Telemedicine for Stroke: Quantifying the Long-Term National Costs and Health Benefits

Abstract

Objective: Few countries have established national programs to maximize access and reduce operational overheads. We aimed to use patient-level data up to 12 months to model the potential long-term costs and health benefits attributable to implementing such a program for Australia.

Methods: A Markov model was created for Australia with an inception population of 10,000 people with stroke presenting to non-urban or suburban hospitals without stroke medical specialists that could receive stroke telemedicine under a national program. Seven Markov states represented the seven modified Rankin Scale (mRS) scores (0 no disability to 6 dead) plus an absorbing state for all other causes of death. The literature informed inputs for the model; for the telemedicine program (including program costs and effectiveness) and patients, these were extrapolated from the Victorian Stroke Telemedicine (VST) program with the initial status of patients being their health state at day 365 as determined by their mRS score. Costs (2018 Australian dollars, healthcare, non-medical, and nursing home) and benefits were reported for both the societal and healthcare perspectives for up to a 25 years (lifetime) time horizon.

Results: We assumed 4,997 to 12,578 ischemic strokes would arrive within 4.5 h of symptom onset at regional hospitals in 2018. The average per person lifetime costs were $126,461 and $127,987 from a societal perspective or $76,680 and $75,901 from a healthcare system perspective and benefits were 4.43 quality-adjusted life years (QALYs) and 3.98 QALYs gained, respectively, for the stroke telemedicine program and practice without such program. The stroke telemedicine program was associated with a cost saving of $1,526 (from the societal perspective) or an additional $779 (from the healthcare system perspective) and an additional 0.45 QALY gained per patient over the lifetime. The incremental costs of the stroke telemedicine program ($2,959) and management poststroke ($813) were offset by cost savings from rehospitalization (-$552), nursing home care (-$2178), and non-medical resource use (-$128).

Conclusion: The findings from this long-term model provide evidence to support ongoing funding for stroke telemedicine services in Australia. Our estimates are conservative since other benefits of the service outside the use of intravenous thrombolysis were not included.

Keywords: cost-effectiveness analysis; ischemic stroke; long-term; stroke telemedicine; thrombolysis (tPA).

Figure 1

Model structure.

Figure 2

A Tornado diagram for the one-say sensitivity analyses. Red bar means the value increases from the base case; blue bar represents the value decreases from the base case. mRS, modified Rankin Scale; mgmt, management; mort, mortality; c_mRS3_1y_int, cost of management for the intervention group for the first year post stroke (mRS3); c_mRS2_1y_int, cost of mangement for the intervention group for the first year post stroke (mRS2); p_backmortality, probability of background mortality; c_mgmt_mRS1, cost of long-term management post stroke (mRS1); timehorizon, modeled time horizion; p_recurStroke, probability of having recurrent stroke; c_mgmt_mRS5, cost of long-term management post stroke (mRS5); c_mgmt_mRS3, cost of long-term management post stroke (mRS3); start_Age, onset age of the index stroke; disc_rate, discount rate for both costs and QALYs; c_nursinghome_accommodationSupplement, cost of nursing home care for the accommodation supplement; c_nursinghome_dailyfee, cost of nursing home care for the daily fee; utility_mRS0, utility weights post stroke (mRS0); utility_mRS2, utility weights post stroke (mRS2); c_mgmt_mRS4, cost of long-term management post stroke (mRS4); p_mort_recurStroke, probabiltiy of death following a recurrent stroke; utility_mRS1, utility weights post stroke (mRS1); utility_mRS3, utility weights post stroke (mRS3); utility_mRS5, utility weights post stroke (mRS5); utility_mRS4, utility weights post stroke (mRS4).

Figure 3

Incremental cost-effectiveness plane from the probabilistic sensitivity analysis. AUDs, Australian dollars; QALYs, quality-adjusted life years. 100% of all the iterations suggest that the stroke telemedicine program is cost saving and more effective over the lifetime of patients.