Abstract
BACKGROUND:
There are currently many approved agents for the treatment of metastatic melanoma (MM), the most aggressive form of skin cancer. Treatments may include systemic therapies such as ipilimumab, dacarbazine, temozolomide, high-dose interleukin 2, interferon α, dacarbazine- or temozolomide-based combination chemotherapy/biochemotherapy, paclitaxel, paclitaxel/cisplatin, and paclitaxel/carboplatin, as well as the targeted therapies vemurafenib, dabrafenib, and trametinib for patients with BRAF V600 mutation. However, all treatment options are associated with different adverse events (AEs) and, in some instances, considerable toxicity. The occurrence of such treatment-related AEs can lead to higher health care resource utilization and increasing treatment and patient management costs. An understanding of the economic burden of these AEs will therefore enable better management of health care expenditures, not just for existing therapies, but also for new and novel treatments in development.
OBJECTIVE:
To estimate the incremental health care costs of specific AEs among patients with MM treated with paclitaxel, vemurafenib, ipilimumab, dacarbazine, temozolomide, high-dose interleukin 2, or interferon α, along with AEs known to be associated with dabrafenib and trametinib.
METHODS:
This cohort study employed a retrospective administrative claims-based analysis of MarketScan commercial and Medicare supplemental databases from July 1, 2004, to April 30, 2012. Patients included those aged ≥ 18 years who had diagnosed melanoma (ICD-9-CM code 172.xx)with ≥ 1 diagnosis of metastasis and ≥ 1 claim for any of the 7 study treatments. Health care encounters for AEs of interest were based on ICD-9-CM diagnosis/procedure codes. Incremental cost per AE was determined by comparing the 30-day expenditures in patients with the event to patients without the event based on a shadow event date. Multivariate generalized linear models (GLMs) with a log-link function and gamma distribution were utilized to control for baseline differences between groups.
RESULTS:
A total of 2,621 patients with MM were included. Mean age was 56.0 years (SD ± 13.0); 64% were male; and 24% had a diagnosis of primary or secondary brain cancer at the time of MM diagnosis. GLM-based estimate of 30-day incremental costs by AE category were metabolic, $9,135 (95% CI = $6,404-$12,392); hematologic/lymphatic, $8,450 (95% CI = $6,528-$10,633); cardiovascular, $6,476 (95% CI = $4,667-$8,541); gastrointestinal, $6,338 (95% CI = $4,740-$8,122); skin/subcutaneous, –$900 (95% CI = –$1,899-$237); central nervous system/psychiatric, $5,903 (95% CI = $3,842-$8,313); and pain, $5,078 (95% CI = $3,392-$7,012).
CONCLUSIONS:
Incremental costs associated with many MM treatment-related AEs are substantial. New approaches to prevent and/or better manage these events may reduce overall health care costs.
What is already known about this subject
Metastatic melanoma is the most aggressive form of skin cancer, with a median overall survival of less than 1 year.
Despite the wide armamentarium of systemic and targeted therapies available for the treatment of metastatic melanoma, all are associated with different adverse event profiles and toxicities, including skin-related toxic effects, fatigue, anemia, peripheral neuropathy, nausea, elevated serum glutamate oxaloacetic transaminase levels, vomiting, and stomatitis, among others.
As a consequence of these treatment-associated adverse events, higher utilization of health care resources and expenditures may be incurred, and understanding the full economic burden associated with different treatments will help inform pharmacoeconomic models and treatment decisions.
What this study adds
This retrospective cohort study is the first to estimate the health care costs of specific treatment-related adverse events among patients with metastatic melanoma receiving commonly used therapies.
The main findings suggest that the incremental costs associated with the specific treatment-related adverse events (with the exception of skin and subcutaneous tissue events) were substantially high, with the 2 most expensive categories being metabolic and nutritional disorders and hematologic and lymphatic disorders and effects.
Data from this pharmacoeconomic analysis suggest that prevention of these adverse events may be an important factor in decreasing the financial burden of existing and new therapies for metastatic melanoma.
Metastatic melanoma (MM) is the most aggressive form of skin cancer, with a median overall survival of less than 1 year.1 The National Comprehensive Cancer Network guidelines include the following systemic therapies as options for patients with advanced or MM: ipilimumab, vemurafenib, dabrafenib, trametinib, dabrafenib + trametinib, dacarbazine, temozolomide, high-dose interleukin 2 (IL-2), dacarbazine- or temozolomide-based combination chemotherapy/biochemotherapy (including cisplatin and vinblastine with or without high-dose IL-2, interferon α), paclitaxel, paclitaxel/cisplatin, and paclitaxel/carboplatin.2 Vemurafenib, dabrafenib, and trametinib are recently approved targeted agents indicated for patients with mutations of the BRAF V600 gene.3-5 All melanoma agents are associated with different adverse event (AE) profiles. These AEs may include skin-related toxic effects, pyrexia, hypertension, fatigue, anemia, peripheral neuropathy, nausea, elevated serum glutamate oxaloacetic transaminase levels, vomiting, and stomatitis, among others.3-11 AEs may lead to higher utilization of health care resources and expenditures (physician, emergency department, or hospitalization) and, in some extreme cases, may lead to significant disability or mortality.12 As new therapies continue to be studied and become available for the treatment of MM, it is important to fully understand the economic burden associated with AEs of existing and new therapies. A more complete understanding of these AEs will better inform economic models to more accurately characterize the overall budgetary impact of new therapies on patients and payers.
Thus, the objective of this study was to estimate the incremental health care costs of specific AEs among patients with MM treated with paclitaxel, vemurafenib, ipilimumab, dacarbazine, temozolomide, high dose IL-2, or interferon α and those known to be associated with dabrafenib and trametinib.
Methods
Study Design and Data Source
This retrospective cohort study used health care administrative claims from the MarketScan Commercial Claims and Encounters (Commercial) Database, MarketScan Medicare Supplemental and Coordination of Benefits (COB; Medicare) Database, and MarketScan Early View Database for July 1, 2004, to April 30, 2012. The Commercial Database reflects the health care experiences of the covered populations (individuals aged ≤ 64 years) of more than 100 employer and health plan clients throughout the United States. The Medicare Database contains the health care experiences of approximately 4 million retirees annually with Medicare supplemental insurance and includes the COB and the employer-paid portion of health care costs. Patients who changed from commercial to Medicare primary insurance during the study period (e.g., by turning age 65) were tracked through both databases and classified according to the first database (Commercial).
All database records were de-identified and fully compliant with U.S. patient confidentiality requirements (Health Insurance Portability and Accountability Act); therefore, the study was exempt from institutional review board approval.
Study Population
The study population was composed of patients with MM treated with 1 of 7 drug therapies (paclitaxel, vemurafenib, ipilimumab, dacarbazine, temozolomide, high dose IL-2, or interferon α monotherapy) between January 1, 2005, and April 30, 2012. The newer drug therapies, dabrafenib and trametinib, were not included in this analysis because their recent approval (May 2013) precluded their availability in the databases under study; however, AEs potentially associated with dabrafenib and trametinib were included in this study, specifically fever and hypertension, respectively. Patients were assigned to 1 of 7 mutually exclusive treatment cohorts based on their study drugs. Each patient entered the study once. For inclusion in the final study population, the following criteria were required:
One or more nondiagnostic health care claim with a diagnosis of MM (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] code 172.xx—malignant melanoma of skin) between July 1, 2004, and April 30, 2012.
A diagnosis of metastasis within 30 days before or 60 days after any MM diagnosis on nondiagnostic health care claims during the same period.
One or more prescription claims or outpatient service claims with a drug code (National Drug Code [NDC] number, Healthcare Common Procedure Coding System [HCPCS], or ICD-9-CM procedure code) for paclitaxel, vemurafenib, ipilimumab, dacarbazine, temozolomide, or interferon α, or 1 inpatient service claim for high dose IL-2 (ICD-9-CM procedure 00.15) on or within 365 days after the first diagnosis of MM between January 1, 2005, and April 30, 2012.
Additional inclusion criteria were (a) aged ≥ 18 years on the index date and (b) continuous enrollment with medical and pharmacy benefits during the 6 months prior to the index date.
Patients were excluded from the final study sample if (a) they had a diagnosis of nonmelanoma primary malignancy in the 6 months preceding the index date; (b) they were pregnant at any time during the entire study period; and (c) they initiated treatment on any combination regimen that included ≥ 1 of the study drugs, with the exception of combinations involving interferon α. Combinations of study drugs and other chemotherapies (e.g., paclitaxel and cisplatin) were allowed.
To maximize the sample size of the newest drugs, patients were assigned to mutually exclusive treatment groups based on the first observed claim of a study drug using the following hierarchical order: vemurafenib, ipilimumab, dacarbazine, temozolomide, high dose IL-2, paclitaxel, and interferon α (monotherapy).
Study Period
The study period comprised a pre-index period, index date, and follow-up period (Figure 1A). The date of the patient’s first prescription/infusion of the cohort’s drug after the MM diagnosis date from January 1, 2005, to April 30, 2012, was set as the index date. The 6-month period prior to the index date was designated as the baseline (pre-index) period, during which demographic and clinical characteristics were assessed. The variable length follow-up period began on the index date and continued to the end of the index treatment. Measures of health care costs and treatment characteristics were evaluated during the follow-up period.
FIGURE 1.
Study Period and Final Study Population
For the AE cost analyses, the 30 days following the AE served as the time period over which health care costs were evaluated (Figure 1A). The date of the first specific AE claim served as the beginning of the 30-day period. For patients without the specific AEs, a shadow AE date was assigned by randomly sampling from the distribution of number of days from index to event for patients with the AE and then adding that number of days to the control’s index date. This ensured that controls were on the study drug of interest for a comparable amount of time as patients who developed an AE. For example, if a patient developed an AE of interest 16 days after initiating Study Drug A and was followed for an additional 30 days to ascertain costs, then the control patient was also followed for 46 days after initiating Study Drug A with cost analysis beginning after day 16.
The follow-up period (covering the study treatment episode) began on the index date for the therapy of interest and ended with the earlier of (a) a gap in therapy of ≥ 45 days, (b) switch to a new therapy, (c) inpatient mortality, (d) end of patient’s insurance eligibility, or (e) end of study period (April 30, 2012).
Outcome Variables
Adverse Events.
Treatment-related AEs included those that have been associated with the 7 study drugs, as well as dabrafenib and trametinib. The selected AEs were identified from a review of each therapy’s package insert and in consultation with one of the co-authors, who is a clinical expert. The criteria for selection of specific AEs were those occurring in ≥ 20% of patients for any grade event or occurring in ≥ 5% of patients for grades 3 and 4. Incident AEs were identified by following patients after their index dates on 1 of the 7 study drugs. Conditions were identified by a primary or secondary diagnosis on any nondiagnostic inpatient or outpatient claim. AEs were grouped into 1 of 8 categories: (1) cardiovascular, (2) central nervous system (CNS) and psychiatric, (3) gastrointestinal, (4) hematologic and lymphatic, (5) metabolic and nutritional, (6) pain, (7) skin and subcutaneous tissue, and (8) others. A detailed description of AE categories is provided in Table 1. Because the study objective was to ascertain mean 30-day event-level AE costs, it was not necessary to require that individuals be AE-naïve prior to inclusion into the study. Patients, regardless of AE history prior to drug therapy, were included, and AEs that occurred subsequent to the index date were included in the costing analysis.
TABLE 1.
Categories of Adverse Events
| Category | Adverse Event |
|---|---|
| Cardiovascular | Secondary hypertension, hypertension complications, hypotension, tachycardia (including supraventricular) |
| CNS and psychiatric | Anxiety/depression, confusion, convulsions, hemiparesis, somnolence |
| Gastrointestinal | Abdominal pain, colitis, constipation, diarrhea, mucositis/stomatitis, nausea/vomiting |
| Hematologic and lymphatic | Anemia, leukopenia, neutropenia, pulmonary embolism, thrombocytopenia |
| Metabolic and nutritional | Acute renal failure; abnormal renal or liver function test; bilirubinemia; elevation of transaminase, lactate dehydrogenase, phosphatase, amylase, or lipase; hyponatremia; peripheral edema |
| Pain | Headache, myalgia/arthralgia/musculoskeletal/back/other pain, peripheral neuropathy |
| Skin and subcutaneous tissue | Alopecia, diaphoresis (sweating), hyperkeratosis, benign neoplasms of the skin (including papilloma), photosensitivity reaction, pruritus (itching), rash, other malignant neoplasms of the skin (squamous-cell carcinoma identified separately after October 1, 2011) |
| Other | Anaphylaxis, anuria/oliguria, asthenia/fatigue, fever and/or chills, decreased appetite/anorexia, infections (including folliculitis); some AEs were combined after further review of the ICD-9-CM codes that map to each AE |
CNS = central nervous system; ICD-9-CM = International Classification of Diseases, Ninth Revision, Clinical Modification.
Incremental AE Costs.
Incremental AE costs were defined as the difference in 30-day costs (excluding costs for study drugs and other cancer therapies) for patients with the specific AE (“cases”) and patients without the AE (“controls”). For the evaluation of each AE, all patients in the same drug cohort who did not have the event were selected as controls. The costs of each AE were evaluated across all drug cohorts.
Health care costs were the total adjudicated amount paid to all providers for inpatient and outpatient services and drugs, with the exception of study drugs and other cancer therapies. Expenditures included those made by the insurer, patient (deductible, copayment, coinsurance), and any coordination of benefits as indicated on the claim. All costs were standardized to January 2012 dollars using the medical component of the Consumer Price Index.
Baseline Characteristics.
Demographic, insurance, and physician characteristics were assessed using the index drug claims. Baseline comorbidities and other clinical characteristics were assessed during the 6-month pre-index period and included the Charlson Comorbidity Index (CCI), National Cancer Comorbidity Index (NCCI), and indicator variables for baseline cancer therapies and other comorbidities of interest.
Statistical Analysis
Descriptive and multivariate analyses were conducted for this study. Unadjusted incremental AE health care costs across all patients are reported as the mean and 95% confidence interval (CI). The adjusted incremental costs of each AE were computed by estimating a multivariate regression on costs during the 30 days following the AE. Generalized linear models with a gamma-distributed error and log-link were used. Age, sex, primary payer (Medicare or commercial), insurance plan type, insurance capitation status, and NCCI were used as independent variables in the multivariate regression models. The incremental cost was estimated by the method of recycled predictions.9-11 In this method, the regression model was used to calculate a predicted 30-day cost for every patient based on the covariate values assuming the patient experienced an AE (case) and again assuming the patient did not (control); then taking the difference at the patient level, which was the estimated incremental cost; and finally averaging the incremental cost across patients.
Results
Study Sample
A total of 2,621 patients with MM initiated 1 of the 7 study medications and met all study criteria. Figure 1B depicts each of the study sample selection steps. After the hierarchy was applied for cohort selection, 4.5% (n = 119) initiated with vemurafenib, 5.8% (n = 152) with ipilimumab, 9.7% (n = 254) with dacarbazine, 32.3% (n = 847) with temozolomide, 8.7% (n = 227) with high dose IL-2, 5.8% (n = 153) with paclitaxel, and 33.2% (n = 869) with interferon α.
Patient Demographic and Clinical Characteristics
Table 2 presents the demographic characteristics of patients with MM who met the study selection criteria. Of the 2,621 patients, the mean age (± standard deviation ([SD]) was 56.0 years (± 13.0), and 64% were male. The majority of study patients in all treatment groups lived in the North Central (28.7%) or South (36.9%) U.S. Census regions at the time of the study and carried a point-of-service insurance plan type. Nearly 90% of patients had noncapitated health plans, and approximately one-quarter of study patients were covered by Medicare. Patients in cohorts for the newer drugs vemurafenib and ipilimumab had index dates in 2011 and 2012. The remaining 5 treatment cohorts had similar index year distributions concentrated in 2005-2010, with large declines in 2011 and 2012 partly due to the selection strategy, which gave priority to the newer drugs. The mean (± SD) NCCI and CCI scores during the pre-index period were 0.32 (± 0.65) and 7.89 (± 1.85), respectively (Table 3). Of the 2,621 total patients, 24% had brain cancer in the pre-index period; 34% had claims for excision surgery in the pre-index period; and approximately 43% and 34% had pre-index period hospitalization and emergency room visits, respectively. In the pre-index period, diabetes (11.7%), cardiovascular disease (9.3%), and cerebrovascular disease (8.8%) were the most common comorbid conditions.
TABLE 2.
Demographic Characteristics (Measured on Index Date, N = 2,621)
| Age, mean ± SD, years | 56.0 ± 13.0 |
| Age group, years | |
| 18-24 | 31 (1.2) |
| 25-34 | 108 (4.1) |
| 35-44 | 333 (12.7) |
| 45-54 | 686 (26.2) |
| 55-64+ | 876 (33.4) |
| 65-74 | 352 (13.4) |
| 75-84 | 205 (7.8) |
| 85+ | 30 (1.1) |
| Sex | |
| Male | 1,677 (64.0) |
| Female | 944 (36.0) |
| Insurance plan type | |
| Comprehensive | 321 (12.2) |
| EPO or PPO | 30 (1.1) |
| HMO | 599 (22.9) |
| Point-of-service | 1,463 (55.8) |
| CDHP or HDHP | 127 (4.8) |
| Unknown | 81 (3.1) |
| U.S. Census geographic region | |
| Northeast | 332 (12.7) |
| North Central | 752 (28.7) |
| South | 967 (36.9) |
| West | 546 (20.8) |
| Unknown | 24 (0.9) |
| Urbanicity | |
| Urban | 2,150 (82.0) |
| Rural | 447 (17.1) |
| Unknown | 24 (0.9) |
| Specialty of physician who prescribed index drug | |
| Oncologist/hematologist | 944 (36.0) |
| Dermatologist | 2 (0.1) |
| General surgeon | 0 |
| Primary care provider | 229 (8.7) |
| Radiologist/nuclear medicine | 85 (3.2) |
| Other specialist | 283 (10.8) |
| Other provider | 57 (2.2) |
| Facility | 696 (26.6) |
| Unknown | 325 (12.4) |
| Place of index treatment | |
| Inpatient hospital | 227 (8.7) |
| Outpatient hospital | 316 (12.1) |
| Independent clinic | 3 (0.1) |
| Physician office | 951 (36.3) |
| Other | 1,124 (42.9) |
| Payer | |
| Commercial | 2,013 (76.8) |
| Medicare | 608 (23.2) |
| Year of index treatment | |
| 2005 | 263 (10.0) |
| 2006 | 244 (9.3) |
| 2007 | 279 (10.6) |
| 2008 | 372 (14.2) |
| 2009 | 471 (18.0) |
| 2010 | 429 (16.4) |
| 2011 | 405 (15.5) |
| 2012 | 158 (6.0) |
| Capitation status | |
| Capitated | 287 (11.0) |
| Noncapitated | 2,334 (89.0) |
Note: Data are presented as n (%) unless otherwise indicated.
CDHP = consumer-driven health plan; EPO = exclusive provider organization; HDHP = high-deductible health plan; HMO = health maintenance organization; PPO = preferred provider organization; SD = standard deviation.
TABLE 3.
Clinical Characteristics (Measured During 6-Month Pre-Index Period)
| NCCI, mean ± SD | 0.32 ± 0.65 |
| Minimum | 0 |
| Median | 0 |
| Maximum | 7 |
| CCI, mean ± SD | 7.89 ± 1.85 |
| Minimum | 0 |
| Median | 8 |
| Maximum | 14 |
| Baseline cancer therapy | |
| Excision surgery | 887 (33.8) |
| Brain cancer (primary or secondary) | 630 (24.0) |
| Chemotherapy or biological therapy-study drug | 291 (11.1) |
| Chemotherapy or biological therapy-nonstudy drug | 123 (4.7) |
| Interferon α | 111 (4.2) |
| Pre-index hospitalization | 1,117 (42.6) |
| Pre-index emergency room visit | 889 (33.9) |
| Comorbid condition | |
| Anxiety | 182 (6.9) |
| Cardiovascular disease | 243 (9.3) |
| Cerebrovascular disease | 231 (8.8) |
| COPD | 97 (3.7) |
| Diabetes | 307 (11.7) |
| Depression | 165 (6.3) |
| Adverse event category | |
| Cardiovascular | 936 (35.7) |
| CNS and psychiatric disorders | 406 (15.5) |
| Gastrointestinal | 543 (20.7) |
| Hematologic and lymphatic disorders and effects | 393 (15.0) |
| Metabolic and nutritional disorders | 257 (9.8) |
| Pain | 788 (30.1) |
| Skin and subcutaneous tissue | 1,035 (39.5) |
| Other | 503 (19.2) |
Note: Data are presented as n (%) unless otherwise indicated.
CCI = Charlson Comorbidity Index; CNS = central nervous system; COPD = chronic obstructive pulmonary disease; NCCI = National Cancer Comorbidity Index; SD = standard deviation.
Cost of Adverse Events
In descending order, the adjusted incremental AE costs were significantly greater for metabolic and nutritional disorders, hematologic and lymphatic disorders, cardiovascular, gastrointestinal, CNS, and psychiatric disorders, and pain compared with individuals without the AE (Table 4). Incremental costs associated with skin and subcutaneous tissue AEs were not significantly different between individuals with and without AEs. Detailed descriptions of costs of AEs are presented in Table 4.
TABLE 4.
Incremental Costs Among Individuals With and Without Adverse Events
| Adverse Event Category | With AE | Without AE | Adjusted Mean a | Incremental Costs b | 95% CI for Adjusted Incremental Costs | |||||
|---|---|---|---|---|---|---|---|---|---|---|
| n | Mean ($) | n | Mean ($) | With AE ($) | Without AE ($) | Observed ($) | Adjusted ($) | Lower 95% CI ($) | Upper 95% CI ($) | |
| Cardiovascular | 516 | 14,693 | 1,404 | 8,077 | 14,614 | 8,137 | 6,615 | 6,476 | 4,667 | 8,541 |
| CNS and psychiatric disorders | 315 | 14,110 | 1,571 | 8,215 | 14,207 | 8,305 | 5,895 | 5,903 | 3,842 | 8,313 |
| Gastrointestinal | 603 | 15,249 | 1,461 | 8,890 | 15,312 | 8,975 | 6,359 | 6,338 | 4,740 | 8,122 |
| Hematologic and lymphatic disorders and effects | 488 | 16,309 | 1,410 | 7,602 | 16,165 | 7,715 | 8,707 | 8,450 | 6,528 | 10,633 |
| Metabolic and nutritional disorders | 238 | 16,864 | 1,567 | 7,712 | 16,936 | 7,800 | 9,151 | 9,135 | 6,404 | 12,392 |
| Pain | 418 | 13,312 | 1,358 | 7,961 | 13,230 | 8,151 | 5,351 | 5,078 | 3,392 | 7,012 |
| Skin and subcutaneous tissue | 513 | 8,255 | 1,268 | 8,878 | 8,180 | 9,080 | -624 | -900 | -1,899 | 237 |
a Models adjusted for age, sex, primary payer (Medicare or commercial), insurance plan type, insurance capitation status, and NCCI.
b Incremental costs are the difference in health care expenditures (excluding cancer drugs) between patients with the specific AEs (cases) and patients without AEs (controls).
AE = adverse event; CI = confidence interval; CNS = central nervous system; NCCI = National Cancer Comorbidity Index.
Discussion
Current treatment options for MM may be associated with considerable toxicity, which can lead to higher health care resource and associated expenditures. This retrospective cohort study is the first to estimate the health care costs of specific treatment-related AEs among patients who received commonly used therapies for MM. It is important to fully understand the economic burden of treatment-related AEs to better manage health care expenditures for existing therapies, as well as therapies in development.
The main findings from this study suggested that the incremental costs associated with the specific treatment-related AEs (with the exception of skin and subcutaneous tissue AEs) were substantially high. The 2 most expensive AE categories were (1) metabolic and nutritional disorders, and (2) hematologic and lymphatic disorders and effects. In addition, the adjusted mean costs for the AE categories analyzed, except for skin and subcutaneous tissue, among patients with AEs were at least 1.6 times higher compared with patients without AEs.
To our knowledge, no other studies have evaluated the incremental cost of specific treatment-related AEs in melanoma. However, AE costing studies have been carried out in other cancers and geographies.13-15 Two of these studies compared the economic burden between 2 anticancer agents in renal cell carcinoma13 and non-small cell lung cancer.14 Direct comparability of these results to the current analysis is difficult due to the costing approaches employed (total costs vs. mean event costs) and AE types included. These studies, however, confirmed that the economic impact of AEs should be considered when the appropriate treatment is selected for a patient.
Limitations
The results from this study should be interpreted with caution. Given the small sample size for many of the AEs, this study had limited ability to detect incremental costs. The incremental cost method estimates the cost of an AE using the difference in cost for patients with and without a particular AE and could lead to negative incremental cost estimates. There may also be unmeasured confounding factors that influenced the difference in cost between those with and without AEs. Causality of AEs to specific drug therapies in this study cannot be attributed because AEs could have been manifestations of chronic conditions or treatments not included in this study. Because the study drew upon a sample of AEs to determine their costs rather than incidence, patients with evidence of these events prior to drug therapy were included.
Moreover, the diagnosis code for MM (ICD-9-CM 172.xx) does not specify whether metastases are present. ICD-9-CM diagnosis codes are available for use for metastases, but they are not always coded on health care claims. A sensitivity analysis tested the effect on the sample size of requiring metastases diagnosis codes as an element of sample selection. When metastases codes were added as a criterion to patient selection, the size of the overall sample decreased by approximately 700 (19%). Ultimately, the requirement of a metastatic diagnosis code within 30 days prior to or 60 days after any MM diagnosis was included as part of patient selection.
In claims data, ICD-9-CM codes were used to identify AEs. These codes were not the same as those used in clinical trials; thus, the specific events that were evaluated may not directly correspond to those in trials. Furthermore, due to the specificity of ICD-9-CM codes, AEs may have been undercoded or incorrectly coded in administrative claims. In addition, AEs of mild severity may not have been reported by a provider on a medical claim. Therefore, the study may underestimate the complete cost of AEs. Findings in this study may not be representative of the whole U.S. MM population because uninsured patients and those covered by Medicaid or military-based insurance were not included in this study.
Due to the limitation of small sample sizes at the treatment group level, it was not possible to estimate specific AE costs by type of treatment. The results reported in this analysis rely on the assumption that the mean cost of specific grade 3/4 AEs are comparable regardless of the treatment that precipitated them. This remains an assumption and warrants further evaluation.
Lastly, rather than including only treatment-naïve patients (e.g., first-line patients only), the study assigned patients to treatment groups of interest by utilizing a hierarchical ordering approach in order to ensure adequate representation of the newer agents (e.g., vemurafenib and ipilimumab). As a result, approximately 15% of the patients’ index events represented second or later lines of therapy. If the risk for AE development and/or severity were to increase with later lines of therapy, there is a possibility that the mean cost of AEs could be overestimated. Given that the majority of included patients were on a first-line therapy of interest, this risk is likely minimized. Additionally, it is not a certainty that switching from one treatment to another increases risk of AE occurrence or severity. Given that different therapies often have different toxicity profiles, it is equally as likely for a patient to experience new and different AEs.
Conclusions
These analyses showed that the incremental costs associated with specific AEs of the most commonly used therapies to treat patients with MM can be substantial. Skin and subcutaneous AEs, however, were not a major driver of AE costs. Prevention of these AEs may be important for decreasing the financial burden of existing therapies and new approaches for the treatment of patients with MM.
ACKNOWLEDGMENTS
Editorial support was provided by Clinical Thinking and was funded by GlaxoSmithKline. Programming support was provided by Robert Sedgley.
Funding Statement
Funding for this study was provided by GlaxoSmithKline (HO-12-928). Arondekar, Monberg, Mirakhur, and Oglesby are employees of GlaxoSmithKline (GSK). Arondekar, Monberg, Mirakhur, and Oglesby own stock in GSK. Meyer is employed by Truven Health, who had a contract with GSK for conducting this study, as was Curkendall, at the time of this study. Oglesby attended the American Society of Clinical Oncology (ASCO) annual meeting in 2013. Lenhart has nothing to disclose.
Arondekar, Monberg, and Mirakhur obtained funding, and Arondekar and Monberg provided administrative support. Analysis and interpretation of data and final approval of the submitted manuscript were contributed by all authors.
Editorial support was provided by Clinical Thinking and was funded by GlaxoSmithKline. Programming support was provided by Robert Sedgley.
REFERENCES
- 1.Finn L, Markovic SN, Joseph RW. Therapy for metastatic melanoma: the past, present, and future. BMC Med. 2012;10:23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.National Comprehensive Cancer Network. NCCN clinical practice guidelines in oncology. Melanoma version 4.2014. Revised 2013. Available at: http://www.nccn.org/professionals/physician_gls/f_guidelines.asp. Accessed November 17, 2014.
- 3.Zelboraf (vemurafenib) tablets. Prescribing information. Roche. Revised November 2014. Available at: http://www.gene.com/download/pdf/zelboraf_prescribing.pdf. Accessed December 4, 2014. [Google Scholar]
- 4.Tafinlar (dabrafenib) capsules. Prescribing information. GlaxoSmithKline. Revised January 2014. Available at: http://www.accessdata.fda.gov/drugsatfda_docs/label/2014/202806s002lbl.pdf. Accessed November 17, 2014. [Google Scholar]
- 5.Mekinist (trametinib) tablets. Prescribing information. GlaxoSmithKline. Revised January 2014. Available at: http://www.accessdata.fda.gov/drugsatfda_docs/label/2014/204114s001lbl.pdf. Accessed November 17, 2014. [Google Scholar]
- 6.Bedikian AY, DeConti RC, Conry R, et al. Phase 3 study of docosahexaenoic acid-paclitaxel versus dacarbazine in patients with metastatic malignant melanoma. Ann Oncol. 2011;22(4):787-93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Chapman PB, Hauschild A, Robert C, et al. Improved survival with vemurafenib in melanoma with BRAF V600E mutation. N Engl J Med. 2011;364(26):2507-16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.O’Day S, Gonzalez R, Lawson D, et al. Phase II, randomized, controlled, double-blinded trial of weekly elesclomol plus paclitaxel versus paclitaxel alone for stage IV metastatic melanoma. J Clin Oncol. 2009;27(32):5452-58. [DOI] [PubMed] [Google Scholar]
- 9.Einzig AI, Hochster H, Wiernik PH, et al. A phase II study of taxol in patients with malignant melanoma. Invest New Drugs. 1991;9(1):59-64. [DOI] [PubMed] [Google Scholar]
- 10.Hauschild A, Grob JJ, Demidov LV, et al. Dabrafenib in BRAF-mutated metastatic melanoma: a multicentre, open-label, phase 3 randomised controlled trial. Lancet. 2012;380(9839):358-65. [DOI] [PubMed] [Google Scholar]
- 11.Flaherty KT, Robert C, Hersey P, et al. Improved survival with MEK inhibition in BRAF-mutated melanoma. N Engl J Med. 2012;367(2):107-14. [DOI] [PubMed] [Google Scholar]
- 12.Fick DM, Cooper JW, Wade WE, Waller JL, Maclean JR, Beers MH. Updating the Beers criteria for potentially inappropriate medication use in older adults: results of a US consensus panel of experts. Arch Intern Med. 2003;163(22):2716-24. [DOI] [PubMed] [Google Scholar]
- 13.Mickisch G, Gore M, Escudier B, Procopio G, Walzer S, Nuijten M. Costs of managing adverse events in the treatment of first-line metastatic renal cell carcinoma: bevacizumab in combination with interferon-alpha2a compared with sunitinib. Br J Cancer. 2010;102(1):80-86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Banz K, Bischoff H, Brunner M, et al. Comparison of treatment costs of grade 3/4 adverse events associated with erlotinib or pemetrexed maintenance therapy for patients with advanced non-small-cell lung cancer (NSCLC) in Germany, France, Italy, and Spain. Lung Cancer. 2011;74(3):529-34. [DOI] [PubMed] [Google Scholar]
- 15.Borovicka JH, Calahan C, Gandhi M, et al. Economic burden of dermatologic adverse events induced by molecularly targeted cancer agents. Arch Dermatol. 2011;147(12):1403-09. [DOI] [PubMed] [Google Scholar]