Pharmacoeconomic analysis for pemetrexed as a maintenance therapy for NSCLC patients with patient assistance program in China

Abstract

Objective: This study is to evaluate the costs, clinical efficacy, and social benefits of a patient assistance program (PAP) implemented by the China Primary Healthcare Foundation for the use of pemetrexed as a first-line non-squamous non-small cell lung cancer (NSCLC) maintenance therapy in China.

Methods: A survival analysis was conducted on the clinical data of 1,366 patients who participated in the PAP. The progression-free survival (PFS) and median maintenance treatment cycle of pemetrexed were analyzed. A 36-month Markov model from a payer’s perspective was constructed to analyze the cost and effectiveness associated with the PAP for pemetrexed. The inputs of the model were sourced from the PAP clinical database and published literature. The study estimated the incremental quality adjusted life-years (QALYs) (pemetrexed plus best supportive care [BSC] vs BSC only), the cost saving of the PAP, the impact on the percentage of catastrophic health expenditures (CHE), and poverty headcount ratio (HCR).

Results: The median of PFS and maintenance treatment cycles were 187 days and five cycles (total nine cycles, which included four cycles of induction therapy), respectively. The pemetrexed plus BSC treatment with PAP resulted in an additional 0.12 QALYs over BSC only. The total cost was $48,034.46 and $96,191.57 for the patients who had or had not joined the PAP in 3 years, respectively. Compared to the patients without PAP, the percentage of CHE and HCR with PAP was reduced from 98.39% to 19.91% and 66.98% to 4.89%, respectively, indicating that the PAP substantially decreased the number of patients who had CHE and fallen into poverty.

Conclusion: The study concluded that the pemetrexed PAP generated noticeable clinical and economic benefits to society and to patients. The program also increased patients’ compliance with chemotherapy by allowing patients, for whom the pemetrexed treatment was unaffordable, to continue to receive it.

Keywords:

• Non-small cell lung cancer
• pemetrexed
• maintenance therapy
• patient assistance program
• cost-effectiveness

Introduction

Lung cancer is the leading cause of cancer death among men in both developed and developing countries¹. In China, lung cancer is the most commonly diagnosed and the deadliest form of cancer, with 7.33 million new cases and 6.10 million deaths in 2015². Non-small-cell lung cancer (NSCLC) accounts for more than 85% of all lung cancer cases; and ∼40% of patients with NSCLC are at an advanced stage when they are diagnosed with the disease^3,4.

There is superior efficacy and reduced toxicity for cisplatin/pemetrexed in NSCLC patients with non-squamous histology; in comparison to cisplatin/gemcitabine⁵. Maintenance therapy has become a strategy that can lead to an improvement in progression-free survival (PFS) and overall-survival (OS) in select patients with advanced NSCLC with non-progressive disease after four-to-six cycles of induction chemotherapy⁶. Pemetrexed was recommended as category 1 (according to NCCN Clinical Practice Guidelines in Oncology: Non-Small Cell Lung Cancer Version 5. 2017) for patients with advanced non-squamous NSCLC⁵. The additional therapy during this period has been shown to have improvements in PFS and OS; but with significantly increased financial cost. This could potentially adversely impact the accessibility to this advanced treatment, especially for the low-income patients^7,8.

The China Primary Healthcare Foundation is currently implementing a Patient Assistance Program (PAP); which allows qualified first-line non-squamous NSCLC patients who had pemetrexed as a maintenance therapy to receive a 100% discount after a self-funded 4-cycle induction pemetrexed therapy. Over 1,700 patients in China had participated in the PAP; the evaluation of the clinical and economic benefits of the program in China is still proceeding. This study aims to evaluate the costs, clinical efficacy, and social benefits of the PAP implemented by the China Primary Healthcare Foundation for the use of pemetrexed in first-line non-squamous NSCLC maintenance therapy in China.

Materials and methods

Survival analysis

A survival analysis was conducted on the clinical data (including the demographic characteristics, dosage of pemetrexed, and some medical information) of 1,366 patients who participated in the PAP and were treated with pemetrexed from January 1, 2015 to June 30, 2016. The PFS and median maintenance treatment cycle of pemetrexed were analyzed. The follow-up cut-off date was October 26, 2016.

In analyzing the PFS, the patients who were still in the PAP were regarded as censored, for their progress or death time was uncertain. For the patients who had already quit from the PAP (principally due to the disease progress or death), the PFS was defined as the time from the date when they joined the PAP to the last time receiving medicine date plus 21 days (the length of a treatment cycle: assuming the progress or death occurs when the patient just runs out of the received medications).

Cost-effectiveness analysis

A cost-effectiveness analysis was conducted with the aid of a three-state Markov model in Microsoft Excel from a payer’s perspective to analyze the incremental cost per QALY (quality adjusted life year) of the maintenance treatment of pemetrexed plus best standard care (BSC) vs BSC only, as well as the monetary value of the PAP program. The Markov model captured the costs and outcomes over 3 years. Also, 3.5% annual discounting rates were used for cost and effectiveness.

Model structure

The Markov model structure is shown in Figure 1. Both the patients from the pemetrexed plus BSC and BSC only groups comprised of three mutually exclusive health states: PFS state, progressed survival (PS) state, and death. The cycle length of the Markov model was 3 weeks, corresponding to the pemetrexed treatment cycles. A 2-parametric Weibull survival analysis by using STATA/SE version 12.0 (Stata Corp LLC; TX College Station, TX) was fitted to build time dependency into the Markov transition probabilities. Model inputs included the efficacy data, resource use, and cost data, which were derived from the survival analysis, clinical trials, and published literature.

Figure 1. Markov model of non-squamous non-small-cell lung cancer.

Transition probability

The transition probabilities used in the Markov model were estimated based on the survival analysis of the PAP’s participants and a multi-center, randomized, double-blind, positive controlled phase III clinical trial (PARAMOUNT)⁹. In order to predict the transition beyond the PAP follow-up time, the survival data were fitted with either a Weibull distribution or an exponential distribution, depending on the shape parameter of the Weibull distribution (if the shape parameter of a Weibull distribution is 1, it is equivalent to an exponential distribution). After fitting the distributions, the transition probabilities were assumed to remain following the distributions in the cost-effectiveness analysis.

Efficacy data

The PFS of the NSCLC patients treated with pemetrexed plus BSC was abstracted from the survival analysis. The PFS of NSCLC patients treated with BSC only, the OS, and adverse events data of both groups were obtained from PARAMOUNT⁹. Health state utility values of the PFS and PS states, and utility decrements due to adverse events were derived from the published literature^10–12 (Table 1). The pemetrexed was administered until disease progression, and/or intolerable toxicity in the PAP group, so we assumed that the utility decrements associated with the adverse events were only applied in the PFS state.

Table 1. Clinical inputs.

Input	Base case (Median)	Range (95% CI)	Source
Patients
Patient BSA	1.67 m^2	1.40–2.03	Survival analysis
Effectiveness
PFS in pemetrexed group	187 days	162.61–207.39	Survival analysis
Hazard ratio of PFS pemetrexed vs BSC	0.60	0.50–0.73	PARAMOUNT
OS in pemetrexed group	417 days	384–480	PARAMOUNT
Hazard ratio of OS pemetrexed vs BSC	0.78	0.64–0.96	PARAMOUNT
Utility
Utility, progression free	0.67	0.26–0.87	Zeng et al.^13 (2013)
Utility, progressed	0.47	0.19–0.56	Zeng et al.^13 (2013)
Utility decrements
Anemia	0.22	0.18–0.26*	Beusterien et al.^12 (2010)
Neutropenia	0.09	0.07–0.11*	Nafees et al.^11 (2008)
Fatigue	0.12	0.10–0.14*	Lloyd et al.^10 (2006)
Adverse event rates
Rate of anemia (Pemetrexed)	6.40%	5.12–7.68%	PARAMOUNT
Rate of neutropenia (Pemetrexed)	5.80%	4.64–6.96%	PARAMOUNT
Rate of fatigue (Pemetrexed)	4.70%	3.76–5.64%	PARAMOUNT
Rate of anemia (BSC)	0.60%	0.48–0.72%	PARAMOUNT
Rate of neutropenia (BSC)	0.00%	0.00–0.00%	PARAMOUNT
Rate of fatigue (BSC)	1.10%	0.88–1.32%	PARAMOUNT

BSA, body surface area.

* Not 95% confidence interval, the range of utility decrements were fluctuated ±20% by base case value.

Resource use and cost data

Cost of adverse events treatment, BSC, and chemotherapy if progressed were derived from the literature^13,14. The average price of pemetrexed in China is 1,811.57 USD/500 mg (average retail price of the hospitals which joined the PAP); the average usage of pemetrexed for one patient per cycle is 1.67 m², 1 m² needs 500 mg pemetrexed for treatment. However, the standard dosage of pemetrexed is 500 mg in China; hence, every treatment cycle needs use two standard dosages of pemetrexed, and the average cost of pemetrexed per cycle is 3,623.14 USD. All costs are presented in USD (6.57 RMB = 1 USD) and correspond to the CPI in 2016 (Table 2).

Table 2. Medical costs and resource use.

Item	Unit cost (USD)	Source
Best supportive care	874.80 per cycle	Li et al.^14 (2015)
Pemetrexed	3,623.14 per cycle	Average retail price
Anemia	618.87 per cycle	Zeng et al.^13 (2013)
Neutropenia	537.16 per cycle	Zeng et al.^13 (2013)
Fatigue	134.32 per cycle	Zeng et al.^13 (2013)
Chemotherapy if processed	2,563.50 per cycle	Zeng et al.^13 (2013)

One-way sensitivity analysis

A one-way sensitivity analysis on the net benefit of pemetrexed plus BSC over BSC only was conducted in order to reflect substantial uncertainty of input parameters. Each key parameter was fitted with high/low values in our model. The net benefit was calculated using the following formula, in which the threshold was set to be 3-times the GDP (gross domestic product) per capita in China:

Social benefits evaluation

In this part, we compared the total maintenance therapy cost (including the pemetrexed cost and the BSC cost) of patients with PAP to those without PAP. The cost saving of the PAP in 3 years (for the patients who join the PAP group the cost of pemetrexed maintenance therapy was $0), the impact of PAP on the percentage of catastrophic health expenditures (CHE), and poverty headcount ratio (HCR) of low-income NSCLC patients were estimated. Catastrophic health expenditure was defined as an out-of-pocket payment for healthcare ≥40% of a household’s capacity to pay by the WHO¹⁵. HCR is the percentage of residents whose annual net income is below the poverty line. China’s poverty lines for rural residents and urban residents were 3,000 RMB (456.62 USD) and 3,894 RMB (592.69 USD) per annum, respectively¹⁶.

Results

Survival analysis

A total of 1,366 NSCLC patients who joined the PAP were included in the survival analysis. Among the patients, 57% were male and 43% were female, the median age was 60 years and 88% of them had stage IV disease (Table 3). The median PFS and maintenance treatment cycles were 187 days (95% CI = 169–204) and five cycles (total nine cycles, which included four cycles of induction therapy), respectively. As summarized in Figure 2, 68% of the patients in the pemetrexed group received ≥4 cycles of maintenance therapy.

Figure 2. Summary of maintenance cycles administered. The percentage of PAP patients per maintenance cycle.

Table 3. Characteristics of the patients in the PAP [n (%)].

Characteristics	Patients in PAP
Gender
Female	593 (43)
Male	773 (57)
Age (years)
<65	678 (49)
≥65	688 (51)
Stage
III	170 (12)
IV	1,196 (88)

Cost-effectiveness analysis

Table 4 presents the base-case results for a duration of 3 years. Overall, in the PAP group, NSCLC patients with pemetrexed plus BSC cost $1,898.54 more than BSC only, and resulted in 0.12 more QALYs, giving an ICER value of $16,667.34/QALY. The WHO’s recommended threshold for intervention cost-effectiveness is 3-times GDP per capita per QALY¹⁷. In this study, the ICER of pemetrexed plus BSC with PAP vs BSC only in non-squamous NSCLC patients is less than 3-times China’s per capita GDP ($8,216; National Bureau of Statistics of China, 2016¹⁸) and, thus, is considered cost-effective according to the WHO recommendation.

Table 4. Outcome of the PAP cost-effectiveness model.

Group	Total cost (per patient)	Total QALYs (per patient)	ICER
PAP group (Pemetrexed)	$48,034.46	0.73	—
BSC only	$46,135.92	0.61	—
Incremental	$1,898.54	0.12	16,667.34 USD/QALY

The cost and QALYs was calculated during 3-year maintenance therapy.

Social benefits evaluation

Compared to the patients treated with pemetrexed who did not join the PAP, the total cost of patients who joined the PAP was $48,034.46; resulting in an average $48,157.11 cost saving during the 3-year maintenance therapy. Due to the PAP, the percentage of CHE and HCR was reduced substantially from 98.39% to 19.91% and 66.98% to 4.89%, respectively (Table 5).

Table 5. Summary of PAP’s impact on the total cost, CHE, and HCR.

Item	NSCLC patients with PAP	NSCLC patients without PAP	Incremental
Total cost (in 3 years)	$48,034.46	$96,191.57	$48,157.11
Percentage of catastrophic health expenditures (CHE)	98.39%	19.91%	–78.48%
Percentage of poverty headcount ratio (HCR)	66.98%	4.89%	–62.09%

One-way sensitivity analysis

A tornado diagram was developed to illustrate the top five populated parameters with regard to the net benefit (Figure 3). The most sensitive, influential variable in the cost-effectiveness analysis was hazard ratio of mortality (Pemetrexed vs BSC). The second to the fifth most sensitive and influential variables are hazard ratio of transition probabilities from the PFS to the PS or death state (Pemetrexed vs BSC), the utility value of stable disease state cost discount rate, and the event rate of anemia (Pemetrexed).

Figure 3. Tornado diagram for univariate analysis showing the top five variables’ influence on the net benefit. BSC, best supportive care; PFS, progression-free survival; QALY, quality-adjusted life-years; X value = 909.06 (base case net benefit).

Discussion

One of the limitations of Zeng et al.’s¹³ cost-effectiveness study of pemetrexed is that the clinical data used was not sourced from the Chinese population. By conducting a survival analysis, we could estimate the PFS of the Chinese NSCLC patients based on the clinical data of patients who participated in the PAP. In China, illness-related poverty was a common situation; therefore, we also estimate the impact of the PAP on the CHR and HCR of NSCLC in China.

The median PFS of the patients in the PAP (treated with pemetrexed plus BSC) was 187 days (6.23 months), which was longer than the pemetrexed treatment group (4.4 months) in PARAMOUNT⁹ and similar to the PFS 6.79 months in Xu et al.’s¹⁹ study. The median maintenance treatment cycle was 5, which is also more than the PARAMOUNT²⁰. A cost-effectiveness study conducted by Zeng et al.¹³ showed that pemetrexed plus BSC is not a cost-effective strategy compared to the BSC only strategy when a threshold of 3-times China’s average GDP per capita is used. In this research, the ICER of patients in the PAP group (pemetrexed plus BSC) was lower than 3-times GDP per capita per QALY which indicated that pemetrexed plus BSC with PAP was considered as a cost-effective strategy compared to BSC only. The total cost-saving of the PAP group during the 3-year maintenance therapy was $48,157.11. During the 1-year maintenance therapy, the PAP prevented 78.48% of the patients' family from CHE and 62.09% of patients from falling below the poverty level.

A key barrier to providing care to indigent populations is the high cost of medications²¹. Patients assistance programs (PAPs) are designed to provide access to medications for those who are unable to pay for their medications²². Patients could have access to more medications after they were enrolled in the PAPs than when they were paying for their prescriptions themselves. Providing free prescription drugs helps to improve medication adherence and avoid the lives lost caused by being unable to pay for the medicines²².

There are some limitations in this study. The PFS is estimated by calculating the duration from patients joining the PAP until last receiving medicine. For the patients who were still in the PAP, we could insure the cancer is not progress based on their medical data. However, for the patients who already quit from the PAP, since it is unlikely that we could continue to follow-up, the last receiving medicine time plus a 21-day extra period was used as estimation for PFS, instead of the exact time the patients’ health status is progressed or dead. The PFS in this study is nearly 1.8 months longer than the PARAMOUNT study, indicating the improvement of patients’ PFS, probably due to the improved compliance. The cost and utilization of the BSC were derived from the published literature due to lack of available costs of BSC for Chinese patients with advanced non-squamous NSCLC. Utilities data generated in China were not yet available, so this is derived from the published literature with a western country utility value. Further study is needed to verify this utility data of the Chinese population.

Conclusion

Pemetrexed plus standard care with PAP resulted in noticeable clinical and economic benefits to society and patients. This PAP program also increased patients’ compliance with chemotherapy by allowing patients, for whom the pemetrexed treatment was unaffordable, to continue to receive it. Lastly, because of the PAP program, a considerable number of patients were able to avoid to fall into poverty level due to their cancer treatments.

Transparency

Declaration of funding

This research was funded by China Primary Healthcare Foundation.

Declaration of financial/other relationships

JC and YC are employees of China Primary Healthcare Foundation. YJL and ZYL received funding from China Primary Healthcare Foundation. Peer reviewers on this manuscript have received an honorarium from JME for their review work, but have no other relevant financial relationships to disclose.

Acknowledgement

The authors wish to acknowledge the medical writing support of Weiwei Feng and Maodong Ren.