Evaluation of a risk-sharing agreement for atezolizumab treatment in patients with non-small cell lung cancer: a strategy to improve access in low-income countries

Abstract

Background

Using immune checkpoint inhibitors (IO) is a promising approach to maximize clinical benefits for patients with non-small cell lung cancer (NSCLC). PD-L1 expression serves as a predictive factor for treatment outcomes with IO. However, the high cost of this treatment creates significant barriers to access. Substantial evidence demonstrates the sustained clinical benefits experienced by patients who respond to immunotherapy. While IOs show promise in NSCLC treatment, their high cost poses access barriers.

Aim

This study focused on a prospective cost analysis conducted at a high-specialty health facility to assess the economic implications of implementing a risk-sharing agreement (RSA) for atezolizumab in NSCLC.

Methods

The study included 30 patients with advanced NSCLC, with the pharmaceutical company funding the initial cycles. If patients responded, a government program covered costs until disease progression.

Results

A median progression-free survival of 4.67 months across populations, rising to 9.4 months for responders. The 2-year overall survival rate for the response group was 64%, significantly higher than for non-response. Without an RSA, a total treatment cost of $881 859.36 ($29 395.31/patient) was reported, compared to $530 467.12 ($17 682.24/patient) with an RSA, representing a 40% cost reduction. In responders, the average cost per year of life per patient dropped by 22%. Risk-sharing, assessed through non-parametric tests, showed a statistically significant difference in pharmacological costs (P < .001).

Conclusion

Implementing RSAs can optimize resource allocation, making IO treatment more accessible, especially in low-income countries.

Introduction

Risk-sharing agreements (RSA) involve a plan in which the performance of a product in a defined population is monitored over a specific time. The amount or level of reimbursement is based on health outcomes and costs achieved, reducing the associated uncertainty regarding the launch of a new product concerning its clinical and economic performance in the real world.¹

A systematic review conducted by Yu et al² identified 26 specific risk-sharing agreements in the United States. Thirteen of these were related to cardiometabolic indications, 5 to oncological indications, and 8 to other indications. This review reported that the implementation of these strategies generated greater access to medication for patients and the possibility of these schemes functioning as an adequate coverage or reimbursement mechanism.² Carlson et al³ also identified the use of these agreements in different countries. Various structures can be identified among the different types of agreements, such as performance-based payment, graduated repayment, or conditional discounts. These agreements may include clinical outcome measures, such as reducing adverse events or improving patient survival, and process measures, such as treatment compliance or fewer hospitalizations.³

In this scenario, performance-based RSA allows conditional reimbursement of the drug’s cost, which depends on the drug’s clinical practice results. This approach restricts its use to a subpopulation of patients through selection criteria or more effective eligibility criteria, thus managing the budgetary impact of treatment.

Holleman et al⁴ analyzed how risk-sharing policies affect access and outcomes for patients with non-small cell lung cancer (NSCLC) based on actual drug performance in clinical practice. The analysis retrospectively identified resource use, costs, and clinical outcomes associated with the selected medications without risk sharing. Study results demonstrate that implementing RSA can significantly impact outcomes for patients with NSCLC. Differences in overall survival and medical care costs were found considering the different scenarios evaluated.⁴

These types of agreements between pharmaceutical companies and payers have gained prominence in recent practice as tools that enable better budget utilization and risk reduction in spending on drugs that lack complete evidence of clinical benefit, especially for innovative medicines in the field of oncology.⁵ According to a study conducted through semi-structured interviews with hospital pharmacy, laboratory, and hospital oncology professionals, there is a predisposition to sign this type of agreement. There is a particular interest in the payment-by-results modality, as it contributes to increasing knowledge of new drugs and improving the health of patients.⁶ Consequently, these contracts will become more necessary due to the broader application of personalized medicine, value-based pricing, constant prices, and budget limitations.⁷

Lung cancer represents the leading cause of death associated with neoplasms worldwide. In 2022, Mexico reported 8257 new cases and 7808 deaths, making it the ninth most common neoplasm and the third leading cause of cancer-related deaths in both sexs.⁸ Among lung cancer subtypes, NSCLC accounts for 85% of diagnoses and is associated with high mortality.⁹

Atezolizumab is a monoclonal antibody directed against the PD-L1 ligand. It was approved in 2017 in the United States and Europe as a second-line treatment for NSCLC based on the results of the phase II POPLAR studies and the phase III OAK study.¹⁰ As a first-line treatment, it has demonstrated superior survival outcomes compared to platinum-based chemotherapy in patients with NSCLC with high expression of PD-L1, regardless of the histological subtype.¹¹ In Mexico, it has been available since 2018 and was included in the Compendium of National Health Supplies in 2019 as a second-line treatment for patients with NSCLC.¹² Based on real-life data, a previous study in our country showed the positive therapeutic effect of atezolizumab and other immunotherapy drugs used as second-line treatments.¹³ Some studies from different nations have reported that atezolizumab and other immune checkpoint inhibitors are cost-effective options.^14-16 However, the high cost of these novel drugs represents a barrier to accessing treatment. A recent study concluded that the evidence generated by such cost-effectiveness studies “serves as evidence for pharmaceutical enterprises to properly and deeply consider the pricing strategy based on effectiveness and safety in the real-world condition”¹⁷

This study aims to evaluate the profitability of an RSA for atezolizumab treatment in patients with NSCLC treated at a National Cancer Institute (NCI). This evaluation aims to determine the potential savings for the institution resulting from its implementation and to contribute to the body of evidence for evaluating future agreements for access to cutting-edge therapies.

Materials and methods

Study design and recruitment

The institutional ethics and research committees approved a prospective cost analysis of atezolizumab as a second- or third-line monotherapy for advanced-stage NSCLC, conducted at a third-level public NCI in Mexico City from March 2019 to December 2022.

Patients aged 18 years or older, recently admitted to the Instituto Nacional de Cancerología (INCan) with a confirmed diagnosis of NSCLC and a PD-L1 level of ≥1%, who demonstrated the ability to comprehend the study, signed consent form and expressed willingness to participate, were eligible for inclusion.

Data was collected on various clinical-demographic variables, including sex, age, wood smoke exposure, presence of comorbidities, ECOG performance status, clinical stage, histology, adenocarcinoma subtype, mutations, PD-L1 expression level, smoking index, death, and determination of brain metastasis at the time of diagnosis. Data collection involved reviewing clinical records.

Definitions

For purposes of this paper, after a consensus among participating clinicians, the clinical response was defined, according to the Response Evaluation Criteria in Solid Tumours (RECIST, version 1.1),¹⁸ by either of the following: a complete response (CR), a partial response (PR) or stable disease (SD). Disease progression (DP) were not deemed a response due to the absence of discernible therapeutic benefits in patients.

A CR was defined as the complete disappearance of all target lesions or reduction of target lesion diameter in lymph nodes to less than 10 mm. At the same time, a PR was characterized by a decrease in the sum of target lesion diameters exceeding 30%. SD was defined as a tumor that has stayed the same size; it hasn’t gotten better or worse. DP was defined as cancer growing by at least a fifth (20%) or there are new areas of cancer.¹⁸

Overall response rate^19,20 (ORR) was defined as the proportion of patients who have a PR, CR, or SD to therapy, is taken as a direct measure of the atezolizumab tumoricidal activity. Patient outcomes were elucidated through the definition of overall survival²¹ (OS), measured as the time between atezolizumab treatment initiation to the last follow-up or death from any cause. Similarly, progression-free survival (PFS) was measured from the time between atezolizumab treatment initiation and either objective tumor progression or death.²⁰

RSA scheme

The execution of the RSA scheme solely unfolded with the pharmaceutical company sponsoring the initial treatment of atezolizumab (up to 4 cycles). This phase was conducted with patients receiving Atezolizumab as an intravenous 1200 mg fixed dose every three weeks (21-day cycle). This strategic arrangement was designed to provide an avenue for patients to access PD-L1 therapies without incurring immediate financial burdens. After the fourth cycle, patients were stratified into two groups based on their clinical response to treatment. Response (RP) group for “responders” (patients showcased discernible therapeutic benefits, either a CR, PR, or SD) and non-response (non-RP) group for “non-responders” (patients with PD). The response gauged at the fourth cycle was identified as a predictive factor for the ORR, signifying the treatment’s effectiveness. After the fourth cycle’s evaluation phase and after confirming therapeutic efficacy, the baton of covering treatment costs seamlessly passed on to the INCan program. This dynamic transition ensured that patients, who had exhibited positive responses to the treatment, continued receiveing the necessary support and medical interventions without shouldering the financial implications. Responders continued with atezolizumab treatment until discontinuation. Simultaneously, treatment was discontinued for non-responders.

Costs

The analysis orbit is delimited from the perspective of the public health sector in Mexico, with particular emphasis on the institutional vision of the INCan. Within the framework of this analysis, it focused exclusively on the direct medical costs linked to the pharmacological treatment of the intervention under evaluation in RSA, encompassing the costs associated with the use of pharmacological therapy by patients throughout the follow-up period until progression. The cost of atezolizumab was solely based on the payer’s (INCan) drug cost. Yet, additional costs of Healthcare Resource Usage (HRU) as hospitalization services, laboratory and imaging studies, among others were taken into consideration but not included into analyses to maintain a clear focus on the drug´s direct costs and clinical outcomes of the drug itself.

The bottom-up costing methodology was chosen to approach the cost of pharmacological treatment.²² Following a rigorous analytical framework, all costs were denominated in US dollars (USD) and adjusted to the average annual USD exchange rate in 2022. This accuracy in monetary considerations provided a standardized basis for cross-comparisons and facilitated a comprehensive understanding of the economic impact associated with prescribed pharmacological interventions.

The calculation for the drug´s acquisition derives from a payer´s price set at USD 3378.77 for atezolizumab administered for one cycle. Based on the information obtained during data collection, the dose and number of cycles required for each patient were considered for cost calculation. This was used to estimate a full-payment scenario (non-RSA scheme), which served as a baseline for comparing the calculated costs of patients with those under the RSA scheme (Table S1). This approach ensured a thorough assessment of all relevant costs related to pharmacological treatment, providing detailed information about the financial implications from the payer’s perspective.

The results of the cost analysis are outlined, emphasizing the difference of the obtained average cost per patient with RSA against the previouslly estimated avarage cost of the non-RSA scheme, along with the average cost-effectiveness ratio (ACER) per year of life earned for all patients with and without the RSA. ACER was obtained by calculating the average cost per patient for effectiveness (expressed as life years obtained by the restricted mean), adapted from previous literature.^23,24 This comprehensive approach provides a holistic view of the economic aspects associated with the pharmacological interventions evaluated.

Statistical analysis

Descriptive statistics were used to analyze the clinical-demographic variables, employing medians and interquartile ranges for quantitative variables based on their distributions. For qualitative variables, a comprehensive examination involved the use of frequencies and percentages. The OS and PFS endpoints elucidated the temporal aspect of patient outcomes. In particular, another set of adjusted OS and PFS was established starting from the fourth cycle. This adjustment is justified by the use of the landmark analysis, which allows for a more accurate comparison between the RP and non-RP group.^25-29

A predefined follow-up duration of 36 months was applied, ensuring survival times exceeding this temporal threshold were duly censored. After this, patients were stratified into RP and non-RP groups. The restricted mean survival time (RMST) was calculated to provide a more nuanced perspective on survival outcomes. Calculating RMST involved measuring the area under the survival curve. RMST serves as a valuable metric, estimating the average survival time or life expectancy within a specified period and offering a clinically significant measure to interpret the contrast in survival between different response groups.³⁰

The analytical framework for this study was executed using R version 4.1.2 through the RStudio interface.³¹ The results were reported with 95% confidence intervals, and statistical significance was determined at a threshold of P ≤ .05. This stringent criterion underscored the robustness and reliability of the findings, ensuring that any observed differences between groups were deemed statistically significant.

Results

Baseline characteristics are reported in Table 1. Thirty patients were included, with 11 (37%) in the RP group and 19 (63%) in the non-RP group. Of these, 23 (77%) and 7 (23%) received second and third-line treatment, respectively. No patient received chemotherapy in combination with atezolizumab, the drug was given as monotherapy according with the National Comprehensive Cáncer Network (NCCN, version 5.2024) international guideline. Among the patients, 13 (43.3%) were male, with a median age of 62 ([7.3-68.5) years. Smoke exposure was reported for 20 (66.7%) patients, and 15 (50%) reported comorbidities. Of all, 28 (93.3%) had an ECOG performance status ≤1 and were in stage IV. Adenocarcinoma (86.7%) was the predominant histology, with a solid or micropapillary subtype in 15 (50%) cases. Mutations were absent in 24 (80%) patients, and 4 (13.3%) presented with an EGFR mutation. The median PD-L1 expression was 10 (2.0-–38.0). The smoking index had a median of 2.9 (0-23.3). The mortality rate was 23 (77%), and brain metastasis was present at the time of diagnosis in 8 (27%) of cases. No significant differences were observed between the clinical response groups in the above variables. However, in the fourth cycle, overall mortality in the RP group was significantly lower (36% vs 100%, P < .001) compared to the non-RP group.

RMST and life years

The PFS for the entire sample was 4.67 (95% CI, 3.4-7.8) months, and the OS was 7.59 (95% CI, 5.1-17.2) months (Figure 1A and B). After the fourth cycle, patients in the RP group achieved a median PFS of 6.6 (95% CI, 4.4-NA) months, in contrast to a median of 1.17 (95% CI, 0.5-NA) months in the non-RP group (n = 10). This difference in PFS was statistically significant (P < .0001), Figure 1C. Additionally, 9 (30%) patients were excluded from this analysis due to DP or death before the fourth cycle.

Moreover, according to landmark analysis after the fourth cycle, the RP group did not reach the median OS (11.0; 95% CI, 16.24—NA). In contrast, the non-RP group had a median OS of 3.7 (95% CI, 2.1-12.7) months. This difference was also statistically significant (P < .0001), with 5 (17%) of patients not considered due to death before the fourth cycle (Figure 1D). Based on the OS curves, we calculated the average years lived for each group (RP vs non-RP) starting from the fourth cycle. The RP group had an estimated RMST of 2.17 years, while the non-RP group had 0.36 years, resulting in a difference of 1.81 years (Table 2)

Cost results

The total cost of care for atezolizumab in all patients without implementing the RSA, was $881 859.36. In contrast, when implementing the RSA scheme, it amounted to $530 467.12, representing a 39.85% reduction. The mean cost per patient without RSA was $29 395.31, which decreased to $17 682.24 with the RSA scheme’s implementation. When considering the years of life calculated by RMST in each group starting from the fourth cycle and the mean cost per patient, the RP group had an ACER of $28 406.27 per year of life without the RSA scheme. With the scheme’s implementation, this ratio decreased to $22 187.98, resulting in a 22.0% cost saving. For the non-RP, the ACER without RSA was $29 383.96, which dropped to no expenses with the scheme’s implementation, representing a 100% cost reduction. Table 2

Discussion

To the best of our knowledge, the present study is the first to report the implementation of an RSA scheme in our country, providing insights into this type of agreement between the pharmaceutical industry and health services.

Implementing an SRA scheme has a substantial impact, with a 40.0% reduction in the total cost of treatment. For patients who respond to atezolizumab therapy, the cost is reduced by 22.0%, and for those who don’t respond, 100% coverage is available. The findings significantly impact patient care and resource utilization within the NCI.

On the other hand, identifying patients who respond to treatment after the fourth cycle allows the selection of individuals who can derive more significant benefits from atezolizumab. This approach places each patient’s outcome measure at the forefront of healthcare decision-making while optimizing the utilization of hospital resources across multiple stakeholders, ultimately delivering value-based healthcare (VBHC).³²

Understanding predictive response factors to immunotherapy, combined with our study’s results, could lead to better candidate selection in limited accessibility. This data would promote resource optimization and expand coverage to patients with the most significant potential for treatment benefit. Previous research has linked immunotherapy effectiveness to clinical, histopathological, genetic traits, and biomarker expression. In Hispanic populations, poorly differentiated solid tumors often show higher PD-L1 expression.³³ PD-L1 is a key biomarker in predicting NSCLC patients’ response to immunotherapy. In Latin American populations, PD-L1 expression independently predicts survival in immunotherapy-treated patients.^34,35 However, PD-L1’s utility is limited by intratumoral heterogeneity. Mutations in STK11 and KEAP1 are associated with reduced response to immunotherapy, particularly in Latin American cohorts, leading to shorter progression-free survival.³⁶

The results of our analysis align with those reported in the existing literature. Holleman and colleagues⁴ study also reported favorable outcomes for implementing risk-sharing schemes. Linking drug prices to health outcomes led to a reduction in payer drug costs ranging from 2.5% to 26.7%. At the same time, there has been a disconnection between the theoretical rationale for choosing Outcomes-based risk-sharing agreements (OBRSA) in cancer treatments and the practical reasons inferred from their implementation. The administrative burden emerged as a significant barrier, impacting payers, manufacturers, and healthcare providers alike. This burden is especially pronounced in lower- and middle-income countries (LMICs), where resources are often limited.³⁷ High-income countries (HICs) have robust frameworks for managing pharmaceuticals, featuring universal health coverage, external and internal price referencing, indirect price-cost controls, and health technology assessments (HTAs) to balance costs with clinical benefits. Conversely, low- and middle-income countries (LMICs) struggle with inadequate coverage and poor access to essential medicines, relying on restrictive state controls that exacerbate health inequalities. The disparity between HICs and LMICs underscores significant global health equity challenges. High-income nations manage technological advancements and demographic changes while developing countries strive for basic access to medicines.³⁸ While all stakeholders show interest in OBRSAs, health plans generally regard them as a lower priority compared to manufacturers. Overcoming operational barriers, alongside policy and regulatory challenges, is crucial for aligning efforts to advance OBRSAs. Achieving this alignment requires collaboration to improve decisions about when and how to pursue OBRSAs, focusing on data management, modeling and piloting OBRSAs, and information sharing. These findings are relevant to companies operating in the HICs as the United States and likely extend to certain value-based arrangements in other countries.³⁹ However, the impact on LMICs can be significant, as these regions may face additional challenges in implementing such agreements due to limited resources and infrastructure. On this behalf, successful implementation depends on effective stakeholder engagement and the establishment of mutual trust among key groups.³⁷ Addressing these challenges requires increased investment, international cooperation, innovative policies, and pharmaceutical companies’ commitment to low-income markets, aiming for universal access to essential medicines.³⁸

Increasing the number of risk-sharing schemes within the institution, such as the one described in this study, would provide more precise estimates of benefits for both stakeholders (the NCI and the pharmaceutical company). These results could encourage broader adoption of innovative treatments that prove effective within the patient population served by the institution. Such a move would align with Mexican legislation for these institutions⁴⁰ and help alleviate inequities in access to innovative treatments within the Mexican health system, as previously highlighted.⁴¹

Our analysis reflects similar economic benefits, consistent with findings elsewhere. Nonetheless, one notable methodological difference is that our study used various treatment schemes, further validating the positive outcomes of implementing shared risk by comparing therapeutic alternatives. Additionally, this approach represents an exciting treatment option that should be considered by all stakeholders involved in managing this condition within the institution.

This study, while comprehensive, acknowledges certain limitations that could impact the robustness of our findings. One limitation pertains to the relatively modest sample size utilized in the analysis. The segmentation of data necessitated this size into distinct response groups and the consideration of the entire sample population. Due to this segmentation strategy, the sample size may have been smaller than desired, potentially introducing variability in our outcome measure estimates. The reimbursement of cancer drugs by publicly funded drug programs varies worldwide, and the reasons behind it vary globally, particularly in LMICs. Hence, the data gathered in this study may be different from other populations. Despite this constraint, the insights garnered from this study remain valuable and contribute to this field’s existing body of knowledge.

Conclusions

The implementation of RSA in pharmaceutical treatments, especially in the case of innovative and high-cost therapies such as atezolizumab for NSCLC, represents significant progress in addressing the challenges inherent in clinical uncertainty and economic outcomes. These agreements are an effective tool to mitigate the uncertainties associated with treatment effectiveness in real-world situations and to efficiently manage the budgetary impact of introducing new drugs. As a result, it is seen as a promising solution to improve patient access while effectively managing the financial risks faced by healthcare institutions.

Adopting RSA schemes facilitates identifying patients who are more likely to obtain significant benefits, thus ensuring that government resources are directed towards those with more substantial survival potential. This approach not only optimizes the allocation of resources but also generates significant savings for the State. In our study, implementing a risk-sharing plan resulted in a 40% reduction in the total cost of treatment, thus evidencing the effectiveness of this cost-effective strategy. These findings suggest that this approach could have a positive impact on improving patient care, especially in contexts in countries with limited resources.

Despite the limitations inherent in the study, such as the relatively small sample size, the results provide valuable information on the potential of risk-sharing arrangements to optimize patient care and efficient resource management. In this regard, this study highlights the importance of continuing the exploration of innovative agreements between pharmaceutical companies and healthcare institutions to address the complexities associated with costs, access, and effectiveness in contemporary healthcare. These schemes emerge as tools that facilitate more informed and agile decision-making, thus contributing to improving efficiency and effectiveness in the management of cutting-edge treatments.

Supplementary material

Supplementary material is available at The Oncologist online.

Acknowledgments

All authors would like to recognize the work of the team of the Thoracic Oncology Unit at our institution in supporting and caring for patients. We acknowledge the pharmaceutical company that donated atezolizumab during the first four cycles of the patient’s treatment.

Author contributions

Conceptualization and funding acquisition: Oscar Arrieta. Data curation and formal analysis: Herman Soto-Molina, Ana Pamela Gómez-García, and Silvia Guzmán-Vázquez. Investigation, Maritza Ramos-Ramírez, Luis A. Cabrera-Miranda, Ana Pamela Gómez-García, Ángel Valencia-Velarde, and Marco Gálvez-Niño. Methodology: Oscar Arrieta, Homero Garcés-Flores, Herman Soto-Molina, Ana Pamela Gómez-García, and Andrés F. Cardona. Project administration: Oscar Arrieta, and Maritza Ramos-Ramírez. Resources: Oscar Arrieta, Homero Garcés-Flores, Maritza Ramos-Ramírez, Luis A. Cabrera-Miranda, Ana Pamela Gómez-García, Ángel Valencia-Velarde, Marco Gálvez-Niño , Andrés F. Cardona. Software, Silvia Guzmán-Vázquez. Supervision: Oscar Arrieta. Validation: Homero Garcés-Flores, Herman Soto-Molina, and Silvia Guzmán-Vázquez. Visualization: Homero Garcés-Flores, Herman Soto-Molina, Ana Pamela Gómez-García, Andrés F. Cardona, and Marco Gálvez-Niño. Writing—original draft: Maritza Ramos-Ramírez, Luis A. Cabrera-Miranda, and Ángel Valencia-Velarde. Writing—review and editing: Luis A. Cabrera-Miranda, Ana Pamela Gómez-García, Ángel Valencia-Velarde, Marco Gálvez-Niño, and Andrés F. Cardona. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Conflicts of interest

The authors declare no financial conflicts of interest.

Institutional review board statement

The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Research Board of Instituto Nacional de Cancerologia (INCan) under protocol number 2023/012, approved on February 21, 2023.

Informed consent statement

All subjects who participated in the study provided informed consent.

Data availability

Due to ethical and privacy restrictions, the data are not publicly available but will be made available at the corresponding author’s request.

References