Cost-Minimization Analysis of Adjuvant Chemotherapy Regimens Given To Patients With Colorectal Cancer in Japan
Cost-Minimization Analysis of Adjuvant Chemotherapy Regimens Given To Patients With Colorectal Cancer in Japan
Cost-Minimization Analysis of Adjuvant Chemotherapy Regimens Given To Patients With Colorectal Cancer in Japan
Abstract
Background: Consideration of medical costs as well as effectiveness and adverse events is rapidly been
becoming an important factor in the selection of chemotherapy regimens. However, practical data on the costs
of chemotherapy are scarce. We clinically estimated the medical costs of 6 adjuvant chemotherapy regimens for
colorectal cancer on the basis of clinical and cost-related data and compared their cost-effectiveness by cost-
minimization analyses.
Methods: All patients who received adjuvant chemotherapy for colorectal cancer between April 2012 and May
2015 at four hospitals affiliated with Showa University were studied retrospectively. Clinical and cost data related
to adjuvant chemotherapy were collected from medical records and medical fee receipt data, respectively. Six
adjuvant chemotherapy regimens were studied: capecitabine and oxaliplatin (CapeOX); 5-fluorouracil (5-FU),
ℓ-leucovorin (LV), and oxaliplatin (modified FOLFOX6 [mFOLFOX6]); 5-FU and LV (5-FU/LV); tegafur and uracil
(UFT), and LV (UFT/LV); capecitabine; and tegafur, gimeracil and oteracil (S-1). The regimens were divided into
2 groups according to whether or not they contained oxaliplatin because of the difference in effectiveness.
Cost-minimization analyses, where relative costs of regimens showing equivalent effectiveness were simply
compared, were performed to evaluate the cost-effectiveness of the regimens in each group.
Results: A total of 154 patients with colorectal cancer received adjuvant chemotherapy during the study period.
Fifty-seven patients were treated with CapeOX, 10 with mFOLFOX6, 38 with UFT/LV, 20 with capecitabine, and 29
with S-1. No patient received 5-FU/LV. The total costs of oxaliplatin-containing regimens were significantly higher
than those of oxaliplatin non-containing regimens. The high cost of oxaliplatin, but not the costs of drugs or
various tests for the treatment of adverse events, was the primary reason for the higher costs of the oxaliplatin-
containing regimens. The cost-effectiveness of the oxaliplatin-containing regimens CapeOX and mFOLFOX6 were
comparable. Among the oxaliplatin non-containing regimens, the cost-effectiveness of S-1 and capecitabine was
superior to that of UFT/LV.
Conclusion: Thus, we provided the cost-effectiveness data of 5 adjuvant chemotherapy regimens for colorectal
cancer based on practical clinical and cost data from Japanese patients. The results can be included as a factor in
regimen selection because these results would represent the real world.
Trial registration: This study is a retrospective observational study and does not include any health care
interventions. Therefore, we did not register the protocol of this study.
Keywords: Cost-minimization analysis, Cost-effectiveness, Colorectal cancer, Adjuvant chemotherapy, Regimen selection
* Correspondence: [email protected]
2
Institute of Molecular Oncology, Showa University, 1-5-8, Hatanodai,
Shinagawa-ku, Tokyo 142-8555, Japan
Full list of author information is available at the end of the article
© The Author(s). 2016 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Takata et al. Journal of Pharmaceutical Health Care and Sciences (2016) 2:30 Page 2 of 9
Chemotherapeutic regimens cost according to the DPC. This analysis was performed
CapeOX consisted of a 2-h intravenous infusion of oxali- from the perspective of the health care payer. We
platin (130 mg/m2) on day 1 and oral capecitabine described the unit of all costs by Japanese yen and US
(1000 mg/m2) twice daily on days 1 to 14, repeated every dollars, assuming that 1 US dollar was equivalent to
3 weeks for 8 cycles [8]. mFOLFOX6 consisted of LV 100 Japanese yen.
(200 mg/m2) given as a 2-h infusion and oxaliplatin
(85 mg/m2) given as a 2-h infusion, followed by a bolus Cost-minimization analyses
infusion of 5-FU (400 mg/m2) and a 46-h continuous Cost-minimization analysis is one of methods to evalu-
infusion of 5-FU (2400 mg/m2). This regimen was re- ate cost-effectiveness of therapeutic options [19], in
peated every 2 weeks for 12 cycles [10]. Brand-name which relative costs of therapeutic options showing
oxaliplatin was used in CapeOX and mFOLFOX6. 5-FU/ equivalent outcomes of interventions are simply com-
LV comprised a 2-h infusion of LV (250 mg/m2) and a pared. We performed cost-minimization analyses for the
bolus infusion of 5-FU (500 mg/m2) given 1 h after start- oxaliplatin-containing regimens (CapeOX and mFOL-
ing the LV infusion, repeated weekly for 6 weeks followed FOX6) and the oxaliplatin non-containing regimens
by a 2-week rest [11]. This regimen was given for 3 cycles. (5-FU/LV, UFT/LV, capecitabine, and S-1) because of
UFT/LV consisted of oral UFT (300 mg/m2) and LV the following reasons:
(75 mg/patient) given 3 times daily on days 1 to 28 1) Because there was no direct comparison between
followed by a 7-day rest, repeated for 5 cycles [12]. Cape- CapeOX and mFOLFOX6, we compared the effective-
citabine was given orally in a dose of 1250 mg/m2 twice ness of these regimens based on the following consider-
daily on days 1 to 14, followed by a 7-day rest, repeated ations. As demonstrated by 2 international phase 3
for 8 cycles [13]. S-1 was administered orally twice daily trials, 16968 [8] and MOSAIC [9], the effectiveness of
for 28 consecutive days, followed by a 2-week rest. S-1 CapeOX and FOLFOX4 was significantly superior to
was given in a fixed dose based on the patient’s BSA that of 5-FU/LV and LV5FU2, respectively (Table 1 and
according to the dose recommendations of the manufac- Fig. 1a)). Because the effectiveness of LV5FU2 and 5-
turer’s package insert in Japan. The dose was 80 mg/day FU/LV [20, 21] and that of FOLFOX4 and mFOLFOX6
for patients with a BSA of less than 1.25 m2, 100 mg/day were comparable [10] (Table 1), the 3-year disease-free
for those with a BSA of 1.25 to 1.5 m2, and 120 mg/day survival (DFS) rates of both CapeOX and mFOLFOX6
for those with a BSA of more than 1.5 m2. This regimen were comparable and approximately 5 % higher than
was given for 4 cycles [14]. that of 5-FU/LV. 2) Two international phase 3 trials,
NSABP C-06 [12] and X-ACT [13] (Table 1), showed
Data collection that UFT/LV and capecitabine were noninferior to 5-FU/
Patient background data, such as age and disease stage, LV in terms of 5-year overall survival (OS). In addition,
as well as data during adjuvant chemotherapy, including the ACTS-CC international phase 3 trial demonstrated
laboratory tests, prescribed drugs, and adverse events, that S-1 was noninferior to UFT/LV with respect to the 3-
were collected from the patients’ medical records. year DFS rate [14] (Table 1 and Fig. 1a)). On the basis of
Cost data related to adjuvant chemotherapy were these results, we assumed that the effectiveness of these 3
extracted from medical fee receipt data. Costs for out- regimens was comparable and nearly equivalent to the
patient visits, laboratory tests, imaging tests for tumor effectiveness of 5-FU/LV.
diagnosis, and prescription fees for administered drugs
were collected. The cost of each administered drug was Statistical analyses
calculated by multiplying the drug dose prescribed by its Differences in quantitative variables, including cost data,
unit price according to the Japanese National Health were tested using the nonparametric Wilcoxon rank-
Insurance fee-for-service system in 2014. The summa- sum test. Differences in qualitative variables were tested
tion of these costs was defined as total cost. Since all hos- using the χ2 test. Two-tailed P values of less than 0.05
pitals in Showa University have adopted the diagnosis were considered to indicate statistical significance. All
procedure combination (DPC) system [18], hospitalization analyses were carried out with the use of JMP version
costs were constant regardless of the number of drugs ad- 12.0 software (SAS Institute, Cary, NC).
ministered and laboratory tests performed. When the total
hospitalization costs calculated by the DPC included the Results
cost of drugs related to adjuvant chemotherapy, the drug Patient characteristics
costs were calculated by the method described above From April 2012 through May 2015, a total of 154 pa-
(the drug dose prescribed x its unit price), and the tients with colorectal cancer received adjuvant chemo-
hospitalization cost was calculated by subtracting the cost therapy in hospitals affiliated with Showa University.
of chemotherapy-related drugs from the hospitalization Fifty-seven patients were treated with CapeOX, 10 with
Takata et al. Journal of Pharmaceutical Health Care and Sciences (2016) 2:30 Page 4 of 9
a)
b)
Fig. 1 Comparisons of a) effectiveness and b) total costs among adjuvant chemotherapy regimens for colorectal cancer. a Three-year DFS rates
of CapeOX and FOLFOX4 were superior to that of 5-FU containing regimens [8, 9], whereas those of UFT/LV and capecitabine showed non-
inferiority to 5-FU containing regimens [12, 13] (see Methods session). S-1 was non-inferior to UFT/LV [14] (see Methods session). b The total costs
included anticancer drug costs, hospitalization costs, laboratory and imaging test costs, prescription fees for administered drugs, supportive care
drug costs, and other costs. The total costs of oxaliplatin-containing regimens were significantly higher than those of oxaliplatin non-containing
regimens (P < 0.001). Mean ± standard deviation, n = 57 for CapeOX, n = 10 for mFOLFOX6, n = 38 for UFT/LV, n = 20 for capecitabine, n = 29
for S-1
Takata et al. Journal of Pharmaceutical Health Care and Sciences (2016) 2:30 Page 5 of 9
S-1, P < 0.001; mFOLFOX6 vs. UFT/LV, P < 0.001; mFOL- cost of oxaliplatin was about 900,000 yen (9000 dollars),
FOX6 vs. capecitabine, P < 0.001; mFOLFOX6 vs. S-1, P < which was equivalent to approximately 40 % of the total
0.001) (Fig. 1b). The total costs of CapeOX and mFOL- cost. The total cost of mFOLFOX6 also included
FOX6 did not differ significantly (P = 0.374). hospitalization costs (400,000 yen [4000 dollars]), such
Among the oxaliplatin non-containing regimens, the as the fee required to prepare a central venous port for
total cost of UFT/LV was significantly higher than that administration of 5-FU, LV, and oxaliplatin. Thus, the
of capecitabine (P < 0.001). The cost of capecitabine was hospitalization costs required for mFOLFOX6 increased
significantly higher than that of S-1 (P = 0.003). the total cost of this regimen to a level comparable to
the cost of CapeOX. The costs of drugs for supportive
Factors causing the higher costs of oxaliplatin-containing care required to administer CapeOX and mFOLFOX6
regimens were approximately equivalent to 10 % of the total costs.
To address the causes of the higher total costs of The breakdown of the costs of supportive care drugs is
oxaliplatin-containing regimens, the breakdown of the shown in Fig. 3. The costs of the drugs prescribed to
costs for each regimen was calculated (Fig. 2). The cost treat peripheral sensory neuropathy, which is frequently
of oxaliplatin in CapeOX was about 1,150,000 yen associated with oxaliplatin-related chemotherapy, were
(11,500 dollars), which was equivalent to approximately approximately 7500 yen (75 dollars) for CapeOX and
60 % of the total cost. In the case of mFOLFOX6, the 4300 yen (43 dollars) for mFOLFOX6, which comprised
Fig. 2 Breakdown of the total costs for each regimen. Supportive care drugs included drugs used as premedication to prevent nausea and vomiting,
drugs used to treat adverse events, and infusion solutions (see Fig. 3)
Takata et al. Journal of Pharmaceutical Health Care and Sciences (2016) 2:30 Page 6 of 9
Fig. 3 Breakdown of the costs for drugs prescribed for supportive care in each regimen. Representative therapeutic drugs included in Others for
CapeOX were ELENTAL®, KRESTIN®, levofloxacin, loxoprofen, and Posterisan® forte
only 0.4 and 0.2 % of the total costs of CapeOX and Cost-minimization analyses
mFOLFOX6, respectively. We considered the possibility Because the effectiveness (Methods session and Fig. 1a))
that a lower frequency of peripheral sensory neuropathy and the total costs (Fig. 1b)) of CapeOX and mFOL-
in the present study than in previous studies led to the FOX6 were comparable, the cost-effectiveness of these
lower cost of prescriptions for this adverse event. The regimens was judged to be similar (Table 4). As described
frequency of peripheral sensory neuropathy of CapeOX in the Methods session and Fig. 1a), the effectiveness of
in the present study was lower than the results of previ- the oxaliplatin non-containing regimens was comparable.
ous study (Table 3). However, in the case of mFOL- Therefore, on the basis of the total costs of these regimens
FOX6, the frequency and grade of peripheral sensory (Fig. 1b)), the cost-effectiveness of S-1 was superior to that
neuropathy in the present study were not necessarily of UFT/LV, and the cost-effectiveness of capecitabine was
lower than those of previous studies (Table 3). On the superior to that of UFT/LV, which were caused by the
other hand, the costs of antiemetics were approximately high cost of LV.
118,000 yen (1180 dollars) for CapeOX and 116,000
yen (1160 dollars) for mFOLFOX6, accounting for Discussion
about 6 % of the total costs. Antiemetics such as aprepi- The present study compared the cost effectiveness of 5
tant, azasetron, domperidone, granisetron, metoclopra- regimens of adjuvant chemotherapy given to patients
mide, ondansetron, palonosetron, prochlorperazine and with colorectal cancer. The total costs were calculated
ramosetron were prescribed in CapeOX and mFOLFOX6 with the use of clinical and cost data obtained from Jap-
regimens. The parentages of patients who used palonose- anese patients who received each regimen of adjuvant
tron and aprepitant were 100 and 26 % in CapeOX, and chemotherapy in clinical practice. This is in contrast to
60 and 40 % in mFOLFOX6, respectively. most previous studies assessing the costs of adjuvant
chemotherapy for colorectal cancer in Japan, which
based the costs of treatment on clinical data obtained
Table 3 Comparison of the frequency of peripheral sensory from large phase 3 clinical trials [15–17].
neuropathy between present study and phase 3 trials To date, three studies of cost-effectiveness employ-
Regimen Grade Present study Phase 3 trials ing clinical data from phase 3 clinical trials have been
CapeOX All Grade 54.4 % 78.0 %a performed: Hisashige et al. [15] analyzed the cost-
effectiveness of UFT by comparing clinical and cost
≥ Grade 3 1.80 % 11.0 %a
data between patients who received or did not receive
mFOLFOX6 All Grade 90.0 % 92.0 %b
UFT in the NSAS CC trial [22]. In other Japanese
≥ Grade 3 40.0 % 12.5 %b studies, the cost-effectiveness of 5-FU/LV and capecit-
Grade of neuropathy was evaluated according to the Common Terminology abine [16] was evaluated with the use of clinical data
Criteria for Adverse Events version 3.0.
a
Data from reference [8]; bResult of FOLFOX4 [9]. Effectiveness and safety of
from X-ACT trial [13], and that of 5-FU/LV and FOL-
mFOLFOX6 were comparable to those of FOLFOX4 [10]. FOX4 [17] was evaluated with the use of data from the
Takata et al. Journal of Pharmaceutical Health Care and Sciences (2016) 2:30 Page 7 of 9
MOSAIC trial [9]. We compared the costs required for estimated cost of laboratory tests for UFT regimens in a
the following 3 categories between the present study and previous study (about 180,100 yen [1801 dollars]) [15] was
previous studies based on large international phase 3 tri- approximately 3 times higher than that calculated in our
als: 1) anticancer drugs, 2) drugs used for supportive care, practical study (about 65,500 yen [655 dollars]). On the
and 3) laboratory tests. 1) The previously estimated cost other hand, the laboratory test costs in patients who re-
of 1 year of treatment with UFT (about 393,700 yen [3937 ceived FOLFOX4 regimens in a previously reported study
dollars]) [15] was generally similar to the cost calculated (76,800 yen [768 dollars]) [17] was lower than that in our
by us (i.e., about 360,200 yen [3602 dollars], equivalent to present study (about 106,500 yen [1065 dollars]). These
twice the cost of 6 months’ treatment with UFT in our findings indicate that the costs of 1) anticancer drugs, 2)
study). However, the cost of capecitabine calculated in drugs prescribed for supportive care, and 3) laboratory
a previous study (540,000 yen [5400 dollars]) [16] was tests calculated on the basis of clinical data from phase 3
higher than that estimated by us (about 420,500 yen trials differ from those calculated on the basis of data from
[4205 dollars]). The reason for the higher cost of cape- actual clinical practice. Because the costs calculated from
citabine in the previous study is considered to be the patient data in clinical practice would precisely represent
difference in relative dose intensity (RDI) of capecita- the actual situation, cost-effectiveness data thus obtained
bine between the two studies. The previous study used can be used for regimen selection.
a theoretical RDI of 100.0 %, whereas our study used In Japan, a system of the public health insurance for
the clinically observed RDI of 75.4 %. The cost of cape- the entire nation has been adopted. Patients have to
citabine estimated by Shiroiwa et al. [16] would have pay for medical costs according to their age and in-
been about 407,200 yen (4072 dollars) if an RDI of come. The cost borne by the patient ranges from 10.0
75.4 % had been adopted, which is nearly comparable to 30.0 % of total medical costs. In addition, the pa-
to our estimated cost. 2) The costs of agents prescribed tient’s financial burden is maintained below specified
for supportive care in previous studies of UFT and cap- limits under the high-cost medical care benefit system.
ecitabine [15, 16] were about 300 yen (3 dollars) and The specified limits are determined by the patient’s in-
7000 yen (70 dollars), respectively, while those in the come. If this system is applied, the costs for adjuvant
present study were about 8400 yen (84 dollars) for chemotherapy that would be actually paid by the pa-
UFT/LV and about 17,500 yen (175 dollars) for capecit- tient could be lower. Data from Showa University Hos-
abine, demonstrating clearly higher costs for supportive pital indicate when the public health insurance was
care in our study. The primary reason first considered applied to a patient, the cost of oxaliplatin-containing
for the higher supportive care costs in our study was a regimens was approximately 550,000 yen (5500 dollars),
higher incidence of adverse events in the present study and that of UFT/LV was 263,000 yen (2630 dollars).
than in previous studies. However, the incidence of bili- The difference was 287,000 yen (2870 dollars). How-
rubin increase in the NSAS CC trial was 60.0 % [22], as ever, when the specified limits were applied, the cost of
compared with 10.5 % in the present study. The inci- oxaliplatin-containing regimens was approximately
dence of hand-foot syndrome associated with capecita- 448,000 yen (4480 dollars), and that of UFT/LV was
bine regimens was 60.0 % in the X-ACT trial [13] and approximately 262,000 yen (2620 dollars), leading to a
30.0 % in our study. Thus, the incidences of adverse events difference of 186,000 yen (1860 dollars). Thus, the speci-
were not necessarily higher in our study as compared with fied limits might lower the medical costs of oxaliplatin-
previous phase 3 trials. As shown in Fig. 3, patients given containing regimens to a greater extent than the costs of
UFT/LV were mainly prescribed drugs to manage gastro- UFT/LV, although the specified limits system is not neces-
intestinal symptoms, such as proton pump inhibitors sarily applicable to all patients because application of this
and histamine-2 blockers. In patients who received cap- system depends on the income of each patient. It is plaus-
ecitabine, Chinese herbal drugs such as Juzentaihoto ible that patients who derive an economic benefit tend to
and Hochuekkito were predominantly prescribed. The select oxaliplatin-containing regimens over other regi-
costs of these drugs might have contributed to the higher mens. The medical costs are supplemented with taxes
costs for supportive care drugs in our study. 3) The from Japanese citizens. To maintain the patient’s financial
Takata et al. Journal of Pharmaceutical Health Care and Sciences (2016) 2:30 Page 8 of 9