI = 1), language (English), time trend.
Among patients with a chief complaint consistent with ACS and a troponin sent, the intervention was associated with increased odds of downstream echocardiograms (1.21; [1.13, 1.30]), cardiology consults (1.07; [1.00, 1.15]), and cardiac catheterization (1.36; [1.19, 1.55]), with no difference in rates of stress testing (1.13; [0.97, 1.31]). The yield of troponin testing increased postimplementation with increased diagnosis of AMI (1.17; [1.001, 1.36]). Abbreviated results available in Table 4 with full regression results available in Table S4. Additionally, there was no significant change in the rate of delayed ACS diagnosis (0.14% pre vs 0.16% post, P = .71)
Table 4.
Multivariable analysis of association between intervention and secondary outcomes, including downstream testing, EHR-integrated clinical pathway use among patients with chief complaints consistent with acute coronary syndrome and troponin testing ordered.
| Conventional troponin (pre) | High-sensitivity troponin (post) | ORa | 95% CI | |
|---|---|---|---|---|
| AMI diagnosis (yield)d | 869 (2.63%) | 1684 (3.21%) | 1.17 | 1.001, 1.36 |
| Echocardiogram | 6257 (18.93%) | 10,775 (20.57%) | 1.21 | 1.13, 1.30 |
| Stress testing | 1187 (3.59%) | 1445 (2.76%) | 1.13 | 0.97, 1.31 |
| Cardiology consult | 7447 (22.53%) | 11,795 (22.52%) | 1.07 | 1.004, 1.15 |
| Coronary catheterization | 1287 (3.89%) | 2177 (4.16%) | 1.36 | 1.19, 1.55 |
| Among those with clinical pathway used | ||||
| Discharge disposition | 0 (0%) | 3907 (57.4%) | 0.84 | 0.79, 0.89 |
| CPC dispositionb | 2 (4.3%)c | 356 (5.2%) | 9.16 | 7.82, 10.7 |
AMI, acute myocardial infarction; CPC, Chest pain center; ED, emergency department; EHR, electronic health record.
All models adjusted for (reference range) age (<65), sex (male), Charlson comorbidity index (CCI = 1), language (English), time trend.
Chest pain center was calculated on the proportion of nondischarged patients and only included the 3 EDs with access to this resource.
Two pathway uses in the preintervention period were related to testing in the EHR production environment but not applied to actual patient care.
AMI diagnosis based on final hospital ICD10 code.
3.4. Sensitivity Analysis
Given chief complaints are an imperfect marker of patients with concern for ACS, a sensitivity analysis was performed utilizing all chief complaints without change in result. Of note, odds of CPC disposition was markedly increased with the use of the ADP (15.8 [13.6, 18.4]) (Table S5).
3.5. Secondary Analysis: Physician-Level Analysis
Given the educational efforts to implement the AHA/ACC guidelines and the associated ADP, we examined the primary outcomes by conducting a mixed model with random effect for physician. Of the 211 total attending physicians in the study period, 147 (69.6%) met inclusion criteria. Results were similar to the primary analysis with an observed decrease in troponin ordering (0.81 [0.77, 0.84]) and no difference in discharge disposition (0.97 [0.93, 1.02]) or use of the CPC (0.86 [0.70, 1.06]). When analyzing the subset of patients where the clinical pathway was used, there was no difference in discharge disposition (0.99 [0.93, 1.06]) and an increase in use of the CPC (8.62 [7.26, 10.23]). Full regression results available in Table S6.
Additionally, physicians with high preintervention discharge rates decreased their discharge rate by 3.06% (95% CI, −3.2%, −2.9%) relative to their peers. Physicians with low preintervention discharge rates increased their discharge rates by 3.11% (2.9%, 3.3%) relative to their peers.
Physicians with high preintervention CPC rates decreased their observation rate by 5.4% (−5.5%, −5.4%) relative to their peers. Physicians with low preintervention CPC rates increased their observation rates by 2.5% (2.4%, 2.5%) relative to their peers (Fig 2).
Figure 2.
Physician practice patterns. Secondary analysis displaying physician behavior before and after hs-cTnT implementation for troponin ordering (A), discharge rate (B), and chest pain center observation rate (C). Darker color represents overlap between pre- and postinterventions.
4. Limitations
Our study results should be interpreted in the context of its limitations. First, although the 9 EDs in our study included academic, community, and free-standing sites, all EDs were part of a single health system in the Northeast United States, and therefore, findings may not be generalizable to other regions. That said, this is one of the largest studies on the ED implementation of hs-cTnT to date, and the broad inclusion of physicians from different practice environments reflects real-world implementation of hs-cTnT. Second, given the multimodal implementation of the order panel, ADP, and automated reflex testing, we cannot isolate their individual effects and have collectively studied these in aggregate. That said, we attempted to subanalyze “ADP users” using access log data, but pathway users are still exposed to the order panel and automated reflex testing. Third, the population of ED patients in whom ACS should be considered is inherently challenging to identify, given both the variety of presentations of ACS and the variety of disease processes that can cause symptoms typical of ACS, such as chest pain and shortness of breath. Although we relied on AHA/ACC guidelines and ED physician consensus to identify high yield chief complaints, it is possible additional chart information would better classify these patients. Among a very large sample, individual chart review would be infeasible, and our sensitivity analysis of all chief complaints did not change the significance of the outcomes. Fourth, regarding downstream testing, cardiology consultation was based on the presence of an EHR note. Anecdotally, the cardiology service discussed cases on the phone without leaving documentation for a proportion of cases, and, therefore, this outcome may underestimate the true presence of consultations, which were nearly unchanged based on unadjusted numbers after hs-cTnT implementation. Fifth, the observed decreased in hs-cTnT ordering after the implementation may reflect nurse, trainee, and/or non-physician practitioner hesitation to order hs-cTnT in triage or prior to discussing the case with an attending. However, the persistent drop in hs-cTnT observed even many months after implementation suggest low-value testing was avoided as clinician comfort with new interventions is expected to increase with time. Sixth, as had been noted previously, the outcome of AMI diagnosis may rely in part on troponin testing results and therefore subject to incorporation bias leading to an overestimation of the postintervention diagnostic yield.28,29 Finally, the 2021 AHA/ACC guidelines were released 3 months before the implementation of hs-cTnT, and although they were publicly accessible, there was no formal educational campaign or quality improvement initiative related to the guidelines or to conventional troponin in our health system during that period. All educational efforts around these guidelines, including presentations, workflow discussions, and clinician training, were conducted specifically in preparation for the hs-cTnT implementation and introduction of the ADP. Thus, it is unlikely that exposure to the guidelines alone meaningfully altered physician behavior prior to the intervention period.
5. Discussion
In this real-world evaluation of an EHR-integrated implementation of hs-cTnT for ED patients presenting with ACS-like symptoms, we found notable improvement in diagnostic efficiency alongside minimal changes in utilization. We found that overall cardiac enzyme testing decreased and, in cases in which patients were not discharged, rates of AMI diagnosis increased, and the rate of CPC observation increased especially when the ADP was directly accessed. Furthermore, we also found that the entire implementation of hs-cTnT was associated with reduced physician-level variation in discharge rates and their use of CPC observation.
This study represents a contemporary, post-COVID implementation of hs-cTnT that extends prior work conducted before the pandemic while reflecting the current acute care environment. Our results are consistent with prior work demonstrating decreased cardiac enzyme testing after implementation.30, 31, 32 We hypothesize this decrease is partly attributable to discontinuation of POC troponin from pre-existing order sets and standing nursing orders, which may decrease low-value testing. In fact, POC troponin was used a total of 12 times in the immediate aftermath of hs-cTnT implementation and within 2 months was retired altogether. Alternatively, anecdotally, clinicians were apprehensive about indeterminate hs-cTnT values or lack of specificity in elderly patients or those with renal dysfunction which may have contributed to more selective testing approaches.33, 34, 35
As part of the hs-cTnT implementation process, and consistent with AHA/ACC guidelines, we implemented an EHR-integrated clinical pathway, which provided guidance on interpretation and management of indeterminate values. Although direct clinical pathway usage was low overall, the order panel-generated automatic hs-cTnT orders based on results and deltas, so patients received appropriate hs-cTnT serial testing independent of pathway usage but consistent with pathway recommendations. The decrease in clinical pathway usage with time (Fig 1) suggests that clinical staff may have retained knowledge from the educational sessions, and/or there was a learning effect from the pathway. This might have occurred through informal36 or experiential learning on shift37 where one pathway usage is remembered and applied over multiple patients (one physician working multiple shifts in a row or attending to resident teaching). Among the patients for whom the ADP was used, there was an increase in use of the CPC observation unit. However, across the entire cohort, we observed a modest decrease in overall observation unit utilization after hs-cTnT implementation. This finding directly challenges the concern that hs-cTnT would lead to more conservative decision-making and unnecessary observation due to detectable low-level troponin elevations. This suggests that the ADP helped risk stratify patients appropriate for expedited cardiac testing compared with potentially discharging or placing an inpatient admission. The use of observation can help preserve already limited inpatient beds for sicker patients, reduce length of stay, improve patient satisfaction, and better match resources to patients.38, 39, 40, 41, 42
Of note, our findings contrast somewhat those of 2 recent pre-post analyses examining the association between adoption of hs-cTnT and downstream testing in United States EDs.43,44 Specifically, prior work has reported no significant difference in initial troponin testing and decrease or no change in downstream testing. These differences in testing are likely a result of more restrictive inclusion criteria of just patients presenting with chief complaint of chest pain which make up 17% of our data, and thus, our work may be a more generalizable assessment. Furthermore, post-COVID access to care is more difficult with average cardiology appointment wait times increasing 26% since 2017,45 which may increase pressure on clinicians to perform further cardiac risk stratification at the index visit.
With respect to ED admission decisions, we did not see a significant change in ED discharge rate despite the overall decrease in troponin testing. We hypothesize there is a co-occurring increase in identification of myocardial injury in some patients resulting in increased admission alongside a decrease in low-value testing. This is supported by the increased odds of AMI diagnosis associated with the intervention and consistent with prior work.46 Taken in context with no change in overall hospitalizations, these findings could be interpreted as showing an increase in appropriate hospitalization and identification of more people with ACS who may have previously been discharged. Conversely, the decrease in overall troponin testing without an increase in delayed ACS diagnosis suggests avoidance of low-value testing. It is also possible that the decreased testing rate was offset somewhat by hs-cTnT’s relatively poor specificity for AMI, particularly in elderly patients and those with end-stage renal disease, resulting in some unnecessary observations or admissions in patients with noncoronary etiologies of hs-cTnT elevations.33, 34, 35 Until validated age- and disease-specific hs-cTnT thresholds are established, it may be prudent to err on the side of caution and pursue further diagnostic testing in patients with hs-cTnT elevations and suspected ACS.
Although overall ED discharge rates remained unchanged, our secondary analysis revealed notable improvement in reducing physician-level variation in testing and ED admission decisions. Specifically, there was movement toward the average for physicians who, before the intervention, had been more likely to admit or more likely to discharge. Prior research has shown that in evaluated patients with suspected ACS, ED physicians both over-test low-risk patients and under-test high-risk patients and that there is a significant practice variation across physicians.47,48 The postintervention trend of physicians at the extremes moving toward the average practice pattern may represent a meaningful improvement in diagnostic efficiency represented by better testing of appropriate patients.
In summary, the transition to hs-cTnT and introduction of a clinical pathway was associated with a decrease in troponin testing, increase in some downstream testing, and a concurrent increase in diagnostic yield (AMI diagnosis) without appreciable impact upon ED disposition. Physicians with preintervention outlier practice moved toward department norms postintervention. Together, these findings suggest the intervention promoted care standardization and diagnostic quality for suspected ACS in the ED.
Author Contributions
RBS, MI, AKV, DJR, and SP performed the study concept and design. RBS, MI, and CR acquired and analyzed the data and drafted the initial manuscript. RBS, MI, CR, SP, RAT, BS, NRD, DJR, and AKV performed the data interpretation and critical revision of the manuscript for important intellectual content.
Funding and Support
By JACEP Open policy, all authors are required to disclose any and all commercial, financial, and other relationships in any way related to the subject of this article as per ICMJE conflict of interest guidelines (see www.icmje.org). The authors have stated that no such relationships exist.
Conflict of Interest
All authors have affirmed they have no conflicts of interest to declare.
Acknowledgments
We thank the Yale New Haven Health System members of clinical consensus groups, Laboratory Medicine, Cardiovascular Medicine, and Emergency Medicine, for their involvement building the clinical pathway. Thank you to the members of the Joint Data Analytic Team (JDAT) for their assistance with data collection. Thank you to Wendy Sun, Constantin Radu, Hazar Khidir, and Ryan Koski-Vacirca for their participation as a consensus group determining chief complaints to be considered part of acute coronary syndrome.
Footnotes
Presented at ACEP Scientific Assembly on October 10, 2023.
Supervising Editor: Bory Kea, MD, MCR
Supplementary material associated with this article can be found in the online version at https://doi.org/10.1016/j.acepjo.2025.100202
Supplementary Materials
References
- 1.National Center for Health Statistics: National Hospital Ambulatory Medical Care Survey: 2022 Emergency Department Summary Table https://ftp.cdc.gov/pub/Health_Statistics/NCHS/Dataset_Documentation/NHAMCS/doc22-ed-508.pdf
- 2.Gulati M., Levy P.D., Mukherjee D., Amsterdam E., Bhatt D.L., Birtcher K.K., et al. 2021 AHA/ACC/ASE/CHEST/SAEM/SCCT/SCMR guideline for the evaluation and diagnosis of chest pain: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation. 2021;144(22):e368–e454. doi: 10.1161/CIR.0000000000001029. [DOI] [PubMed] [Google Scholar]
- 3.Virani S.S., Alonso A., Benjamin E.J., Bittencourt M.S., Callaway C.W., Carson A.P., et al. Heart disease and stroke statistics-2020 update: a report from the American Heart Association. Circulation. 2020;141(9):e139–e596. doi: 10.1161/cir.0000000000000757. [DOI] [PubMed] [Google Scholar]
- 4.Collinson P. High sensitivity troponin, analytical advantages, clinical benefits and clinical challenges—an update. Clin Biochem. 2021;91:1–8. doi: 10.1016/j.clinbiochem.2021.02.001. [DOI] [PubMed] [Google Scholar]
- 5.Reichlin T., Hochholzer W., Bassetti S., Steuer S., Stelzig C., Hartwiger S., et al. Early diagnosis of myocardial infarction with sensitive cardiac troponin assays. N Engl J Med. 2009;361(9):858–867. doi: 10.1056/NEJMoa0900428. [DOI] [PubMed] [Google Scholar]
- 6.Sherwood M.W., Kristin Newby L. High-sensitivity troponin assays: evidence, indications, and reasonable use. J Am Heart Assoc. 2014;3(1) doi: 10.1161/jaha.113.000403. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Roffi M., Patrono C., Collet J.-P., Mueller C., Valgimigli M., Andreotti F., et al. 2015 ESC Guidelines for the management of acute coronary syndromes in patients presenting without persistent ST-segment elevation: task force for the management of acute coronary syndromes in patients presenting without persistent ST-segment elevation of the European Society of Cardiology (ESC) Eur Heart J. 2016;37(3):267–315. doi: 10.1093/eurheartj/ehv320. [DOI] [PubMed] [Google Scholar]
- 8.Boeddinghaus J., Nestelberger T., Twerenbold R., Wildi K., Badertscher P., Cupa J., et al. Direct comparison of 4 very early rule-out strategies for acute myocardial infarction using high-sensitivity cardiac troponin I. Circulation. 2017;135(17):1597–1611. doi: 10.1161/circulationaha.116.025661. [DOI] [PubMed] [Google Scholar]
- 9.Kaier T.E., Twerenbold R., Lopez-Ayala P., Nestelberger T., Boeddinghaus J., Alaour B., et al. A 0/1h-algorithm using cardiac myosin-binding protein C for early diagnosis of myocardial infarction. Eur Heart J Acute Cardiovasc Care. 2022;11(4):325–335. doi: 10.1093/ehjacc/zuac007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Nowak R.M., Christenson R.H., Jacobsen G., McCord J., Apple F.S., Singer A.J., et al. Performance of novel high-sensitivity cardiac troponin I assays for 0/1-hour and 0/2- to 3-hour evaluations for acute myocardial infarction: results from the HIGH-US study. Ann Emer Med. 2020;76(1):1–13. doi: 10.1016/j.annemergmed.2019.12.008. [DOI] [PubMed] [Google Scholar]
- 11.McCarthy C., Li S., Wang T.Y., Raber I., Sandoval Y., Smilowitz N.R., et al. Implementation of high-sensitivity cardiac troponin assays in the United States. J Am Coll Cardiol. 2023;81(3):207–219. doi: 10.1016/j.jacc.2022.10.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Bates D.W., Kuperman G.J., Wang S., Gandhi T., Kittler A., Volk L., et al. Ten commandments for effective clinical decision support: making the practice of evidence-based medicine a reality. J Am Med Inform Assoc. 2003;10(6):523–530. doi: 10.1197/jamia.M1370. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Lomas J., Sisk J.E., Stocking B. From evidence to practice in the United States, the United Kingdom, and Canada. Milbank Q. 1993;71(3):405–410. doi: 10.2307/3350408. [DOI] [PubMed] [Google Scholar]
- 14.Sampson U.K.A., McGlynn E.A., Perlin J.B., Frisse M.E., Arnold S.B., Benz E.J., Jr., et al. Advancing the science of healthcare service delivery: the NHLBI corporate healthcare leaders’ panel. Glob Heart. 2018;13(4):339–345. doi: 10.1016/j.gheart.2018.09.508. [DOI] [PubMed] [Google Scholar]
- 15.Cabana M.D., Rand C.S., Powe N.R., Wu A.W., Wilson M.H., Abboud P.A., et al. Why don’t physicians follow clinical practice guidelines? A framework for improvement. JAMA. 1999;282(15):1458–1465. doi: 10.1001/jama.282.15.1458. [DOI] [PubMed] [Google Scholar]
- 16.Davis D.A., Taylor-Vaisey A. Translating guidelines into practice. A systematic review of theoretic concepts, practical experience and research evidence in the adoption of clinical practice guidelines. CMAJ. 1997;157(4):408–416. [PMC free article] [PubMed] [Google Scholar]
- 17.Seckler E., Regauer V., Rotter T., Bauer P., Müller M. Barriers to and facilitators of the implementation of multi-disciplinary care pathways in primary care: a systematic review. BMC Fam Pract. 2020;21(1):113. doi: 10.1186/s12875-020-01179-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Le Grand Rogers R., Narvaez Y., Venkatesh A.K., Fleischman W., Hall M.K., Taylor R.A., et al. Improving emergency physician performance using audit and feedback: a systematic review. Am J Emerg Med. 2015;33(10):1505–1514. doi: 10.1016/j.ajem.2015.07.039. [DOI] [PubMed] [Google Scholar]
- 19.Osheroff J.A., Teich J.M., Middleton B., Steen E.B., Wright A., Detmer D.E. A roadmap for national action on clinical decision support. J Am Med Inform Assoc. 2007;14(2):141–145. doi: 10.1197/jamia.M2334. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Ostropolets A., Zhang L., Hripcsak G. A scoping review of clinical decision support tools that generate new knowledge to support decision making in real time. J Am Med Inform Assoc. 2020;27(12):1968–1976. doi: 10.1093/jamia/ocaa200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Rotter T., Kinsman L., James E., Machotta A., Gothe H., Willis J., et al. Clinical pathways: effects on professional practice, patient outcomes, length of stay and hospital costs. Cochrane Database Syst Rev. 2010;(3) doi: 10.1002/14651858.CD006632.pub2. [DOI] [PubMed] [Google Scholar]
- 22.Braithwaite J., Ludlow K., Testa L., Herkes J., Augustsson H., Lamprell G., et al. Built to last? The sustainability of healthcare system improvements, programmes and interventions: a systematic integrative review. BMJ Open. 2020;10(6) doi: 10.1136/bmjopen-2019-036453. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Burke R.E., Marang-van de Mheen P.J. Sustaining quality improvement efforts: emerging principles and practice. BMJ Qual Saf. 2021;30(11):848–852. doi: 10.1136/bmjqs-2021-013016. [DOI] [PubMed] [Google Scholar]
- 24.Iqbal U., Siddique O., Jameel A., Anwar H., Chaudhary A. Prognostic significance of elevated cardiac troponin in acute gastrointestinal bleeding. Gastroenterol Res. 2017;10(4):238–243. doi: 10.14740/gr893w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Vasile V.C., Chai H.S., Abdeldayem D., Afessa B., Jaffe A.S. Elevated cardiac troponin T levels in critically ill patients with sepsis. Am J Med. 2013;126(12):1114–1121. doi: 10.1016/j.amjmed.2013.06.029. [DOI] [PubMed] [Google Scholar]
- 26.Sangal R.B., Liu R.B., Cole K.O., Rothenberg C., Ulrich A., Rhodes D., et al. Implementation of an electronic health record integrated clinical pathway improves adherence to COVID-19 hospital care guidelines. Am J Med Qual. 2022;37(4):335–341. doi: 10.1097/jmq.0000000000000036. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Charlson M., Szatrowski T.P., Peterson J., Gold J. Validation of a combined comorbidity index. J Clin Epidemiol. 1994;47(11):1245–1251. doi: 10.1016/0895-4356(94)90129-5. [DOI] [PubMed] [Google Scholar]
- 28.Keller T., Zeller T., Ojeda F., Tzikas S., Lillpopp L., Sinning C., et al. Serial changes in highly sensitive troponin I assay and early diagnosis of myocardial infarction. JAMA. 2011;306(24):2684–2693. doi: 10.1001/jama.2011.1896. [DOI] [PubMed] [Google Scholar]
- 29.Zhelev Z., Hyde C., Youngman E., Rogers M., Fleming S., Slade T., et al. Diagnostic accuracy of single baseline measurement of Elecsys Troponin T high-sensitive assay for diagnosis of acute myocardial infarction in emergency department: systematic review and meta-analysis. BMJ. 2015;350:h15. doi: 10.1136/bmj.h15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Aloe R., Lippi G., Di Pietro M., Bonfanti L., Dipalo M., Comelli I., et al. Improved efficiency and cost reduction in the emergency department by replacing contemporary sensitive with high-sensitivity cardiac troponin immunoassay. Acta Biomed. 2019;90(4):614–620. doi: 10.23750/abm.v90i4.8769. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Hodgson N.R., Kunze K.L., Lim E.S., Maher S.A., Traub S.J. Adoption of high-sensitivity troponin testing and emergency physician ordering behavior. West J Emerg Med. 2022;23(3):439–442. doi: 10.5811/westjem.2022.2.54242. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Furmaga J., McDonald S.A., Hall H.M., Muthukumar A., Willett K., Basit M., et al. Impact of high-sensitivity troponin testing on operational characteristics of an urban emergency department. Acad Emerg Med. 2021;28(1):114–116. doi: 10.1111/acem.13956. [DOI] [PubMed] [Google Scholar]
- 33.Freund Y., Chenevier-Gobeaux C., Bonnet P., Claessens Y.E., Allo J.C., Doumenc B., et al. High-sensitivity versus conventional troponin in the emergency department for the diagnosis of acute myocardial infarction. Crit Care. 2011;15(3) doi: 10.1186/cc10270. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Kumar N., Michelis M.F., DeVita M.V., Panagopoulos G., Rosenstock J.L. Troponin I levels in asymptomatic patients on haemodialysis using a high-sensitivity assay. Nephrol Dial Transplant. 2011;26(2):665–670. doi: 10.1093/ndt/gfq442. [DOI] [PubMed] [Google Scholar]
- 35.Lowry M.T.H., Doudesis D., Wereski R., Kimenai D.M., Tuck C., Ferry A.V., et al. Influence of age on the diagnosis of myocardial infarction. Circulation. 2022;146(15):1135–1148. doi: 10.1161/circulationaha.122.059994. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Marsick V.J., Volpe M., Watkins K.E. Theory and practice of informal learning in the knowledge era. Adv Dev Hum Resour. 1999;1(3):80–95. doi: 10.1177/152342239900100309. [DOI] [Google Scholar]
- 37.Eraut M. Learning from other people in the workplace. Oxf Rev Educ. 2007;33(4):403–422. doi: 10.1080/03054980701425706. [DOI] [Google Scholar]
- 38.Baugh C.W., Bohan J.S. Estimating observation unit profitability with options modeling. Acad Emerg Med. 2008;15(5):445–452. doi: 10.1111/j.1553-2712.2008.00082.x. [DOI] [PubMed] [Google Scholar]
- 39.Farkouh M.E., Smars P.A., Reeder G.S., Zinsmeister A.R., Evans R.W., Meloy T.D., et al. A clinical trial of a chest-pain observation unit for patients with unstable angina. Chest Pain Evaluation in the Emergency Room (CHEER) Investigators. N Engl J Med. 1998;339(26):1882–1888. doi: 10.1056/nejm199812243392603. [DOI] [PubMed] [Google Scholar]
- 40.Gomez M.A., Anderson J.L., Karagounis L.A., Muhlestein J.B., Mooers F.B. An emergency department-based protocol for rapidly ruling out myocardial ischemia reduces hospital time and expense: results of a randomized study (ROMIO) J Am Coll Cardiol. 1996;28(1):25–33. doi: 10.1016/0735-1097(96)00093-9. [DOI] [PubMed] [Google Scholar]
- 41.Mikhail M.G., Smith F.A., Gray M., Britton C., Frederiksen S.M. Cost-effectiveness of mandatory stress testing in chest pain center patients. Ann Emerg Med. 1997;29(1):88–98. doi: 10.1016/s0196-0644(97)70314-7. [DOI] [PubMed] [Google Scholar]
- 42.Roberts R.R., Zalenski R.J., Mensah E.K., Rydman R.J., Ciavarella G., Gussow L., et al. Costs of an emergency department-based accelerated diagnostic protocol vs hospitalization in patients with chest pain: a randomized controlled trial. JAMA. 1997;278(20):1670–1676. doi: 10.1001/jama.1997.03550200046030. [DOI] [PubMed] [Google Scholar]
- 43.Ford J.S., Chaco E., Tancredi D.J., Mumma B.E. Impact of high-sensitivity cardiac troponin implementation on emergency department length of stay, testing, admissions, and diagnoses. Am J Emerg Med. 2021;45:54–60. doi: 10.1016/j.ajem.2021.02.021. [DOI] [PubMed] [Google Scholar]
- 44.Ganguli I., Cui J., Thakore N., Orav E.J., Januzzi J.L., Baugh C.W., et al. Downstream cascades of care following high-sensitivity troponin test implementation. J Am Coll Cardiol. 2021;77(25):3171–3179. doi: 10.1016/j.jacc.2021.04.049. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Hawkins M. Survey of physician appointment wait times and medicare and medicaid acceptance rates. 2022. https://www.merritthawkins.com/uploadedFiles/MerrittHawkins/Content/News_and_Insights/Articles/mha-2022-wait-time-survey.pdf
- 46.Reichlin T., Twerenbold R., Reiter M., Steuer S., Bassetti S., Balmelli C., et al. Introduction of high-sensitivity troponin assays: impact on myocardial infarction incidence and prognosis. Am J Med. 2012;125(12):1205–1213.e1. doi: 10.1016/j.amjmed.2012.07.015. [DOI] [PubMed] [Google Scholar]
- 47.Mullainathan S., Obermeyer Z. Diagnosing physician error: a machine learning approach to low-value health care. Q J Econ. 2022;137(2):679–727. doi: 10.1093/qje/qjab046. [DOI] [Google Scholar]
- 48.Natsui S., Sun B.C., Shen E., Redberg R.F., Ferencik M., Lee M.S., et al. Higher emergency physician chest pain hospitalization rates do not lead to improved patient outcomes. Circ Cardiovasc Qual Outcomes. 2021;14(1) doi: 10.1161/circoutcomes.119.006297. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.