|Home | About | Journals | Submit | Contact Us | Français|
ST-segment elevation myocardial infarction (STEMI) care is time-dependent. Many STEMI patients require inter-hospital helicopter transfer for percutaneous coronary intervention (PCI) if ground emergency medical services (EMS) initially transport the patient to a non-PCI center. This investigation models potential time savings of ground EMS requests for helicopter EMS (HEMS) transport of a STEMI patient directly to a PCI center, rather than usual transport to a local hospital with subsequent transfer.
Data from a multicenter retrospective chart review of STEMI patients transferred for primary PCI by a single HEMS agency over 12 months were used to model medical contact to balloon times (MCTB) for two scenarios: a direct-to-scene HEMS response, and hospital rendezvous after ground EMS initiation of transfer.
Actual MCTB median time for 36 hospital-initiated transfers was 160 minutes (range 116 to 321 minutes). Scene response MCTB median time was estimated as 112 minutes (range 69 to 187 minutes). The difference in medians was 48 minutes (95% CI = 33 to 62 minutes). Hospital rendezvous MCTB median time was estimated as 113 minutes (range 74 to 187 minutes). The difference in medians was 47 minutes (95% CI = 32 to 62 minutes). No patient had an actual MCTB time of less than 90 minutes; in the scene response and hospital rendezvous scenarios, two out of 36 (6%) and three out of 36 (8%), respectively, would have had MCTB times under 90 minutes.
In this setting, ground EMS initiation of HEMS transfers for STEMI patients has the potential to reduce MCTB time, but most patients will still not achieve MCTB time of less than 90 minutes.
Care of the ST-segment elevation myocardial infarction (STEMI) patient is time-critical, and the goal of achieving reperfusion within 90 minutes of initial medical contact has been established.1-11 Evaluation of the medical contact to balloon time (MCTB), instead of the door to balloon time, acknowledges the role that emergency medical services (EMS) plays in achieving timely reperfusion, and prompts measures to minimize prehospital time delays that may otherwise go unnoticed. Timely fibrinolytic therapy with follow-up percutaneous coronary intervention (PCI) is accepted, yet many individuals presenting to non-PCI capable facilities are transferred to a PCI-capable facility for definitive reperfusion without receiving fibrinolytic treatment.12 Helicopter EMS (HEMS) agencies often make this transfer in an attempt to meet the 90 minute MCTB goal.13-18 However, hospital-to-hospital HEMS transfer does not guarantee timely PCI, thereby placing the patient at risk of delayed reperfusion.12 Alternative strategies to reduce the time to reperfusion in patients being transferred to a PCI-capable facility are needed. Studies have shown the interval between STEMI diagnosis and HEMS request can be longer than 30 minutes,12,19 and that such delays affect mortality,20 making this interval a possible target for intervention.
Currently, many ground-based EMS systems minimize time from diagnosis to PCI by transmitting an electrocardiogram (ECG) from the scene to a physician for rapid interpretation, and in those patients confirmed to have ST-segment elevation, travel directly to a PCI-capable facility even if it means bypassing closer non-PCI hospitals.9,21-23 However, if transport time is expected to be lengthy (e.g. more than 30 minutes longer than transport to the closest hospital) alternative approaches may be needed. Some have proposed a “direct-to-scene” HEMS activation by ground EMS, analogous to the care of trauma patients.24 Another alternative would be for ground EMS to rendezvous with HEMS at a local non-PCI capable facility for immediate transfer of the patient to a PCI-capable facility. Evidence to support implementation of this care-delivery model, however, is not reported. It is unknown if ground EMS activation of HEMS would increase the proportion of patients in which the target of 90 minutes to reperfusion can be achieved.
Our study uses HEMS flight mission data combined with hospital operational data to model the MCTB times that could be achieved under direct-to-scene response and hospital rendezvous scenarios. We hypothesized that ground EMS request for HEMS transport to a PCI center could offer significant time savings.
This project was a planned secondary analysis of data from a multi-center retrospective chart review. This study was approved by the institutional review boards or ethics committees of each of the participating hospitals.
The greater Cincinnati area has a dense urban core, but the surrounding areas of northern Kentucky, southwest Ohio, and southeast Indiana quickly transition to rural (using United States Census, Office of Management and Budget, and Economic Research Service definitions25-27). Transferring facilities were rural and suburban emergency departments (EDs) without on-site interventional cardiology. Receiving hospitals were suburban and urban facilities having interventional cardiology capability. Three competing regional hospital systems were represented. No pre-determined referral relationships or regionally accepted triage of STEMI transfers were established at the time of the study.
Subjects in the present analyses include those who did not receive fibrinolytics prior to HEMS transfer and who had complete prehospital, transferring hospital, and receiving hospital records. Complete EMS records are required to model the theoretical response scenarios, and only those receiving primary PCI are included to allow MCTB calculations.
All subjects were transported by a single hospital-based helicopter EMS system (University AirCare, Cincinnati, OH) providing daily coverage with two Eurocopter BK117 aircraft. One aircraft was centrally located at the sponsor hospital, an urban Level I trauma center with PCI capability, and was staffed 24 hours daily. The second aircraft was based approximately 25 miles away at an affiliated suburban hospital without PCI capability, and was staffed 12 hours daily between noon and midnight. The system provided coverage for a 71,000 square-mile area (150 mile radius), encompassing urban, suburban, and rural populations. Approximately 60% of the system’s flights were inter-hospital critical care transfers. An emergency physician and a critical care flight nurse staffed all flights.
Decisions regarding transport modality (HEMS versus ground EMS) and method of reperfusion (primary fibrinolysis versus PCI) were made by the treating ED and cardiology clinicians.
We abstracted charts from STEMI patients requiring helicopter EMS inter-hospital transfer for primary or rescue PCI, the methods for which have been previously reported, and follow accepted recommendations for chart reviews.12,28 Briefly, EMS run sheets and both transferring and receiving hospital charts of all inter-hospital STEMI transfers by the HEMS service during 2007 were reviewed and underwent dual data abstraction using structured case report forms with pre-specified data definitions. Data collected included that necessary to determine treatment and transport intervals, including ground EMS, referring hospital, HEMS course, and PCI segments.17 Abstractors were trained by a senior investigator; discrepancies in case report forms underwent an adjudication process whereby a senior investigator independently review the chart to clarify data queries. Patients who were transferred from an ED at a non-PCI hospital to another hospital’s cardiac catheterization lab for PCI were included.
The sequences of events required for transfer are depicted in Figure 1. The helicopter location at the time of request, ground EMS scene, transferring hospital, and receiving hospital locations were geocoded, and straight-line distances for each flight segment were calculated.29 Flight speeds were calculated using actual flight times and distances obtained from the HEMS computer-assisted dispatch system (AeroMed, Innovative Engineering, Lebanon, NH). Flight speed and distance data were then used to compute flight times for theoretical flight segments under the modeled scenarios. To account for variables such as wind and weather conditions, actual flight path taken, lift-off and landing delays, and aircraft used, flight times were computed independently for each patient using the flight speeds for that patient’s encounter. Each actual flight was modeled under both scenarios.
The direct-to-scene model was designed to simulate the following sequence: ground EMS diagnoses STEMI and requests HEMS to respond directly to the location of the patient, where ground EMS would await HEMS arrival. HEMS would then directly transport the patient to a PCI center for intervention. In this model, the estimated time taken for the helicopter to reach the scene was computed based on the actual flight speed from helicopter location at dispatch to the transferring hospital. The estimated time taken for the helicopter to travel from scene to a receiving hospital was based on the actual flight speed of the inter-hospital transfer. Actual helicopter ground times for activation and patient transfers, as well as actual PCI time intervals, were used.
The hospital rendezvous model was designed to simulate the following sequence: ground EMS diagnoses STEMI and requests HEMS to meet at the local non-PCI hospital’s helipad. Ground EMS would drive to the local hospital while the helicopter flew to rendezvous, but the patient would not be taken into the local ED. In this scenario, actual ground transport and flight times were used. In instances where ground EMS would arrive at the local hospital before the helicopter, a wait time was included in the model. Actual helicopter ground times for activation and patient transfers, as well as actual PCI time intervals, were used.
For all computations, medical contact was defined as the time at which ground EMS arrived on the scene.1,30 Balloon time was defined as the time of initial reperfusion during PCI intervention, which could include balloon inflation, thrombectomy, and stent deployment.
In this exploratory analysis, actual missions flown served as internal controls for each modeled scenario. More complex situations, such as helicopter non-availability, were not factored into the models.
Differences in medians were computed, and 95% confidence intervals (CI) of the differences in medians were estimated using the method of Bonett and Price.31 Data were managed within Microsoft Excel (Microsoft Corp., Redmond, WA), and statistical analyses were conducted using SPSS 18.0 for Windows (SPSS Inc., Chicago, IL).
Of 187 patients with a STEMI transferred by HEMS during the study period, 69 used ground EMS. Of those, 15 had missing EMS times, 11 underwent primary fibrinolysis, and 7 did not receive an intervention during cardiac catheterization (four no obstructive lesion, two referred for coronary artery bypass graft, and one procedure terminated early due to patient instability). The remaining 36 received primary PCI and had complete EMS records for review to allow computation of MCTB times, and thus met criteria for inclusion in this analysis (Figure 2). The mean age of the included subjects was 61 years (SD ±10 years); 31 of 36 (86%) were white, and 24 of 36 (67%) were male (Table 1).
Nine referring non-PCI capable hospitals and six receiving PCI capable facilities were included. The median straight-line distances between helicopter base and transferring hospital was 19.6 miles (range 5.5 to 50.8 miles), between hospitals was 19.6 miles (range 4.5 to 50.8 miles), between helicopter base and ground EMS scene was 21.9 miles (range 6.0 to 45.6 miles), and between ground EMS scene and receiving PCI hospital was 21.3 miles (range 4.0 to 42.6 miles).
For patients included in this analysis, the median actual MCTB time was 160 minutes (range 116 to 321 minutes); no patients achieved reperfusion within 90 minutes. Actual and calculated time intervals for the various journey segments are shown in Table 2. Figure 3 demonstrates potential time savings for both theoretical models.
The projected median time from MCTB for the direct-to-scene response model was 112 minutes (range 69 to 187 minutes). The difference in medians vs actual data was 48 minutes (95% CI = 33 to 62 minutes). Using the direct-to-scene approach, two subjects (6%) would have achieved reperfusion within 90 minutes.
The projected median time from MCTB for the hospital rendezvous model was 113 minutes (range 74 to 187 minutes). The difference in medians vs actual data was 47 minutes (95% CI = 32 to 62 minutes). Using the hospital rendezvous approach, three subjects (8%) would have achieved reperfusion within 90 minutes.
Our modeling shows that HEMS activation by ground EMS personnel has the potential to facilitate the timely transfer of a STEMI patient to a PCI center. While the concept of implementing such a strategy might seem straightforward, there may be local and regional operational and logistical considerations. For example, when multiple PCI centers are available, it might be necessary to prioritize transfer routes. Similarly, operational constraints may mean an interventional cardiologist is unable to accept a patient from an EMS or HEMS agency, and alternatives would need to be sought. There is also likely an economic effect of bypassing local non-PCI hospitals. Potential cost concerns both for these hospitals and for patients should be clearly outweighed by the potential benefits of reduced time to reperfusion, and rigorous oversight would be needed to ensure appropriate HEMS activation. These challenges, though, are not unlike those that have been overcome in ongoing endeavors to regionalize care of STEMI patients.9,13,21,32-34
One concern might be that with a direct transfer of any kind, the patient arrives at the receiving hospital prior to arrival of the catheterization laboratory team, increasing the need for short-term boarding of the patient in the ED prior to catheterization. Based on our data, the projected median time from HEMS request to arrival in the catheterization laboratory was greater than 30 minutes. After normal hours, cardiac catheterization lab personnel are expected to respond within the same time frame, therefore having the patient arrive earlier than the interventional cardiology team may not be common.
Certain assumptions were made in designing our theoretical models of alternate helicopter dispatch strategies for STEMI patients to be taken directly to a PCI center for primary PCI. By using actual times for each flight segment (out/back for each flight), we are able to capture the flight-specific variables that are otherwise not quantifiable. These variables include headwind/tailwind, day/night assessments of landing zone, weather-based delays in lift-off, and route taken. We apply the actual time taken to a standard calculation of the straight-line distance to allow comparisons; in this case, if the actual distance flown was 10% longer, the time would be increased and depict a slower flight speed. Applying this slower speed to the modeled distances will account for similarly affected routes. We have elected not to apply other correction factors to the models; the primary purpose of this work is not to most accurately estimate distances or speeds, but to use the real data to estimate time intervals.
Both of the modeled strategies could result in meaningful time savings, but neither strategy was superior to the other. The stable landing zone offered by the hospital rendezvous model may make this a preferred method over direct-to-scene response. Community hospital landing pads are currently used without Emergency Medical Treatment and Active Labor Act (EMTALA) implications in trauma and other specialized care, which would avoid requiring EMS to proceed into the ED unnecessarily. However, in cases of unexpected cancellation or delay of the aircraft, or change in patient condition, the patient could be diverted into the community hospital without the delay in care that would occur if ground EMS waited on scene for helicopter arrival.
Ground EMS transport of a STEMI patient directly to a PCI facility is not commonly done in the suburban and rural areas that form the basis for these models, but could represent another strategy to improve reperfusion times. Accounting for the many unknown variables in modeling a ground transport scenario (e.g. traffic patterns, choice of route taken, use of lights and sirens, road and weather conditions) is beyond the scope of the present study. However, using Google’s driving direction software for an overly simplified estimate shows a potential to achieve even more rapid medical contact to balloon times. Median ground transport time from the scene to the PCI hospital is 41 minutes (range 12 to 64 minutes), and median MCTB is 98 minutes (range 57 to 169 minutes); 12 (33%) subjects could have achieved MCTB < 90 minutes. Although this strategy may be even quicker than HEMS utilization, only a minority of patients could receive perfusion within guideline recommendations, further underscoring the role of primary fibrinolysis with subsequent non-emergent transfer.
Individual communities should determine which reperfusion strategy is best based on unique community characteristics and then strive to minimize any delays to therapy as much as possible. Settling the debate of primary fibrinolysis versus primary PCI, as well as ground vs. air transport for primary PCI, is well beyond the scope of this exploration. It is clear that timely reperfusion, by either fibrinolysis or PCI, reduces morbidity and mortality.7,10 In some cases, prehospital fibrinolysis may be an option for treatment,35 although this strategy is not frequently employed in the United States. Follow-up PCI after primary fibrinolysis is indicated, which will require ultimate transport of STEMI patients to PCI centers; early post-fibrinolysis PCI may also convey better outcomes than routine PCI performed the next day.36 We offer our findings to give decision-makers objective information to design the best care delivery model for their communities, and our data add new knowledge about systems for transport for STEMI, even if we do not resolve the debate.
It should be noted that while time to reperfusion would likely be reduced if ground EMS were to activate HEMS, most subjects were still predicted to exceed the guideline-recommended 90 minute time to reperfusion. Therefore, continued consideration of fibrinolytic therapy at a non-PCI capable hospital as a mechanism for early reperfusion is critical. However, if primary PCI is preferred, ground EMS activation for HEMS transport directly to the PCI center may be a feasible strategy to facilitate timely reperfusion.
This study is a secondary analysis of existing data, and therefore is limited by the quality of that data. We used rigorous chart review methodology to minimize this limitation.28 In addition to this primary limitation, our modeling involved estimation of times based upon various assumptions. For example, helicopter ground time at a scene was assumed to be the same as actual time spent on the ground at a transferring hospital. In practice, this time may be shorter or longer because of differences in the complexity of patient care transfers at a scene versus in a hospital.
We chose to include only those patients who initially presented to the non-PCI hospital’s ED via EMS and underwent primary PCI so as to make the best estimates of potential MCTB improvements. While this does not allow pure intention-to-treat analysis by excluding those who were transferred for intended PCI but were ultimately not treated, our primary outcome of MCTB matches the outcome of interest by which many EMS agencies, EDs, and PCI centers are judged.
The data used to form these models were collected in 2007. Since that time, there have been growing initiatives to improve STEMI systems of care. However, many regions have yet to fully optimize a regional approach to expedient reperfusion, and our results will help inform planners responsible for such optimization. We are not aware of significant technological advances (i.e. faster helicopters, quicker PCI techniques) that would invalidate our models.
Retrospective analysis precludes re-creation of decision-making by ground EMS to transport to a non-PCI hospital, by the treating physician to choose primary PCI over fibrinolysis, or in choice of one PCI hospital over another. Because of this, some selection bias may be present in the studied cohort.
The sample size for this modeling study was small, limiting generalization to other geographic regions and clinical situations; small sample sizes may also compound any errors in estimating times and distances. To maximize effect, each actual flight served as an internal control for each of two scenarios. Transport time intervals in this cohort are similar to those previously reported in a review of almost 3,200 helicopter EMS inter-facility transfers of STEMI patients, suggesting that the studied HEMS agency performs similarly to others.17 However, this is the first report in the literature evaluating potential alternative dispatch scenarios, and the striking difference between actual and modeled times deserves further evaluation.
When considering wider regionalization of the care for patient with ST-segment elevation myocardial infarction, ground EMS activation of helicopter EMS for direct transport to a PCI-capable hospital may decrease the time to reperfusion when compared to inter-facility transfer. However, achieving a medical contact to balloon time within 90 minutes could remain elusive.
Funding Sources/Disclosures: This project was supported in part by an Institutional Clinical and Translational Science Award, NIH/NCRR Grant Number 5UL1RR026314-02
Prior Presentations: none