Modeling the potential efficiency of a blood biomarker-based tool to guide pre-hospital thrombolytic therapy in stroke patients

Abstract

Objectives: Stroke treatment with intravenous tissue-type plasminogen activator (tPA) is effective and efficient, but as its benefits are highly time dependent, it is essential to treat the patient promptly after symptom onset. This study evaluates the cost-effectiveness of a blood biomarker test to differentiate ischemic and hemorrhagic stroke to guide pre-hospital treatment with tPA in patients with suspected stroke, compared with standard hospital management. The standard care for patients suffering stroke consists mainly in diagnosis, treatment, hospitalization and monitoring.

Methods: A Markov model was built with four health states according to the modified Rankin scale, in adult patients with suspected moderate to severe stroke (NIHSS 4-22) within 4.5 hours after symptom onset. A Spanish Health System perspective was used. The time horizon was 15 years. Quality-adjusted life-years (QALYs) and life-years gained (LYGs) were used as a measure of effectiveness. Short- and long-term direct health costs were included. Costs were expressed in Euros (2022). A discount rate of 3% was used. Probabilistic sensitivity analysis and several one-way sensitivity analyses were conducted.

Results: The use of a blood-test biomarker compared with standard care was associated with more QALYs (4.87 vs. 4.77), more LYGs (7.18 vs. 7.07), and greater costs (12,807€ vs. 12,713€). The ICER was 881€/QALY. Probabilistic sensitivity analysis showed that the biomarker test was cost-effective in 82% of iterations using a threshold of 24,000€/QALY.

Conclusions: The use of a blood biomarker test to guide pre-hospital thrombolysis is cost-effective compared with standard hospital care in patients with ischemic stroke.

Keywords: Biomarkers; Cost-effectiveness; Stroke; tPA.

Fig. 1

Model Structure. A Short-term decision analytic tree structure of clinical trial outcomes. Outcomes were: non-disabled patient, only requiring secondary prevention strategies (modified Rankin scale [mRS 0–1]), disabled patients also requiring rehabilitation (mRS 2–3), severely disabled patients, requiring long-term nursing care (mRS 4–5) and death (mRS 6) from stroke onset to 90 days. Patients enter the model with suspected stroke with < 4.5 h from symptom onset and with a National Institutes of Health Stroke Scale score between 4 and 22 (moderate or severe stroke). B Long-term Markov model used to simulate lifetime patient outcomes. Patients transition between the different health states as indicated by the arrows

Fig. 2

Cost-effectiveness acceptability curve