Thank you for having me. My name is Dean Barone. My research is cost-effectiveness of patient selection based on advanced imaging for patients with acute ischemic stroke. Since 1996, the treatment of stroke has been based on time. Recent studies utilizing advanced imaging have extended the time to treating acute ischemic stroke in such populations as wake-up strokes using deltaplase and in patients having acute ischemic strokes within 24 hours using mechanical thrombectomy. These recent studies are leaning toward treatment based on physiology with advanced imaging and not time alone, with all of them showing improved outcomes. This research evaluated if it is cost-effective and if there are overall improved outcomes with patient selection for stroke treatment using advanced imaging instead of the American Heart Association guidelines based on time. The methods, which will be detailed in the following slides, were to develop a decision tree, configure the parameters of the tree for a primary analysis of cost-effectiveness, configure the tree for a secondary analysis based on outcomes, obtain and collate the data to input into the tree, configure the tree for a probability sensitivity analysis to support or reject the base case. A decision tree was developed using TreeAge Pro. This is a representation with an open source picture to discuss the tree secondary to the actual tree being too large to reject and clearly read on the slide, as you can see to the right. The decision tree node was stroke, as depicted by the number 1. Chance nodes were the scenario based on time, named CT, depicted as A, which were the current guidelines for the American Heart Association. And the scenario based on advanced imaging, B, named CT, CTA, CTP, was an algorithm based on the most recent studies utilizing advanced imaging to determine treatment of stroke. The decision tree node number of strokes was taken from the report by the American Heart Association. In the chance nodes for the time-based data was also from a report by the American Heart Association with the percentage of patients receiving alteplase, number of patients having endoascler procedures, along with the reported outcomes. The outcomes were all reported with modified Rankin scale score. The terminal node in QALY was the outcomes, modified Rankin scale score, combined with the utility based on a 2015 study that assigned utility to the modified Rankin scale score. All costs with treatments and or procedures and hospitalization were taken from CMS data from the year 2017 and 2018. The configuration parameters of the tree were cost-effective calculations with cost and payoff of U.S. dollars, effectiveness and payoff in life years. The willingness to pay was set at $50,000. And global discounting was not used in this scenario. A secondary analysis configuration parameter was performed, showing the breakdown of the modified Rankin scale scores for each scenario to show if there was improved outcomes other than just life years. The data outcomes for the chance nodes of time in the form of modified Rankin scale scores were weighted averages with the information taken from major studies, NINDS, ECAS-2, ECAS-3, Hepathat, Atlantis A and B. The data outcomes in the chance node of time for mechanical thrombectomy patients with the patients that received Alteplase were weighted averages with the information taken from major studies Mr. Clean, Xtend IA, Escape, RevScat, and Swift Prime. The data outcomes for the chance node of advanced imaging in the form of modified Rankin scale scores were weighted averages of the patients that did not receive Alteplase, were taken from the major studies DAWN, Diffuse-3, and Escape. The data outcomes for the chance node of advanced imaging in the form of modified Rankin scale scores of patients that received Alteplase were taken from the only study, the wake-up trial. This chart is a graphic representation of all the data, its location in the tree, and its source from where the information was procured. The decision tree was as well configured to perform a probability sensitivity analysis using a Monte Carlo simulation of 10,000 iterations with a triangular distribution, secondary to the overall limited data and allowing for the lower, likeliest, and maximum limits to be determined by the researcher. However, considering beta distribution is more commonly used and understood in the literature, a second probability sensitivity analysis was performed, again with 10,000 iterations, and a mean of 0.5 and standard deviation of 0.2. The simulation was performed, and as seen here, the results with a triangular distribution with willingness of $50,000 to pay showed a base case with the advanced imaging being more cost-effective than the scenario based on time. The cost of the advanced imaging was more expensive with an incremental cost of $17,049. However, the effectiveness was higher in the advanced imaging scenario with an incremental effectiveness of 0.58 life years. The incremental cost-effectiveness ratio, or ICER, favored the advanced imaging group at $29,149, and the net monetary benefit as well favored the advanced imaging scenario by $12,196. Although it is a higher cost, there is greater effectiveness showing it to overall be cost-effective at the willingness to pay $50,000. As previously discussed, beta distribution is more widely accepted, so a second base case scenario was performed with a beta distribution. As seen here, the results with a beta distribution showed an incremental cost of $13,265, an incremental effectiveness of 0.44 life years, an ICER of $29,971, and a net monetary benefit of $8,865. Again, as seen in the triangular distribution, favoring the advanced imaging group with a higher cost as well as a higher effectiveness, but overall cost-effective with a willingness to pay $50,000. In the probability sensitivity analysis with 10,000 iterations and a willingness to pay a threshold of $50,000, the scenario based on advanced imaging is favored significantly as seen in this willingness to pay graph. The graph shows that 78% of the beta distribution and 74% of the triangular distribution of the 10,000 iterations agree with the base case that advanced imaging being more cost-effective compared to time. This is the second analysis that was performed showing the breakdown of the modified Rankin scale scores for each scenario in a triangular distribution to show if there are improved outcomes other than just life years. As you can see here, there was a higher percentage of patients that would have a modified Rankin score of zero, 21.4% for the advanced imaging group compared to 16.4% in the group utilizing time. There was also less deaths in the advanced imaging group compared to the time group with 12.2% compared to 15.4% respectively. Considering it is cost-effective and the outcomes are improved using advanced imaging for the decision to treat patients, further studies based on decision to treat patients using advanced imaging alone and not from less known time well might benefit stroke patients as well as payers. Thank you very much for your time. Enjoy the rest of the conference.

Dean Barone

research

cost-effectiveness

advanced imaging

patient selection