FormalPara Key Points for Decision Makers

Patients with advanced epidermal growth factor mutated (EGFRm) non-small cell lung cancer (NSCLC) initiating first line (1L) osimertinib or chemotherapy incurred total costs of care per patient per month of US$27,174 and US$23,343, respectively.

Descriptive differences in the type of spending and health resource use were observed between 1L osimertinib and 1L chemotherapy populations, with higher inpatient costs and inpatient days observed for osimertinib therapy and higher outpatient costs observed for chemotherapy.

Unmet needs may remain for therapeutic options that balance benefits, risks, and total costs of care among patients with advanced EGFRm NSCLC.

1 Introduction

Lung cancer (LC) is the second most common type of cancer and the leading cause of cancer death for men and women, accounting for 25% of all cancer deaths in the United States (US) [1]. Over 236,000 new cases are expected to be diagnosed in 2022 in the US [1]. While screening has improved, nearly 70% of cases are diagnosed late in the disease course, at the stage of locally advanced or metastatic disease [2]. Non-small cell lung cancer (NSCLC) is the most common type of LC, representing up to 85% of cases [2]. Despite progress in therapeutic options, survival remains low, with 5-year survival rates of Stage III and IV NSCLC at 15% and 10%, respectively [3, 4]. Depending on the stage of the cancer and other factors, treatment options for NSCLC include surgery, radiofrequency ablation, radiation therapy, chemotherapy, targeted therapy, and immunotherapy.

Approximately 10–15% of patients with NSCLC present with epidermal growth factor receptor (EGFR) mutations (EGFRm), with a greater frequency in women and nonsmokers [3]. For these patients, targeted therapy with EGFR tyrosine kinase inhibitors (TKIs) is recommended as first-line (1L) treatment [5]. At the time of the present analysis, five EGFR TKIs were approved by the US Food and Drug Administration (FDA) for the treatment of NSCLC: erlotinib, afatinib, gefitinib, osimertinib, and dacomitinib. As a class, TKIs have been found to result in better progression-free survival (PFS) than chemotherapy for EGFRm NSCLC [6]. Osimertinib specifically has been found to result in significantly prolonged PFS and improved response relative to platinum-based chemotherapy [7].

In addition to differences in efficacy, osimertinib-based and chemotherapy-based regimens have notable differences in their administration and tolerability profiles. Osimertinib is more commonly associated with side effects such as diarrhea and has warnings for serious adverse events such as interstitial lung disease/pneumonitis, QT-prolongation, keratitis, erythema multiforme, Stevens-Johnson syndrome, cutaneous vasculitis, and cardiomyopathy [8], while more common adverse events for platinum doublet chemotherapy include neutropenia, leukopenia, and anemia [9, 10]. Recent trial data reported that serious (i.e., grade 3 or greater) adverse events were reported by 48% of patients receiving chemotherapy and 37% of patients receiving osimertinib alone [11]. Additionally, chemotherapy administration is usually via infusion for a fixed number of cycles, whereas osimertinib is an oral medication taken until treatment progression. The differences between osimertinib and chemotherapy treatment may have an impact on patient care, healthcare resource utilization (HRU), and cost.

Prior research has found that NSCLC is associated with notable economic burden, with greater direct and indirect costs per patient associated with advanced NSCLC compared with less-advanced disease [12]. While favorable clinical trial data has led to the rapid emergence of osimertinib as the 1L standard of care, the full impact of this evolved treatment landscape on quality of care, patient quality of life, and economic burden, remains to be fully elucidated. To contribute to the elucidation of the economic impact, the present study sought to describe the HRU and cost associated with 1L osimertinib or 1L chemotherapy for treatment of advanced EGFRm NSCLC, not only by quantifying total cost of care, but also by describing the types of healthcare spending associated with both therapy options. Using TKI administration as a proxy for EGFRm status, this descriptive study reports HRU and healthcare costs among patients in the United States with advanced EGFRm NSCLC initiating 1L therapy, both overall and separately among osimertinib and chemotherapy recipients.

2 Methods

2.1 Data Source

IBM MarketScan Research Databases (Commercial and Medicare Supplemental Databases [January 1, 2017 through April 30, 2020] and Multi-State Medicaid Database [January 1, 2017 through December 31, 2018]) were used. Data were de-identified and complied with the patient requirements of the Health Insurance Portability and Accountability Act (HIPAA). Therefore, no Institutional Review Board exemption was needed.

The Commercial and Medicare Supplemental databases contain information on medical and prescription drug experience of beneficiaries. The Commercial database consists of employer- and health plan-sourced data of beneficiaries, comprising employees, their spouses, and dependents who are covered by employer-sponsored private health insurance. The Medicare Supplemental database profiles retirees with Medicare Supplemental insurance paid by employers. Both the Commercial and Medicare Supplemental databases cover all US census regions, with a higher concentration in the South and North Central (Midwest) regions. The databases contain actual payer costs (paid amount and coordination of benefits) from both the Commercial Claims and Encounters database and the Medicare Supplemental and Coordination of Benefits database. The Multi-State Medicaid database contains the pooled health care experience of approximately 7 million Medicaid enrollees from multiple states. The data includes medical and prescription drug claims, standard demographic variables, and monthly information on health plan enrollment.

2.2 Study Design

A retrospective study design was used to assess patient characteristics, HRU, and cost of care associated with 1L EGFRm NSCLC. Given that there are no specific International Classification of Diseases, Ninth/Tenth Revision, Clinical Modification (ICD-9 CM/ICD-10 CM) diagnosis codes for EGFRm or advanced NSCLC, a claims-based algorithm was used to identify these patients. First, patients with an LC diagnosis were selected. Then, those receiving treatment typically used for small cell lung cancer (SCLC; i.e., etoposide, irinotecan, or topotecan) at any time were excluded. The remaining patients were considered as having NSCLC and were required to have a washout period of ≥ 12 months of continuous insurance eligibility prior to the first observed LC diagnosis (to confirm the first LC diagnosis). To be considered as having advanced NSCLC, patients had to initiate 1L antineoplastic therapy or have a diagnosis for metastatic disease within 30 days following the first LC diagnosis. The first claim for any type of antineoplastic agent was identified as the initiation of 1L therapy and was defined as the index date. All agents received in a window of 21 days following the initiation of 1L therapy were included in the 1L therapy regimen. The end of 1L therapy was defined as the day prior to the initiation of a new antineoplastic agent that was not part of the 1L regimen or at the resumption of the same treatment regimen after a gap of more than 90 days (re-treatment). To proxy EGFRm advanced NSCLC, patients were required to initiate a TKI (i.e., erlotinib, afatinib, gefitinib, osimertinib, or dacomitinib) in any line of therapy. Finally, to capture the current treatment landscape, TKI initiation was required to be in 2018 or later, a period during which osimertinib was approved as a 1L treatment for EGFRm NSCLC.

2.3 Study Population

As shown in Fig. 1, patients were included in the study if they met the following inclusion criteria: two or more diagnoses for LC, initiation of 1L therapy following the first LC diagnosis, one or more diagnosis for metastatic cancer within 30 days following the first LC diagnosis or initiation of 1L therapy within 30 days following the first LC diagnosis, ≥ 18 years of age as of the index date, ≥ 12 months of continuous insurance eligibility prior to the first LC diagnosis, and one or more claim for a TKI in 2018 or later and on or following the first LC diagnosis. Patients were excluded if they met one of the following criteria: one or more claim at any time for any antineoplastic medication primarily used to treat SCLC (i.e., etoposide, irinotecan, or topotecan) and one or more diagnosis for any other cancer prior to the first LC diagnosis.

Fig. 1
figure 1

Study population selection. aMetastatic cancer was identified based on ICD-9-CM codes 196, 197, 198, 209.7 or ICD-10-CM codes C77, C78, C79, C7B. bTKI-based therapy included erlotinib, afatinib, gefitinib, osimertinib, and dacomitinib. cCancers other than LC were identified based on ICD-9-CM codes 140-161, 162.0, 163-209 and ICD-10-CM codes C00-C33, C35-C96. dIncluded 43 patients treated with platinum-based chemotherapy and 2 patients treated with non-platinum-based chemotherapy. eIncluded 15 patients treated with afatinib, 8 patients treated with erlotinib, 2 patients treated with gefitinib, 1 patient treated with pembrolizumab, and 1 patient treated with nivolumab. 1L first line, EGFR epidermal growth factor mutation, ICD-9-CM/ICD-10-CM International Classification of Disease, Ninth/Tenth Revision, Clinical Modification, LC lung cancer, NSCLC non-small cell lung cancer, SCLC small-cell lung cancer, TKI tyrosine kinase inhibitor

Full size image

Two subgroups of patients were subsequently identified from this population: patients initiating osimertinib in 1L and patients initiating any chemotherapy-based regimen (including platinum-based chemotherapy) in 1L. HRU and costs during 1L therapy were assessed for the overall sample and for these two subgroups.

2.4 Study Measures

Patient demographics and clinical characteristics were evaluated during the 12-month period before 1L initiation, referred to as the baseline period. All-cause HRU, laboratory service utilization, supportive care utilization, and healthcare costs were reported during 1L therapy. HRU included outpatient, inpatient, emergency room (ER), and other services (i.e., durable medical equipment and dental and vision care). The number of days with laboratory service use was also reported. Supportive care was defined as the number of days with claims for erythropoiesis-stimulating agents (ESAs), granulocyte or granulocyte-macrophage colony-stimulating factor (G-CSF/GM-CSFs), pain management agents, or use of respiratory support care. All-cause healthcare costs, including total costs (sum of medical and pharmacy costs), medical costs (sum of outpatient, inpatient, ER, and other costs), and pharmacy costs, as well as laboratory service-related and supportive care-related costs were reported.

2.5 Statistical Analysis

Baseline characteristics and study outcomes during 1L therapy were described for the overall sample, the osimertinib cohort, and the chemotherapy cohort using means, standard deviations (SDs), and medians for continuous variables and frequencies and proportions for categorical variables. To account for different lengths of observation periods among study patients, HRU and healthcare costs were reported per-patient-per-month (PPPM). Healthcare costs were assessed from the payers’ perspective and adjusted for inflation (to 2020 USD) using the medical care component of the Consumer Price Index [13].

3 Results

3.1 Baseline Demographics and Clinical Characteristics

A total of 213 patients with advanced EGFRm NSCLC met the study criteria (Fig. 1). Baseline demographics, clinical characteristics, and HRU/costs are shown in Table 1. Among the study population, 141 patients (66.2%) initiated osimertinib as 1L therapy, while 45 patients (21.1%) initiated chemotherapy (platinum-based chemotherapy: 43 patients; non-platinum-based chemotherapy: 2 patients) and 27 patients (12.7%) initiated another regimen (afatinib: 15 patients; erlotinib: 8 patients; gefitinib: 2 patients; pembrolizumab: 1 patient; nivolumab: 1 patient). The average duration of 1L therapy for the overall sample, the osimertinib subgroup, and the chemotherapy subgroup were 8.6 months, 8.8 months, and 7.6 months, respectively.

Table 1 Baseline demographics and clinical characteristics
Full size table

At 1L initiation, the mean age was 60.9 years, 69.0% were female, and the majority (70.9%) were covered by commercial insurance, followed by Medicare Supplemental (25.4%) and Medicaid (3.8%). Demographic characteristics were similar for the subgroup initiating 1L osimertinib and the subgroup initiating 1L chemotherapy. During the baseline period, the mean Quan-Charlson comorbidity index (CCI) score was 7.1 for the total population and on average, 2.7 months elapsed between the first LC diagnosis and initiation of 1L therapy. Patients’ mean monthly all-cause healthcare costs during the year prior to 1L initiation were US$5776, predominantly driven by medical costs (US$5623).

3.2 Healthcare Resource Utilization

All-cause monthly HRU, laboratory service utilization, and supportive care utilization during 1L therapy are reported for the total population, for the osimertinib subgroup, and the chemotherapy subgroup in Table 2. An inpatient admission was observed among 26.8% of the total population, 27.7% of the osimertinib subgroup, and 22.2% of the chemotherapy subgroup. An ER visit was observed for 39.4% of the total population, 39.7% of the osimertinib subgroup, and 31.1% of the chemotherapy subgroup. Among all three groups, over 99% of patients had an outpatient visit. The mean PPPM number of days with outpatient visits was 4.44 days for the total population, 3.60 days for the osimertinib subgroup, and 7.05 days for the chemotherapy subgroup. The mean number of PPPM days of inpatient stay was 0.53 days for the total population, 0.64 days for the osimertinib subgroup, and 0.23 days for the chemotherapy subgroup. The mean PPPM number of days with laboratory services was 2.61 days for the total population, 2.15 days for the osimertinib subgroup, and 4.60 days for the chemotherapy subgroup. The mean PPPM number of days with supportive care was 0.71 days for the overall sample, 0.39 days for the osimertinib subgroup, and 1.83 days for the chemotherapy subgroup.

Table 2 HRU and costs during 1L therapy
Full size table

3.3 Healthcare Costs

All-cause monthly healthcare costs incurred during 1L therapy are presented for the three groups in Table 2. Cost breakdowns for the major cost drivers are presented in Fig. 2. Total PPPM all-cause costs during 1L therapy for the total population were US$25,499; drug-related costs (including pharmacy costs as well as antineoplastic drug costs and their administration) were US$15,356, other outpatient costs were US$5400, and inpatient costs were US$4167. For the osimertinib subgroup, PPPM all-cause costs totaled US$27,174, of which US$16,673 were drug-related costs, US$4432 were other outpatient costs, and US$5462 were inpatient costs. For the chemotherapy subgroup, PPPM all-cause costs totaled US$23,343, of which US$13,883 were drug-related costs, US$7734 were other outpatient costs, and US$1166 were inpatient costs. ER costs and other services costs were numerically similar for the total population, the osimertinib subgroup, and the chemotherapy subgroup. Among the total population, PPPM laboratory service utilization costs were US$474. Laboratory costs were numerically higher in the chemotherapy subgroup (US$886 PPPM). Total supportive care costs were US$596 PPPM (ESAs: US$0 PPPM; G-CSFs/GM-CSFs: US$145 PPPM; pain management agents: US$15 PPPM; and respiratory support care: US$436 PPPM) in the total population and were similar for the osimertinib and chemotherapy subgroups. G-CSF/GM-CSF costs were numerically higher among the chemotherapy subgroup (osimertinib: US$28 PPPM; chemotherapy: US$598 PPPM), while respiratory support care costs were numerically higher in the osimertinib subgroup (osimertinib: US$643 PPPM; chemotherapy: US$3 PPPM).

Fig. 2
figure 2

All-cause healthcare costs per patient per month during 1L therapy. 1L first line, ER emergency room

Full size image

4 Discussion

This descriptive study identified 213 patients with advanced EGFRm NSCLC initiating treatment in recent years, most of whom initiated osimertinib (141 patients) or chemotherapy (45 patients) in 1L. Regardless of the therapy received, significant HRU and costs primarily attributable to treatment regimen, inpatient services, and outpatient services were incurred. However, total costs of care PPPM, on average, were numerically higher in 1L osimertinib-treated patients than in patients receiving 1L chemotherapy. There were notable descriptive differences in the types of costs incurred; osimertinib-treated patients incurred higher inpatient costs while chemotherapy-treated patients incurred higher outpatient costs (consistent with chemotherapy infusions occurring in the outpatient setting). Non-drug-related costs were similar for 1L osimertinib and 1L chemotherapy. The descriptive differences observed in inpatient admissions are noteworthy and unexpected. Given the clinical benefit data that has established osimertinib as the current 1L standard of care, the intuitive expectation is that other aspects of patient care, including quality of care, non-pharmacy HRU, or patient quality of life would accordingly also be improved relative to other therapies (e.g. chemotherapy). The findings of higher inpatient costs and inpatient days for 1L osimertinib use relative to chemotherapy seem to counter this assumption. Additional research is indicated to better understand how these 1L treatment options compare in terms of their overall quality and value for population health management and ability to fully address unmet needs in EGFRm NSCLC.

The findings of this study should be interpreted in the context of the respective efficacy and tolerability profiles of osimertinib- and chemotherapy-based regimens. Recent clinical trial data suggests that patients treated with osimertinib in second line have significantly longer PFS relative to patients treated with platinum-based chemotherapy in second line (10.1 months vs 4.4 months) and significantly better response rate (71% vs 31%) [7], which can potentially be cost effective [14]. However, second-line osimertinib treatment was not associated with significantly different overall survival than second-line chemotherapy, although the authors noted high (73%) crossover from chemotherapy to osimertinib during the trial [7, 11]. In addition to efficacy, osimertinib and chemotherapy’s different tolerability profiles were highlighted in these trials; grade 3 adverse events occurred in 37% of patients treated with osimertinib, compared with 48% of patients treated with platinum-based chemotherapy and 36% of crossover patients [11]. The type of adverse events experienced also differed by treatment, with common adverse events for osimertinib including diarrhea, rashes, dry skin, and nail effects whereas platinum doublet chemotherapy recipients more frequently reported anemias, low platelet counts (thrombocytopenia), neutropenia, decreased appetite, fatigue, and nausea [7, 11]. These differences in clinical profile are reflected in the total costs of care associated with each type of treatment.

While there have been studies evaluating the healthcare costs [15,16,17] and cost effectiveness [18, 19] of previous generation TKIs such as erlotinib and afatinib, research remains limited on the magnitude of real-world total costs of care associated with osimertinib and chemotherapy as 1L therapies. This may be due to the limited time osimertinib has been approved for 1L treatment of metastatic NSCLC (since 2018) [8], the specific population studied (advanced EGFRm NSCLC), and the difficulty associated with combining cost information with mutation data (without relying on proxies) using real-world data sources. Nevertheless, in the current study, 1L osimertinib recipients incurred numerically higher costs (US$27,174) relative to 1L chemotherapy recipients (US$23,343). Consequentially, unmet needs may remain for 1L treatment for advanced NSCLC, despite advances in targeted care. As data continues to accumulate, mutation testing continues to be more prevalent [20], and the ability to merge various data sources improves, additional real-world evidence will emerge to confirm these findings in larger patient samples with longer follow-up.

The results of this study should be interpreted in light of certain limitations. First, the results were descriptive in nature and no statistical tests or adjustments were made to determine the statistical significance of the differences observed between patient cohorts, given the small sample size. In addition, findings may not be generalizable to patients without health insurance or patients covered by other types of insurance. In the absence of staging and mutation information in administrative claims data, patients with EGFRm advanced NSCLC were identified via proxy measures. For instance, NSCLC patients were identified based on the exclusion of patients using common SCLC treatments and advanced disease stage was proxied by metastatic cancer diagnosis or initiation of 1L therapy within 30 days following LC diagnosis. Similarly, the use of TKIs was used as a proxy for EGFRm; therefore; it is possible that patients may have been misclassified as having or not having EGFRm NSCLC. It also means that the results should be interpreted with caution since patients with certain EGFR mutations, such as the exon-20 insertion mutation which imparts a level of TKI resistance, may have differences in cost and HRU which were not assessed separately from the population of patients with EGFRm NSCLC sub-types that are responsive to TKIs. Furthermore, costs evaluated as part of this study only included direct healthcare costs captured in claims data. Thus, other types of (indirect) costs, such as the less tangible value that treatments may provide in terms of positive quality-of-life outcomes that are not associated with direct cost implications, were not observable in this database. Lastly, as with all claims-based studies, there are possible inaccuracies due to coding errors and missing data.

5 Conclusions

The current descriptive study found that among patients with advanced EGFRm NSCLC, average total costs of care were numerically higher in 1L osimertinib-treated patients than in patients receiving 1L chemotherapy and descriptive differences in type of spending and HRU were also observed between the patient populations. Similar total non-drug-related costs of care (i.e., excluding treatment costs) during 1L therapy were observed for the osimertinib and chemotherapy subgroups. These findings indicate that unmet needs may remain for 1L treatment of patients with EGFRm advanced NSCLC. Additional therapies are needed to balance benefits, risks, and total cost of care for this patient population.