Oh for fucks sake, the stupidity of measuring cost effectiveness rather than 100% recovery effectiveness. I'd fire the lot of you. Do any of you actually think about what survivors want?
Cost-effectiveness of mechanical thrombectomy for acute ischemic stroke in Brazil: Results from the RESILIENT trial
Ana Claudia de Souzahttps://orcid.org/0000-0002-8722-99881, Sheila O Martins1, Carisi Anne Polanczyk2, Denizar Vianna Araújo3, Ana Paula BS Etgeshttps://orcid.org/0000-0002-6411-34802, Bruna Stella Zanotto2, Jeruza Lavanholi Neyeloff2, Leonardo Augusto Carbonerahttps://orcid.org/0000-0002-9313-79671, Márcia Lorena Fagundes Chaves1, João José Freitas de Carvalho4,5, Letícia Costa Rebello6, Daniel Giansante Abud7, Lucas Scotta Cabral8, Fabrício O Lima4, Francisco Mont’Alverne4,9,10, Pedro SC Magalhães11, Henrique Diegolihttps://orcid.org/0000-0002-3970-019X11, Juliana Safanellihttps://orcid.org/0000-0003-1924-827911, Thales André Silveira Salvetti12, Bruno de Sousa Mendes Parente13, Michel Eli Frudit14, Gisele Sampaio Silva15, Octávio M Pontes-Netohttps://orcid.org/0000-0003-0317-843X16, and Raul G Nogueira17
Background
Background
The RESILIENT trial demonstrated the clinical benefit of mechanical thrombectomy in patients presenting acute ischemic stroke secondary to anterior circulation large vessel occlusion in Brazil.
Aims
Aims
This economic evaluation aims to assess the cost-utility of mechanical thrombectomy in the RESILIENT trial from a public healthcare perspective.
Methods
Methods
A cost-utility analysis was applied to compare mechanical thrombectomy plus standard medical care (n = 78) vs. standard medical care alone (n = 73), from a subset sample of the RESILIENT trial (151 of 221 patients). Real-world direct costs were considered, and utilities were imputed according to the Utility-Weighted modified Rankin Score. A Markov model was structured, and probabilistic and deterministic sensitivity analyses were performed to evaluate the robustness of results.
Results
Results
The incremental costs and quality-adjusted life years gained with mechanical thrombectomy plus standard medical care were estimated at Int$ 7440 and 1.04, respectively, compared to standard medical care alone, yielding an incremental cost-effectiveness ratio of Int$ 7153 per quality-adjusted life year. The deterministic sensitivity analysis demonstrated that mRS-6 costs of the first year most affected the incremental cost-effectiveness ratio. After 1000 simulations, most of results were below the cost-effective threshold.
Conclusions
Conclusions
The intervention's clear long-term benefits offset the initially higher costs of mechanical thrombectomy in the Brazilian public healthcare system. Such therapy is likely to be cost-effective and these results were crucial to incorporate mechanical thrombectomy in the Brazilian public stroke centers.
Keywords
Stroke, mechanical thrombectomy, cost-effectiveness, economic evaluation
1Department of Neurology, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil
2National Institute of Science and Technology for Health Technology Assessment (IATS), Porto Alegre, Brazil
3Medical School, State University of Rio de Janeiro, Rio de Janeiro, Brazil
4Department of Neurology, Hospital Geral de Fortaleza, Fortaleza, Brazil
5Department of Neurology, University of Fortaleza, Fortaleza, Brazil
6Department of Neurology, Hospital de Base do Distrito Federal, Brasilia, Brazil
7Department of Interventional Neuroradiology, Ribeirão Preto Medical School, University of São Paulo, Ribeirão Preto, Brazil
8Department of Interventional Neuroradiology, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil
9Department of Interventional Neuroradiology, Hospital Geral de Fortaleza, Fortaleza, Brazil
10Department of Post-Graduation Medical Sciences, University of Fortaleza, Fortaleza, Brazil
11Stroke Neurology Division, Hospital Municipal de Joinville, Joinville, Brazil
12Administration Department, Ribeirão Preto Medical School, University of São Paulo, Ribeirão Preto, Brazil
13Department of Neurosurgery, Hospital de Base do Distrito Federal, Brasilia, Brazil
14Department of Interventional Neuroradiology, Universidade Federal de São Paulo (UNIFESP), São Paulo, Brazil
15Department of Neurology and Neurosurgery, Universidade Federal de São Paulo (UNIFESP), São Paulo, Brazil
16Stroke Service Neurology Division, Ribeirão Preto Medical School, University of São Paulo, Ribeirão Preto, Brazil
17Department of Neurology, Marcus Stroke & Neuroscience Center, Grady Memorial Hospital, Emory University School of Medicine, Atlanta, GA, USA
Corresponding author(s):
Raul G Nogueira, Grady Memorial Hospital, 80 Jesse Hill Drive SE, Room 8D108A, Atlanta, GA 30303, USA. Email: raul.g.nogueira@emory.edu
Introduction
Stroke is a leading cause of mortality and long-term disability worldwide.1,2 Low- and middle-income countries have higher rates of healthy years of life lost to stroke than high-income countries.3 In Brazil, stroke corresponds to 107,658 deaths per year, and almost 80% of stroke patients are treated by the public healthcare system.4
Over the last 25 years, intravenous (IV) thrombolysis with recombinant tissue plasminogen activator (rtPA) for acute ischemic stroke (AIS) has considerably improved outcomes in stroke.5–7 Since 2012, IV rtPA has been offered by the Brazilian public healthcare system for AIS, a strategy that has been proved cost-effective in the country’s reality.8,9
Since 2014, several randomized clinical trials have supported the use of mechanical thrombectomy (MT) in addition to IV rtPA. This procedure has dramatically improved the functional outcomes of patients with AIS due to LVO of the anterior circulation and substantially modified short- and long-term consequences of ischemic stroke.10–17
Despite its higher clinical benefits, MT is costly and requires specific infrastructure and highly trained multidisciplinary team. The RESILIENT trial recently showed the efficacy of MT in a reality that deals with all the typical economic constraints found in developing countries.18
Previous studies demonstrating the cost-effectiveness of MT were exclusively performed in high-income countries.19–25 Therefore, there is a paucity of data concerning the cost-effectiveness of MT in low- and middle-income countries.
Aims
This study aimed to perform a cost-utility analysis of MT from the perspective of the Brazilian public healthcare system and to give sufficient data to support stakeholders for better resource allocation in this setting.
Methods
Study design and patients
The RESILIENT study was a randomized clinical trial that assessed the safety and efficacy of MT in patients presenting AIS due to LVO of the anterior circulation up to 8 h after the last time seen well in the Brazilian public healthcare system. A total of 221 patients were randomly assigned to MT plus standard medical care (SMC) (n = 111) or SMC alone (n = 110) and the primary outcome favored the intervention (common odds ratio, 2.28; 95% confidence interval, 1.41–3.70; p = 0.001).18
For this analysis, data from a one-year follow-up (2017–2018) of a subset sample of the RESILIENT trial (151 of 221 patients) was obtained. The four public hospitals with available cost data had the highest volume of recruitment and from distinct geographic regions were selected for the analysis: Hospital de Clínicas de Porto Alegre (n = 25), Hospital das Clínicas de Ribeirão Preto (n = 26), Hospital Base de Brasília (n = 27), and Hospital Geral de Fortaleza (n = 73). The population’s baseline characteristics included in the economic analyses and in the overall RESILIENT trial are described in Table 1 – Supplementary Material. The subset sample included more non-white subjects than the full trial sample (53.7% vs. 46.6%, p = 0.002). Other variables were similar across groups.
Model design
A cost-utility analysis was performed to compare both treatment arms in terms of incremental costs, incremental effectiveness, and incremental cost-effectiveness ratio (ICER).
Costs and utilities for the MT plus SMC (n = 78) vs. SMC alone (n = 73) groups were estimated and compared along the first year after stroke and projected for a lifetime horizon. Direct healthcare costs were considered from one-year follow-up, and utilities were imputed according to each patient's Utility-Weighted modified Rankin Score (UW-mRS) to calculate Quality-Adjusted Life Years (QALYs). A discount of 5% a year was applied for costs and outcomes.
A Markov model was developed to capture the natural history of the acute stroke phase and the probabilities between the different stages of severity of post-stroke disability (based on mRS scores) for 30 years (Figure 1). Accordingly, we admitted a lifetime horizon for this population.
Figure 1. Markov Model. Patients with acute ischemic stroke due to anterior large vessel occlusion enter the model on hospital admission and were allocated to receive one of the treatment arms. Consequences of recurrent stroke for a mRS 3 with an annual probability of recurrent stroke, death or stable mRS was exemplified. || Decision node, v terminal node, chance node, and M Markov node.
A patient with AIS entered the model with 64 years old (mean age for the overall RESILIENT cohort) on hospital admission and was allocated in one of the treatment arms: MT plus SMC or SMC alone. Subsequently, the patient entered one of the seven possible health states according to the mRS score. The health state transition probabilities in the first year after stroke and health state probabilities after one year were estimated according to the patient health state from the RESILIENT study18 (Table 2 – Supplementary Material).
In the first year, three-month cycles were considered, and at the end of each cycle, patient's health condition was classified again according to the mRS score to calculate utilities. Patient health condition assessments were performed at 3, 6, and 12 months in the trial. Thus, we assumed that health status at the ninth month was equal to the values observed at 12 months. After the first year, one-year cycles were considered, where the patient remained in the same health state of one year after stroke or died. This model was based on a previous Markov model to simulate the cost-effectiveness of intravenous thrombolysis in Brazil.9
First-year survival was obtained from the trial follow-up and modeled for a lifetime horizon adjusted by national mortality data.26 The model also considered that after the index stroke, the mortality rate would be 2.67 times higher after the index stroke than the general population.27 We did not consider symptomatic intracerebral hemorrhages in the model because both treatments had similar rates. The perspective was the Brazilian public healthcare system.
Costs and use of resources
A micro-costing, individual patient level, Time-Driven Activity-Based Costing (TDABC) method was used to estimate real-world resources used from 151 patients.28–31 Costs were obtained from one-year follow-up from RESILIENT trial patients and included the total costs of primary hospitalization services, readmissions, outpatient visits, and rehabilitation, including infrastructure and staff cost.
All costs were discounted at 5%.32,33 Costs in local country currency (Brazilian Reals – R$) were converted to International Dollars (I$) using Purchasing Power Parity (PPP) 2018.34
More descriptions about costs and use of resources are described in Supplementary Material. Underlying data can be provided upon reasonable request.
Effectiveness measure
Utilities were imputed according to each patient’s UW-mRS. The UW-mRS allows the conversion of seven levels of the mRS into a health utility (Table 2 – Supplementary Material).35
The utility values derived in the first year from the patient’s mRS score of three-month cycles. After the first year, the values derived from patients’ mRS score of one year were used and extrapolated for a lifetime horizon. All QALYs were discounted 5% a year,32,33 adjusted for half cycle corrections.
Statistical analysis
Continuous variables were reported as medians and interquartile range, and categorical data were reported as frequencies and constant variables. Sample data, Markov Model, and Sensitivity analysis were consolidated in a Microsoft® Excel spreadsheet for Mac 2019. Descriptive data analysis was performed in IBM SPSS® for Mac 2019.
Keywords
Stroke, mechanical thrombectomy, cost-effectiveness, economic evaluation
1Department of Neurology, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil
2National Institute of Science and Technology for Health Technology Assessment (IATS), Porto Alegre, Brazil
3Medical School, State University of Rio de Janeiro, Rio de Janeiro, Brazil
4Department of Neurology, Hospital Geral de Fortaleza, Fortaleza, Brazil
5Department of Neurology, University of Fortaleza, Fortaleza, Brazil
6Department of Neurology, Hospital de Base do Distrito Federal, Brasilia, Brazil
7Department of Interventional Neuroradiology, Ribeirão Preto Medical School, University of São Paulo, Ribeirão Preto, Brazil
8Department of Interventional Neuroradiology, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil
9Department of Interventional Neuroradiology, Hospital Geral de Fortaleza, Fortaleza, Brazil
10Department of Post-Graduation Medical Sciences, University of Fortaleza, Fortaleza, Brazil
11Stroke Neurology Division, Hospital Municipal de Joinville, Joinville, Brazil
12Administration Department, Ribeirão Preto Medical School, University of São Paulo, Ribeirão Preto, Brazil
13Department of Neurosurgery, Hospital de Base do Distrito Federal, Brasilia, Brazil
14Department of Interventional Neuroradiology, Universidade Federal de São Paulo (UNIFESP), São Paulo, Brazil
15Department of Neurology and Neurosurgery, Universidade Federal de São Paulo (UNIFESP), São Paulo, Brazil
16Stroke Service Neurology Division, Ribeirão Preto Medical School, University of São Paulo, Ribeirão Preto, Brazil
17Department of Neurology, Marcus Stroke & Neuroscience Center, Grady Memorial Hospital, Emory University School of Medicine, Atlanta, GA, USA
Corresponding author(s):
Raul G Nogueira, Grady Memorial Hospital, 80 Jesse Hill Drive SE, Room 8D108A, Atlanta, GA 30303, USA. Email: raul.g.nogueira@emory.edu
Introduction
Stroke is a leading cause of mortality and long-term disability worldwide.1,2 Low- and middle-income countries have higher rates of healthy years of life lost to stroke than high-income countries.3 In Brazil, stroke corresponds to 107,658 deaths per year, and almost 80% of stroke patients are treated by the public healthcare system.4
Over the last 25 years, intravenous (IV) thrombolysis with recombinant tissue plasminogen activator (rtPA) for acute ischemic stroke (AIS) has considerably improved outcomes in stroke.5–7 Since 2012, IV rtPA has been offered by the Brazilian public healthcare system for AIS, a strategy that has been proved cost-effective in the country’s reality.8,9
Since 2014, several randomized clinical trials have supported the use of mechanical thrombectomy (MT) in addition to IV rtPA. This procedure has dramatically improved the functional outcomes of patients with AIS due to LVO of the anterior circulation and substantially modified short- and long-term consequences of ischemic stroke.10–17
Despite its higher clinical benefits, MT is costly and requires specific infrastructure and highly trained multidisciplinary team. The RESILIENT trial recently showed the efficacy of MT in a reality that deals with all the typical economic constraints found in developing countries.18
Previous studies demonstrating the cost-effectiveness of MT were exclusively performed in high-income countries.19–25 Therefore, there is a paucity of data concerning the cost-effectiveness of MT in low- and middle-income countries.
Aims
This study aimed to perform a cost-utility analysis of MT from the perspective of the Brazilian public healthcare system and to give sufficient data to support stakeholders for better resource allocation in this setting.
Methods
Study design and patients
The RESILIENT study was a randomized clinical trial that assessed the safety and efficacy of MT in patients presenting AIS due to LVO of the anterior circulation up to 8 h after the last time seen well in the Brazilian public healthcare system. A total of 221 patients were randomly assigned to MT plus standard medical care (SMC) (n = 111) or SMC alone (n = 110) and the primary outcome favored the intervention (common odds ratio, 2.28; 95% confidence interval, 1.41–3.70; p = 0.001).18
For this analysis, data from a one-year follow-up (2017–2018) of a subset sample of the RESILIENT trial (151 of 221 patients) was obtained. The four public hospitals with available cost data had the highest volume of recruitment and from distinct geographic regions were selected for the analysis: Hospital de Clínicas de Porto Alegre (n = 25), Hospital das Clínicas de Ribeirão Preto (n = 26), Hospital Base de Brasília (n = 27), and Hospital Geral de Fortaleza (n = 73). The population’s baseline characteristics included in the economic analyses and in the overall RESILIENT trial are described in Table 1 – Supplementary Material. The subset sample included more non-white subjects than the full trial sample (53.7% vs. 46.6%, p = 0.002). Other variables were similar across groups.
Model design
A cost-utility analysis was performed to compare both treatment arms in terms of incremental costs, incremental effectiveness, and incremental cost-effectiveness ratio (ICER).
Costs and utilities for the MT plus SMC (n = 78) vs. SMC alone (n = 73) groups were estimated and compared along the first year after stroke and projected for a lifetime horizon. Direct healthcare costs were considered from one-year follow-up, and utilities were imputed according to each patient's Utility-Weighted modified Rankin Score (UW-mRS) to calculate Quality-Adjusted Life Years (QALYs). A discount of 5% a year was applied for costs and outcomes.
A Markov model was developed to capture the natural history of the acute stroke phase and the probabilities between the different stages of severity of post-stroke disability (based on mRS scores) for 30 years (Figure 1). Accordingly, we admitted a lifetime horizon for this population.
Figure 1. Markov Model. Patients with acute ischemic stroke due to anterior large vessel occlusion enter the model on hospital admission and were allocated to receive one of the treatment arms. Consequences of recurrent stroke for a mRS 3 with an annual probability of recurrent stroke, death or stable mRS was exemplified. || Decision node, v terminal node, chance node, and M Markov node.
A patient with AIS entered the model with 64 years old (mean age for the overall RESILIENT cohort) on hospital admission and was allocated in one of the treatment arms: MT plus SMC or SMC alone. Subsequently, the patient entered one of the seven possible health states according to the mRS score. The health state transition probabilities in the first year after stroke and health state probabilities after one year were estimated according to the patient health state from the RESILIENT study18 (Table 2 – Supplementary Material).
In the first year, three-month cycles were considered, and at the end of each cycle, patient's health condition was classified again according to the mRS score to calculate utilities. Patient health condition assessments were performed at 3, 6, and 12 months in the trial. Thus, we assumed that health status at the ninth month was equal to the values observed at 12 months. After the first year, one-year cycles were considered, where the patient remained in the same health state of one year after stroke or died. This model was based on a previous Markov model to simulate the cost-effectiveness of intravenous thrombolysis in Brazil.9
First-year survival was obtained from the trial follow-up and modeled for a lifetime horizon adjusted by national mortality data.26 The model also considered that after the index stroke, the mortality rate would be 2.67 times higher after the index stroke than the general population.27 We did not consider symptomatic intracerebral hemorrhages in the model because both treatments had similar rates. The perspective was the Brazilian public healthcare system.
Costs and use of resources
A micro-costing, individual patient level, Time-Driven Activity-Based Costing (TDABC) method was used to estimate real-world resources used from 151 patients.28–31 Costs were obtained from one-year follow-up from RESILIENT trial patients and included the total costs of primary hospitalization services, readmissions, outpatient visits, and rehabilitation, including infrastructure and staff cost.
All costs were discounted at 5%.32,33 Costs in local country currency (Brazilian Reals – R$) were converted to International Dollars (I$) using Purchasing Power Parity (PPP) 2018.34
More descriptions about costs and use of resources are described in Supplementary Material. Underlying data can be provided upon reasonable request.
Effectiveness measure
Utilities were imputed according to each patient’s UW-mRS. The UW-mRS allows the conversion of seven levels of the mRS into a health utility (Table 2 – Supplementary Material).35
The utility values derived in the first year from the patient’s mRS score of three-month cycles. After the first year, the values derived from patients’ mRS score of one year were used and extrapolated for a lifetime horizon. All QALYs were discounted 5% a year,32,33 adjusted for half cycle corrections.
Statistical analysis
Continuous variables were reported as medians and interquartile range, and categorical data were reported as frequencies and constant variables. Sample data, Markov Model, and Sensitivity analysis were consolidated in a Microsoft® Excel spreadsheet for Mac 2019. Descriptive data analysis was performed in IBM SPSS® for Mac 2019.
More at link.
No comments:
Post a Comment