Cost-effectiveness of deucravacitinib versus apremilast of moderate-to-severe plaque psoriasis in China

Abstract

Objectives

The study aimed to assess the cost-effectiveness of deucravacitinib versus apremilast for treating moderate-to-severe plaque psoriasis from the Chinese healthcare system’s perspective.

Methods

The treatment efficacy of deucravacitinib was compared with apremilast using response rates derived from the head-to-head phase 3 clinical trials, POETYK PSO-1 and PSO-2. A decision-tree (first 24-week)/ Markov model (later period) was constructed to estimate the incremental cost per quality-adjusted life-year (QALY) gained over a lifetime horizon. The efficacy inputs were based on randomized controlled trials, while adverse event rates, discontinuation probabilities, costs, and utility data were obtained from relevant literature and Chinese sources. A 5% annual discount rate was used for the analysis of outcomes and costs. Model outcomes were characterized by quality-adjusted life-years (QALYs) and incremental cost-effectiveness ratio (ICER). One-way sensitivity analysis and probability sensitivity analysis (PSA) were performed to examine the robustness of the results.

Results

According to the assumed lifetime horizon and model, the ICER of deucravacitinib 6 mg once daily compared with apremilast 30 mg twice daily was 140,047 CNY per QALY. Deucravacitinib was more cost-effective than apremilast at the willingness-to-pay (WTP) threshold of 287,247 CNY per QALY. In the One-way sensitivity analysis, the cost of deucravacitinib was identified as the parameter exerting the greatest impact on the base-case results. The results of PSA showed the probability of deucravacitinib being cost-effective was 99.4%.

Conclusion

At the WTP threshold of 287,247 CNY, deucravacitinib 6 mg once daily was a cost-effective treatment strategy for moderate-to-severe plaque psoriasis compared with apremilast 30 mg twice daily from the Chinese healthcare system perspective over a lifetime horizon.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12913-025-13403-3.

Key points

1. This study constitutes the first comprehensive economic evaluation of deucravacitinib versus apremilast for treating moderate-to-severe plaque psoriasis in China.

2. Following the decision-tree/Markov model, the economic evaluation indicates that deucravacitinib is a new cost-effective treatment for adult patients with moderate-to-severe plaque psoriasis.

3. This empirical evidence can be used to inform debates about the use of QALYs in drug price negotiations, reimbursement, and coverage policy making.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12913-025-13403-3.

Introduction

Psoriasis is a chronic, immune-mediated and recurrent inflammatory skin disorder, characterized by scaly erythema or plaques with localized or widespread distribution [1]. The pathogenesis of psoriasis involves genetics, immunologic disarrangement, autoantigens, and multiple environmental factors. Recently, it has been linked to a theory that this disease is primarily driven by pathogenic T-cells; specifically, psoriasis-related pro-inflammatory factors, IL-23, and Th-17 cells are believed to play a central role [2–4]. The prevalence rate of psoriasis ranges from 0.51 to 11.43% worldwide [5]. In China, epidemiological analyses have shown an age-standardized incidence of 59.7 per 100,000 person-years, with a persistent upward trend [6]. This disease burden has demonstrated a significant demographic predominance among young and middle-aged males [7]. Psoriasis severity is affected by some factors, such as depression, psoriatic arthritis, and cardiometabolic syndrome, with approximately 20% of patients progressing to moderate-to-severe psoriasis [8, 9]. Plaque psoriasis is one of the most common types of psoriasis, accounting for approximately 80–90% of cases [10]. This disease is incurable and unpredictable, with onset possible at any age. It imposes profound detrimental effects on patients’ quality of life (QoL), with 61.8% of Chinese patients reporting severe to very severe QoL impairment in national cohort studies [11, 12]. Most patients with moderate-to-severe psoriasis require lifelong treatment [13]. The economic burden of psoriasis has become a main problem that the World Health Organization (WHO) needs to address. Therefore, it is essential to identify and implement currently available and cost-effective treatment options for psoriasis patients.

There are several treatment options for patients. In China, as in other countries, topical vitamin D3 derivatives, topical corticosteroids (TCS), and skin emollients are first-line options, and phototherapy is considered an effective approach for treating refractory psoriasis. Based on current Chinese clinical practice guidelines, following traditional drugs such as methotrexate and cyclosporine, systemic treatments including biologics and oral small-molecule drugs have emerged as the principal therapeutic agents for moderate-to-severe psoriasis and its special subtypes, demonstrating established clinical efficacy [14, 15]. Among the new generation of oral systemic therapies, two small-molecule drugs with distinct mechanisms, apremilast and deucravacitinib, have gained clinical prominence in the treatment of moderate-to-severe plaque psoriasis. Notably, apremilast and deucravacitinib are both covered under China’s National Reimbursement Drug List (NRDL), which can significantly reduce the financial burden on patients associated with these specific oral agents. While most biologics exhibit comparable efficacy to apremilast and deucravacitinib, they are typically associated with higher costs and require subcutaneous administration, which may negatively impact patient adherence [16]. Given that deucravacitinib and apremilast hold the same status as oral systemic drugs and target similar patient populations, it is imperative to compare their accessibility.

Deucravacitinib exerts therapeutic effects by modulating immune signaling pathways involving IL-12, IL-23, and type I interferons. Its approval in China was fast-tracked following the POETYK PSO trials in 2023. Clinical trial data have demonstrated significant and sustained efficacy of deucravacitinib in moderate-to-severe plaque psoriasis, with a favorable safety profile observed in Asian patients [17]. Apremilast, the first approved small-molecule targeted therapy for psoriasis in China, is an oral phosphodiesterase-4 (PDE4) inhibitor. Long-term study has suggested that improvements in Psoriasis Area and Severity Index (PASI) scores, itch, nail involvement, and QoL indicators for plaque psoriasis patients treated with apremilast may be sustained for up to 5 years [18]. A network meta-analysis has indicated that patients treated with deucravacitinib achieve superior treatment responses and greater gains in quality-adjusted life-years (QALYs) compared to those treated with apremilast [19]. However, comparative cost-effectiveness research evaluating deucravacitinib versus apremilast is currently limited in China. This study aims to use a decision-tree (first 24-week)/Markov model to assess the cost-effectiveness of deucravacitinib versus apremilast in the treatment of moderate-to-severe plaque psoriasis from a Chinese healthcare perspective.

Methods

Target population and interventions

The clinical data for the target population used in the economic analysis were from two global phase 3, double-blind, randomized trials in moderate-to-severe plaque psoriasis (POETYK PSO-1 and PSO-2), which assessed the efficacy and safety of deucravacitinib compared to apremilast [20]. Both trials enrolled adults aged ≥ 18 years with moderate-to-severe plaque psoriasis for at least 6 months, of whom 65.9% were male. The patients were randomized in a 2:1 ratio to receive deucravacitinib 6 mg once daily (n = 842) or apremilast 30 mg twice daily (n = 422). 79.0% of the population suffered from sPGA 3 (moderate), while 21.0% had sPGA 4 (severe), with an average duration of approximately 18 years.

Model structure and assumptions

A decision-tree (first 24-week)/Markov model based on clinical pathways and disease progression was constructed using TreeAge Pro 2022 to evaluate the cost-effectiveness of deucravacitinib versus apremilast in China (Fig. 1a, b). The model was adapted from previously published economic analyses of plaque psoriasis [16, 21]. The Markov model adopted a lifetime horizon of 32 years, structured into 4-week cycles. This time horizon was based on the average age of patients at enrollment (46 years) and an assumed life expectancy of 78 years, consistent with the general Chinese population [22].

Fig. 1.

a Model structure: short-term decision tree model. PASI Psoriasis Area Severity Index, SoC standard of care. b Model structure: long-term Markov model

The model was initiated with a 24-week induction period, and the PASI scores at week 24 were used to stratify patients into health states. Patients were assigned to two health states based on PASI scores: PASI ≥ 75 (responder) and PASI < 75 (non-responder) [23]. At week 24, responders continued the same treatment and entered the “active treatment” state in the Markov model, while non-responders switched to the “standard of care (SoC)” or “death” states.

The Markov model comprised four health states defined by the treatment response and treatment approach: initial state (decision-tree period), active treatment, SoC, and death. In each 4-week cycle, patients in a given health state received corresponding treatment and were reassessed at the end of the cycle, at which point they could either remain on the current treatment or transition to the next treatment until death or the end of the model horizon. Subsequent transitions to second-line biologics following deucravacitinib or apremilast was excluded from the model. Standard of care included methotrexate, cyclosporine, topical corticosteroids, and UVB phototherapy [14]. Different clinical probabilities were applied to each state transition. A half-cycle correction was implemented in the model to avoid overestimating expected survival.

Model input parameters

Efficacy parameters

For the decision-tree model, the baseline PASI response rates at week 24 were derived from the phase 3 POETYK PSO-1 and PSO-2 trials [20].

Discontinuation and mortality parameters

For the Markov model, the discontinuation rate was defined as the probability of transitioning from the “Active treatment” state to the “SoC” state. For the deucravacitinib arm, a 3-year discontinuation rate of 35.7% was estimated based on data from the phase 3 POETYK PSO-1, PSO-2, and long-term extension (LTE) trials [24]. For comparator interventions, apremilast exhibited a substantially higher discontinuation rate within 4 years, at 59% [25]. The 4-week transition probabilities between the “Active treatment” and “SoC” states were shown in Table 1.

Table 1.

Transition probabilities inputs

Treatment	Base case	Range	Distribution	Sources
Deucravacitinib
Baseline to Responder per 24w	0.623	0.561–0.685	Beta	PSO-1 + PSO-2 [20]
Active treatment to SoC per 4w	0.0119	0.0107–0.0131	Beta	Long-term study [24]
Apremilast
Baseline to Responder per 24w	0.379	0.341–0.417	Beta	PSO-1 + PSO-2 [20]
Active treatment to SoC per 4w	0.0170	0.0153–0.0187	Beta	Long-term study [25]

SoC standard of care

This study assumed that the mortality rate was minimally impacted by deucravacitinib and apremilast over the 24-week period. The transition probabilities from each state to death were based on the dynamic mortality rates published in the China Population Census Yearbook, adjusted by the mortality hazard ratio (HR: 1.41) for patients with psoriasis (Table 2) [26, 27]. Both the discontinuation and mortality probabilities were adjusted for the 4-week cycle length according to the standard exponential conversion formula: r = -[ln(1-P₁)]/t₁, P₂ = 1-exp(-rt₂) [28].

Table 2.

The occurrence rate inputs (4 weeks)

Model parameter	Deucravacitinib	Apremilast	Source
Alive of 16-week	100%	100%	PSO-1 + PSO-2 [20]
Mortality
45–49	0.021%		National data [26]
50–54	0.032%
55–59	0.049%
60–64	0.081%
65–69	0.128%
70–74	0.222%
75–79	0.393%
Adverse events
Nasopharyngitis	1.41%	1.05%	PSO-1 + PSO-2 [20]
Upper respiratory tract infection	0.73%	0.51%	PSO-1 + PSO-2 [20]
COVID-19 (including COVID-19 Pneumonia)	0.008%	–	PSO-1 + PSO-2 [20]
Pneumonia	0.015%	0.008%	PSO-1 + PSO-2 [20] long-term study [20, 25]
MACE	0.018%	0.002%	Long-term study [24, 29]
Malignant carcinoma	0.046%	0.013%	Long-term study [24, 30]

MACE major adverse cardiovascular events, COVID-19 corona virus disease 2019

Adverse events parameters

Common adverse reactions associated with deucravacitinib were reported to include nasopharyngitis and upper respiratory tract infection. Severe adverse reactions considered in this analysis included major adverse cardiovascular events (MACE), malignant carcinoma, and severe infections (including pneumonia and coronavirus disease 2019 [COVID-19]) [20, 24, 25, 29, 30].

Health utility parameters

All health utilities for each state were obtained from the published cost-effectiveness analyses of moderate-to-severe plaque psoriasis [16, 21, 31]. Based on the data, the health utility values corresponding to different PASI improvement levels were estimated as follows: 0.660 for baseline state, 0.737 for SoC, and 0.934 for active treatment. Adverse events were assumed to reduce QoL, with a disutility value of −0.05 modeled over a duration of two weeks [21]. Costs and utilities were discounted at 5.0% annually over the lifetime horizon. The health utilities were presented in Table 2.

Cost parameters

In this study, the model included only direct medical costs, categorized into the targeted drug costs, other medical support costs, standard care costs, and adverse event costs per cycle. All these resource data were from some published literature and the Yaozhi database in China [16, 26, 31–36]. All cost inputs were inflated to 2024 using the Chinese consumer price index [26]. The price of deucravacitinib was 371.00 CNY (7 tablets per box, 6 mg per tablet) in 2025, and apremilast was 942.00 CNY (Otelza^®, 60 tablets per box, 30 mg per tablet). The initial dosage of apremilast was 10 mg on the first day, with daily increments of 10 mg until the target dosage of 30 mg twice daily on day 6 was tolerated by patients. For each treatment cycle, the cost of deucravacitinib remained constant, whereas the cost of apremilast was 800.70 CNY for the first 4 weeks and 879.00 CNY for each subsequent 4-week cycle. The total costs for the analysis model were shown in Table 3 and Supplementary Table S1.

Table 3.

Costs and utility parameters (4 weeks)

Model parameter	Base case	Range	Distribution	Source
Cost inputs
Cost for drugs
Deucravacitinib	1484.00	1113.00–1855.00	Gamma	Bidding price [36]
Apremilast	879.20	659.40–1099.00	Gamma	Bidding price [36]
Medical support costs
Pre-treatment assessment	517.19	387.89-646.49	Gamma	Zhang [16]
Monitoring	225.31	168.98–281.64	Gamma	Zhang [16]
Regular physician visit	27.65	20.74–34.56	Gamma	Zhang [16]
Skin cancer screening	137.24	102.93–171.54	Gamma	Zhang [16]
Adverse event costs
Nasopharyngitis	12.99	9.74–16.24	Gamma	An [32]
Upper respiratory tract infection	174.40	130.80–217.99	Gamma	National Bureau [26]
COVID-19 (including COVID-19 Pneumonia)	749.44	562.08–936.80	Gamma	Zhang [33]
Pneumonia	1121.36	841.02–1401.7	Gamma	Zhang [16]
MACE	1126.02	844.51–1407.52	Gamma	China data [35]
Malignant carcinoma	882.43	661.83–1103.04	Gamma	Li [34]
Costs for the standard of care
Regular physician visit	2.13	1.60–2.66	Gamma	Zhang [16]
UVB phototherapy	47.03	35.27–58.79	Gamma	Zhang [16]
Monitoring labwork: methotrexate or cyclosporine	10.48	7.86–13.10	Gamma	Zhang [16]
Health utility (per year)
Baseline	0.660	0.616–0.704	Beta	Udkoff [31]
Active Treatment	0.934	0.921–0.946	Beta	Zhang [16]
SoC	0.737	0.687–0.787	Beta	Udkoff [31]
Adverse event disutility	−0.05	None	Beta	Foley [21]

MACE major adverse cardiovascular events, SoC standard of care

Outcomes

In this economic evaluation, the primary outcome was the incremental cost-effectiveness ratio (ICER), calculated as the difference in costs divided by the difference in QALYs. The cost-effectiveness of interventions was evaluated by comparing the ICER to a predefined willingness-to-pay (WTP) threshold. An intervention was deemed an “advantageous option” when its ICER fell below the WTP threshold, even if total costs were higher, provided it yielded greater QALY gains relative to the comparator. The WTP threshold for China was set at three times the gross domestic product (GDP) per capita in 2024, equivalent to 287,247 CNY per QALY [26]. The discount rate for the cost and utility was 5.0% annually.

Sensitivity analyses

One-way sensitivity analysis and probabilistic sensitivity analysis (PSA) were conducted to assess the impact of model parameter uncertainty on base-case results. The parameter variation ranges used in the one-way sensitivity analysis were shown in Tables 1 and 3. The costs of deucravacitinib and apremilast, as well as medical support, standard of care, and adverse events, were assumed to vary by ± 25% from baseline. The transition probabilities were varied by ± 10%. The feasible ranges of utility values for baseline, active treatment and SoC states were derived from the literature (Table 3) [16, 31]. The discount rate ranged from 0 to 8%. The Monte Carlo simulation of PSA was used to simulate the distribution of uncertain parameters over 1000 iterations. The parameters considered in the PSA included the drug costs, other costs for medical support, standard of care, and adverse events, the utility values for each health state, and transition probabilities. The results were presented as an incremental cost-effectiveness (ICE) scatter plot and a cost-effectiveness acceptability curve (CEAC).

Results

Model validation and base case

The results of total costs, QALYs, and the ICER for patients with moderate-to-severe plaque psoriasis were presented in Table 4. Over a lifetime horizon, deucravacitinib was associated with greater effectiveness and higher costs than apremilast. The incremental cost of deucravacitinib was 44,233 CNY, while the incremental QALY gain was 0.32. Compared to apremilast, the ICER of deucravacitinib was 140,047 CNY per QALY, which fell below the WTP threshold of 287,247 CNY per QALY in China.

Table 4.

The results of cost-effectiveness analysis of Deucravacitinib vs. apremilast

Costs and QALYs per intervention	Deucravacitinib	Apremilast
Total costs	251,699	207,466
Total QALYs	11.86	11.55
Incremental Cost	44,233
Incremental QALY	0.32
ICER	140,047

ICER incremental cost-effectiveness ratio, QALYs quality adjusted life years

Sensitivity analyses

The results of the one-way sensitivity analysis were presented in Fig. 2. Health utility values and transition probabilities were assumed to follow beta distributions, while costs were modeled using a gamma distribution. The tornado ICER chart displayed only the parameters with a significant impact on ICER in this study. The results indicated that it was largely affected by changes in parameters related to the cost of deucravacitinib and the utility values of active treatment. The results were robust according to the one-way sensitivity analysis.

Fig. 2.

Tornado graphs of one-way sensitivity analysis

The ICER scatter plot was obtained through 1,000 iterations of Monte Carlo simulation (Fig. 3). The horizontal axis indicated the incremental effectiveness of deucravacitinib versus apremilast, and the vertical axis indicated the incremental cost. With the WTP threshold of 287,247 CNY as the boundary line, the majority of points fell to the right of this threshold. The results demonstrated that deucravacitinib was more likely to be cost-effective.

Fig. 3.

The incremental cost-effectiveness (ICER) scatter plot

The CEAC was shown in Fig. 4, where the horizontal axis represented the range of WTP thresholds, and the vertical axis represented the probability that the treatment option was cost-effective. As the WTP threshold increased, the probability of cost-effectiveness of deucravacitinib rose steadily. At the WTP threshold of 287,247 CNY per QALY, deucravacitinib had a 99.4% probability of being cost-effective. The results remained robust according to the sensitivity analyses.

Fig. 4.

C-E acceptability curve. PASI psoriasis area and severity index, ICER incremental cost-effectiveness ratio

Discussion

To the best of our knowledge, this study was the first to evaluate the cost-effectiveness of deucravacitinib versus apremilast for treating moderate-to-severe plaque psoriasis using a decision-tree/Markov model from a healthcare system perspective in China, based on data from the POETYK PSO trials. Incorporating efficacy data from the POETYK PSO-1 and PSO-2 trials, as well as long-term clinical inputs, the base-case analysis demonstrated that deucravacitinib resulted in an additional 0.32 QALYs at an incremental cost of 44,233 CNY compared to apremilast. The ICER of deucravacitinib versus apremilast was 140,047 CNY per QALY, falling below the WTP threshold of 287,247 CNY per QALY, establishing that deucravacitinib was a cost-effective option. The robustness of the study was collectively demonstrated through both PSA and one-way sensitivity analyses. Results from the one-way sensitivity analysis indicated that the cost of deucravacitinib had the greatest influence on the results. The favorable cost-effectiveness of deucravacitinib may be largely attributed to a substantial price reduction following its inclusion in the NRDL in China in 2024. Given that these costs may vary across different payers, the results could differ when analyzed from alternative perspectives. In addition to its efficacy and economic advantages, when compared with a twice-daily administration of apremilast, the once-daily dosage of deucravacitinib may offer a lower dosing frequency and better adherence.

From the US payer perspective, deucravacitinib demonstrated a lower cost per PASI 75 response compared with apremilast at both 16 and 24 weeks [37]. Consistent with the prior finding, our study extended the analysis to a lifetime horizon, which enabled a comprehensive assessment of treatment efficacy over the entire treatment period [37]. Based on the dosing frequency of the medication, a 4-week model cycle was used to evaluate the cost-effectiveness, which was more informative for clinical applications.

Nonetheless, the study was subject to several limitations. First, the model projected lifetime efficacy based on short-term (24-week) PASI response rather than 52-week data, owing to the crossover of treatment arms beyond week 24^24,25. As data beyond 24 weeks were unavailable, results from long-term trials were used as inputs, which might have introduced a degree of increased bias. Second, the treatment of SoC was different. Patients with different initial treatments could receive varying subsequent therapies after disease progression. In our study, it was assumed that patients who did not respond to treatment would transfer to SoC. However, due to data limitations, the SoC state in the Markov model was not allowed to transition to other states, which did not fully reflect real-world disease progression. Finally, although apremilast was considered a suitable comparator, an increasing number of biologics and biosimilars have become available as treatment options for moderate-to-severe plaque psoriasis. Therefore, future research should focus more on economic evaluations of these subsequenttargeted therapies.

Conclusions

Economic modeling of moderate-to-severse plaque psoriasis was conducted to evaluate the cost-effectiveness of deucravacitinib versus apremilast. Based on our analysis, deucravacitinib was associated with higher costs but yielded more QALYs than apremilast, resulting in an ICER well below the WTP threshold of 287,247 CNY per QALY. The results suggested that deucravacitinib could be a potentially cost-effective treatment option for adult patients with moderate-to-severe plaque psoriasis.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

 Quality

Adjusted life-year

QALY

Quality-adjusted life-year

QoL

Quality of life

TCS

Topical corticosteroids

TNFα

Tumor necrosis factor alpha

TYK2

Tyrosine kinase 2

PASI

Psoriasis Area and Severity Index

SoC

Standard of care

WTP

Willingness to pay

PSA

Probability sensitivity analysis

ICE

Incremental-effectiveness

ICER

Incremental cost-effectiveness ratio

CEAC

Cost-effectiveness acceptability curve

NRDL

National Reimbursement Drug List

COVID-19

Corona virus disease 2019

GDP

Gross domestic product

Author contributions

HSF were involved in the design of the study. TBY and LRM was involved in the design of the economic model and collected the data. TBY provided the economic analysis. All authors contributed to the development of the manuscript.

Funding

None.

Data availability

No datasets were generated or analyzed during the current study. This is a secondary study conducted using available evidence and all data and evidence obtained are included in article.

Declarations

Ethics approval and consent to participate

Not applicable. This article is based on mathematical modeling with inputs informed primarily by previously conducted studies and does not contain any new studies with human participants or animals performed by any of the authors.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.