Cost-Effectiveness Analysis of Current Treatment Options for Eosinophilic Esophagitis

Abstract

Introduction: The management strategies for eosinophilic esophagitis include proton pump inhibitors (PPIs), swallowed topical corticosteroids (tCSs), elimination diets, and the biologic agent dupilumab, although there remains little guidance on the selection of initial treatment. We performed cost-effectiveness analyses to compare these approaches of first-line therapy.

Methods: A Markov model was constructed from a payer perspective to evaluate the cost-effectiveness of first-line therapies for eosinophilic esophagitis, including PPI, tCS, and 6-food elimination diet (SFED), with crossover in treatments for primary and secondary nonresponse. The primary outcome was incremental cost-effectiveness ratio at 2 and 5-year time horizons. Secondary analyses included modeling from a societal perspective that also accounted for patient-specific costs, as well as a separate simplified model comparing dupilumab with tCS and PPI.

Results: In the base-case scenario (5-year time horizon), the average costs were SFED: $15,296.81, PPI: $16,153.77, and tCS: $20,975.33 as initial therapy, with SFED being the dominant strategy (more effective/less costly), while PPI offered the lowest cost on a 2-year time horizon. From a societal perspective, PPI was the dominant initial strategy on both 2 and 5-year time horizons. Among pharmacologic therapies, PPI was the most cost-effective first-line option. Dupilumab was not cost-effective relative to tCS, unless the quarterly cost is reduced from $7,311 to $2,038.50 per price threshold analysis under permissive modeling conditions.

Discussion: SFED was the most effective/least costly first-line therapy from the payer perspective while PPI was more cost-effective from the societal perspective. PPI is also the most cost-effective pharmacologic strategy. Dupilumab requires substantial cost reductions to be considered cost-effective first-line pharmacotherapy.

Figure 1:

(A) Simplified overview of Markov model including up to third line of treatment. (B) Simplified Markov model comparing Dupilumab 1^st to tCS 1^st strategies allowing for crossover to second line therapy. PPI: proton pump inhibitor; tCS: topical corticosteroid; HR: histologic response; NR: non-response.

Figure 1:

(A) Simplified overview of Markov model including up to third line of treatment. (B) Simplified Markov model comparing Dupilumab 1^st to tCS 1^st strategies allowing for crossover to second line therapy. PPI: proton pump inhibitor; tCS: topical corticosteroid; HR: histologic response; NR: non-response.

Figure 2:

Comparisons of cost ($) vs. effectiveness (QALY) between treatment strategies in the primary (payer perspective) model (blue dots) and the societal perspective models (orange dots) at the 5-year and 2-year time horizon. (A) From the payer perspective, elimination diets 1^st are the most cost-effective strategies compared to PPI 1^st and tCS 1^st over a 5-year horizon. However, from the societal perspective taking into account patient-specific costs, including days lost from work and dietary costs, PPI 1^st became the most cost-effective strategy. (B) Over a 2-year horizon, PPI 1^st is the most cost-effective strategy from both the payer and the societal perspectives.

Figure 2:

Comparisons of cost ($) vs. effectiveness (QALY) between treatment strategies in the primary (payer perspective) model (blue dots) and the societal perspective models (orange dots) at the 5-year and 2-year time horizon. (A) From the payer perspective, elimination diets 1^st are the most cost-effective strategies compared to PPI 1^st and tCS 1^st over a 5-year horizon. However, from the societal perspective taking into account patient-specific costs, including days lost from work and dietary costs, PPI 1^st became the most cost-effective strategy. (B) Over a 2-year horizon, PPI 1^st is the most cost-effective strategy from both the payer and the societal perspectives.

Figure 3:

Cost-effectiveness acceptability curve of probabilistic sensitivity analysis modeling 10,000 simulations using the primary (payor perspective) model at (A) a 5-year time horizon and (B) a 2-year time horizon, and using the societal perspective model at (C) a 5-year time horizon and (D) a 2-year time horizon. Payor perspective models favored the SFED 1^st strategy at the $100,000 willingness-to-pay (WTP) threshold. Conversely, PPI 1^st was favored at this WTP threshold in the societal perspective model.

Figure 3:

Cost-effectiveness acceptability curve of probabilistic sensitivity analysis modeling 10,000 simulations using the primary (payor perspective) model at (A) a 5-year time horizon and (B) a 2-year time horizon, and using the societal perspective model at (C) a 5-year time horizon and (D) a 2-year time horizon. Payor perspective models favored the SFED 1^st strategy at the $100,000 willingness-to-pay (WTP) threshold. Conversely, PPI 1^st was favored at this WTP threshold in the societal perspective model.

Figure 3:

Cost-effectiveness acceptability curve of probabilistic sensitivity analysis modeling 10,000 simulations using the primary (payor perspective) model at (A) a 5-year time horizon and (B) a 2-year time horizon, and using the societal perspective model at (C) a 5-year time horizon and (D) a 2-year time horizon. Payor perspective models favored the SFED 1^st strategy at the $100,000 willingness-to-pay (WTP) threshold. Conversely, PPI 1^st was favored at this WTP threshold in the societal perspective model.

Figure 3:

Cost-effectiveness acceptability curve of probabilistic sensitivity analysis modeling 10,000 simulations using the primary (payor perspective) model at (A) a 5-year time horizon and (B) a 2-year time horizon, and using the societal perspective model at (C) a 5-year time horizon and (D) a 2-year time horizon. Payor perspective models favored the SFED 1^st strategy at the $100,000 willingness-to-pay (WTP) threshold. Conversely, PPI 1^st was favored at this WTP threshold in the societal perspective model.

Figure 4:

Incremental cost-effectiveness scatterplot comparing pharmacologic therapy 1^st strategies on probabilistic sensitivity analysis. The PPI 1^st strategy was more cost-effective relative to the tCS 1^st strategy in the majority of simulations (indicated by green dots) at a $100,000 willingness-to-pay (WTP) threshold.