Letter by Lahr et al Regarding Article, “Promoting Thrombolysis in Acute Ischemic Stroke”

With great interest, we read the article by Dirks et al¹ published in the March 2011 issue of Stroke. The authors showed that a multidimensional implementation strategy slightly increased the thrombolysis treatment rate for the intervention group (13%) compared to the control group (12%) without actually increasing clinical outcome at 3 months.

Apparently, an evaluation of a complex multifaceted intervention such as thrombolysis is considered difficult, whereas power to detect a difference in clinical outcome may have been low a priori. The fact that a multitude of interventions was applied introduces considerable potential for dilution of effect. This raises the question whether classical experimental designs are the way forward when the aim is to improve implementation of thrombolysis. Is the result obtained relevant, ie, does it help care providers or policy makers to adopt and implement a new therapy? Or might we consider the gains too limited for the time and money invested? In other words, might other research methodology be warranted?

The acute stroke pathway can be considered as a chain of linked events or actions between which complex interaction occurs. All activities along the chain should be coordinated and planned. The weakest links or bottlenecks may be identified, leading to specific solutions. Industrial engineering yielded simulation-based approaches² that, in addition to current “trial and error” methodology, likely will contribute to improving implementation. Simulation enables assessing the entirety of the patient pathway without having to already perform a costly clinical experiment. There have been examples of successful implementation of industrial methods for health care issues. For instance, a discrete event simulation model was conducted of the process of stroke care from symptom onset through treatment with intravenous tissue plasminogen activator and it was found that a National Institutes of Neurological Disorders and Stroke-compliant treatment strategy resulted in a higher proportion of patients treated while remaining cost-effective.³ In a different study, combining an experimental design with a simulation model in infant ultrasound screening for developmental dysplasia of the hip led to selection of optimized implementation strategies.⁴ We advocate that implementation of thrombolytic therapy for acute ischemic stroke may benefit from such an industrial approach. First, it allows investigating the pathway as a whole, thereby offering the opportunity to study links and possible interactions in detail. Second, it could identify barriers or bottlenecks that have the greatest impact on undertreatment and patient outcomes. Third, it could directly assess the result of changes in the process in a time-saving and relatively inexpensive manner.

Applying research methodology such as simulation modeling apparently constitutes a drastic change in the way health care professionals look at a problem. It would appear, however, that such methods immensely increase the chance of efficiently finding a solution to health care problems.