Healthcare Costs in Men with Metastatic Castration-Resistant Prostate Cancer: An Analysis of US Medicare Fee-For-Service Claims

In 2022, the estimated number of new prostate cancer (PC) cases in the US was 268,000 (14% of all new cancer diagnoses), with an estimated 34,500 deaths, making PC the most common cancer and one of the most lethal in men [1]. While some men present with metastatic castration-sensitive PC (mCSPC), others present with nonmetastatic castration-sensitive PC (nmCSPC). For those with nmCSPC, many will undergo primary definitive therapy, after which approximately one-third of men will experience biochemical recurrence [2]. Over time, many of the men with mCSPC and those with biochemically recurrent nmCSPC will progress to castration-resistant PC (CRPC), a form of advanced disease associated with a very high risk of mortality [3].

Currently, there are multiple classes of US Food and Drug Administration-approved life-prolonging therapy for metastatic CRPC (mCRPC). These include chemotherapy (docetaxel and cabazitaxel), immunotherapy (sipuleucel-T), novel hormonal therapy (abiraterone and enzalutamide), poly(adenosine diphosphate-ribose) polymerase (PARP) inhibitors (olaparib and rucaparib), and radiopharmaceuticals (radium 223 and lutetium Lu 177 vipivotide tetraxetan) [4, 5]. Depending on the patient’s treatment prior to mCRPC diagnosis, mCRPC treatments are generally sequenced through multiple lines of therapy by the clinician, with an emphasis on recommending therapy with an alternative mechanism of action [4]. As new therapies for the management of mCRPC emerge, it is important to assess treatment patterns and their impact on healthcare resource utilization (HRU) and healthcare costs in a real-world setting.

Previous real-world studies have reported healthcare costs before and after mCRPC diagnosis in various databases, including commercial insurance claims databases, or have compared HRU and healthcare costs between patients initiating different treatments [6‐15]. However, limited real-world research in mCRPC has been conducted in the fee-for-service US Medicare population [16, 17]. This is significant, as around 61% of incident PC cases are diagnosed in men aged 65 years or older, and Medicare is an important insurer in this age group [1, 18]. To our knowledge, a comprehensive analysis of HRU and healthcare costs with progression through different lines of therapy after mCRPC diagnosis has not been reported, and further insights into the economic burden of mCRPC are needed.

The purpose of this analysis was to describe the HRU and healthcare costs of patients with mCRPC in the US Medicare fee-for-service population before and after diagnosis, and during progression through lines of life-prolonging therapy. We hypothesized that HRU and healthcare costs would be higher and continue to rise through multiple lines of therapy after mCRPC diagnosis, compared with before mCRPC diagnosis.

This non-interventional, retrospective cohort study used claims data from the Medicare fee-for-service research identifiable files, sourced from the Centers for Medicare & Medicaid Services, between January 1, 2014, and December 31, 2019. The study design was reviewed by the WIRB-Copernicus Group^® institutional review board (Princeton, NJ, USA) and was determined to be exempt from institutional review board review. This study was conducted in accordance with legal and regulatory requirements, and followed the research practices described in the Guidelines for Good Pharmacoepidemiology Practices issued by the International Society for Pharmacoepidemiology.

Data were included for beneficiaries aged 65 years or older who were continuously enrolled with insurance coverage for ≥ 1 year before and ≥ 6 months after index mCRPC diagnosis, or until patients died for those with < 6 months post-index follow-up. A modified version of a claims-based algorithm was used to identify men with mCRPC to include in the analysis population (Fig. 1) [16]. Modifications were made to the criteria for medications for mCRPC identifying castration resistance. Briefly, male patients were required to have ≥ 2 claims between 2014 and 2019 with a diagnosis code for PC, ≥ 1 claim with a diagnosis code for metastatic disease on or after the first PC diagnosis, no metastasis or other cancer diagnosis prior to the first observed PC diagnosis, and to have evidence of castration resistance. Patients were excluded if they were located outside the US during the baseline and follow-up periods, were enrolled in Medicare Advantage (Medicare Part C) during the baseline or follow-up period, or had ≥ 1 hospice claim during the baseline period. Additional selection criteria and study population numbers can be found in Fig. 1.

HRU and healthcare costs of men with mCRPC in the Medicare population were investigated by studying claims data. These included Medicare insurance coverage for institutional (Part A), non-institutional (Part B), and pharmacy (Part D) claims. Institutional claims from hospitals included inpatient hospital admissions, outpatient visits, and skilled nursing facility (SNF) admissions. Non-institutional claims (physician and other visits) included fee-for-service claims submitted by professional providers, including physicians, physician assistants, clinical social workers, and nurse practitioners. Claims for some organizational providers, such as free-standing facilities, are also found in non-institutional claims; examples of these include independent clinical laboratories, ambulance providers, free-standing ambulatory surgical centers, and free-standing radiology centers.

We described the annualized, unadjusted, all-cause, and PC-related medical HRU and healthcare costs for the following study periods: the year before mCRPC diagnosis, after diagnosis until end of follow-up or death, and from the start of first-line (1L), second-line (2L), and third-line (3L) therapy with mCRPC life-prolonging treatments to the start of subsequent therapy or end of follow-up/death. For these study periods, HRU was measured in days per patient per year with a claim (PPPY; emergency department visits, inpatient hospitalization, outpatient visits, physician and other visits, and SNF). Healthcare costs paid by Medicare were inflated to 2019 US dollars (USD) using the Gross Domestic Product Price Index, and included all-cause healthcare costs (medical and pharmacy) PPPY, as well as PC-related medical costs PPPY [19]. PC-related HRU and healthcare costs were identified via claims with a diagnosis code for PC in any position. The start of 1L therapy was defined as the date of the first claim for a life-prolonging treatment of interest (including novel hormonal therapy, chemotherapy, immunotherapy, radiotherapy, and PARP inhibitors) up to 14 days prior to the index date. The 1L regimen included treatments started within 28 days of the start date for 1L treatment. The start of second or later lines of therapy was defined as the start of a new life-prolonging therapy more than 28 days after the previous line of therapy start date or a treatment claim after a 90-day gap in treatment. The end date of each line of therapy was defined as the earliest of the end of follow-up, the day before the start of subsequent therapy, or patient death.

Descriptive statistics (means, standard deviations, and medians for costs) were reported for HRU and cost outcomes over the different periods. No statistical comparisons were conducted for HRU and cost outcomes between the different periods. Data analysis was performed using SAS software v.9.4 (SAS Institute, Cary, NC, USA).

The results of this analysis demonstrate that annualized HRU and healthcare costs (including non-pharmacy costs) increased dramatically after mCRPC diagnosis and continued to rise with progression through multiple lines of life-prolonging therapy. Mean annualized all-cause healthcare costs increased more than four fold from before to after mCRPC diagnosis. In addition, healthcare costs continued to increase with each line of therapy, with mean PPPY costs of $148,325 at 1L, $160,118 at 2L, and $165,186 at 3L. Moreover, mean inpatient PPPY costs were higher in untreated patients ($64,429) relative to those who received 1L ($19,848), 2L ($20,808), and 3L ($27,504) treatment.

In general, the results of this analysis were consistent with previously published analyses [6, 8‐10]. In the overall fee-for-service Medicare population (including treated and untreated patients), annualized HRU (inpatient days per year) increased from 0.4 days to 2.5 days after mCRPC diagnosis and with each subsequent line of therapy (1L, 1.6; 2L, 1.7; 3L, 2.3 days). Emergency department days per year also increased from 1.0 day to 3.3 days after mCRPC diagnosis and with each subsequent line of therapy (1L, 2.4; 2L, 2.6; 3L, 3.3 days). Similarly, in a retrospective claims analysis of commercially insured and Medicare patients from the Truven Health MarketScan database conducted from 2009 to 2015, mean PPPY inpatient admissions increased from 0.2 pre-diagnosis to 1.4 and emergency department visits increased from 0.6 pre-diagnosis to 1.6 following mCRPC diagnosis [6]. In contrast, lower values were reported in another analysis of the Truven Health MarketScan database of commercially insured and Medigap cohorts conducted from 2012 to 2016 [10]; with mean numbers (PPPY) of hospitalization events of 0.8 for the commercially insured cohort and 1.4 for the Medigap cohort, and emergency room visits of 0.5 and 0.6, respectively. While overall trends were similar, we noted higher overall rates of HRU, which possibly reflects differences in the patient populations studied compared with these prior studies.

Surprisingly, our analysis showed that 3252 patients (22% of the overall cohort) did not receive any life-prolonging treatment for mCRPC after mCRPC diagnosis. Patients who did not receive life-prolonging therapy were older and had more comorbidities at baseline compared with patients who received treatment. These patients also had greater all-cause medical HRU (number of days with claims) after mCRPC diagnosis across all subcategories, except outpatient care. Together, these findings suggest that patients who did not receive life-prolonging treatment after mCRPC diagnosis may have had poorer overall health, as evidenced by higher baseline comorbidity and older age, compared with those who received treatment after mCRPC diagnosis. This could explain the higher inpatient PPPY costs among patients who did not receive treatment after mCRPC diagnosis compared with patients who received treatment. Ultimately, it is unclear whether the lack of treatment for these patients reflects a shorter follow-up, refusal of treatment by patients, or treatment not being offered to patients due to comorbidities/poorer health; this warrants further analysis.

The overall four fold increase in Medicare fee-for-service population spending after mCRPC diagnosis that was observed in our study was consistent with similar spending trends (1.9–6.2-fold increase) noted in analyses using the Humana national research database (commercially insured and Medicare Advantage cohorts), and the Truven Health MarketScan database (commercially insured and Medigap cohorts) [9, 10]. In another analysis, mean monthly pharmacy costs increased with each line of therapy in a managed-care population consisting of commercially insured and Medicare patients between January 2008 and March 2018 [8]. However, in our Medicare fee-for-service population analysis, we saw the highest pharmacy costs in 2L, with costs decreasing in 3L ($64,242, $71,511, and $52,431 in 1L, 2L, and 3L, respectively). Pharmacy costs may have decreased in 3L because a greater proportion of patients were treated with chemotherapy in 3L (37%) compared with 1L (17%) and 2L (20%). Chemotherapy is often administered in a physician's office and covered under Medicare Part B. Therefore, chemotherapy costs would not have been captured under pharmacy costs, but under medical costs, and may be reflected in the costs for physician and other visits or outpatient care. Additionally, typical chemotherapy is generic and less costly than novel hormonal therapy. Finally, all-cause PPPY healthcare costs were comparable with those reported in the Truven Health MarketScan database, where mean annualized spending per patient (2015 USD) significantly increased from $35,103 in the pre-index period (nonmetastatic CRPC) to $156,500 after metastasis diagnosis [6].

Medicare is an important insurer of beneficiaries aged ≥ 65 years, and approximately 61% of men with PC are diagnosed in this age group [1, 18]. Therefore, a strength of the study was that the Medicare fee-for-service claims database provided a large analysis population (n = 14,780). In addition, the database tracks payments across a very broad set of medical services. Furthermore, this analysis evaluated annualized HRU and healthcare costs not only before and after mCRPC diagnosis but also with progression through different lines of life-prolonging therapy after mCRPC diagnosis, and among mCRPC patients who did not receive life-prolonging therapy. In advanced prostate cancer, delaying disease progression while maintaining or improving quality of life is an important goal of care. This study provides information on the economic burden associated with disease progression. A limitation of the Medicare claims data is that not all services received by a patient may be present. Even for services covered by Medicare, it is possible that total payments for those services may exceed what is observable in claims data (e.g., a secondary payer covers the difference). As such, our estimates of cost may underestimate the true economic burden of mCRPC.

There are several other limitations of this analysis. Firstly, all estimates presented here are unadjusted and do not account for possible differences in patient characteristics. Secondly, this study was not designed to assess the causal impact of increasing use of or spending on one type of treatment on the use of or spending on another type. Thirdly, our results may not be generalizable outside of the fee-for-service Medicare population. Finally, the claims-based algorithm used to identify mCRPC patients has not been validated using data from medical charts and may have misclassified some mCSPC patients as having mCRPC; we expect this misclassification would lead to attenuation of the HRU and cost burden associated with mCRPC, and thus our results may further underestimate the true impact and cost of mCRPC care.

HRU and healthcare costs (including non-pharmacy costs) increased substantially following mCRPC diagnosis and continued to increase with progression from 1L through 3L mCRPC therapy. Annualized mean healthcare costs increased more than four fold from before mCRPC diagnosis to after mCRPC diagnosis. These findings help to quantify the economic burden of mCRPC and to contextualize the economic value of treatments that delay disease progression.