The scope for improvement in hyper-acute stroke care in Scotland

Discrete-Event Simulation to Model the Thrombolysis Process for Acute Ischemic Stroke Patients at Urban and Rural Hospitals

Background: Effective treatment with tissue plasminogen activator (tPA) critically relies on rapid treatment. Door-to-needle time (DNT) is a key measure of hospital efficiency linked to patient outcomes. Numerous changes can reduce DNT, but they are difficult to trial and implement. Discrete-event simulation (DES) provides a way to model and determine the impact of process improvements.

Methods: A conceptual framework was developed to illustrate the thrombolysis process; allowing for treatment processes to be replicated using a DES model developed in ARENA. Activity time duration distributions from three sites (one urban and two rural) were used. Five scenarios, three process changes, and two reductions in activity durations, were simulated and tested. Scenarios were tested individually and in combinations. The primary outcome measure is median DNT. The study goal is to determine the largest improvement in DNT at each site.

Results: Administration of tPA in the imaging area resulted in the largest median DNT reduction for Site 1 and Site 2 for individual test scenarios (12.6%, 95% CI 12.4–12.8%, and 8.2%, 95% CI 7.5–9.0%, respectively). Ensuring that patients arriving via emergency medical services (EMS) remain on the EMS stretcher to imaging resulted in the largest median DNT improvement for Site 3 (9.2%, 95% CI 7.9–10.5%). Reducing both the treatment decision time and tPA preparation time by 35% resulted in a 11.0% (95% CI 10.0–12.0%) maximum reduction in median DNT. The lowest median and 90^th percentile DNTs were achieved by combining all test scenarios, with a maximum reduction of 26.7% (95% CI 24.5–28.9%) and 17.1% (95% CI 12.5–21.7%), respectively.

Conclusions: The detailed conceptual framework clarifies the intra-hospital logistics of the thrombolysis process. The most significant median DNT improvement at rural hospitals resulted from ensuring patients arriving via EMS remain on the EMS stretcher to imaging, while urban sites benefit more from administering tPA in the imaging area. Reducing the durations of activities on the critical path will provide further DNT improvements. Significant DNT improvements are achievable in urban and rural settings by combining process changes with reducing activity durations.

Modified Needleman-Wunsch algorithm for clinical pathway clustering

Clinical pathways are used to guide clinicians to provide a standardised delivery of care. Because of their standardisation, the aim of clinical pathways is to reduce variation in both care process and patient outcomes. When learning clinical pathways from data through data mining, it is common practice to represent each patient pathway as a string corresponding to their movements through activities. Clustering techniques are popular methods for pathway mining, and therefore this paper focuses on distance metrics applied to string data for k-medoids clustering. The two main aims are to firstly, develop a technique that seamlessly integrates expert information with data and secondly, to develop a string distance metric for the purpose of process data. The overall goal was to allow for more meaningful clustering results to be found by adding context into the string similarity calculation. Eight common distance metrics and their applicability are discussed. These distance metrics prove to give an arbitrary distance, without consideration for context, and each produce different results. As a result, this paper describes the development of a new distance metric, the modified Needleman-Wunsch algorithm, that allows for expert interaction with the calculation by assigning groupings and rankings to activities, which provide context to the strings. This algorithm has been developed in partnership with UK's National Health Service (NHS) with the focus on a lung cancer pathway, however the handling of the data and algorithm allows for application to any disease type. This method is contained within Sim.Pro.Flow, a publicly available decision support tool.

The positive predictive value of an ambulance prealert for stroke and transient ischaemic attack

OBJECTIVE

Therapeutic options for ischaemic stroke, such as thrombolysis or thrombectomy, are time sensitive. Multiple innovations have been established to reduce the symptom-to-needle time. One such innovation is the prealerting of emergency department (ED) or stroke unit staff by prehospital personnel of suspected stroke patients. The diagnosis of stroke can sometimes be difficult, with stroke mimics being a recognized issue. The prealert mobilizes ED, stroke and imaging personnel, which, for a true-positive, improves door-to-needle times. However, there are a proportion of false-positive prealerts (nonstrokes) that have a significant resource activation implication. The aim of this study was to evaluate the positive predictive value of a prealert for stroke and transient ischaemic attack (TIA).

METHODS

Ambulance service prealert forms for stroke and TIA collated by the ED were compared with the Scottish Stroke Audit database findings, ED electronic notes and imaging reports to establish whether the prealert was a true-positive or a false-positive.

RESULTS

A prealert was obtained for 77 patients as query stroke/TIA. The true-positive rate was 52 and the false-positive rate was 25. The positive predictive value was 0.675. The median symptom-to-arrival time for prealerted patients was 97 min and the door-to-needle time for thrombolysis (n=17 patients) was 38 min.

CONCLUSION

The diagnosis of true-positive stroke can be difficult in the prehospital environment. Although prealert has been shown to improve the patient's journey in terms of door-to-thrombolysis times, we have identified that the prealert has a significant false-positive rate that has important resource allocation and activation consequences. Further analysis of this may inform paramedic training and improve protocols for information handover.