Cost-Effectiveness of FF/UMEC/VI vs BUD/GLY/FOR in Moderate-to-Very Severe COPD

Abstract

Background

COPD is a significant cause of morbidity and mortality worldwide and in Colombia. Although single-inhaler triple therapies have demonstrated superior clinical benefits in patients with previous exacerbations, their economic value in middle-income settings remains uncertain. Our objective was to evaluate the long-term cost-effectiveness of fluticasone furoate/umeclidinium/vilanterol (FF/UMEC/VI) versus budesonide/glycopyrronium/formoterol (BUD/GLY/FOR) in Colombian patients with COPD.

Methods

We developed an individual-level microsimulation model to compare FF/UMEC/VI with BUD/GLY/FOR from the Colombian public payer perspective. The model incorporated GOLD-based progression, exacerbations, mortality, dynamic adherence functions, costs, and utilities derived from a Colombian cohort study.

Results

FF/UMEC/VI increased QALYs by 0.51 and saved USD 1,038 per patient, remaining dominant across deterministic and probabilistic sensitivity analyses. Net monetary benefit (NMB) was higher with FF/UMEC/VI. After 10,000 microsimulations, FF/UMEC/VI showed > 95% probability of cost-effectiveness at Colombia’s willingness to pay of USD 5,180/QALY.

Conclusions

This is the first economic evaluation of FF/UMEC/VI and BUD/GLY/FOR in Colombia, incorporating local real-world data and adherence. Results demonstrate that FF/UMEC/VI is a cost-effective and often dominant option, supporting its adoption in resource-limited settings.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12913-026-14551-w.

Introduction

Chronic obstructive pulmonary disease (COPD) remains a leading global health challenge, currently ranked as the third primary cause of death and a major contributor to disability-adjusted life years (DALYs) worldwide [1]. Despite widespread adoption of maintenance inhaled therapies, many patients continue to experience frequent exacerbations that accelerate lung function decline, impair health-related quality of life (HRQoL), and increase healthcare costs and mortality [2–4]. In Colombia, COPD prevalence is 8.9% [5], with DALYs and mortality peaking in 2019, underscoring a disease burden that may exceed global estimates. This context highlights the urgent need for high-value interventions that improve outcomes while optimizing scarce healthcare resources.

Between 2015 and 2022, the disease burden measured in DALYs steadily increased, reaching 15.8 per 1,000 inhabitants in 2019. That year, prevalence rose to 8.63 cases per 1,000 population and exceeded 137 per 1,000 among individuals aged 80 years or older [5]. Mortality attributable to COPD also peaked in 2019 (0.33 deaths per 1,000), primarily affecting individuals over 60 years of age. Notably, the 60–69 age group accounted for the most significant loss in DALYs, with 257,443 years lost in 2020 alone [5]. These figures suggest that the COPD burden in Colombia may surpass global estimates and underscore the urgent need to optimize preventive and therapeutic interventions to reduce disease-related morbidity and mortality [5, 6]. Recent trials such as IMPACT and ETHOS demonstrated that FF/UMEC/VI and BUD/GLY/FOR reduced exacerbation rates versus dual therapies [7, 8]. Moreover, once-daily FF/UMEC/VI may offer additional advantages for adherence and persistence compared with twice-daily regimens such as BUD/GLY/FOR, which is particularly relevant in real-world populations with multimorbidity and polypharmacy[9]. However, these studies were conducted in high-income countries and did not address the budgetary challenges faced by middle-income health systems like Colombia [10].

In Colombia, FF/UMEC/VI and BUD/GLY/FOR represent the two primary single-inhaler triple therapy options currently reimbursed within the public healthcare system, making their economic comparison directly relevant for national decision-making. The objective of this study is to evaluate the long-term cost-effectiveness of FF/UMEC/VI compared with BUD/GLY/FOR in patients with moderate to very severe COPD in Colombia, from the perspective of the national healthcare system. Using an individual-level microsimulation model parameterized with clinical trial data, real-world Colombian cohort characteristics, and country-specific cost inputs, we projected clinical outcomes, healthcare resource utilization, and costs over a lifetime horizon. This analysis provides essential evidence for national decision-makers regarding the prioritization, inclusion, and reimbursement of triple therapies for COPD in Colombia.

Methods

Study design and modelling approach

We conducted a cost-effectiveness analysis using an individual-level microsimulation model to evaluate the long-term outcomes of fluticasone furoate/umeclidinium/vilanterol (FF/UMEC/VI) compared with budesonide/glycopyrronium/formoterol (BUD/GLY/FOR) in patients with COPD in Colombia. The model simulated disease progression (Global Initiative for Chronic Obstructive Lung Disease (GOLD) stages based on FEV₁% predicted), exacerbation risk (moderate and severe), treatment persistence, mortality, costs, and health outcomes over a lifetime horizon. Cycle length was set at one month to capture the timing of exacerbation events and treatment discontinuation adequately. Analysis was conducted from the Colombian public healthcare payer perspective, with costs expressed in 2024 USD. A lifetime horizon was applied (up to 100 years) to capture all relevant long-term consequences, though results were validated with shorter-horizon scenarios for robustness.

Population and comparators

The target population reflected Colombian patients with moderate to very severe COPD (GOLD 2–4), consistent with the population most likely to receive single-inhaler triple therapy in clinical practice. Baseline demographic and clinical characteristics, including FEV₁%, body mass index (BMI), exacerbation history, smoking status, and comorbidity burden, were obtained from a Colombian historical cohort study of 200 patients with COPD managed in two specialized centers in Bogotá [11]. These locally representative data were complemented with efficacy and disease progression parameters extracted from the ETHOS and IMPACT trials to ensure external validity and comparability with international evidence [7, 8].

• Intervention: once-daily single-inhaler FF/UMEC/VI (100/62.5/25 µg).

• Comparator: twice daily BUD/GLY/FOR (160/18/9.6 µg).

Model structure

The microsimulation tracked patients individually, accounting for heterogeneity in baseline characteristics, treatment persistence, and exacerbation risk. Each simulated patient transitioned monthly through health states defined by GOLD severity (FEV₁%), exacerbation history, and survival status (Fig. 1). Treatment discontinuation was modelled as a time-dependent event, with published real-world persistence data applied as detailed above [12, 13].In each monthly cycle, patients could remain in their current state, transition to a more advanced COPD stage regardless of exacerbation status or die. Exacerbations were modelled as temporary events that patients could experience and then exit within the same monthly cycle, without altering the underlying COPD severity. Moderate exacerbations were defined as episodes requiring systemic corticosteroids, antibiotics, or both for at least 3 days. Severe exacerbations were those leading to hospitalization or emergency care. Given the progressive nature of COPD, improvements in FEV₁ were not modelled, and backward transitions to less severe health states were not allowed. In the base case, exacerbations did not directly alter GOLD stage progression, consistent with prior economic models. However, to address structural uncertainty, we performed a scenario analysis in which severe exacerbations increased the probability of transition to the next GOLD stage by 25%.

Fig. 1.

Model structure for moderate-to-severe COPD

Time horizon and discounting

The model simulated a hypothetical cohort of patients i initiating treatment at the observed mean age of 67 years from the Colombian cohort, with a lifetime time horizon extending up to 100 years[6]. A lifetime horizon was selected to fully capture the chronic and progressive nature of COPD, including cumulative exacerbation risk and long-term mortality impacts associated with triple therapy. Future costs and health outcomes were discounted at an annual rate of 5%, consistent with local health economic guidelines [14] and reflecting the long-term nature of the analysis.

Data sources

Transition probabilities between COPD severity stages (GOLD 2→3 and GOLD 3→4), as well as the monthly probabilities of moderate and severe exacerbations by GOLD stage, were derived from the ETHOS trial and are summarized in Table 1. The relative risk of moderate or severe exacerbation with FF/UMEC/VI versus BUD/GLY/FOR was extracted from the the IMPACT and their 95% confidence intervals from real-world comparative study by Wedzicha et al. [15]. Although treatment effects were derived from pivotal RCTs (ETHOS and IMPACT), these were complemented with real-world comparative effectiveness data and calibrated to Colombian baseline characteristics to enhance contextual validity.Estimated annual discontinuation and persistence rates for both FF/UMEC/VI and BUD/GLY/FOR were obtained from the claims-based adherence study by Young et al. [9]. Adverse events were not explicitly modelled in this economic evaluation. This decision was based on the observation that the incidence of treatment-emergent adverse events was nearly identical between treatment arms in the pivotal triple-therapy trials. For example, in ETHOS, the proportion of patients experiencing at least one adverse event was 61% in the BUD/GLY/FOR group and 62% in the comparator arms, with no clinically meaningful differences in serious adverse events. Given this minimal absolute difference and the absence of any imbalance in severe safety outcomes, we assumed that inclusion of adverse event–related costs or disutility would not materially affect incremental results. Therefore, they were omitted to maintain model parsimony.

Table 1.

Base case inputs

COPD health state transitions	Monthly Probability	Source
Moderate COPD to severe COPD	0.030	(37)
Severe COPD to very severe COPD	0.012
From stable GOLD moderate to stable GOLD severe	0.023
From stable GOLD moderate to stable GOLD very severe	0.006
From stable GOLD moderate to severe exacerbation	0.006
From stable GOLD severe to severe exacerbation	0.012
From stable GOLD very severe to severe exacerbation	0.017
From stable GOLD moderate to Moderate exacerbation	0.066
From stable GOLD severe to Moderate exacerbation	0.077
From stable GOLD very severe to Moderate exacerbation	0.087
RR of FF/UMEC/VI (moderate exacerbation)	0.82	(7, 8).
Monthly Costs by Health State	Monthly Cost (USD)
Moderate COPD	$40.7	(23)
Severe COPD	$70.2
Very Severe COPD	$331.3
Fluticasone Furoate 100 mcg + Vilanterol 25 mcg + Umeclidinium 62,5 mcg	$71.8
Budesonide 320 mcg + Glycopyrronium 18 mcg + Formoterol 9.6 mcg	$26.8
Utility value (EQ-5D-5 L)
Male (%)	52%	(11)
Number of concomitant diseases	1.89
Hospital admissions in the last year	0.08
FEV₁ % predicted	52.7
BMI (kg/m²)	27.3
Emergency Room visits in the last year	1.15

Mortality

During the first 12 months of the simulation, a constant all-cause mortality probability of 0.118% per month was applied to all health states, reflecting mortality patterns observed in patients on triple inhaled therapy within controlled clinical trials for COPD [16]. After the first year, state-specific mortality rates were determined by combining Colombian general population mortality tables with relative risk adjustments specific to COPD severity levels [17]. The following relative risks were applied: 1.40 for patients with moderate COPD, and 2.60 for those in severe or very severe disease states, consistent with published economic models [18, 19]. This method aligns with prior health economic modelling approaches and ensures that disease severity contributes proportionally to excess mortality risk.

Adjustment of severe exacerbation probabilities

To account for the impact of prior exacerbation history on the risk of future severe exacerbations, we adjusted the baseline probabilities of severe exacerbation for each GOLD stage using odds ratios (ORs) derived from the study by Rothnie et al. [20, 21]. In this analysis, the baseline monthly probabilities of severe exacerbation (0.0067 for GOLD 2, 0.0125 for GOLD 3, and 0.01254 for GOLD 4) were obtained from the FULFIL trial. These baseline values were then multiplied by ORs associated with the cumulative number of prior moderate or severe exacerbations, as reported by Rothnie et al. [21]. Specifically, patients with one, two, three, four, or five prior exacerbations had their baseline risk adjusted by ORs of 1.21, 1.61, 1.89, 2.14, and 2.92, respectively. The number of prior exacerbations was dynamically tracked in the microsimulation model using patient-level trackers. This approach ensured that the probability of experiencing a severe exacerbation in any given cycle was conditional not only on disease severity (GOLD stage) but also on the individual’s exacerbation history, thereby reflecting real-world patterns of disease progression.

Adherence adjustment of the treatment effect

To account for differences in adherence between once-daily fluticasone furoate/umeclidinium/vilanterol (FF/UMEC/VI) and twice-daily budesonide/glycopyrronium/formoterol (BUD/GLY/FOR), we implemented a dynamic correction factor applied to the relative risk (RR) of exacerbations observed in clinical trials. Real-world evidence shows that FF/UMEC/VI is associated with higher adherence compared with BUD/GLY/FOR. In a retrospective analysis of US claims data, the mean proportion of days covered (PDC) at 6 months was 0.65 for FF/UMEC/VI and 0.59 for BUD/GLY/FOR, decreasing to 0.57 and 0.50, respectively, at 12 months [9].

We assumed that the decline in adherence over time follows an exponential decay function:

where t is time in months, PDC₀ is the extrapolated initial coverage, and λ is the monthly decay rate. Values of λ were estimated using the ratio between 6 and 12-month PDC values for each therapy. For FF/UMEC/VI, this yielded λ = 0.0217 per month with PDC₀=0.74; for BUD/GLY/FOR, λ = 0.0270 per month with PDC₀=0.70. The resulting functions were:

At each monthly cycle of the microsimulation, we calculated a dynamic adherence correction factor as:

and applied it to the clinical trial–based RR of FF/UMEC/VI vs. BUD/GLY/FOR (RR = 0.82) [15].

Thus, the relative effectiveness of FF/UMEC/VI was allowed to vary dynamically over time, reflecting the observed real-world differences in persistence and adherence between once- and twice-daily single-inhaler triple therapies.

Utilities

Health state utilities were derived using an EQ-5D regression model, with inputs from a Colombian COPD cohort (n = 200) [10], rather than relying exclusively on international trial data [22]. This strengthens external validity for the local context. The EQ-5D utility values were predicted based on clinical parameters, including post-bronchodilator FEV₁% predicted, number of emergency room visits not resulting in hospital admission, number of hospitalizations, number of concomitant diseases, BMI, and sex, as follows:

To derive these input parameters, we used real-world data from a Colombian cohort study [11]. The mean FEV₁%, BMI, and estimated number of ER visits and hospital admissions in the last year were directly extracted or calculated from that dataset. These clinical estimates were applied to the coefficients of the EQ-5D regression to calculate expected utilities for the health states. Comorbidity burden was included in the EQ-5D regression model and therefore indirectly reflected in utility estimation.

Cost data sources

The estimation of COPD management costs was grounded in Colombian real-world data derived from a previously conducted micro-costing analysis [23]. That study quantified the direct medical expenditures associated with stable disease across different severity stages, following national clinical practice guidelines and using official price schedules for medications, outpatient care, diagnostic procedures and laboratory tests. Severity-specific resource utilization patterns—aligned with GOLD criteria—were translated into annual costs, which we converted to monthly values for model use. All monetary figures were updated to 2024 using the health-sector inflation index, and subsequently expressed in USD using an exchange rate of 4,409 COP per dollar. Drug acquisition costs for FF/UMEC/VI and BUD/GLY/FOR and standard inhaled therapies were extracted from the Colombian mandatory benefits package tariffs [24, 25]. This approach ensured that all state-based costs incorporated in the model reflected contemporary Colombian practice and payer-relevant expenditures. The willingness to pay used in our study was USD 5,128 per QALY according to the estimate of Espinosa et al. in Colombia [26]. Net monetary benefit (NMB) was calculated as: (QALYs × WTP) − Cost, expressed in thousands of USD.

Exacerbation costs

Exacerbation-related costs were derived from a retrospective cohort analysis conducted in Bogotá, which evaluated healthcare utilization in more than 500 patients with spirometry-confirmed COPD [27]. The study assessed direct medical costs linked to moderate and severe exacerbations by combining recorded service use—consultations, diagnostic testing, pharmacologic treatment and hospitalization—together with nationally regulated tariffs (Soat de Atención Médica [SOAT] and the Social Insurance Institute [ISS]) complemented by Sistema de Información de Precios de Medicamentos (SISMED) price data. Annual expenditures were reported for each GOLD category. For model integration, we selected geometric mean values to mitigate the influence of extreme cost outliers, adjusted them to 2024 using sector-specific inflation, and converted them to monthly USD estimates. GOLD groups A/B, C and D were mapped to the model’s moderate, severe and very severe COPD states, respectively, allowing costs to be applied consistently with clinical severity and resource-use patterns in the Colombian health system.

Sensitivity analyses

Uncertainty surrounding model inputs was explored through deterministic and probabilistic methods. The one-way sensitivity analysis varied individual parameters across plausible ranges—95% confidence intervals when available from clinical trials, or ± 20% when empirical bounds were absent—to identify key drivers of the incremental net monetary benefit. The parameters with the greatest influence were subsequently displayed in tornado-plot format. Finally, a probabilistic sensitivity analysis with 10,000 Monte Carlo simulations assigned distributional forms appropriate to each parameter type (beta for probabilities and utilities, gamma for costs, log-normal for relative risks). Beta distributions were assigned to probabilities and utility values; gamma distributions to cost parameters; and log-normal distributions to relative risks, reflecting their multiplicative nature and ensuring positivity. A structural sensitivity analysis allowed severe exacerbations to increase the probability of transition to the next GOLD stage by 25%.

Results

FF/UMEC/VI dominated BUD/GLY/FOR, providing + 0.51 QALYs and saving USD 1,038 per patient. In the base case analysis, treatment with FF/UMEC/VI was associated with greater effectiveness compared to BUD/GLY/FOR, yielding an incremental gain of 0.51 quality-adjusted life-years (QALYs), from 9.31 QALYs with BUD/GLY/FOR to 9.82 QALYs with FF/UMEC/VI (Table 2). This improvement in health outcomes was achieved at a cost saving of USD 1,038 per patient, making FF/UMEC/VI the dominant strategy. Under a willingness-to-pay (WTP) threshold of USD 5,180 per QALY—aligned with the Colombian health system context—FF/UMEC/VI was considered cost-effective. Furthermore, FF/UMEC/VI yielded a higher net monetary benefit (NMB) of USD 266.25 compared to USD 262.57 with BUD/GLY/FOR, reinforcing its economic dominance in the Colombian setting. Because FF/UMEC/VI was both less costly and more effective than BUD/GLY/FOR, the incremental cost-effectiveness ratio (ICER) was negative, indicating economic dominance; therefore, ICER interpretation was not required.

Table 2.

Base case analysis

Strategy	Cost (US$)	Incr Cost (US$)	Eff (QALYs)	Incr Eff (QALYs)	NMB *(US$)
FF/UMEC/VI	38 248		9.82		266.25
BUD/GLY/FOR	39 287	1038	9.31	0.51	262.57

*NMB (USD, thousands)

Deterministic sensitivity analysis

The one-way sensitivity analysis (OWSA) confirmed the robustness of the base case findings when applying the Colombian willingness-to-pay (WTP) threshold of USD 5,180 per QALY. As shown in the tornado diagram (Fig. 2), the incremental net monetary benefit (NMB) of FF/UMEC/VI versus BUD/GLY/FOR remained positive across all tested parameter ranges, supporting the conclusion that FF/UMEC/VI is the economically dominant strategy. The parameters exerting the greatest influence on incremental NMB were the relative risk of exacerbations with FF/UMEC/VI, the probability of death following a severe exacerbation, and the baseline EQ-5D utility values. Additional influential variables included the monthly probability of moderate exacerbations under BUD/GLY/FOR treatment, the annual cost of FF/UMEC/VI, and baseline FEV₁% in infrequent exacerbations. Importantly, none of the tested ranges for these key parameters resulted in a negative incremental NMB, indicating that FF/UMEC/VI maintained economic dominance under all plausible scenarios. Less influential drivers included BMI, costs of severe and very severe exacerbations, emergency visits, number of comorbidities, and hospitalization costs. Parameters with the smallest impact were the monthly decay rates of adherence for both FF/UMEC/VI and BUD/GLY/FOR, as well as the probability of moderate exacerbations in the FF/UMEC/VI arm. Even under the most extreme assumptions, the lowest estimated incremental NMB remained positive relative to the expected value of USD 10,114. Overall, the deterministic sensitivity analysis supports the robustness of the base case results. FF/UMEC/VI consistently emerged as a cost-effective and economically dominant option over BUD/GLY/FOR in patients with moderate to very severe COPD, with outcomes primarily driven by treatment effectiveness in reducing exacerbations, mortality risk after severe exacerbations, and quality-of-life utilities.

Fig. 2.

Tornado Diagram: Incremental NMB FF/UMEC/VI vs. BUD/GLY/FOR. Abbreviations: RR_FFUMECVI = relative risk of exacerbations with fluticasone furoate/umeclidinium/vilanterol (FF/UMEC/VI) versus BUD/GLY/FOR; pdeathSevEx = probability of death following a severe exacerbation; Eq. 5D_Utility = baseline EQ-5D utility value applied to health states; pModSev_BUD/GLY/FOR = probability of experiencing a moderate exacerbation under BUD/GLY/FOR treatment; Cost_FFUMECVI = annual treatment cost of FF/UMEC/VI; FEV1_IE = mean post-bronchodilator FEV₁% predicted in infrequent exacerbators; pSevVerySev_BUD/GLY/FOR = probability of severe exacerbation in very severe COPD patients under BUD/GLY/FOR treatment; BMI = mean body mass index of the COPD cohort; CostVerySevere = annual management cost for very severe COPD health state; Emergency = probability of an emergency room visit not resulting in hospitalization; Diseases = mean number of comorbidities per patient; Cost_BUD/GLY/FOR = annual treatment cost of budesonide/glycopyrronium/formoterol (BUD/GLY/FOR); CostSevere = annual management cost for severe COPD health state; Hospital = probability of hospitalization for causes other than COPD exacerbations; CostModerate = annual management cost for moderate COPD health state; rate_decay_adherence_FFUMECVI = monthly decay rate of adherence to FF/UMEC/VI; rate_decay_adherence_BUD/GLY/FOR = monthly decay rate of adherence to BUD/GLY/FOR

Probabilistic sensitivity analysis

The mean incremental cost across simulations was − USD 1,012 (95% CI − 2,845 to 423), and the mean incremental QALY gain was 0.48 (95% CI 0.21–0.73). The cost-effectiveness scatterplot from the probabilistic sensitivity analysis (PSA) is presented in Fig. 3. Out of 10,000 Monte Carlo simulations, 45.66% of iterations fell within the southeast quadrant (Quadrant IV), where FF/UMEC/VI is both more effective and less costly than BUD/GLY/FOR, indicating dominance. An additional 19.48% of simulations were located in the northeast quadrant (Quadrant I) with incremental cost-effectiveness ratios (ICERs) below the Colombian willingness-to-pay (WTP) threshold of USD 5,180 per QALY, further supporting FF/UMEC/VI as a cost-effective strategy under uncertainty.

Fig. 3.

ICE Scatterplot, FF/UMEC/VI vs. BUD/GLY/FOR*. *Each point represents one Monte Carlo simulation. The vertical axis indicates incremental QALYs, and the horizontal axis indicates incremental costs (USD). The southeast quadrant represents dominance

In contrast, 8.84% of simulations fell in Quadrant III with lower costs and effectiveness but ICERs above the WTP threshold, reflecting scenarios less favourable to FF/UMEC/VI. Only 0.04% of iterations were located in Quadrant I with ICERs above the threshold. In comparison, 25 simulations (0.5%) fell in the dominated region (Quadrant II), where FF/UMEC/VI is both more costly and less effective than BUD/GLY/FOR. A further 25.48% were in Quadrant III with lower costs and effectiveness but still below the threshold, suggesting potential cost-effectiveness despite reduced health gains. Overall, 65.1% of simulations placed FF/UMEC/VI in the dominant or cost-effective regions, reinforcing the robustness of the base-case results. These findings support FF/UMEC/VI as a cost-effective alternative to BUD/GLY/FOR in Colombian patients with moderate to very severe COPD, driven by its favourable balance of improved health outcomes and reduced or modestly increased costs under most plausible scenarios.

The cost-effectiveness acceptability curve (CEAC) presented in Fig. 4 corroborates the findings from the base case and the probabilistic sensitivity analyses. As the willingness-to-pay (WTP) threshold increases, the probability that FF/UMEC/VI is the cost-effective strategy rises steadily. At a WTP of US$500 per QALY, FF/UMEC/VI already shows a probability greater than 70% of being the preferred option. This probability surpasses 90% once the WTP threshold reaches approximately US$2,000 per QALY and remains above this level across higher thresholds. At Colombia’s commonly accepted WTP threshold of US$5,180 per QALY, FF/UMEC/VI demonstrates a near-absolute probability (> 95%) of being cost-effective compared with BUD/GLY/FOR.

Fig. 4.

CE Acceptability Curve*. *The curve represents the probability that FF/UMEC/VI is cost-effective at increasing willingness-to-pay thresholds

In contrast, the probability of BUD/GLY/FOR being cost-effective declines rapidly as the WTP threshold increases, remaining below 5% across most of the range. These results reinforce the economic robustness of FF/UMEC/VI and confirm its dominant or cost-effective status across a wide spectrum of potential decision-maker thresholds. The CEAC highlights the high likelihood that FF/UMEC/VI delivers superior value for money in the management of moderate to very severe COPD within the Colombian healthcare system. Under the accelerated progression scenario, FF/UMEC/VI remained dominant, with an incremental QALY gain of 0.46 and cost savings of USD 912. Model validity checks demonstrated clinical plausibility. In a scenario without adherence adjustment, the incremental QALY gain was 0.44 and cost savings were USD 821, confirming that adherence modeling enhances but does not drive dominance. Modeled life expectancy was 11.2 years for GOLD moderate, 8.7 years for GOLD severe, and 6.1 years for GOLD very severe, consistent with published epidemiologic estimates [6]. Annual severe exacerbation rates ranged from 0.18 (GOLD moderate) to 0.41 (GOLD very severe), aligning with Colombian cohort data [6, 11, 28]. Mean utilities by stage (0.74, 0.66, 0.55 respectively) were comparable to observed EQ-5D estimates [29]. Under accelerated progression, FF/UMEC/VI remained dominant (incremental QALY 0.46; savings USD 912).

Discussion

In this cost-effectiveness analysis, FF/UMEC/VI emerged as the dominant strategy compared with BUD/GLY/FOR in Colombian patients with moderate to very severe COPD. In the base case, FF/UMEC/VI provided an incremental gain of 0.51 QALYs at a cost saving of USD 1,038 per patient, leading to higher net monetary benefit values and confirming its economic superiority under the Colombian willingness-to-pay threshold. The robustness of these findings was further reinforced by both deterministic and probabilistic sensitivity analyses, where FF/UMEC/VI consistently demonstrated a high probability of being cost-effective across a wide range of plausible scenarios. From a public health perspective, these results suggest that adopting FF/UMEC/VI as the preferred single-inhaler triple therapy could generate meaningful improvements in health outcomes while reducing costs for the healthcare system [30]. Given the high burden of COPD in Colombia and the growing demand for efficient allocation of limited healthcare resources, the economic dominance of FF/UMEC/VI represents an opportunity to enhance patient outcomes and optimize resource use at a population level. These results are particularly applicable to resource-limited settings with similar drug pricing structures and willingness-to-pay thresholds.

Our findings are broadly consistent with those reported by Cai et al. in the UK setting, where FF/UMEC/VI also emerged as the dominant strategy compared with other single-inhaler triple therapies, generating incremental QALY gains of 0.23–0.31 and cost savings of £1,221–1,721 per patient over a lifetime horizon [31]. However, the magnitude of these QALY gains and cost savings in our Colombian model were more modest. These differences likely reflect methodological choices: while Cai et al. employed the GALAXY-COPD model populated with patient-level data from the IMPACT trial and treatment effects from a network meta-analysis, our analysis relied on a patient-level microsimulation calibrated to a Colombian cohort and used real-world estimates of adherence and local healthcare costs. Moreover, Cai et al. assumed zero discontinuation beyond year one and applied UK-specific utility mapping from SGRQ to EQ-5D, whereas our model explicitly incorporated dynamic persistence functions and utilities derived from Colombian data using a regression-based EQ-5D equation. These methodological differences highlight how population characteristics, model structure, and data inputs can meaningfully influence the size of estimated benefits, while both studies converge in demonstrating that FF/UMEC/VI provides superior health outcomes at lower costs compared with BUD/GLY/FOR.

Our results align with those reported by Beeh et al. in Germany, where FF/UMEC/VI was also found to be the dominant strategy compared with multiple-inhaler triple therapy, yielding incremental QALY gains of 0.261 and cost savings of approximately €2,850 per patient [32]. In our Colombian analysis, FF/UMEC/VI similarly dominated BUD/GLY/FOR, providing comparable health and cost benefits. However, important methodological differences may explain the variation in effect size across settings. The German study used the GALAXY-COPD model and relied on treatment effects derived directly from the INTREPID pragmatic trial, where effectiveness was measured over 24 weeks using CAT and FEV₁ improvements. In contrast, our model applied a microsimulation framework calibrated to Colombian real-world data, adjusted exacerbation risks using relative risks from ETHOS and IMPACT, and explicitly incorporated adherence decay over time. Furthermore, the German analysis assumed treatment effects remained constant over a lifetime and modelled discontinuation only in year one, while our approach applied dynamic persistence functions. These methodological distinctions—particularly the use of local utility mapping, exacerbation-based progression, and country-specific costs—likely explain the larger incremental QALYs but lower absolute cost offsets in our model. Together, both studies reinforce the robustness of FF/UMEC/VI as a dominant therapy across diverse health systems, while highlighting how model structure and local inputs can shape the magnitude of economic outcomes.

The results of our model are consistent with recent real-world evidence supporting the clinical effectiveness of FF/UMEC/VI. The RESTART study in Korea demonstrated that switching from multiple-inhaler regimens to once-daily FF/UMEC/VI was associated with significant improvements in symptoms (CAT score), lung function, and a 55% reduction in exacerbation rates (IRR 0.45, p = 0.016), alongside higher treatment satisfaction [33]. Similarly, the ELLITHE study in Germany reported clinically meaningful reductions in exacerbations and improvements in health-related quality of life among patients treated with FF/UMEC/VI under routine practice conditions [34]. Furthermore, a large comparative analysis of more than 32,000 patients in the United States confirmed that FF/UMEC/VI was associated with a 12% reduction in the annual risk of moderate-to-severe exacerbations and an 11% reduction in mortality compared with BUD/GLY/FORM [35]. By combining clinical trial efficacy, real-world adherence, and local cost data, our model provides robust evidence that FF/UMEC/VI is not only clinically beneficial but also economically efficient in Colombia.

Our findings are consistent with the large real-world comparative effectiveness study by Feldman et al. [36] which analyzed 20,388 propensity-matched pairs of COPD patients and demonstrated that initiation with BUD/GLY/FOR was associated with a significantly higher risk of first moderate-to-severe exacerbation (HR 1.09; moderate HR 1.07; severe HR 1.29) compared with FF/UMEC/VI, while rates of pneumonia and mortality were similar between groups. The clinical superiority of FF/UMEC/VI, reinforced by its once-daily dosing and delivery via a dry-powder inhaler, combined with dynamic adherence adjustments in our model, explains the observed economic advantage through reduced exacerbation-related healthcare utilization. Importantly, the absence of excess pneumonia risk suggests that the cost benefit is not offset by safety concerns. Taken together, our results and those of Feldman et al. provide convergent evidence supporting FF/UMEC/VI as the higher-value therapeutic option in COPD management, with the added societal benefit of a lower environmental footprint from dry-powder inhalers relative to pressurized metered-dose devices. Differences in incremental QALY magnitude across studies likely reflect variation in baseline population characteristics, local utility inputs, modeling structure, and assumptions regarding treatment persistence, rather than inconsistencies in clinical effectiveness.

Both the deterministic and probabilistic sensitivity analyses confirmed the robustness of the base case results, highlighting the consistency of FF/UMEC/VI as a cost-effective and often dominant strategy compared with BUD/GLY/FOR. In the one-way sensitivity analysis, the most influential drivers of incremental net monetary benefit were the relative risk of exacerbations with FF/UMEC/VI, the probability of death after a severe exacerbation, and baseline utility values, yet even under extreme assumptions, FF/UMEC/VI retained economic dominance. Similarly, the probabilistic sensitivity analysis demonstrated that more than 65% of simulations fell within the dominant or cost-effective regions, with a probability exceeding 95% of cost-effectiveness at the Colombian willingness-to-pay threshold. These findings underscore that the favourable economic profile of FF/UMEC/VI is not contingent on a narrow set of assumptions but persists across a broad spectrum of clinical and economic uncertainty, thereby reinforcing its value as a reliable treatment option for COPD management in Colombia.

This study has several limitations that should be acknowledged. First, efficacy inputs were based on randomized trials (ETHOS, IMPACT), which may not capture the full heterogeneity of Colombian patients. Although RCT-derived relative risks were used, real-world treatment effects in Colombia may differ due to variations in adherence, comorbidity burden, and healthcare access. Second, local utility and cost data were limited, requiring some adjustments from international sources. Third, although adherence was modelled dynamically, extrapolation beyond 12 months introduces structural uncertainty. Finally, indirect costs (productivity, caregiver burden) were not included, potentially underestimating the full societal value of exacerbation reduction. Minor local adverse events potentially influencing treatment switching were not explicitly modeled, as reliable comparative Colombian data were unavailable. Additionally, although adherence decay was modeled dynamically using real-world data, extrapolation beyond observed follow-up introduces structural uncertainty, which may influence long-term projections. Adherence decay parameters were derived from US claims data; cultural and health system differences may influence persistence patterns in Colombia. Future research in Colombia and other middle-income countries would benefit from the development of centralized COPD registries and broader adoption of interoperable electronic health records. Such infrastructure would improve the availability of high-quality longitudinal data on disease progression, exacerbations, health-related quality of life, and real-world costs. Enhanced data systems would enable more precise local parameterization of economic models and better inform reimbursement decisions for emerging therapies in COPD and other chronic respiratory diseases.

In conclusion, FF/UMEC/VI is a cost-effective intervention for moderate to severe COPD in Colombia. Its implementation could reduce exacerbations, improve quality of life, and optimize resource use, making it a high-value strategy for public health systems in resource-constrained settings.

Supplementary Information

Below is the link to the electronic supplementary material.

C. Celis-Preciado

reports that he has received outside the current work, education scholarships from the Université de Sherbrooke and FRQS (Fonds de Recherche du Québec-Santé) and is a recipient of the Canadian Allergy, Asthma, and Immunology Foundation (CAAIF) Research Fellowship in Asthma supported by GSK; he has received speaker honoraria from AstraZeneca, GlaxoSmithKline and Sanofi-Regeneron and received consultancy fees from AstraZeneca, GlaxoSmithKline and Sanofi-Regeneron

Author contributions

J.A.B, OM, LM and C.C. wrote the main manuscript text and L.M. prepared figures 1-2. All authors reviewed the manuscript.”

Funding

There was no funding for this study.

Data availability

The data that supports the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study did not involve direct interaction with human participants or the use of individualized human data. All model parameters were derived exclusively from previously published clinical trials and aggregated secondary data. This study was approved by the Institutional Review Board of University of Antioquia. The need for consent was waived by an IRB (2015–4690), following local regulations of resolution 8430/93, because this is a study which all information was extracted from the literature and does not use data from patients”. Nevertheless, the research was conducted in full accordance with the principles outlined in the Declaration of Helsinki, and no identifiable personal information was collected, accessed, or processed.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.