Door to thrombolysis: ER reorganization and reduced delays to acute stroke treatment

Immediate initiation of thrombolysis is imperative in acute ischemic stroke. Attention has been devoted to symptom recognition and rapid prehospital transport to decrease the symptom-to-needle time (SNT). Less attention has been paid to in-hospital delays (door-to-needle time [DNT]). No controlled studies exist of improving the access to stroke thrombolysis,¹ but increased use of emergency medical services (EMS) and direct referral to the stroke center have reduced pre- and in-hospital delays.^2–5 Awareness campaigns have promoted thrombolysis,^1,6,7 but their effect may be transient.⁷ Specific code stroke protocols and stroke teams have improved the efficiency of in-hospital care.^8,9

In 1999, we had a stroke code for EMS, a stroke team, and a stroke fellow serving the ER, but the DNT still accounted for >50% of SNT. CT took generally more than an hour to start, not 25 minutes as recommended by National Institute of Neurological Disorders and Stroke. This led us to restructure our ER, to streamline management, and facilitate CT access.

Methods.

The Helsinki University Central Hospital (HUCH) serves a catchment area of >1.4 million and has received all ischemic stroke patients in Helsinki since 1994, altogether 1,000 annually. In 1998, IV thrombolytic therapy was adopted as a standard management protocol. This was preceded by a training program among the EMS to improve symptom recognition.¹⁰ This increased the number of thrombolysis candidates transported to our ER and reduced the prehospital delay (symptom-to-door time [SDT]).

In 1999, stroke patients were triaged and managed by nurses responsible for all acute medical conditions. CT was performed in the Department of Radiology, one floor down and >100 meters from the ER, and transporting was done by lay personnel after phoned requests. Imaging needs competed with those of in-hospital patients. We undertook a guided quality management project, which led to identification of several bottlenecks.

The three main organizational steps were as follows: 1) The patients were triaged to be managed by a nurse team caring for only neurologic ER patients placed together. The nurses were trained in stroke monitoring. 2) The ambulance staff prenotified the ER of a thrombolysis candidate, which allowed initiation of CT preparation and laboratory tests. 3) A major rebuilding of our ER started in 2000, and a small CT unit was placed inside the ER to be accessed for priority ER patients. These steps were implemented stepwise, the first from 2000 to 2001 and the second from 2001 to 2003. Key performance indexes suggested only modest improvement during the reconstruction phase in 2000 to 2003, when the ER functioned in a temporary site. It was not until the final step in August 2003 when the ER was moved to the rebuilt premises housing the new CT unit that a major improvement occurred. Simultaneously, several novel digital patient data record systems were implemented.

Treatment delays were determined from patient records. In-hospital time points included ordering and starting CT. We also recorded the time of reporting the radiologic findings, although throughout the period, scans were evaluated by a stroke fellow directly from the CT console, often supported by a neuroradiologist. Retrospective analysis was performed for 100 consecutive patients admitted in October 1999 to determine delays in nonselected acute stroke patients. Because thrombolysed patients may represent a selected population with the least delays, we evaluated from those 100 patients data also for those considered ad hoc as “thrombolysis candidates” (acutely hemiparetic/dysphasic patient presumed to meet the 3-hour window). Similar data were collected from nonselected consecutive acute stroke patients from October 2003 (n = 30), right after the opening of the rebuilt ER, and again in September 2004 (100). Reanalysis in 2004 proved necessary because in October 2003, the whole ER infrastructure and digital patient data record systems were revised, and employees were still on learning curves. To finally check the delays in the real target group, the actual DNT and SNT were analyzed for all thrombolysed patients in 1999 (n =17) and for 50 consecutive thrombolysed patients in 2004 (IV to XII). During the observation period, the European Agency for the Evaluation of Medicinal Products required setting up a registry to store individual clinical data (treatment delays, laboratory parameters, hemodynamic data, drug doses) of all thrombolysed stroke patients within the approved 3-hour alteplase therapeutic window, which was implemented by the Web-based SITS-MOST (Safe Implementation of Thrombolysis in Stroke–Monitoring Study) registry (www.acutestroke.org). The criteria for administering thrombolytic therapy have remained essentially unchanged.¹⁰ Data were compared using Student t test and expressed as mean ± SEM.

Results.

The delays for nonselected acute stroke patients remained virtually unchanged (figure 1A) from 1999 to 2004. However, the CT delay for thrombolysis candidates dropped from 1 hour 5 minutes ± 12 minutes to 12 ± 4 minutes in 2003 and further to 7 ± 2 minutes 1 year later (figure 1B). CT thus preceded the full clinical examination. In 2004, the CT was ordered 15 ± 4 minutes before admission and could therefore start right after admission and immediate blood tests.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Delays in the management of stroke patients from 1999 to 2004. (Top) In-hospital delays in the management of all consecutive patients with general acute stroke symptoms. The recorded time points were the time to examine the patients clinically, to order CT scanning, to start CT scanning, and to get the final written radiologic report. (Bottom) Corresponding delays for consecutive ad hoc thrombolysis candidates during the same period. Based on prenotification by emergency medical services of incoming patients, CT could be ordered for a number of patients before their admission, as reflected by the negative bar in 2004. SDT = symptom-to-door time (± SE) for the thrombolysis candidates in the observed time periods is indicated. **p < 0.01; *p < 0.05.

Although the SDT remained constant for thrombolysed patients throughout 1999 to 2004 (table), the DNT was cut by 38 minutes (−43%) and SNT by 39 minutes (−24%). The SDT for thrombolysed patients in 1999 was somewhat less (1 hour 15 minutes ± 8 minutes) than for ad hoc thrombolysis candidates (1 hour 37 minutes ± 18 minutes) (figure 1B). Data stored in the SITS-MOST registry for thrombolysed stroke patients indicates that in the period 2003 to 2005, the CT delay in our center has stabilized to 6 to 9 minutes, and the DNT was 49 minutes during the latter half of 2004 and decreased further to 36 minutes during 2005. This was accompanied by a decrease in SNT from 130 to 103 minutes.

Although 17 patients received alteplase in 1999, this number increased initially after implementing the EMS prenotification system and streamlining of triage. However, in 2000, the DNT remained at 1 hour 31 minutes ± 4 minutes, and the use of thrombolysis did not increase thereafter until the CT relocation in 2003 (figure 2). After the completed ER restructuring, the 50 consecutive patients thrombolysed in 2004 demonstrated highly significant delay reductions (table). In 2004, our ER offered thrombolysis to a total of 83 hemispheric stroke patients (100 patients including basilar occlusions). In 2005, the number of thrombolysed patients was 183 (figure 2), corresponding roughly to 20% of all admitted patients with ischemic stroke, which also represents the top in Europe in terms of the number of thrombolysed patients and shortness of in-hospital delay in providing thrombolysis in the SITS-MOST registry.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. A decade of stroke thrombolysis in Helsinki. The number of stroke patients actually treated with alteplase after 1998, when thrombolysis became a part of routine management of acute hemispheric stroke at the Helsinki University Central Hospital. Thrombolysis of basilar artery occlusions had started already in 1995. Arrows indicate the period of the described project of reorganizing triage and renovation of the ER and the relocation of the CT scanner to the emergency room in 2003 followed by increased access to thrombolytic therapy. In 2005, 183 patients were treated, which corresponds roughly to 20% of all admitted patients with ischemic stroke in this primary stroke center. SDT = symptom-to-door time.

Discussion.

This retrospective report describes three organizational steps that led to major reductions in the management delays of stroke patients considered for thrombolysis. After streamlining the triage and coordinating pre- and in-hospital management and relocating the CT unit to the ER, CT delay was reduced by 1 hour. DNT was reduced by 38 minutes, and because SDT did not change during 1999 to 2004 (table), this translated to an equal reduction in SNT for those who actually received thrombolysis. Because there was substantial expediting in ad hoc thrombolysis candidates, but not in general stroke patients (figure 1A and B), the prenotification by EMS of patients needing urgent therapy decisions led to successful triage, creating a genuine selective chain of recovery. Relocation of the CT unit with its corollary effects to the general ER workflow probably was the single most important step.

In 2003, we started performing perfusion CT (PCT) and CT angiography (CTA) for patients with unequivocal thrombolytic indications to support therapy decisions. In 2004, PCT was performed in 42% and CTA in 48% of the 50 analyzed thrombolysis patients. That the CT delay cuts did not fully translate to a reduced SNT probably reflected the additional imaging time for PCT and CTA (about 15 to 20 minutes). Although these examinations may help to target thrombolysis in certain patients, we expedited imaging protocols and reduced their routine use in 2005 by about 25%; thereafter, the median DNT decreased to <40 minutes and SNT to 103 minutes (SITS-MOST registry). Clearly, additional studies are necessary to validate how to best target these examinations.

We believe that ER restructuring contributed to the fact that five- to 10-fold more patients could be treated with thrombolysis after the renovation period than before it (figure 2). The eligibility criteria, diagnostic workup for thrombolysis, and patient referral remained essentially unchanged throughout these years. The annual figure of roughly 1,000 stroke patients treated in our ER serving as a primary stroke center for Helsinki has remained stable during recent years. This corresponds to a fraction of 20% being treated with thrombolysis in 2005. Many hospitals face a similar situation similar to that of HUCH in 1999 and may consider interventions described here to improve access to thrombolysis.

Acknowledgment

The authors thank Virve Paavola, RN, for collecting the delay data.