Computing the Cost of Care per Patient per Day for Patients With Metastatic Neuroendocrine Neoplasms

Abstract

PURPOSE

Patients with well-differentiated, low-grade metastatic neuroendocrine neoplasms (NENs) usually have a long median survival and require complex, expensive care over many years at multidisciplinary centers. The cost burden for patients and institutions serves as a barrier to care. Understanding the drivers of these costs and whether intense monitoring adds value will help to optimize value-based care.

METHODS

We adapted the cost of care per patient per day (CCPD) validated methodology to measure cost while accounting for varying follow-up duration. We queried the Stanford NEN Database, which aggregates data from the electronic health record and other electronic sources, to study patients with metastatic NENs receiving regular care at Stanford. Current Procedural Terminology codes for services incurred during the monitoring period for each patient were mapped to the corresponding cost conversion factor and date in the Medicare fee schedule.

RESULTS

Two hundred two patients between 2010 and 2017 were studied with a mean CCPD of $119.11 in US dollars (USD); NEN-specific systemic therapy made up 55% of this cost. Somatostatin analogs were the costliest systemic therapy. Systemic therapy was the driver of cost differences among patients with various primary tumor types, stage of disease, tumor differentiation and grade, and functional hormone status. Patients in the most expensive CCPD group did not have a significant survival benefit (P = .66).

CONCLUSION

The CCPD methodology was effective in studying cancer care value in NENs. Systemic therapy, specifically somatostatin analogs, was the primary driver of cost, and intense monitoring and higher-cost care did not improve survival outcomes.

High-cost care of patients with metastatic NENs is driven by somatostatin analogs and does not improve survival.

INTRODUCTION

Neuroendocrine neoplasms (NENs) are heterogeneous and are composed of the indolent, well-differentiated neuroendocrine tumors (NETs) and the faster growing, poorly differentiated neuroendocrine carcinomas. The large majority of NENs fall into the well-differentiated NET category and are rare by incidence but high by prevalence, given their relative slow growth.¹ In the United States, the age-adjusted incidence rate of NENs increased seven-fold from the 1970s to 2012, and the prevalence increased eight-fold from 0.006% in 1993 to 0.048% in 2012.² In fact, the prevalence of GI NENs, estimated to be 171,321 cases in 2014,² is higher than the prevalence of other more common GI cancers, including esophageal cancer, gastric adenocarcinoma, and pancreatic adenocarcinoma.³

CONTEXT

• Key Objective

• This study aimed to assess the cost burden and value of care for patients with metastatic neuroendocrine neoplasms (NENs) by using the cost of care per patient per day (CCPD) methodology.

• Knowledge Generated

• The mean CCPD for a cohort of 202 patients with metastatic NENs from 2010 to 2017 was $119.11 USD; systemic therapy accounted for 55% of this cost, and somatostatin analogs were the costliest systemic therapy. Most notably, patients in the highest CCPD group did not experience significant survival benefits (P = .66).

• Relevance

• Our findings underscore the importance of optimizing value-based care for patients with metastatic NENs, as intensive monitoring and higher cost care did not improve survival outcomes. One approach to enhancing the value of care in this population involves careful consideration of limiting the use of somatostatin analogues, especially in the upfront setting for patients with low grade disease, given their prominent role in driving costs.

The prevalence of cancer survivors has increased substantially over time and is anticipated to increase further with advances in treatment and early detection.⁴ This remains true for patients with NENs, for whom median overall survival is 9.3 years, ranging from 3.6 years for NENs in the pancreas to >30 years for NENs in the appendix,² with anticipated increases in survivorship because of the rising incidence, relative indolence, and rapid development of new therapies. Given this trend, the economic burden of NENs both for patients and institutions will continue to rise,⁵ especially since these patients tend to require complex multidisciplinary care at an expert center and remain in the health care system for many years. Despite this, there is a paucity of economic evaluations of NENs.^6,7

Understanding the costs associated with care of patients with NENs is relevant to both patients, for whom cost is often a barrier to care and results in a high burden of financial toxicity,^8,9 and to institutions and policymakers when deciding how to allocate resources.^6,7,10 We believe it is especially important to study patients with metastatic NENs, as they are particularly complex and require a higher burden of health care resources, often over many years. The cost of care per patient per day (CCPD) is a validated methodology that was developed to estimate the value of cancer care and to compare costs across health care systems, disease types, and physician practices in a transparent, reproducible, and standardized manner.¹¹ To that end, we aim to calculate the CCPD for patients with metastatic NENs, understand the drivers of these costs, and evaluate whether intense monitoring and higher-cost care add value.

METHODS

Data Source

We queried the Stanford NEN Database, derived from the Stanford Cancer Institute Research Database, which aggregates structured data from the Stanford electronic health record (EHR) and other electronic sources, including the Stanford Cancer Registry, specialized databases in surgical pathology and radiation oncology, and Stanford Research Repository. International Classification of Diseases (ICD)-O-3 and ICD-10 codes were used to define NENs. The Stanford NEN database includes a variety of patient-specific data, such as patient demographics, primary tumor site, tumor characteristics and pathology, date of metastatic disease diagnosis, date of last follow-up, date of death, vital status, NEN-specific medical therapies, surgeries, and radiation.

To explore services that contributed to cost, we searched the Stanford EHR for Current Procedural Terminology (CPT) codes related to various services and grouped these into 10 bins. The bins included services related to NEN-specific systemic therapy, non-NEN treatment-related medications, surgery, radiation, laboratory services, and imaging. For services related to radiation or surgery, these included any service attributable to radiation or surgery at any time in the clinical course of the patient. Other buckets included nonsurgery or radiation-related procedures, inpatient facility and professional services, and outpatient facility and professional services. CPT codes that could not be included elsewhere were placed in a miscellaneous bin.

Study Design

We designed a retrospective study to assess the cost of care for patients with metastatic NENs. We used the CCPD method to measure cost, as it is an evidence-based method that accounts for variation in follow-up duration and variation in tempo of disease progression.¹¹ This method was also chosen for its ease of replication and generalizability across institutions, as it uses the publicly available Medicare physician fee schedule to normalize cost. Through this method, the total cost of care for each patient was calculated across a monitoring period, defined as the time from metastatic disease diagnosis to one of the following three dates: date of death, most recent date of follow-up, or December 31, 2017, whichever occurred first. December 31, 2017 was used as the cutoff date, since at the time of this study, the Medicare fee schedule was only available through 2017. CPT codes for services incurred during the monitoring period were tallied for each patient and mapped to the corresponding cost conversion factor (in US dollars) and year of service from the Medicare physician fee schedule. The total cost was then divided by the number of days in the monitoring period, resulting in the CCPD for each patient. We summed the CCPD of each patient and divided this by the total number of patients to obtain an overall mean CCPD. We also calculated the individual average CCPD for each type of service received.

Study Population

We first selected patients with NENs from the Stanford database diagnosed with metastatic disease between 2010 and 2017 at Stanford. Only patients at Stanford Health Care (SHC) were studied, as their EHR data and corresponding CPT codes were readily available. We included patients who had two or more oncology-related encounters at SHC (ie, surgery, medical oncology, or radiation) within 3 years of metastatic disease diagnosis to ensure we captured a cohort of patients who received regular follow-up and care at SHC. Similar inclusion criteria were used in a previous CCPD study.¹¹ We used prescription of a NEN-specific systemic therapy at SHC as an additional inclusion criterion to further ensure patients were receiving regular care at SHC.

Statistical Analyses

Descriptive statistics were used to summarize demographic and clinical features of the patients included in the cohort. These are presented as means with standard deviations or medians with IQRs for continuous variables and count and percentages for categorical variables.

To assess whether costs differed by various clinical factors, the average overall CCPD and average CCPD for each type of service were summarized by primary site of tumor, TNM stage, functional hormone status, initial tumor grade, and tumor differentiation. We also examined the length of monitoring duration (<1 year, 1-5 years, and 6-7 years) in relation to costs as well as costs by the specific type of systemic therapy used among patients to better understand if certain medications accounted for higher costs. Systemic therapies were categorized as cytotoxic chemotherapy, including both IV and oral forms such as temozolomide and capecitabine; somatostatin analogs, including octreotide and lanreotide; immunotherapy; and biologics, including everolimus and bevacizumab. Of note, sunitinib and sorafenib did not have associated costs in the physician fee schedule at the time of the study. Kaplan-Meier survival curves were generated to evaluate whether survival varied by primary site of tumor and degree of monitoring costs. Patients were categorized into two groups—high cost (top 20% total cost of care per day) and the remaining cost (the other 80%). The log-rank test was used to compare survival curves.

All analyses were performed using SAS Version 9.4 (SAS Institute Inc, Cary, NC) and R Version 3.5.2 (R Core Team, 2018). All statistical tests were two-sided and evaluated at an alpha level of .05. All research was approved by the Stanford Cancer Center Institutional Review Board.

RESULTS

Patient Characteristics

An initial cohort of 435 patients with metastatic NEN diagnosis was identified from the Stanford Database. Of these patients, only 349 had two or more medical oncology–, surgical oncology–, or radiation oncology–related encounters at SHC within 3 years of diagnosis. Ultimately, 202 patients had been prescribed systemic therapy at SHC and were thus included in the final cohort. The ages of patients ranged between 29 and 91 years, with a median age of 64 years (Table 1). Of the patients, 48.5% were female and 51.5% were male. Sixty-six percent of patients were White, 13% other, 12% Asian, 5% unknown, and 4% Black. The most common primary tumor site was pancreas (37%), followed by nonpancreatic GI (30%), unknown primary site (18%), and lung (11%). Most patients had hormonally nonfunctional tumors (54%), while 40% had hormonally functional tumors. Three quarters of the patients had stage IV disease at diagnosis (77%). Most patients had well-differentiated tumors (72%), whereas 16% had poorly differentiated tumors. Similarly, most patients had grade 1 (31%) or grade 2 (44%) tumors, while only 21% had grade 3 tumors. Seventy-two percent of patients underwent a total monitoring duration of 1-5 years, while 23% were monitored for <1 year and 5% were monitored for 6-7 years.

TABLE 1.

Patient Demographics (N = 202)

Clinical Variable	Total (N =202)
Sex, No. (%)
Female	98 (48.5)
Male	104 (51.5)
Age, years
Mean (SD)	63.8 (12.9)
Median (q1, q3)	64 (54, 72)
Ethnicity, No. (%)
Hispanic/Latino	20 (9.9)
Non-Hispanic/Non-Latino	172 (85.1)
Unknown	10 (5.0)
Race, No. (%)
Asian	24 (11.9)
Black or African American	7 (3.5)
Native Hawaiian or Other Pacific Islander	1 (0.5)
Other	25 (12.4)
White	134 (66.3)
Unknown	11 (5.4)
Functional hormone status, No. (%)
Functional	80 (39.6)
Nonfunctional	110 (54.5)
Unknown	12 (5.9)
Primary site of tumor categories, No. (%)
GI (nonpancreas)	61 (30.2)
Lung	22 (10.9)
Other	7 (3.5)
Pancreas	75 (37.1)
Unknown primary site	37 (18.3)
TNM stage, No. (%)
1	8 (4.0)
2	13 (6.4)
3	20 (9.9)
4	156 (77.2)
Unknown	5 (2.5)
Degree of differentiation, No. (%)
Poorly	33 (16.3)
Well	145 (71.8)
Unknown	24 (11.9)
WHO grade, No. (%)
1	63 (31.2)
2	89 (44.1)
3	43 (21.3)
Unknown	7 (3.5)

Abbreviation: SD, standard deviation.

Cost Results

The average overall CCPD for our cohort of patients with metastatic NENs was $119.11 USD. Systemic treatment made up 55% of this cost. Imaging was 14.5% of the total cost, followed by outpatient evaluation and management (outpatient EM), lab, surgery, radiation, miscellaneous, inpatient evaluation and management (inpatient EM), procedures, and medications (non-NEN treatment–specific; Figs 1A and 2A). Of the systemic therapies, somatostatin analogs were the costliest, with a mean CCPD of $53.26 USD and highest percentage of the total systemic therapy cost (Figs 1B and 2B). Of the total somatostatin analog cost, the cost of lanreotide was 22.6% and the cost of octreotide was 77.4%. This was followed by biologics with a mean cost per day of $33.05 USD and then cytotoxic chemotherapy with a mean cost per day of $5.52 USD (Figs 1B and 2B). Similarly, the mean cost per day of intramuscular (IM) and subcutaneous therapies was $53.07 USD compared with $24.56 USD for intravenous (IV) therapies and $14.21 USD for oral therapies.

FIG 1.

CCPD (A) overall and (B) by class of systemic therapy. Boxplots display median CCPD in US dollars, with IQR shown. CCPD, cost of care per patient per day; Chemo, chemotherapy; inpatient EM, evaluation and management services delivered during an inpatient stay; Lab, laboratory; Outpatient EM, evaluation and management services delivered to outpatients.

FIG 2.

CCPD (A) overall and (B) by class of systemic therapy. Boxplots display median CCPD in US dollars, with IQR shown; y-axis truncated at 100 to show relative contribution of the smaller boxplots. CCPD, cost of care per patient per day; Chemo, chemotherapy; Inpatient EM, evaluation and management services delivered during an inpatient stay; Lab, laboratory; Outpatient EM, evaluation and management services delivered to outpatients.

Patients with pancreatic primary tumors had higher total and systemic treatment costs per day compared with nonpancreatic GI, lung, and other primary tumor types (Fig 3A, Appendix Fig A1 [online only]). Those with initial stage I and stage II tumors that later developed metastatic disease incurred higher systemic treatment and total costs per day compared with patients with initial stage III and IV disease (Fig 3B, Appendix Fig A1). Those with hormonally functional tumors had higher costs per day compared with hormonally nonfunctional tumors, driven by higher systemic therapy costs (Fig 3C, Appendix Fig A1). Lower-grade and well-differentiated tumors had higher mean costs per day compared with higher-grade and poorly differentiated tumors, also because of higher systemic therapy costs (Figs 3D and 3E, Appendix Fig A1). With regards to monitoring duration, those monitored for less than 1 year incurred total higher costs of care per day compared with those monitored for 1-5 years or 6-7 years, driven by higher costs in imaging, inpatient EM, lab, radiation, and surgery. Patients monitored for 1-5 years and 6-7 years, however, had higher mean systemic therapy costs per day compared with patients monitored for <1 year.

FIG 3.

CCPD by (A) primary site of tumor; (B) stage; (C) functional hormone status; (D) tumor grade; and (E) tumor differentiation. Boxplots display median CCPD in US dollars, with IQR shown. CCPD, cost of care per patient per day.

Survival

Patients with GI nonpancreatic primary tumors who develop metastatic disease had the longest survival, followed by patients with primary pancreatic tumors, unknown primary site, and lung tumors; this is a significant difference, with a P value of <.0001 (Fig 4). High-cost patients, categorized as those in the top 20% of cost of care per day, did not have a significant difference in survival compared with low-cost patients in the bottom 80% of cost of care per day (Fig 4).

FIG 4.

Survival curves of patients by (A) primary site of tumor and (B) high versus low cost; high-cost patients are those in the top 20^th percentile of CCPD; low-cost patients are those at or below the 80^th percentile of CCPD. The vertical axis displays the percent of surviving patients, and the horizontal axis is the time in days after metastasis. The number of patients at risk at the beginning of each time point is displayed in the bottom table. CCPD, cost of care per patient per day.

DISCUSSION

The costs of cancer care have risen significantly since 2010¹² and are anticipated to rise, driven by the increasing prevalence of cancer overall, the introduction of costly new therapies, and adoption of more expensive diagnostic tests.¹³ As a result, there has been an increasing effort to assess and optimize value in cancer care,^13-15 and quantifying cost is a key component of the value assessment.^13,14 This is especially true of NENs, whose rising prevalence is due to their relative indolent growth and the advent of new and costly NEN-specific diagnostics and therapies such as Lu 177 dotatate peptide receptor radionuclide therapy (PRRT).⁶ Moreover, previous studies have shown that patients with NENs have high 5-year costs and continuing phase of care costs,^16,17 largely driven by the longer survival of patients with NENs compared with other cancers.

Although there is literature aimed at assessing the cost-effectiveness of various NEN-specific therapeutics^18-23 and the economic burden of NENs to patients and families,^9,10,17,24 the overall body of evidence is sparse.^5-7 This study demonstrates a novel and effective approach to assessing the cost of care for metastatic NENs at an academic center via the CCPD methodology. This methodology was initially applied to a cohort of breast cancer, and this study serves as an example of the ease of replicating the CCPD model. Benefits of this methodology include using the easily accessible public Medicare fee schedule data as a proxy for cost and lack of reliance upon charges or amounts paid, which are often arbitrary and proprietary. As a result, this methodology is reproducible across various tumor types at the same institution as well as across various institutions.¹¹ For instance, CCPD for patients with stage 0-III breast cancer during breast cancer relapse at SHC was $13.80 USD¹¹ compared with CCPD of $119.11 USD for patients with metastatic NENs. The higher CCPD for patients with metastatic NENs in this study is driven by higher systemic treatment costs.

The CCPD for patients with metastatic NENs at Stanford is driven by systemic therapy, primarily somatostatin analogs. This is true of the CCPD for the entire cohort but also within all subgroups. For instance, higher costs among patients with pancreatic primaries compared with other primary tumor subtypes is driven by higher systemic therapy costs, which is expected, given the more aggressive nature of pancreatic NENs that merits earlier treatment. Similarly, patients with initial stage I/II disease have higher systemic therapy costs likely because of longer disease course, and thus, duration of treatment with somatostatin analogs. The same can be said of patients with lower-grade and well-differentiated disease. These findings support those from previous studies, both from the United States²⁵ and other countries,^6,24 including a Canadian study comparing cost patterns in patients with NENs to those with colon cancer, concluding that drug-related costs were the predominant drivers of cost in the late phase of care of patients with NENs.²⁶

CCPD, unlike total cost, also accounts for varying durations of monitoring and facilitates the analysis of the intensity of services in relation to their outcome. In our study, we did not detect a survival benefit in patients with metastatic NENs who had more expensive care in the monitoring period.

The large contribution of somatostatin analogs to the CCPD for patients with metastatic NENs as well as lack of survival benefit seen in patients with high-cost care raises the question of the cost-effectiveness of somatostatin analogs, especially when used in the upfront setting for patients with relatively low-grade disease. A study by Barnes et al investigated the cost-effectiveness of upfront lanreotide use in patients with grade 1-2, nonfunctioning gastroenteropancreatic NENs, which showed that upfront lanreotide results in similar quality of life compared with active surveillance followed by lanreotide upon progression.²¹ At current prices, lanreotide was not deemed to be cost-effective.²¹ Our findings and those of the study by Barnes et al²¹ support the need for further research to elucidate the cost-effectiveness of somatostatin analogs.

Our study has several limitations. Our analysis included only patients with metastatic NENs who were receiving systemic therapies at Stanford. As a result, patients with low-grade, indolent metastatic disease who were not receiving systemic therapy were excluded from the study population, thus overestimating cost, including cost contributed by somatostatin analogs. We only captured care delivered at SHC and we estimate one third of unplanned cancer care is not captured in our EHR,²⁷ resulting in an underestimation cost. The potential underestimate is likely to be greater for those with low-grade, indolent disease that may seek more regular care closer to home. Moreover, at the time of the study, PRRT had not yet been FDA approved and did not have an associated CPT code, and thus, its costs were not captured by our study. Only 40 patients were included in the high-cost group of our survival analysis, and this small sample size is a limitation of our analysis. A limitation of the CCPD methodology is reliance upon the Medicare fee schedule, which may not reflect actual costs or charges or reflect inflationary effects despite being updated yearly.¹¹ We did not include sunitinib costs in the study because the Medicare fee schedule does not include costs for sunitinib, and there is a lack of reliable historical cost data. However, only six patients in our study received sunitinib. An exploratory analysis demonstrated that at 2021 wholesale acquisition costs, which likely overestimates the cost of sunitinib to Medicare, the inclusion of sunitinib costs would only add $3.53 USD to the overall mean CCPD. Biologics would have a mean CCPD of $36.58 USD (as opposed to $33.05 USD) and would be 38% of the systemic treatment cost (as opposed to 36%). Finally, our study did not include financial toxicity to patients. This is a critically important area of active research in oncology and should also be examined in chronic cancers such as NENs.

In summary, the CCPD methodology is an effective means of studying cancer care value in NENs. The findings from our study highlight the importance of optimizing value-based care for patients with metastatic NENs, as intense monitoring and higher-cost care do not necessarily improve survival outcomes. We identified systemic therapy, specifically somatostatin analogs, as the primary driver of high cost in this population of patients, in line with findings from previous studies from other countries.^24-26 Our study adds to the growing body of literature^21,24,26,28 that supports careful consideration of limiting the upfront use of somatostatin analogs, particularly in patients with low-grade disease, to optimize high value care. Future studies using this methodology should incorporate the cost of PRRT to reflect the current state of metastatic NEN treatment, study the impact of insurance type on the CCPD and survival, and aim to capture costs at other institutions or health care systems for comparison to identify best practices for delivering high-value care.

APPENDIX

FIG A1.

CCPD by (A) primary site of tumor; (B) stage; (C) functional hormone status; (D) tumor grade; and (E) tumor differentiation. Boxplots display median CCPD in US dollars, with IQR shown. CCPD, cost of care per patient per day; Inpatient EM, evaluation and management services delivered during an inpatient stay; Lab, laboratory; Outpatient EM, evaluation and management services delivered to outpatient.

AUTHOR CONTRIBUTIONS

Conception and design: Divya M. Gupta, James Barnes, Pamela L. Kunz

Financial support: Pamela L. Kunz

Administrative support: Pamela L. Kunz

Collection and assembly of data: Divya M. Gupta, Kathleen Hornbacker

Data analysis and interpretation: Divya M. Gupta, James Barnes, FeiFei Qin, Kristopher Kapphahn, Douglas W. Blayney

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Computing the Cost of Care per Patient per Day for Patients with Metastatic Neuroendocrine Neoplasms

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/op/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).