Quality improvement project leads to better care of all stroke patients

Since the publication of the first IV thrombolysis trial with alteplase¹ and subsequent analyses showing a strong correlation between early treatment and favorable outcome,^2,3 the medical community has been focused on early identification of stroke and rapid delivery of suspected patients to the nearest alteplase-capable hospital. The importance of early stroke treatment created a need for a stroke system of care where there is a widespread network of medical facilities in which IV alteplase can be administered, and subsequent care can take place at another facility of higher level of care. The state of Minnesota has adopted this philosophy, leading to legislation aimed at ensuring the availability of stroke-ready hospitals within 30 minutes of all residents (Minnesota Statutes 144.194 [2013]). This has proved to be quite successful, increasing the percentage of population within 30 minutes from a hospital with available alteplase from 60% in 2013 to 93% in 2019.⁴

More recently, mechanical thrombectomy (MT) was proven to benefit large vessel occlusion (LVO) stroke, and time to treatment remained an important factor determining outcome.^5,6 Because MT is only available in large tertiary centers, the “bypass” debate erupted. The idea is to triage patients with presumed LVO directly to a thrombectomy-capable hospital, even if this means bypassing a closer alteplase-capable hospital. Many arguments and counterarguments have been made, and the uncertainty is reflected in the most recent guidelines, which state “the benefit of bypassing the closest to bring the patient to one that offers a higher level of stroke care, including MT, is uncertain.”⁷

A key counterargument against the bypass concept is the potential for harm by delaying IV thrombolysis, particularly to patients who are subsequently deemed ineligible for MT. For these patients, the bypass is both futile and detrimental. Another important counterargument is that the new concept could deprive alteplase-capable, but MT-incapable, hospitals of their local patients. This might lead to a loss of stroke expertise and interest in those hospitals, the centralization of stroke care within the MT-capable hospitals, and the inevitable decline of the stroke systems built over the last 2 decades.

Critics of the bypass model prefer maintaining the current Drip-and-Ship model, delivering patients to the nearest alteplase-capable hospital for initial evaluation and thrombolysis, and then transferring selected patients who are suitable for MT. This ensures that all alteplase-eligible patients are treated rapidly and all MT-eligible patients are accurately selected, thereby avoiding overwhelming thrombectomy-capable hospital and retaining the stroke patient volume critical for skillset maintenance in the alteplase-capable hospital. The disadvantage of this Drip-and-Ship model is that MT may be inadvertently delayed with longer time to reperfusion for patients with LVO because of the stop-over in the alteplase-capable hospital.

Sablot et al.'s⁸ quality improvement (QI) project, which appears in this issue of Neurology Clinical Practice, demonstrates that reperfusion time can be readily shortened in the Drip-and-Ship model. The QI project consisted of several improvements of the workflow within and between an alteplase-capable hospital and a MT-capable hospital, with the goal of minimizing the onset to treatment time in both hospitals. The authors compared the workflow time metrics and the clinical outcomes of patients who were transferred from the alteplase-capable hospital to the MT-capable hospital for MT before and after implementation of the QI project. The QI project led to an impressive reduction of all the measured time metrics: median door to needle time at the alteplase-capable hospital from 82 to 49 minutes (p = 0.0001); median door-in-door-out (DIDO) at the alteplase-capable hospital from 118 to 78 minutes (p < 0.01); and median onset to reperfusion time from 417 to 367 minutes (p < 0.04). The study did not show a significant difference in 3-month modified Rankin Scale between the QI cohort and the pre-QI cohort. This should not be interpreted as a failure of the QI project to affect outcome. It was not a prospective trial powered to examine the effect of QI on functional outcome. Strong evidence from large randomized trials has already demonstrated the clear positive effect of shortening time to reperfusion on outcome.⁶

The fact that the shorter reperfusion time was primarily mediated by DIDO improvement rather than interhospital time metrics should not be regarded as a failure to improve interhospital workflow. The alteplase-capable hospital in the study is located in a mountainous region that often relies on air transport for urgent transfer. Unique and unmodifiable issues often prevail in air medical retrieval. Instead, the influence of the DIDO should be viewed as an important reminder that improving the fundamental steps of hyperacute stroke management (i.e., efficient patient assessment, imaging, and treatment decision-making) has far-reaching benefits in the subsequent care of the patient.

The authors highlight the elements that contributed most significantly to their metrics gains. A key factor implied in these processes is the active involvement of both centers in collaboration and in tandem. Specifically, the MT-capable hospital is prenotified during initial imaging at the alteplase-capable hospital while preparation for transfer commences pre-emptively before final approval from the thrombectomy-capable hospital. Concurrently, the MT-capable hospital begins preparation for thrombectomy with the patient still in transit, to ensure the entire team is available upon the patient's arrival for direct angiographic suite admission without reimaging or formal reassessment. In doing so, the MT-capable hospital fully relies on the expertise of the fellow alteplase-capable hospital vascular neurologist and commits resources before receiving the patient.

These steps illustrate the foundation of an ideal Drip-and-Ship model where there is considerable trust, mutual respect, and understanding in a partnership where both parties are equal and active participants in the care of the patient. Inevitably, differences in clinical opinion may arise. For example, patients referred for thrombectomy may be declined by the MT-capable hospital despite being initially deemed appropriate candidates by the referring vascular neurologist. Such declined referrals should be formally reviewed in periodic structured conjoint QI meetings with the aim to foster ongoing collaborative dialogue.

The landscape in models of care is rapidly evolving. Telestroke utilization is rapidly increasing, bringing neurology expertise to hospitals without attending neurologists. New organization concepts are actively being developed as emerging evidence expands the time-window for reperfusion therapies. Prehospital LVO triage, mobile stroke units with thrombolysis capability, and transporting the interventionalist to the alteplase-capable hospital are being explored while randomized control trials comparing different models of care are being conducted. Despite the ongoing debate, at times Drip-and-Ship remains the only option due to geographical constraints, as in the case of the current study.

Sablot et al. should be commended on their successful QI project. The most important and practical lesson is that improving workflow efficiency in a Drip-and-Ship model of care is very achievable, especially when there is a collaborative partnership between hospitals with and without endovascular capabilities. Until there is unequivocal evidence demonstrating the superiority of a specific model of care, collaboration with, not the elimination of, hospitals without endovascular capability in the conventional model of care is important and will lead to better care of all stroke patients.

Author contributions

H.M. Hussein: Drafting/revising the manuscript. F.C. Ng: Drafting/revising the manuscript.

Study funding

No targeted funding reported.

Disclosure

The authors report no disclosures relevant to the manuscript. Full disclosure form information provided by the authors is available with the full text of this article at Neurology.org/cp.