Health Care Resource Use for Modern First-Line Treatments in Metastatic Renal Cell Carcinoma

This cohort study investigates health care resource use, costs, and outcomes for first-line pembrolizumab plus axitinib or ipilimumab plus nivolumab among patients with metastatic renal cell carcinoma.

Key Points

Question

What are the health care resource use, associated costs, and clinical outcomes for patients with metastatic renal cell carcinoma (mRCC) receiving first-line ipilimumab plus nivolumab (I+N) or pembrolizumab plus axitinib (P+A)?

Findings

In this cohort study among 507 patients with mRCC, including 126 patients receiving P+A and 381 patients receiving I+N, treatment with P+A was associated with longer time on treatment and time to first emergency department visit and inpatient stay. Total adjusted all-cause per-patient, per-month costs were similar between groups, but 12-month estimated costs were higher for patients receiving P+A.

Meaning

This study found differences in costs associated with clinical use of mRCC treatments. These findings suggest that clinicians and patients should be aware of and consider cost differences in determining treatment selection.

Abstract

Importance

Immuno-oncology agents have changed the treatment paradigm for metastatic renal cell carcinoma (mRCC). Such therapies improve survival but can impose considerable health care resource use (HCRU) and associated costs, necessitating their examination.

Objective

To compare HCRU, costs, and clinical outcomes among patients receiving first-line pembrolizumab plus axitinib (P+A) or ipilimumab plus nivolumab (I+N).

Design, Setting, and Participants

This retrospective cohort study used data from an administrative claims database on patients with mRCC receiving first-line P+A or I+N that was initiated between January 2018 and May 2020. Data were analyzed from February 2021 to July 2022.

Exposure

First-line P+A or I+N.

Main Outcome and Measures

HCRU and costs during the first 90 days, full first-line treatment, and full follow-up periods were assessed. Using Kaplan-Meier analysis, time on treatment, overall survival, time to first emergency department (ED) visit, and time to first inpatient stay were compared.

Results

Among 507 patients, there were 126 patients receiving P+A (91 male [72.2%]; mean [SD] age, 67.93 [9.66] y) and 381 patients receiving I+N (271 male [71.1%]; mean [SD] age, 66.52 [9.94] years). The median time on treatment was longer for the P+A compared with I+N group (12.4 months [95% CI, 8.40 months to not estimable] vs 4.1 months [95% CI, 3.07 to 5.30 months]; P < .001). The median time to first ED visit was longer for the P+A than I+N group (7.2 months [95% CI 3.9 to 11.1 months ] vs 3.3 months [95% CI, 2.6 to 3.9 months]; P = .005), as was time to first inpatient stay (9.0 months [95% CI 6.5 months to not estimable] vs 5.6 months [95% CI, 3.9 to 7.9 months]; P = .02). During the first 90 days, a lower proportion of the P+A than N+I group had ED visits (43 patients [34.1%] vs 182 patients [47.8%] and inpatient stays (24 patients [19.1%) vs144 patients [37.8%]; P < .001). During full follow-up, mean total adjusted costs were similar for P+A and I+N groups, but adjusted 12-month estimated total costs were higher for P+A than I+N groups ($325 574 vs $ 263 803; P = .03).

Conclusions and Relevance

In this study, treatment with P+A was associated with longer time on treatment, time to first ED visit, and inpatient stay, while 12-month estimated costs were higher for the P+A group. This is among the first clinical studies to evaluate economic burden associated with modern treatments for mRCC.

Introduction

Renal cell carcinoma (RCC) is the eighth most common cancer in the US,¹ accounting for 90% of kidney cancer cases.^2,3,4 Approximately 20% to 40% of patients with localized RCC progress to metastatic RCC (mRCC), with a 5-year relative survival of 15% to 16%.^5,6,7 Vascular endothelial growth factor–targeted tyrosine kinase inhibitors (VEGF-TKIs) have been the standard first-line treatment (LOT1) of mRCC since 2005.^{8,9,10,11,12,13} Since 2018, this paradigm has shifted with immuno-oncology (IO)–based therapies in combination with another IO or VEGF-TKI. These IO-based therapies have shown durable response and improved survival in clinical trials and studies of clinical practice.^{14,15,16,17,18,19,20,21,22,23,24,25,26,27,28}

Unfortunately, mRCC is associated with a considerable increase in economic burden among patients relative to premetastatic disease.^29,30, However, a paucity of published studies about modern mRCC treatment in clinical practice leaves much to be known about IO-IO and IO-TKI regimens outside of clinical trials. Given that financial burden and out-of-pocket expenses are major concerns for patients^31,32 and the entire health care infrastructure,^33,34,35 it is critical to study economic outcomes for cancer care in clinical practice. Results may provide ancillary information to optimize treatment for individual patients. We performed a retrospective study of patients with mRCC receiving LOT1 pembrolizumab plus axitinib (P+A; US approval May 2019) or ipilimumab plus nivolumab (I+N; US approval March 2018). We compared all-cause health care resource use (HCRU), associated costs, and clinical outcomes between these groups using health care claims data to determine whether differences were present.

Methods

This retrospective cohort study was approved by the Western Copernicus Group Institutional Review Board. This study was retrospective in nature, where data on processed claims compiled in the form of structured data were used, so the study did not require patient informed consent. It was reviewed for concordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

Study Design and Data Source

This study used deidentified data from the Optum Research Database containing medical and pharmacy claims (including linked enrollment) data for Commercial and Medicare Advantage enrollees. Claims are collected from all sites of health care, including associated paid amounts. Medical claims contain International Statistical Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) diagnosis and procedure codes, as well as Current Procedural Terminology and Healthcare Common Procedure Coding System codes. Pharmacy claims were sourced from outpatient filled prescriptions, with National Drug Codes. Mortality data were accessed via the National Death Index, Centers for Medicare & Medicaid Services, and Social Security Administration Death Master File. In 2021, the Optum Research Database represented 7% of the US commercially insured population and 20% of the Medicare Advantage population.

Patient Selection

To be eligible, adult patients (aged ≥18 years) had to have at least 2 ICD-10-CM codes for RCC diagnosis (C64*) at least 30 days apart and at least 1 diagnosis of secondary metastasis (C77xx-C80.0x) on or after the RCC diagnosis date between January 1, 2017, and August 31, 2020 (the date of mRCC diagnosis). Patients with claims for LOT1 P+A or I+N from 14 days prior to any time after the mRCC date (during January 1, 2018, to May 31, 2020) were identified using codes from the National Drug Codes, Healthcare Common Procedure Coding System, or both (eTable in Supplement 1). The date of the first identified claim for P+A or I+N was considered the index date (start date of LOT1). To be considered combination therapies of P+A or I+N, individual claims had to be within 28 days of each other. LOT1 ended on whichever was earliest: the end of the study period, start of a subsequent LOT, death, disenrollment from health plan, or discontinuation (ie, 120 days after the runout date). Additionally we required continuous enrollment with medical and pharmacy coverage for a minimum of 6 months preceding and at least 3 months after the index date unless follow-up was shortened by death. We excluded patients who received systemic treatment prior to the mRCC diagnosis date, participated in clinical trials, or were pregnant. Demographic and clinical characteristics were described during a baseline period of at least 6 months before the index date; death events were captured as month and year.

HCRU and Costs

For P+A and I+N groups, all-cause HCRU counts by type of use and associated costs were examined from the index date to the end of database availability, patient death, or end of enrollment, whichever occurred first, and were reported for the first 90 days of LOT1, full LOT1 period, and full follow-up period. We examined HCRU associated with medical encounters, including ambulatory (physician office and hospital outpatient) visits, ED visits, inpatient (IP) stay (including length of stay in days), and IP stay with intensive care unit (ICU) admission. Total all-cause costs included pharmacy and medical costs. We obtained costs for axitinib from pharmacy claims and costs for intravenous drugs (pembrolizumab, ipilimumab, and nivolumab) from medical claims. We adjusted costs to reflect inflation between the claim date and 2020 using Consumer Price Index medical care components. Costs were also adjusted by coordination of benefits attributed to Medicare; for other payers, costs were based on coordination of benefits information obtained during the usual course of business. We incorporated estimated amounts paid by other payers for a total paid or allowable amount. We also examined the time to first visits for ED or IP care and for IP stay with ICU admission.

Clinical Outcomes

We defined time on treatment as the time between the index date and discontinuation date. If no discontinuation criteria were met, patients were censored at the last LOT claims date. The time to next treatment was measured as the time between the index date and first date of subsequent LOT administration. For examining overall survival, we calculated survival length; patients alive during the study period were censored for date of death.

Statistical Analysis

Patient characteristics, treatment patterns, and HCRU were analyzed and compared descriptively by treatment group. We used χ² and t tests to assess differences in categorical and continuous variables, respectively. HCRU was reported as per patient per month (PPPM) to account for differences in follow-up time. PPPM counts for medical HCRU, such as ambulatory, ED, and IP visits and IP with ICU stay were assessed using descriptive statistics and reported as means and SDs. PPPM length (in number of days) of IP stays was calculated and presented as mean and SD. Descriptive analysis was used to report all-cause PPPM medical, pharmacy, and total costs during 3 periods described similarly as for HCRU. Adjusted incidence rate ratios and 95% CIs for HCRU counts were also analyzed using generalized linear models with log link and γ distribution accounting for differences in baseline patient characteristics.

To account for variable follow-up in P+A and I+N groups, we performed Lin regression to calculate 12-month estimated costs.³⁶ Kaplan-Meier was used for analysis of time-to-event outcomes, with adjustment for baseline characteristics. A 2-sided P value < .05 was considered statistically significant, and all analyses were conducted using SAS statistical software version 9.4 (SAS Institute). Data were analyzed from February 2021 to July 2022.

Results

Baseline Characteristics

Among 507 patients with mRCC, there were 126 patients receiving P+A (91 male [72.2%]; mean [SD] age, 67.93 [9.66] y) and 381 patients receiving I+N (271 male [71.1%]; mean [SD] age, 66.52 [9.94] years) (eFigure 1 in Supplement 1). Table 1 describes demographic and clinical characteristics, which were similar across groups.

Table 1. Patient Characteristics.

Characteristic	Patients, No. (%)		P value
	P+A (n = 126)	I+N (n = 381)
Index year
2018	0	139 (36.48)	<.001
2019	84 (66.67)	168 (44.09)	<.001
2020	42 (33.33)	74 (19.42)	.001
Age, mean (SD), y	67.93 (9.66)	66.52 (9.94)	.17
Sex
Male	91 (72.22)	271 (71.13)	.81
Female	35 (27.78)	110 (28.87)
Medicare	77 (61.11)	231 (60.63)	.92
Region
Northeast	11 (8.73)	53 (13.91)	.13
Midwest	41 (32.54)	111 (29.13)	.47
South	57 (45.24)	167 (43.83)	.78
West	17 (13.49)	50 (13.12)	.92
Baseline NCI Charlson comorbidity index score
0-2	81 (64.29)	229 (60.11)	.40
≥3	45 (35.71)	152 (39.89)
No. of metastatic sites
0-2	81 (64.28)	219 (57.48)	.18
≥3	45 (35.71)	162 (42.52)
Sites of metastasis
Lung	70 (55.56)	207 (54.33)	.81
Bone	41 (32.54)	128 (33.60)	.83
Lymph nodes	32 (25.40)	106 (27.82)	.60
Liver	15 (11.90)	63 (16.54)	.21

Abbreviations: I+N, ipilimumab plus nivolumab; mRCC, metastatic renal cell carcinoma; NCI, National Cancer Institute; P+A, pembrolizumab plus axitinib.

HCRU Counts

During the first 90 days, the proportion of patients with ED visits was lower for those receiving P+A than those receiving N+I (43 patients [34.1%] vs 182 patients [47.8%]; P < .001), as was the proportion with IP stays (24 patients [19.1%] vs 144 patients [37.8%]; P < .001). Overall, unadjusted mean all-cause PPPM HCRU was lower in the P+A group compared with the I+N group (Table 2). In the first 90 days of LOT1, patients receiving P+A experienced fewer mean (SD) PPPM ambulatory visits (6.7 [3.4] visits vs 7.5 [4.5] visits; P = .03) and IP stays (0.1 [0.2] stays vs 0.2 [0.4] stays; P < .001) and a similar mean (SD) number of ED visits (0.32 [0.74] visits vs 0.4 [0.7] visits; P = .25). In addition, the mean (SD) length of IP stay was shorter for the P+A group than I+N group (0.8 [2.6] days vs 2.2 [4.8] days; P < .001). Findings were similar between total duration and the first 90 days of LOT1. In contrast, during the full follow-up period, the mean (SD) PPPM number of ambulatory and ED visits was similar between groups. The number of IP stays and IP stays with ICU care and the length of IP stays were greater for the I+N group compared with the P+A group (Table 2) during the full follow-up period. In adjusted analyses (Table 3), the P+A group had a lower incidence rate ratio for IP stay across all three time periods compared with the I+N group (0.42 [95% CI, 0.27-0.66]; P < .001 for the first 90 days; 0.56 [95% CI, 0.39-0.81]; P = .002 for LOT1; 0.57 [95% CI, 0.41-0.80]; P = .001 for the full follow-up). In contrast, ED visits and IP stay with ICU visit were similar between groups across all periods.

Table 2. Unadjusted PPPM HCRU and Costs.

Outcome	First 90 d of LOT1			Total duration of LOT1			Full follow-up period
	Mean (SD)		P value	Mean (SD)		P value	Mean (SD)		P value
	P+A	I+N		P+A	I+N		P+A	I+N
HCRU event
Ambulatory visit	6.68 (3.37)	7.52 (4.52)	.03	6.28 (2.80)	7.00 (4.20)	.03	6.16 (2.70)	6.54 (3.89)	.23
ED visit	0.32 (0.74)	0.40 (0.65)	.25	0.32 (0.76)	0.38 (0.58)	.36	0.30 (0.73)	0.38 (0.57)	.27
IP stay	0.09 (0.22)	0.23 (0.38)	<.001	0.11 (0.20)	0.23 (0.36)	<.001	0.11 (0.20)	0.22 (0.35)	<.001
Length of IP stay, d	0.80 (2.55)	2.24 (4.79)	<.001	1.00 (2.25)	2.05 (3.75)	<.001	0.96 (2.04)	2.06 (3.47)	<.001
IP stay with ICU visit	0.05 (0.16)	0.10 (0.29)	.02	0.06 (0.15)	0.10 (0.28)	.02	0.05 (0.15)	0.10 (0.27)	.01
Cost in (US $)a
Total (medical + pharmacy)	36 963 (15 240)	48 939 (37 040)	<.001	31 868 (14 739)	37 115 (31 993)	.01	31 319 (15 002)	33 298 (27 971)	.31
Medical	21 123 (14 737)	48 436 (37 154)	<.001	19 328 (13 573)	36 645 (32 048)	<.001	18 936 (113 616)	30 598 (28 004)	<.001
Pharmacy	15 840 (6150)	502 (2697)	<.001	12 540 (5473)	469 (2630)	<.001	12 382 (5504)	2701 (4643)	<.001

Abbreviations: ED, emergency department; HCRU, health care resource use; I+N, ipilimumab and nivolumab; ICU, intensive care unit; IP, inpatient; LOT1, first line of therapy; P+A, pembrolizumab plus axitinib; PPPM, per patient per month.

^a Adjusted for coordination of benefits for Medicare and Consumer Price Index in 2020 US dollars.

Table 3. Adjusted Risk of HCRU Events.

HCRU event	First 90 d of LOT1			Total duration of LOT1			Full follow-up
	P+A	I+N	P value	P+A	I+N	P value	P+A	I+N	P value
ED visit
Events, No	99	366	NA	219	659	NA	246	958	NA
Person-mo at risk, No.	363	1063	NA	958	2834	NA	1099	4105	NA
Proportion of events/time at risk	0.27	0.34	NA	0.23	0.23	NA	0.22	0.23	NA
Adjusted IRR (95% CI)a	0.74 (0.51-1.08)	1 [Reference]	.12	0.75 (0.53-1.06)	1 [Reference]	.10	0.74 (0.54-1.01)	1 [Reference]	.06
IP stay
Events, No.	28	198	NA	80	338	NA	96	490	NA
Person-mo at risk, No.	363	1063	NA	958	2834	NA	1099	4105	NA
Proportion of events/time at risk	0.08	0.19	NA	0.08	0.12	NA	0.09	0.12	NA
Adjusted IRR (95% CI)a	0.42 (0.27-0.66)	1 [Reference]	<.001	0.56 (0.39-0.81)	1 [Reference]	.002	0.57 (0.41-0.80)	1 [Reference]	.001
IP stay with ICU visit
Events, No.	15	74	NA	38	133	NA	41	180	NA
Person-mo at risk, No.	363	1063	NA	958	2834	NA	1099	4105	NA
Proportion of events/time at risk	0.04	0.07	NA	0.04	0.05	NA	0.04	0.04	NA
Adjusted IRR (95% CI)a	0.68 (0.36-1.32)	1 [Reference]	.25	0.75 (0.45-1.26)	1 [Reference]	.27	0.70 (0.43-1.14)	1 [Reference]	.15

Abbreviations: HCRU, health care resource use; I+N, ipilimumab plus nivolumab; ICU, intensive care unit; IP, inpatient; IRR, incidence rate ratio; LOT1, first line of therapy; NA, not applicable; P+A, pembrolizumab plus axitinib; REF, reference.

^a Adjusted for age (continuous), sex, insurance type, geographical region, baseline National Cancer Institute Charlson comorbidity index score, index date, number of metastatic sites, and site of metastasis.

The median time to first ED visit was 7.2 months (95% CI, 3.9 to 11.1 months) for P+A and 3.3 months (95% CI, 2.6 to 3.9 months) for I+N (log-rank P = .005) (Figure 1A). The median time to IP stay was 9.0 months (95% CI, 6.5 months to not estimable) for P+A and 5.6 months (95% CI, 3.9 to 7.9 months) for I+N (log-rank P = .02) (Figure 1B). There was no difference in the median time to first IP stay with ICU visit for P+A (not estimable) compared with I+N (22.4 months [95% CI, 20.6 months to not estimable]) (log-rank P = .69) (Figure 1C).

Figure 1. Time to Outcomes During Full Follow-Up by Treatment Group.

ED indicates emergency department; I+N, ipilimumab plus nivolumab; ICU, intensive care unit; IP, inpatient; P+A, pembrolizumab plus axitinib.

HCRU Costs

Total (medical + pharmacy) mean (SD) all-cause unadjusted PPPM costs were lower in the P+A group compared with the I+N group during the first 90 days of LOT1 ($36 963 [$15 240] vs $48 939 [$37 040]; P < .001) and the total duration of LOT1 ($31 868 [$14 739] vs $37 115 [$31 993]; P = .01) (Table 2). Mean (SD) costs for the full duration of follow-up were similar between groups ($31 319 [$15 002] vs $33 298 [$27 971]; P = .31). Mean (SD) all-cause unadjusted PPPM medical costs were lower for patients receiving P+A compared with those receiving I+N during the entire LOT1 ($19 328 [$13 573] vs $36 645 [$32 048]; P < .001). However, LOT1 mean (SD) all-cause PPPM pharmacy costs were higher for patients in the P+A group compared with those in the I+N group ($12 540 [$5473] vs $469 [$2630]; P < .001). Similarly, pharmacy costs for P+A and the medical cost for I+N remained higher during the first 90 days of LOT1 and the full follow-up period. These differences in medical and pharmacy costs were associated with benefit coverage of infusions (medical costs) and oral therapies (pharmacy costs).

After adjustment for clinical and demographic factors, all-cause PPPM total costs for LOT1 and the full follow-up period were similar between P+A and I+N groups ($33 421 vs $36 530; P = .20 and $31 720 vs $33 173; P = .53, respectively) (Figure 2A). However, patients receiving P+A had lower all-cause PPPM medical costs in LOT1 ($19 425 vs $36 522; P < .001) and the full follow-up ($18 377 vs $30 902; P < .001). Higher pharmacy costs were observed for P+A than I+N during LOT1 ($17 946 vs $404; P < .001) and the full follow-up ($15 702 vs $2511; P < .001).

Figure 2. Estimated Costs.

I+N indicates ipilimumab plus nivolumab; LOT1, first line of therapy; P+A, pembrolizumab plus axitinib; PPPM, per patient per month.

The 12-month estimated total costs during the full follow-up period using the Lin regression method were higher for the P+A group than the I+N group ($325 574 vs $263 803; P = .03) (Figure 2B). The 12-month estimated medical costs for P+A vs I+N did not differ statistically ($200 856 vs $233 912; P = .19), whereas pharmacy costs were higher ($124 718 vs $29 891; P < .001). Similar findings were observed when 12-month estimated total costs analysis was performed for LOT1-only treatments (Figure 2B).

Clinical Outcomes

The median (IQR) continuous follow-up time for the study population was approximately 8.6 months (8.6 [4.6-14.9] months for the I+N group and 8.5 [4.9-12.6] months for the P+A group). The median time on treatment was longer for the P+A than I+N group (12.4 months [95% CI, 8.40 months to not estimable] vs 4.1 months [95% CI, 3.07 to 5.30 months]; log-rank P < .001) (eFigure 2 in Supplement 1).

The median overall survival and median time to next treatment were not estimable in the population because estimated survival proportions and time to next treatment for the groups remained greater than 50% through the end of follow-up (eFigure 3 in Supplement 1). At 12 months, 36 patients (77.3%) in the P+A group and 146 patients (66.9%) in the I+N group were still alive. Similarly, 28 patients (42.8%) in the I+N group had begun their next treatment at 12 months compared with 94 patients (25.4%) in the P+N group.

Discussion

The IO-IO and IO-VEGF-TKI combination treatments for mRCC are relatively new; thus, their clinical effectiveness outcomes are not well understood. We conducted a retrospective cohort study using US administrative claims data to examine treatment patterns, HCRU, and costs among patients with mRCC initiating LOT1 with I+N or P+A. The analysis of time on treatment has been suggested as an effectiveness end point in clinical studies due to the moderate to high correlation with progression-free survival.^37,38,39,40 We observed longer time on treatment for patients receiving P+A compared with those receiving I+N. These results are consistent with the US oncology study that found that P+A reduced the risk of treatment discontinuation (time on treatment) by 41% compared with I+N.²⁴ Time on treatment values observed were similar to those reported by Shah et al.^41,42 Treatment discontinuation for the I+N group may have been associated with immune-related adverse events.^{43,44,45,46,47}

For the I+N group, the time to first ED visit and first IP stay were shorter than for the P+A group. After adjustment for baseline characteristics, patients receiving P+A had a lower risk of having an IP stay; however, there was no difference in risk of ED visit or IP stay with ICU. Few studies have been published detailing HCRU among patients with mRCC in LOT1. One study⁴⁸ used the US point-in-time survey and reported a mean of 0.2 hospitalizations over the past 6 months in patients treated with IO-IO combinations. We observed a similar mean unadjusted PPPM for IP stays of 0.2 in the I+N group for the full follow-up period. The unadjusted number of IP stays was smaller for patients receiving P+A compared with those receiving I+N (0.1 vs 0.2 stays). Furthermore, after adjustment for baseline characteristics, the incidence of IP stays remained different between study groups. Differences in HCRU may be associated with differences in the tolerability profile of the 2 regimens. However, we did not collect reasons for IP stays.

We observed an increased economic burden associated with these IO and TKI–based mRCC treatments. This is not unexpected given that studies have shown an increase in PPPM costs for cancer drugs and radiation over the last decades. Our 12-month estimated costs for LOT1 IO and TKI combination treatments ranged from $263 803 to $325 574 (2020 US dollars). We also noted significant differences between LOT1 I+N and P+A regimens for total, medical, and pharmacy costs. Treatment with I+N was associated with similar mean adjusted PPPM total costs during the full follow-up period compared with treatment with P+A ($33 173 vs $31 720; P = .53). However, mean estimated 12-month costs were higher among patients receiving P+A compared with those receiving I+N ($325 574 vs $263 803; P = .03). The higher 12-month costs for P+A were probably associated with a significantly longer duration of treatment compared with I+N (median time on treatment, 12.4 months vs 4.1 months).

Strengths and Limitations

Our study has several strengths, including the large sample size, underlying geographic distribution, and completeness of claims available for individual patients. However, several limitations also must be considered. First, while health care claims data present good sources for retrospective studies of health care use and costs, claims data are recorded for reimbursement rather than research. Thus, identification beginning with RCC was subject to incomplete and miscoded claims. Costs were presented as the combined (patient-paid and payer-paid) amounts because the goal was to give a more holistic presentation of costs associated with these modern mRCC treatments. Furthermore, because metastatic disease staging is not available in claims, we must rely on accurate coding of secondary malignant neoplasms for mRCC identification.

Our comprehensive analysis of 3 periods (the first 90 days of treatment, LOT1 duration, and full follow-up period) represents an important strength. These various time analyses provided insights into other unseen factors that may have been associated with treatment outcomes, including treatment-related toxic effects, treatment length, and patient survival. Furthermore, we were unable to calculate cost associated with specific toxic effects or cost of treatment owing to limitations of the study dataset. Our 3-point evaluation may provide some insights into this. In addition, because claims do not include all clinical information, we could not consider relevant data, such as histology or International mRCC Database Consortium risk criteria,⁴⁹ which are correlated with treatment selection and clinical outcomes.⁵⁰ Moreover, although we performed a robust statistical analysis accounting for baseline characteristics and time length bias to adjust for confounding variables, findings may be affected by unobservable confounding, such as treatment dosing schedule (nivolumab 2, 3, or 4 week or pembrolizumab 3 or 6 week). In addition, more IO-TKI–based agents, including nivolumab plus cabozantinib and lenvatinib plus pembrolizumab, are now approved for kidney cancer treatment, and data should not be generalized to these regimens. The study period used for this analysis did not include long-term follow-up; therefore, the overall survival measure and time to next treatment are likely not mature, should be interpreted with caution, and should be replicated by other studies with longer follow-up time available. Additionally, this database includes commercial or private Medicare supplemental insurers; thus, results may not be generalizable to patients with other insurers, and patient-paid costs are not available.

Conclusions

In this cohort study, we present one of the first studies evaluating the economic burden associated with modern IO- and TKI-based combination treatments for patients with mRCC. We observed differences in HCRU and costs between first-line I+N and P+A regimens. Treatment with I+N was associated with a shorter time to first ED or IP stay, a higher risk for an IP stay, and greater total adjusted PPPM costs, whereas P+A was associated with higher 12-month total estimated costs. Our study highlights differences in costs associated with clinical use of mRCC treatments of which clinicians and patients should be aware. Several factors are associated with choice of treatment, including treatment efficacy, toxic effects, access to health care, cancer-related symptoms, and availability of drugs. Therefore, our findings should not be the primary driver for the treatment selection but rather be considered a critical ancillary data point. Our study highlights an underrepresented area in cancer care, and similar retrospective and prospective studies are needed to confirm our findings.

Supplement 1.

eTable. Codes for Identifying Treatments

eFigure 1. Patient Identification and Attrition

eFigure 2. Time on Treatment for Patients With mRCC Initiating L1 P+A or I+N

eFigure 3. Overall Survival and Clinical Time to Next Treatment for Patients With mRCC Initiating LOT1 with P+A or I+N

Supplement 2.

Data Sharing Statement