bs_bs_banner
Asia-Pacific Journal of Clinical Oncology 2015; 11: 178–186
doi: 10.1111/ajco.12354
ORIGINAL ARTICLE
Cost of cancer care for patients undergoing chemotherapy: The Elements of Cancer Care study Robyn L WARD,1 Maarit A LAAKSONEN,1 Kees VAN GOOL,2 Sallie-Anne PEARSON,1* Ben DANIELS,1 Patricia BASTICK,1 Richard NORMAN,2 Changhao HOU,2 Philip HAYWOOD2 and Marion HAAS2 1
Adult Cancer Program, Lowy Cancer Research Centre, Prince of Wales Clinical School, UNSW Australia, and 2Centre for Health Economics Research and Evaluation, University of Technology, Sydney, New South Wales, Australia
Abstract Aim: To determine the monthly treatment costs for each element of cancer care in patients receiving chemotherapy and to apportion the burden of cost by financing agent (Commonwealth, State government, private health insurer, patient). Methods: A cohort of 478 patients (54% breast, 33% colorectal and 13% non-small-cell lung cancer) were recruited from 12 centers representing metropolitan and regional settings in public and private sectors. Primary data were linked to secondary data held in New South Wales state (Admitted Patients and Emergency Department Data) and Commonwealth (Medicare and Pharmaceutical Benefits) databases. The monthly treatment costs of each element of care and the funding agent were calculated from secondary health data. Results: Across all tumor types, the mean monthly treatment cost was $4162 (10%–90% quantiles $1018–$8098; range $2853 [adjuvant colorectal] to $5622 [metastatic lung]), with 54% of this cost borne by Commonwealth government, 26% by private health insurers, 14% by State government and 6% by patients. The mean monthly costs of treating metastatic disease were $1415 greater than those for adjuvant therapy. The mean monthly costs were contributed to by inpatient care ($1657, 40%), chemotherapy prescriptions ($1502, 36%), outpatient care ($452, 11%) and administration of chemotherapy ($364, 9%). Conclusion: All four funders have a shared incentive to reduce absolute monthly treatment costs since their proportional contribution is relatively constant for most tumor types and stages. There are opportunities to reduce cancer care costs by minimizing the risk of inpatient hospital admissions that arise from chemotherapy administration and by recognizing incentives for cost-shifting. Key words: breast neoplasm, colorectal neoplasm, cost and cost analysis, drug therapy, lung neoplasm.
INTRODUCTION The annual health expenditure on cancer is rising at a rate that is widely regarded as unsustainable.1–3 This
Correspondence: Professor Robyn L Ward MD, PhD, Level 1, South Wing, EBB Prince of Wales Hospital, Randwick, NSW 2031, Australia. Email: [email protected] *Present address: University of Sydney, Sydney, Australia. Conflict of interest: none Accepted for publication 1 February 2015.
© 2015 Wiley Publishing Asia Pty Ltd
trend is attributed to the rising prevalence of cancer, the increase in the price of surgery, radiation and drug therapy, and the longer duration of cancer treatments. Of these factors, the largest proportional increase in costs is related to prescription pharmaceuticals. In Australia, expenditure on pharmaceuticals increased from 6% of the health system expenditure in 2000–2001 ($169M) to 12% in 2008–2009 ($540M).4 This trend, in terms of both absolute expenditure and the high proportional contribution of pharmaceutical costs, is likely to have continued over the last five years because of the proliferation of very high-cost anticancer drug
179
Cost of cancer chemotherapy in Australia
treatments, some of which cost up to $10 000 per month of treatment.5 While aggregate costing studies provide useful estimates of the national economic burden of cancer, their very nature masks important variations in personal, clinical or health services delivery. They fail to capture the impact of differences in patient characteristics such as disease stage, socioeconomic status and private health insurance status. Likewise, they do not reflect those services used in alternative treatment pathways, or the volume of care delivered to each person.6 Yet, each of these factors is likely to be an important determinant of cost variability. In contrast, studies of costs to individual patients are able to capture these factors and measure the extent of variation in costs between patients. Such data provide the basis for undertaking real-world costeffectiveness analyses as well as examining the extent to which some population groups face financial barriers to access for specific treatments. Another advantage of individual patient costing studies is that they provide data regarding which of the multiple payers (Commonwealth or State government, patient, third-party insurers) bear the cost of cancer care, and how this may change in different disease states and tumor types. Analysis of these data can be used to demonstrate how funding arrangements create perverse incentives to shift costs to other payers and potentially impede the delivery of effective and efficient care. Despite the advantages of individual patient costing studies, there are few studies of this type in the setting of cancer chemotherapy. Furthermore, given that drug pricing and healthcare systems vary between jurisdictions, it is important that such studies are undertaken in the context in which the information could ultimately be used. The Elements of Cancer Care (EoCC) study was designed to provide a comprehensive analysis of the cost of all components of care for patients undergoing chemotherapy for breast, colorectal and non-small-cell lung cancer (NSCLC) in a broad range of treatment settings. To achieve this goal, primary data were collected through medical record review and monthly patient interviews, and secondary linked data were obtained from New South Wales (NSW) and Australian Commonwealth administrative health data collections. This is the first report from the EoCC study. It describes the proportion of monthly treatment costs borne by four major financing agents (Commonwealth and State governments, private healthcare payers and patients), and it identifies the monthly costs for each element of health care in this cohort.
Asia-Pac J Clin Oncol 2015; 11: 178–186
METHODS Study design This was a prospective cohort study of patients over 18 years of age who were undergoing chemotherapy for either breast, colorectal or NSCLC at one of the 12 cancer treatment sites in NSW (study registration Research Data Australia, Identifier 004:273). These sites represented metropolitan (n = 7), regional and rural settings (n = 5) in both the public (n = 7) and the private sectors (n = 5). The study was approved by Human Research Ethics Committee (University of NSW, approval number 07014) and governance approvals were obtained at the 12 sites. Recruitment commenced in January 2009 and ceased in October 2010. To identify potentially eligible patients, field staff reviewed medical oncology records and then sought confirmation of eligibility by the treating physician. Patients already undergoing chemotherapy or about to commence treatment in the next week were then approached for written consent. Participation involved consent for medical record review, interview and data linkage with NSW and Commonwealth (Medicare Australia) secondary health data collections (Fig. 1a). Medical record reviews and interviews were conducted monthly to collect data on cancer treatment, health service utilization, socioeconomic status, use of complementary therapy and a range of other factors (data set available upon request). Medical record reviews continued until patients either withdrew from the study, ceased chemotherapy and did not recommence within 30 days, died, or the census date was reached (June 11, 2010 for the 2009 cohort and May 5, 2011 for the 2010 cohort). Regular interviews ceased 6 months after the recruitment date or when the above conditions were met. While extensive primary data were collected, the current report uses only four of these data items, specifically cancer type and stage, administered chemotherapy protocol and use of S100 drugs by public hospital patients.
Data analysis Data collected by field staff were cleaned by the investigators and statisticians, and data inconsistencies were resolved through face-to-face meetings between investigators, field staff and data analysts. The calculation of costs was based on secondary health data, specifically PBS (prescription medicines), MBS (Medicare services), NSW Admitted Patients Data Collection (APDC, hospital admissions) and NSW Emergency Department Data Collection (EDDC, emergency visits). The only primary data used for cost
© 2015 Wiley Publishing Asia Pty Ltd
180
RL Ward et al.
(a)
Allocated to cohort (n = 478)
2009 cohort (n = 310)
n =1
n =1
2010 cohort (n = 168)
n =58
n =31
n =137 Record review plus interview plus data linkage
Record review plus data linkage
Record review plus interview
Record review
n =250
Commonwealth data (n = 453) NSW health data (n = 404)
(b) 500 450
Number of parcipants
400 350 300 250 200 150 100 50 0 0
1
2
3
4
Duraon of review (months)
© 2015 Wiley Publishing Asia Pty Ltd
5
6
Figure 1 Number of patients in the Elements of Cancer Care (EoCC) study who consented to medical record review, data linkage and interview (a) and period of observation (b). The EoCC study recruited two patient cohorts, labeled 2009 and 2010, respectively. For the 2009 cohort, consent to either data linkage or interview was optional, while for the 2010 cohort, only consent to interview was optional. Data linkage to the four NSW Health data collections: NSW Central Cancer Registry (CCR),7 NSW Admitted Patients Data Collection (APDC),8 Registry of Births, Deaths and Marriages plus the Australian Bureau of Statistics Mortality Database (ABS-BDM), and the NSW Emergency Department Data Collection (EDDC) was performed by NSW Centre for Health Record Linkage (CHeReL). To perform data linkage to Medicare Australia data (Pharmaceutical Benefit Scheme [PBS] and Medicare Benefit Schedule [MBS]), an extract file of cohort members was sent to Medicare Australia. Of the 476 patients who consented to data linkage, 453 had at least one record in Commonwealth data (MBS or PBS) and 404 patients had linked data in NSW health records (APDC, EDDC, RBDM-ABS) during the follow-up period.
Asia-Pac J Clin Oncol 2015; 11: 178–186
181
Cost of cancer chemotherapy in Australia
calculations were the costs of S100 drugs for public hospital patients (the costs of these drugs were in PBS data for private patients). The total PBS and MBS costs were sums of government benefits and patient out-ofpocket costs. The costs of public patients’ hospital admissions were estimated by linking the Australian Refined Diagnosis Related Groups (AR-DRG) codes to their corresponding total average costs using national cost reports.9,10 The costs of private patients’ hospital admissions were estimated by linking the AR-DRG codes to their corresponding total average costs using Hospital Casemix Protocol annual reports.11–13 The cost of emergency department care was estimated by weighing the average cost of an emergency department presentation by departure status and triage category.14 A summary of the decision rules used to calculate costs for each funding agent (Commonwealth, State government, private health insurer, patient out-of-pocket) and by the type of expenditure is shown in Supporting Information Tables S1 and S2. The individual total cost of each cost component was calculated as the sum of all costs incurred from recruitment until the first of either cessation of chemotherapy, death or the census date. The individual mean monthly cost was calculated by dividing the total cost by the number of 28-day periods accrued during their followup. Total cost and costs by site and stage of cancer were reported as the mean monthly cost and as the 10th and 90th quantiles (to illustrate variability in costs across patients) of the mean cost per patient.
to hospital (APDC) and 350 visits to the emergency ward (EDDC). Using these data, the mean monthly cost of cancer care was calculated as $4162 (10%–90% quantiles $1018–$8098) with 54% of this cost borne by the Commonwealth government and 26% by private health insurers (Table 2). The mean monthly cost of treating metastatic disease was $1415 greater than the costs of adjuvant therapy ($4810, 10%–90% quantiles $1533–$8365 compared with $3395, 10%–90% quantiles $730–$7660). In aggregate, across all tumor types, the distribution of monthly treatment costs for adjuvant disease was 65% Commonwealth government, 8% State government, 21% private health insurers and 6% out-of-pocket expenses. The respective figures for metastatic disease were 47%, 19%, 29% and 5%. The mean monthly treatment cost was highly variable between patients, tumor types and extents of disease, from $5622 (10%–90% quantiles $2210–9338) for metastatic NSCLC to $2853 (10%–90% quantiles $735–8098) for adjuvant colorectal cancer. The distribution of costs by funding across tumor types was similar with two notable exceptions. Over 73% of the monthly treatment costs of adjuvant breast cancer were borne by the Commonwealth government as a consequence of a lesser contribution from the State government (reduced to 6% of the costs). For metastatic NSCLC, 32% of the costs were borne by the state government as a consequence of a reduced proportional contribution from private health insurers. For all tumor types, the out-of-pocket expenses represented only a modest contribution to the monthly treatment costs (∼6%).
RESULTS Profile of study participants
Costs of the individual care elements
A total of 478 patients consented to the study (50.5% of those approached) and over 80% agreed to record review, interview and data linkage. The period of observation for participants is shown in Figure 1b. The characteristics of the patients recruited in the 2009 and 2010 cohorts were very similar (see Table 1 and Supporting Information Table S3). The patients in this study received 148 unique chemotherapy protocols; however, most patients received one of the commonly used protocols for treatment in the adjuvant or metastatic settings (Supporting Information Table S4). The distribution of treatment between public and private centers was 49.6% and 50.4%, respectively.
In total, the largest proportion of mean monthly costs is related to inpatient care ($1657, 40%) and chemotherapy prescriptions ($1502, 36%), with lesser contributions from outpatient care ($452, 11%) and administration of chemotherapy ($364, 9%) (Table 3). The monthly inpatient costs for patients with metastatic disease substantially exceeded the costs incurred for patients on adjuvant treatment ($2269, 10%–90% quantiles $0–$5264, 47% compared with $934, 10%– 90% quantiles $0–$2820, 28%). Overall, the monthly cost of chemotherapy prescriptions was very similar for patients receiving adjuvant ($1539, 10%–90% quantiles $47–$4179) or metastatic treatment ($1471, 10%– 90% quantiles $76–$3135). In both the adjuvant and the metastatic settings, other care elements such as radiology, pathology, specialist and general practitioner visits accounted for only a small percentage of monthly
Costs of cancer care by financing agent Analysis of the linked data on 475 patients identified 46 854 MBS and 18 854 PBS claims, 3917 admissions
Asia-Pac J Clin Oncol 2015; 11: 178–186
© 2015 Wiley Publishing Asia Pty Ltd
182
RL Ward et al.
Table 1 Demographic and clinical baseline characteristics of the 2009 and 2010 cohorts derived from medical record review at recruitment Total cohort (n = 478) Variable Demographic variables Sex Men Women Age (years) ≤45 46–65 66–85 Treatment location Rural/Regional public Rural/Regional private Metropolitan public Metropolitan private Clinical variables Site and stage of cancer Breast Stage I Stage II Stage III Stage IV Colorectal Stage I Stage II Stage III Stage IV Non-small-cell lung cancer Stage I Stage II Stage III Stage IV Previous chemotherapy Yes No †
2009 cohort (n = 310)
†
†
2010 cohort (n = 168)
N (%)
N (%)
N† (%)
126 (26.4) 352 (73.6)
69 (22.3) 241 (77.7)
57 (33.9) 111 (66.1)
61 (12.8) 290 (60.7) 127 (26.6)
42 (13.5) 190 (61.3) 78 (25.2)
19 (11.3) 100 (59.5) 49 (29.2)
37 (7.7) 17 (3.6) 204 (42.7) 220 (46.0)
35 (11.3) Nil Nil 142 (45.8) 133 (42.9)
2 (1.2) 17 (10.1) 62 (36.9) 87 (51.8)
258 (54.0) 24 (5.0) 82 (17.2) 49 (10.3) 103 (21.5) 157 (32.8) 0 (0) 6 (1.3) 48 (10.0) 103 (21.5) 63 (13.2) 4 (0.8) 2 (0.4) 16 (3.3) 41 (8.6)
188 (60.7) 18 (9.6) 65 (34.6) 29 (15.4) 76 (40.4) 83 (26.8) 0 (0.0) 2 (2.4) 24 (28.9) 57 (68.7) 39 (12.6) 3 (7.7) 1 (2.6) 8 (20.5) 27 (69.2)
70 (41.7) 6 (8.6) 17 (24.3) 20 (28.6) 27 (38.5) 74 (44.0) 0 (0.0) 4 (5.4) 24 (32.4) 46 (62.2) 24 (14.3) 1 (4.2) 1 (4.2) 8 (33.3) 14 (58.3)
198 (42.0) 273 (58.0)
135 (44.5) 168 (55.5)
63 (37.5) 105 (62.5)
Due to missing observations, all numbers do not add up to the size of the cohort.
treatment costs (total $452, 10%–90% quantiles $114– $868, 11%). While there was marked variability in the monthly treatment costs between patients and between the three tumor types, the relative difference between adjuvant and metastatic inpatient treatment costs held true, with 1.6- to 2.6-fold higher mean monthly inpatient costs for patients on metastatic treatment. Of all costs examined, the greatest variability was for chemotherapy prescriptions, ranging from around $700 for adjuvant colorectal and NSCLC (10%–90% quantiles $18–$1374 and $0–$1706, respectively) to nearly
© 2015 Wiley Publishing Asia Pty Ltd
$2000 per month for adjuvant breast (10%–90% quantiles) and metastatic NSCLC (10%–90% quantiles $33–$3329).
DISCUSSION This multicenter cohort study showed that the Commonwealth government bore the largest proportion of the cost of care, followed by private health insurers, State government and patients. Admissions to hospital for purposes unrelated to chemotherapy administration were found to be the largest contributor to overall
Asia-Pac J Clin Oncol 2015; 11: 178–186
Commonwealth government costs based on benefits paid through the Pharmaceutical Benefits Scheme and Medicare Benefits Scheme as well as hospital treatment costs for residential aged care (based on Admitted Patients Data Collection data). ‡State government costs based on hospital treatment for public patients (based on Admitted Patients Data Collection) and emergency department visits (based on Emergency Department Data Collection). §PHI (Private health insurers) costs based on Medicare Benefits Scheme and hospital treatment (based on Admitted Patients Data Collection) for private patients covered by private health insurers. ¶OOP (out of pocket) cost based on Pharmaceutical Benefits Scheme, Medicare Benefits Scheme, and hospital costs (based on Admitted Patients Data Collection) paid by patients out-of-pocket. NSCLC, non-small-cell lung cancer.
†
6% 6% 6% 10% 6% 6% 3% 238 (24–545) 199 (26–422) 240 (21–565) 271 (33–622) 311 (27–738) 186 (17–602) 172 (18–465) 26% 15% 30% 36% 34% 35% 17% 1075 (0–3182) 541 (0–1350) 1219 (0–3426) 1032 (0–3694) 1770 (0–3881) 1144 (0–5100) 969 (0–2947) 14% 6% 16% 11% 14% 15% 32% 607 (0–1782) 225 (0–416) 647 (0–2059) 322 (0–548) 718 (0–2193) 493 (0–1123) 1770 (0–5531) 54% 73% 48% 43% 46% 44% 48% 2242 (488–4386) 2639 (563–5984) 1939 (557–3544) 1228 (374–2408) 2393 (548–4626) 1429 (321–2881) 2710 (733–4065) 476 149 108 54 102 13 50 Overall Breast adjuvant Breast metastatic Colorectal adjuvant Colorectal metastatic NSCLC adjuvant NSCLC metastatic
4162 (1018–8098) 3604 (714–7660) 4046 (1279–7633) 2853 (735–8098) 5191 (1846–8928) 3251 (492–7379) 5622 (2210–9338)
Patient out-of-pocket¶ (%) Private health insurers§ State government‡ Commonwealth government† Total N Cancer site and stage
Table 2 Mean (10%–90% quantiles) monthly (28 day) costs (in 2011 Australian dollars) of cancer care by funding agent, presented according to site and stage of cancer
Cost of cancer chemotherapy in Australia
Asia-Pac J Clin Oncol 2015; 11: 178–186
183
costs of care. The absolute monthly treatment costs of cancer care varied by patient, tumor type, the presence of metastatic disease and the extent of inpatient care. With the exceptions of adjuvant breast cancer and metastatic NSCLC settings, the proportional contribution of funders to cancer care was consistent. The variations for breast and NSCLC are related to the frequency of inpatient admissions and the selection of private health insurance at separation (as recorded in secondary health datasets). For instance, in adjuvant breast cancer, the low contribution of private health insurers to monthly treatment costs is related to low rates of inpatient admissions. In the case of NSCLC, the proportional lower contribution from private health insurers (17%) was matched by a greater proportional financing from State government. Patients with metastatic colorectal cancer had absolute monthly treatment and inpatient costs that were similar to patients with metastatic NSCLC, yet the higher proportional contribution by private insurers (34%) reflected the greater use of private health insurance. At the baseline interview, most patients reported having private insurance; however, comparison with secondary data found that selfreported private status was an unreliable measure of the use of private insurance, as this varied with health system encounter. Understanding the drivers for decisions to use private health insurance has important implications for both funders and planners of health services. Like other studies,15 inpatient hospitalizations, unrelated to admission for administration of chemotherapy, were the largest contributor to overall costs (40%) for all tumor types and settings, except adjuvant breast cancer. Inpatient hospitalizations reflected the management of side effects of chemotherapy and active cancer. These findings show that initiating chemotherapy clearly increases the cost of cancer care well beyond the cost of anticancer drugs and their administration alone. Thus there is a strong rationale for collecting real-world data on the risk and predicted duration of hospital admissions for different chemotherapy protocols, so that prescribers and patients can make better treatment choices. The higher costs for patients undergoing metastatic treatment for all tumor types when compared to adjuvant therapy were mostly explained by increased inpatient costs, likely related to management of cancer progression or its complications. Other studies have consistently found that the net costs of cancer care are highest immediately following diagnosis, and in the last year of life.15–17 Given the natural histories of metastatic
© 2015 Wiley Publishing Asia Pty Ltd
© 2015 Wiley Publishing Asia Pty Ltd Total
Breast adjuvant
11 (0–30) 3 (0–12) 27 (0–74)
0% 0% 1%
445 (43–1144)
147 (0–413)
0% 1%
0%
4%
9%
2%
7 (0–20) 52 (0–176)
7 (0–23)
133 (34–253)
80 (0–290)
227 (66–390)
20% 1795 (0–4567) 16% 452 (180–781)
8%
2% 357 (38–990)
135 (14–254)
0% 1%
0%
3%
2%
6%
6 (0–19) 22 (0–57)
7 (0–9)
109 (47–209)
0 (0–0)
149 (38–245)
511 (31–1331)
136 (8–330)
0% 1%
0%
4%
0%
5%
4 (0–14) 61 (0–165)
3 (0–9)
134 (39–236)
24 (0–0)
227 (62–377)
49% 2508 (0–5479) 9% 393 (143–625)
12%
5%
97 (11–344)
260 (26–625)
0% 1%
0%
3%
0%
4%
1 (0–0) 121 (0–395)
12 (0–29)
141 (41–317)
48 (0–109)
171 (20–356)
158 (0–336)
218 (61–468)
0% 4%
0%
4%
1%
5%
6 (0–23) 128 (0–313)
9 (0–17)
129 (57–203)
86 (0–338)
214 (37–376)
52% 2750 (0–6797) 11% 444 (183–902)
3%
8%
0% 2%
0%
2%
2%
4%
49% 8%
3%
4%
34%
NSCLC metastatic 22% 1924 (33–3329)
NSCLC adjuvant 708 (0–1706)
48% 1692 (0–6888) 8% 372 (112–672)
10%
3%
30%
Colorectal metastatic
24% 1583 (30–3838)
Colorectal adjuvant 673 (18–1374)
44% 1395 (0–5322) 11% 270 (128–519)
11%
4%
29%
Breast metastatic 53% 1154 (165–2314)
4162 (1018–8098) 100% 3604 (714–7660) 100% 4046 (1279–7633 100% 2853 (735–8098) 100% 5191 (1846–8928) 100% 3251 (492–7379) 100% 5622 (2210–9338) 100%
127 (41–257)
88 (20–167)
4%
3%
700 (0–1683) 569 (94–1091)
40% 11%
340 (0–824)
298 (14–852)
9%
3%
85 (0–219)
3%
36% 1925 (138–5114)
Costs to public and private patients for cancer treatment. Costs include chemotherapy-related† or non-chemotherapy‡ prescription medicines (based on Pharmaceutical Benefits Scheme); chemotherapy-related§ or non-chemotherapy-related hospitalization¶ (based on Admitted Patients Data Collection); outpatient medical services costs†† (based on Medicare Benefits Scheme); and emergency department visit costs‡‡ (based on Emergency Department Data Collection). GP, general practitioner; NSCLC, non-small-cell lung cancer.
Total
1502 (61–3549) Chemotherapy medicines† Other prescription 134 (0–343) medicines‡ Chemotherapy 364 (18–998) administration§ Inpatient care¶ 1657 (0–4826) 452 (114–868) Out-of-hospital care†† Radiology and 172 (28–340) pathology Radiation 139 (0–589) oncology Medical, GP and 129 (41–253) specialist visits Allied health and 8 (0–17) Nurse visits Surgery 5 (0–16) Emergency 53 (0–170) department care‡‡
Care element
Table 3 Mean (10%–90% quantiles) monthly (28 day) costs (in 2011 Australian dollars) of cancer care by care element, overall and according to site and stage of cancer
184 RL Ward et al.
Asia-Pac J Clin Oncol 2015; 11: 178–186
185
Cost of cancer chemotherapy in Australia
NSCLC and colorectal cancer, a significant proportion of patients entered this study in or close to their final year of life. Like US studies,15,18,19 the current work showed large inter-patient variations in the costs of chemotherapy independent of tumor site. Subsidy of new cancer drugs such as trastuzumab for adjuvant breast cancer, erlotinib for metastatic NSCLC and bevacizumab for metastatic colorectal cancer in part explained the high mean monthly treatment costs and variability in chemotherapy costs in these groups. However, we propose that new medications, tumor type and stage, and the extent of inpatient care did not account fully for the observed inter-patient variations in treatment costs. Differences in treatment practices, site of treatment, price paid for each element of care, and volume and duration of hospital admissions may also contribute to observed differences and will be the subject of future work. One of the strengths of this study is that by using primary and secondary data, costs were able to be reported by tumor stage and type. Second, enrollment was permitted at any point in the treatment course rather than only at initiation. This ensured a broad capture of the costs of adverse events which may vary in time of onset depending on duration of chemotherapy exposure. The study also has a number of limitations, including the fact that treatment costs were likely to be underestimated because only secondary health data sources were used to determine the cost of the elements of care. Omitted costs include direct patient costs, private scripts and drugs such as trastuzumab which were provided outside the PBS. The primary data showed 32 patients with metastatic breast cancer received trastuzumab, yet this drug cost was not captured at all in the secondary data sources. Another limitation of the study was that a participation bias (resulting largely from increased frequency and duration of clinic visits) led to an overrepresentation of breast cancer patients receiving protocols containing trastuzumab. Finally, it was not possible to reliably distinguish the monthly cancer specific treatments from the background of noncancer healthcare costs for all patients in this study. Despite these limitations, to our knowledge, the study provides the only snapshot of the elements and costs of cancer care for patients receiving chemotherapy for three common cancers in Australia. The data collected in this study will be made available to researchers for future studies of patterns of care, resource use and costs.
Asia-Pac J Clin Oncol 2015; 11: 178–186
ACKNOWLEDGMENT This study was funded by the National Health and Medical Research Council (Health Services Research Grant ID 455366).
REFERENCES 1 Warren JL, Yabroff KR, Meekins A, Topor M, Lamont EB, Brown ML. Evaluation of trends in the cost of initial cancer treatment. J Natl Cancer Inst 2008; 100: 888–97. 2 Smith TJ, Hillner BE. Bending the cost curve in cancer care. N Engl J Med 2011; 364: 2060–5. 3 Delivering High-Quality Cancer Care: Charting a New Course for a System in Crisis. The National Academies Press, 2013. 4 Australian Institute of Health and Welfare Health system expenditure on cancer and other neoplasms in Australia 2008–09. Canberra, 2013. 5 Experts in Chronic Myeloid Leukemia. The price of drugs for chronic myeloid leukemia (CML) is a reflection of the unsustainable prices of cancer drugs: from the perspective of a large group of CML experts. Blood 2013; 121: 4439– 42. 6 Lipscomb J. Estimating the cost of cancer care in the United States: a work very much in progress. J Natl Cancer Inst 2008; 100: 607–10. 7 Cancer Institute NSW. NSW Central Cancer Registry Data Access. 2012. Sydney. [Cited Nov 2013.] Available from: http://www.cancerinstitute.org.au/data-and-statistics/ cancer-registries/nsw-central-cancer-registry-data-access. 8 NSW Health. Health Statistics NSW – Admitted Patient Report. 2012. Sydney. [Cited Aug 2014.] Available from: http://www.health.nsw.gov.au/pubs/2012/admitted_patient _rpt2012.html. 9 The Australian Government Department of Health. Round 13. 2008–2009. Cost Report. 10 The Australian Government Department of Health. Round 14. 2009–2010. National Public Cost Weight Tables. 11 The Australian Department of Health. Hospital Casemix Protocol Annual Report. 2009–2010. Canberra. [Cited Aug 2014.] Available from: http://www.health.gov.au/ internet/main/publishing.nsf/Content/health-casemix-datacollections-publications-HCPAnnualReportsArchived. 12 The Australian Department of Health. Hospital Casemix Protocol Annual Report. 2010–2011. Canberra. [Cited Aug 2014.] Available from: http://www.health.gov.au/ internet/main/publishing.nsf/Content/health-casemix-datacollections-publications-HCPAnnualReports. 13 The Australian Department of Health. Hospital Casemix Protocol Annual Report. 2008–2011. Canberra. [Cited Aug 2014.] Available from: http://www.health.gov.au/ internet/main/publishing.nsf/Content/health-casemix-datacollections-publications-HCPAnnualReportsArchived.
© 2015 Wiley Publishing Asia Pty Ltd
186
14 New South Wales Health Costs of Care Standards 2009/ 10. Sydney. [Cited Aug 2014.] Available from: http:// www0.health.nsw.gov.au/policies/gl/2011/pdf/GL2011_ 007.pdf. 15 Yabroff KR, Lamont EB, Mariotto A et al. Cost of care for elderly cancer patients in the United States. J Natl Cancer Inst 2008; 100: 630–41. 16 Riley GF, Potosky AL, Lubitz JD, Kessler LG. Medicare payments from diagnosis to death for elderly cancer patients by stage at diagnosis. Med Care 1995; 33: 828– 41. 17 Langton JM, Blanch B, Drew AK, Haas M, Ingham JM, Pearson SA. Retrospective studies of end-of-life resource utilization and costs in cancer care using health administrative data: a systematic review. Palliat Med 2014; 28: 1167–96. 18 Meropol NJ, Schulman KA. Cost of cancer care: issues and implications. J Clin Oncol 2007; 25: 180–6.
© 2015 Wiley Publishing Asia Pty Ltd
RL Ward et al.
19 Ferro SA, Myer BS, Wolff DA et al. Variation in the cost of medications for the treatment of colorectal cancer. Am J Manag Care 2008; 14: 717–25.
SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article at the publisher’s web-site: Table S1 Decision rules for the calculation of the costs by funding agent. Table S2 Decision rules for the calculation of costs by care element. Table S3 Socioeconomic characteristics of 2009 and 2010 cohorts obtained from baseline recruitment interview. Table S4 The five most frequently administered chemotherapy protocols according to site and stage of cancer.
Asia-Pac J Clin Oncol 2015; 11: 178–186
Copyright of Asia Pacific Journal of Clinical Oncology is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
Asia-Pacific Journal of Clinical Oncology 2015; 11: 178–186
doi: 10.1111/ajco.12354
ORIGINAL ARTICLE
Cost of cancer care for patients undergoing chemotherapy: The Elements of Cancer Care study Robyn L WARD,1 Maarit A LAAKSONEN,1 Kees VAN GOOL,2 Sallie-Anne PEARSON,1* Ben DANIELS,1 Patricia BASTICK,1 Richard NORMAN,2 Changhao HOU,2 Philip HAYWOOD2 and Marion HAAS2 1
Adult Cancer Program, Lowy Cancer Research Centre, Prince of Wales Clinical School, UNSW Australia, and 2Centre for Health Economics Research and Evaluation, University of Technology, Sydney, New South Wales, Australia
Abstract Aim: To determine the monthly treatment costs for each element of cancer care in patients receiving chemotherapy and to apportion the burden of cost by financing agent (Commonwealth, State government, private health insurer, patient). Methods: A cohort of 478 patients (54% breast, 33% colorectal and 13% non-small-cell lung cancer) were recruited from 12 centers representing metropolitan and regional settings in public and private sectors. Primary data were linked to secondary data held in New South Wales state (Admitted Patients and Emergency Department Data) and Commonwealth (Medicare and Pharmaceutical Benefits) databases. The monthly treatment costs of each element of care and the funding agent were calculated from secondary health data. Results: Across all tumor types, the mean monthly treatment cost was $4162 (10%–90% quantiles $1018–$8098; range $2853 [adjuvant colorectal] to $5622 [metastatic lung]), with 54% of this cost borne by Commonwealth government, 26% by private health insurers, 14% by State government and 6% by patients. The mean monthly costs of treating metastatic disease were $1415 greater than those for adjuvant therapy. The mean monthly costs were contributed to by inpatient care ($1657, 40%), chemotherapy prescriptions ($1502, 36%), outpatient care ($452, 11%) and administration of chemotherapy ($364, 9%). Conclusion: All four funders have a shared incentive to reduce absolute monthly treatment costs since their proportional contribution is relatively constant for most tumor types and stages. There are opportunities to reduce cancer care costs by minimizing the risk of inpatient hospital admissions that arise from chemotherapy administration and by recognizing incentives for cost-shifting. Key words: breast neoplasm, colorectal neoplasm, cost and cost analysis, drug therapy, lung neoplasm.
INTRODUCTION The annual health expenditure on cancer is rising at a rate that is widely regarded as unsustainable.1–3 This
Correspondence: Professor Robyn L Ward MD, PhD, Level 1, South Wing, EBB Prince of Wales Hospital, Randwick, NSW 2031, Australia. Email: [email protected] *Present address: University of Sydney, Sydney, Australia. Conflict of interest: none Accepted for publication 1 February 2015.
© 2015 Wiley Publishing Asia Pty Ltd
trend is attributed to the rising prevalence of cancer, the increase in the price of surgery, radiation and drug therapy, and the longer duration of cancer treatments. Of these factors, the largest proportional increase in costs is related to prescription pharmaceuticals. In Australia, expenditure on pharmaceuticals increased from 6% of the health system expenditure in 2000–2001 ($169M) to 12% in 2008–2009 ($540M).4 This trend, in terms of both absolute expenditure and the high proportional contribution of pharmaceutical costs, is likely to have continued over the last five years because of the proliferation of very high-cost anticancer drug
179
Cost of cancer chemotherapy in Australia
treatments, some of which cost up to $10 000 per month of treatment.5 While aggregate costing studies provide useful estimates of the national economic burden of cancer, their very nature masks important variations in personal, clinical or health services delivery. They fail to capture the impact of differences in patient characteristics such as disease stage, socioeconomic status and private health insurance status. Likewise, they do not reflect those services used in alternative treatment pathways, or the volume of care delivered to each person.6 Yet, each of these factors is likely to be an important determinant of cost variability. In contrast, studies of costs to individual patients are able to capture these factors and measure the extent of variation in costs between patients. Such data provide the basis for undertaking real-world costeffectiveness analyses as well as examining the extent to which some population groups face financial barriers to access for specific treatments. Another advantage of individual patient costing studies is that they provide data regarding which of the multiple payers (Commonwealth or State government, patient, third-party insurers) bear the cost of cancer care, and how this may change in different disease states and tumor types. Analysis of these data can be used to demonstrate how funding arrangements create perverse incentives to shift costs to other payers and potentially impede the delivery of effective and efficient care. Despite the advantages of individual patient costing studies, there are few studies of this type in the setting of cancer chemotherapy. Furthermore, given that drug pricing and healthcare systems vary between jurisdictions, it is important that such studies are undertaken in the context in which the information could ultimately be used. The Elements of Cancer Care (EoCC) study was designed to provide a comprehensive analysis of the cost of all components of care for patients undergoing chemotherapy for breast, colorectal and non-small-cell lung cancer (NSCLC) in a broad range of treatment settings. To achieve this goal, primary data were collected through medical record review and monthly patient interviews, and secondary linked data were obtained from New South Wales (NSW) and Australian Commonwealth administrative health data collections. This is the first report from the EoCC study. It describes the proportion of monthly treatment costs borne by four major financing agents (Commonwealth and State governments, private healthcare payers and patients), and it identifies the monthly costs for each element of health care in this cohort.
Asia-Pac J Clin Oncol 2015; 11: 178–186
METHODS Study design This was a prospective cohort study of patients over 18 years of age who were undergoing chemotherapy for either breast, colorectal or NSCLC at one of the 12 cancer treatment sites in NSW (study registration Research Data Australia, Identifier 004:273). These sites represented metropolitan (n = 7), regional and rural settings (n = 5) in both the public (n = 7) and the private sectors (n = 5). The study was approved by Human Research Ethics Committee (University of NSW, approval number 07014) and governance approvals were obtained at the 12 sites. Recruitment commenced in January 2009 and ceased in October 2010. To identify potentially eligible patients, field staff reviewed medical oncology records and then sought confirmation of eligibility by the treating physician. Patients already undergoing chemotherapy or about to commence treatment in the next week were then approached for written consent. Participation involved consent for medical record review, interview and data linkage with NSW and Commonwealth (Medicare Australia) secondary health data collections (Fig. 1a). Medical record reviews and interviews were conducted monthly to collect data on cancer treatment, health service utilization, socioeconomic status, use of complementary therapy and a range of other factors (data set available upon request). Medical record reviews continued until patients either withdrew from the study, ceased chemotherapy and did not recommence within 30 days, died, or the census date was reached (June 11, 2010 for the 2009 cohort and May 5, 2011 for the 2010 cohort). Regular interviews ceased 6 months after the recruitment date or when the above conditions were met. While extensive primary data were collected, the current report uses only four of these data items, specifically cancer type and stage, administered chemotherapy protocol and use of S100 drugs by public hospital patients.
Data analysis Data collected by field staff were cleaned by the investigators and statisticians, and data inconsistencies were resolved through face-to-face meetings between investigators, field staff and data analysts. The calculation of costs was based on secondary health data, specifically PBS (prescription medicines), MBS (Medicare services), NSW Admitted Patients Data Collection (APDC, hospital admissions) and NSW Emergency Department Data Collection (EDDC, emergency visits). The only primary data used for cost
© 2015 Wiley Publishing Asia Pty Ltd
180
RL Ward et al.
(a)
Allocated to cohort (n = 478)
2009 cohort (n = 310)
n =1
n =1
2010 cohort (n = 168)
n =58
n =31
n =137 Record review plus interview plus data linkage
Record review plus data linkage
Record review plus interview
Record review
n =250
Commonwealth data (n = 453) NSW health data (n = 404)
(b) 500 450
Number of parcipants
400 350 300 250 200 150 100 50 0 0
1
2
3
4
Duraon of review (months)
© 2015 Wiley Publishing Asia Pty Ltd
5
6
Figure 1 Number of patients in the Elements of Cancer Care (EoCC) study who consented to medical record review, data linkage and interview (a) and period of observation (b). The EoCC study recruited two patient cohorts, labeled 2009 and 2010, respectively. For the 2009 cohort, consent to either data linkage or interview was optional, while for the 2010 cohort, only consent to interview was optional. Data linkage to the four NSW Health data collections: NSW Central Cancer Registry (CCR),7 NSW Admitted Patients Data Collection (APDC),8 Registry of Births, Deaths and Marriages plus the Australian Bureau of Statistics Mortality Database (ABS-BDM), and the NSW Emergency Department Data Collection (EDDC) was performed by NSW Centre for Health Record Linkage (CHeReL). To perform data linkage to Medicare Australia data (Pharmaceutical Benefit Scheme [PBS] and Medicare Benefit Schedule [MBS]), an extract file of cohort members was sent to Medicare Australia. Of the 476 patients who consented to data linkage, 453 had at least one record in Commonwealth data (MBS or PBS) and 404 patients had linked data in NSW health records (APDC, EDDC, RBDM-ABS) during the follow-up period.
Asia-Pac J Clin Oncol 2015; 11: 178–186
181
Cost of cancer chemotherapy in Australia
calculations were the costs of S100 drugs for public hospital patients (the costs of these drugs were in PBS data for private patients). The total PBS and MBS costs were sums of government benefits and patient out-ofpocket costs. The costs of public patients’ hospital admissions were estimated by linking the Australian Refined Diagnosis Related Groups (AR-DRG) codes to their corresponding total average costs using national cost reports.9,10 The costs of private patients’ hospital admissions were estimated by linking the AR-DRG codes to their corresponding total average costs using Hospital Casemix Protocol annual reports.11–13 The cost of emergency department care was estimated by weighing the average cost of an emergency department presentation by departure status and triage category.14 A summary of the decision rules used to calculate costs for each funding agent (Commonwealth, State government, private health insurer, patient out-of-pocket) and by the type of expenditure is shown in Supporting Information Tables S1 and S2. The individual total cost of each cost component was calculated as the sum of all costs incurred from recruitment until the first of either cessation of chemotherapy, death or the census date. The individual mean monthly cost was calculated by dividing the total cost by the number of 28-day periods accrued during their followup. Total cost and costs by site and stage of cancer were reported as the mean monthly cost and as the 10th and 90th quantiles (to illustrate variability in costs across patients) of the mean cost per patient.
to hospital (APDC) and 350 visits to the emergency ward (EDDC). Using these data, the mean monthly cost of cancer care was calculated as $4162 (10%–90% quantiles $1018–$8098) with 54% of this cost borne by the Commonwealth government and 26% by private health insurers (Table 2). The mean monthly cost of treating metastatic disease was $1415 greater than the costs of adjuvant therapy ($4810, 10%–90% quantiles $1533–$8365 compared with $3395, 10%–90% quantiles $730–$7660). In aggregate, across all tumor types, the distribution of monthly treatment costs for adjuvant disease was 65% Commonwealth government, 8% State government, 21% private health insurers and 6% out-of-pocket expenses. The respective figures for metastatic disease were 47%, 19%, 29% and 5%. The mean monthly treatment cost was highly variable between patients, tumor types and extents of disease, from $5622 (10%–90% quantiles $2210–9338) for metastatic NSCLC to $2853 (10%–90% quantiles $735–8098) for adjuvant colorectal cancer. The distribution of costs by funding across tumor types was similar with two notable exceptions. Over 73% of the monthly treatment costs of adjuvant breast cancer were borne by the Commonwealth government as a consequence of a lesser contribution from the State government (reduced to 6% of the costs). For metastatic NSCLC, 32% of the costs were borne by the state government as a consequence of a reduced proportional contribution from private health insurers. For all tumor types, the out-of-pocket expenses represented only a modest contribution to the monthly treatment costs (∼6%).
RESULTS Profile of study participants
Costs of the individual care elements
A total of 478 patients consented to the study (50.5% of those approached) and over 80% agreed to record review, interview and data linkage. The period of observation for participants is shown in Figure 1b. The characteristics of the patients recruited in the 2009 and 2010 cohorts were very similar (see Table 1 and Supporting Information Table S3). The patients in this study received 148 unique chemotherapy protocols; however, most patients received one of the commonly used protocols for treatment in the adjuvant or metastatic settings (Supporting Information Table S4). The distribution of treatment between public and private centers was 49.6% and 50.4%, respectively.
In total, the largest proportion of mean monthly costs is related to inpatient care ($1657, 40%) and chemotherapy prescriptions ($1502, 36%), with lesser contributions from outpatient care ($452, 11%) and administration of chemotherapy ($364, 9%) (Table 3). The monthly inpatient costs for patients with metastatic disease substantially exceeded the costs incurred for patients on adjuvant treatment ($2269, 10%–90% quantiles $0–$5264, 47% compared with $934, 10%– 90% quantiles $0–$2820, 28%). Overall, the monthly cost of chemotherapy prescriptions was very similar for patients receiving adjuvant ($1539, 10%–90% quantiles $47–$4179) or metastatic treatment ($1471, 10%– 90% quantiles $76–$3135). In both the adjuvant and the metastatic settings, other care elements such as radiology, pathology, specialist and general practitioner visits accounted for only a small percentage of monthly
Costs of cancer care by financing agent Analysis of the linked data on 475 patients identified 46 854 MBS and 18 854 PBS claims, 3917 admissions
Asia-Pac J Clin Oncol 2015; 11: 178–186
© 2015 Wiley Publishing Asia Pty Ltd
182
RL Ward et al.
Table 1 Demographic and clinical baseline characteristics of the 2009 and 2010 cohorts derived from medical record review at recruitment Total cohort (n = 478) Variable Demographic variables Sex Men Women Age (years) ≤45 46–65 66–85 Treatment location Rural/Regional public Rural/Regional private Metropolitan public Metropolitan private Clinical variables Site and stage of cancer Breast Stage I Stage II Stage III Stage IV Colorectal Stage I Stage II Stage III Stage IV Non-small-cell lung cancer Stage I Stage II Stage III Stage IV Previous chemotherapy Yes No †
2009 cohort (n = 310)
†
†
2010 cohort (n = 168)
N (%)
N (%)
N† (%)
126 (26.4) 352 (73.6)
69 (22.3) 241 (77.7)
57 (33.9) 111 (66.1)
61 (12.8) 290 (60.7) 127 (26.6)
42 (13.5) 190 (61.3) 78 (25.2)
19 (11.3) 100 (59.5) 49 (29.2)
37 (7.7) 17 (3.6) 204 (42.7) 220 (46.0)
35 (11.3) Nil Nil 142 (45.8) 133 (42.9)
2 (1.2) 17 (10.1) 62 (36.9) 87 (51.8)
258 (54.0) 24 (5.0) 82 (17.2) 49 (10.3) 103 (21.5) 157 (32.8) 0 (0) 6 (1.3) 48 (10.0) 103 (21.5) 63 (13.2) 4 (0.8) 2 (0.4) 16 (3.3) 41 (8.6)
188 (60.7) 18 (9.6) 65 (34.6) 29 (15.4) 76 (40.4) 83 (26.8) 0 (0.0) 2 (2.4) 24 (28.9) 57 (68.7) 39 (12.6) 3 (7.7) 1 (2.6) 8 (20.5) 27 (69.2)
70 (41.7) 6 (8.6) 17 (24.3) 20 (28.6) 27 (38.5) 74 (44.0) 0 (0.0) 4 (5.4) 24 (32.4) 46 (62.2) 24 (14.3) 1 (4.2) 1 (4.2) 8 (33.3) 14 (58.3)
198 (42.0) 273 (58.0)
135 (44.5) 168 (55.5)
63 (37.5) 105 (62.5)
Due to missing observations, all numbers do not add up to the size of the cohort.
treatment costs (total $452, 10%–90% quantiles $114– $868, 11%). While there was marked variability in the monthly treatment costs between patients and between the three tumor types, the relative difference between adjuvant and metastatic inpatient treatment costs held true, with 1.6- to 2.6-fold higher mean monthly inpatient costs for patients on metastatic treatment. Of all costs examined, the greatest variability was for chemotherapy prescriptions, ranging from around $700 for adjuvant colorectal and NSCLC (10%–90% quantiles $18–$1374 and $0–$1706, respectively) to nearly
© 2015 Wiley Publishing Asia Pty Ltd
$2000 per month for adjuvant breast (10%–90% quantiles) and metastatic NSCLC (10%–90% quantiles $33–$3329).
DISCUSSION This multicenter cohort study showed that the Commonwealth government bore the largest proportion of the cost of care, followed by private health insurers, State government and patients. Admissions to hospital for purposes unrelated to chemotherapy administration were found to be the largest contributor to overall
Asia-Pac J Clin Oncol 2015; 11: 178–186
Commonwealth government costs based on benefits paid through the Pharmaceutical Benefits Scheme and Medicare Benefits Scheme as well as hospital treatment costs for residential aged care (based on Admitted Patients Data Collection data). ‡State government costs based on hospital treatment for public patients (based on Admitted Patients Data Collection) and emergency department visits (based on Emergency Department Data Collection). §PHI (Private health insurers) costs based on Medicare Benefits Scheme and hospital treatment (based on Admitted Patients Data Collection) for private patients covered by private health insurers. ¶OOP (out of pocket) cost based on Pharmaceutical Benefits Scheme, Medicare Benefits Scheme, and hospital costs (based on Admitted Patients Data Collection) paid by patients out-of-pocket. NSCLC, non-small-cell lung cancer.
†
6% 6% 6% 10% 6% 6% 3% 238 (24–545) 199 (26–422) 240 (21–565) 271 (33–622) 311 (27–738) 186 (17–602) 172 (18–465) 26% 15% 30% 36% 34% 35% 17% 1075 (0–3182) 541 (0–1350) 1219 (0–3426) 1032 (0–3694) 1770 (0–3881) 1144 (0–5100) 969 (0–2947) 14% 6% 16% 11% 14% 15% 32% 607 (0–1782) 225 (0–416) 647 (0–2059) 322 (0–548) 718 (0–2193) 493 (0–1123) 1770 (0–5531) 54% 73% 48% 43% 46% 44% 48% 2242 (488–4386) 2639 (563–5984) 1939 (557–3544) 1228 (374–2408) 2393 (548–4626) 1429 (321–2881) 2710 (733–4065) 476 149 108 54 102 13 50 Overall Breast adjuvant Breast metastatic Colorectal adjuvant Colorectal metastatic NSCLC adjuvant NSCLC metastatic
4162 (1018–8098) 3604 (714–7660) 4046 (1279–7633) 2853 (735–8098) 5191 (1846–8928) 3251 (492–7379) 5622 (2210–9338)
Patient out-of-pocket¶ (%) Private health insurers§ State government‡ Commonwealth government† Total N Cancer site and stage
Table 2 Mean (10%–90% quantiles) monthly (28 day) costs (in 2011 Australian dollars) of cancer care by funding agent, presented according to site and stage of cancer
Cost of cancer chemotherapy in Australia
Asia-Pac J Clin Oncol 2015; 11: 178–186
183
costs of care. The absolute monthly treatment costs of cancer care varied by patient, tumor type, the presence of metastatic disease and the extent of inpatient care. With the exceptions of adjuvant breast cancer and metastatic NSCLC settings, the proportional contribution of funders to cancer care was consistent. The variations for breast and NSCLC are related to the frequency of inpatient admissions and the selection of private health insurance at separation (as recorded in secondary health datasets). For instance, in adjuvant breast cancer, the low contribution of private health insurers to monthly treatment costs is related to low rates of inpatient admissions. In the case of NSCLC, the proportional lower contribution from private health insurers (17%) was matched by a greater proportional financing from State government. Patients with metastatic colorectal cancer had absolute monthly treatment and inpatient costs that were similar to patients with metastatic NSCLC, yet the higher proportional contribution by private insurers (34%) reflected the greater use of private health insurance. At the baseline interview, most patients reported having private insurance; however, comparison with secondary data found that selfreported private status was an unreliable measure of the use of private insurance, as this varied with health system encounter. Understanding the drivers for decisions to use private health insurance has important implications for both funders and planners of health services. Like other studies,15 inpatient hospitalizations, unrelated to admission for administration of chemotherapy, were the largest contributor to overall costs (40%) for all tumor types and settings, except adjuvant breast cancer. Inpatient hospitalizations reflected the management of side effects of chemotherapy and active cancer. These findings show that initiating chemotherapy clearly increases the cost of cancer care well beyond the cost of anticancer drugs and their administration alone. Thus there is a strong rationale for collecting real-world data on the risk and predicted duration of hospital admissions for different chemotherapy protocols, so that prescribers and patients can make better treatment choices. The higher costs for patients undergoing metastatic treatment for all tumor types when compared to adjuvant therapy were mostly explained by increased inpatient costs, likely related to management of cancer progression or its complications. Other studies have consistently found that the net costs of cancer care are highest immediately following diagnosis, and in the last year of life.15–17 Given the natural histories of metastatic
© 2015 Wiley Publishing Asia Pty Ltd
© 2015 Wiley Publishing Asia Pty Ltd Total
Breast adjuvant
11 (0–30) 3 (0–12) 27 (0–74)
0% 0% 1%
445 (43–1144)
147 (0–413)
0% 1%
0%
4%
9%
2%
7 (0–20) 52 (0–176)
7 (0–23)
133 (34–253)
80 (0–290)
227 (66–390)
20% 1795 (0–4567) 16% 452 (180–781)
8%
2% 357 (38–990)
135 (14–254)
0% 1%
0%
3%
2%
6%
6 (0–19) 22 (0–57)
7 (0–9)
109 (47–209)
0 (0–0)
149 (38–245)
511 (31–1331)
136 (8–330)
0% 1%
0%
4%
0%
5%
4 (0–14) 61 (0–165)
3 (0–9)
134 (39–236)
24 (0–0)
227 (62–377)
49% 2508 (0–5479) 9% 393 (143–625)
12%
5%
97 (11–344)
260 (26–625)
0% 1%
0%
3%
0%
4%
1 (0–0) 121 (0–395)
12 (0–29)
141 (41–317)
48 (0–109)
171 (20–356)
158 (0–336)
218 (61–468)
0% 4%
0%
4%
1%
5%
6 (0–23) 128 (0–313)
9 (0–17)
129 (57–203)
86 (0–338)
214 (37–376)
52% 2750 (0–6797) 11% 444 (183–902)
3%
8%
0% 2%
0%
2%
2%
4%
49% 8%
3%
4%
34%
NSCLC metastatic 22% 1924 (33–3329)
NSCLC adjuvant 708 (0–1706)
48% 1692 (0–6888) 8% 372 (112–672)
10%
3%
30%
Colorectal metastatic
24% 1583 (30–3838)
Colorectal adjuvant 673 (18–1374)
44% 1395 (0–5322) 11% 270 (128–519)
11%
4%
29%
Breast metastatic 53% 1154 (165–2314)
4162 (1018–8098) 100% 3604 (714–7660) 100% 4046 (1279–7633 100% 2853 (735–8098) 100% 5191 (1846–8928) 100% 3251 (492–7379) 100% 5622 (2210–9338) 100%
127 (41–257)
88 (20–167)
4%
3%
700 (0–1683) 569 (94–1091)
40% 11%
340 (0–824)
298 (14–852)
9%
3%
85 (0–219)
3%
36% 1925 (138–5114)
Costs to public and private patients for cancer treatment. Costs include chemotherapy-related† or non-chemotherapy‡ prescription medicines (based on Pharmaceutical Benefits Scheme); chemotherapy-related§ or non-chemotherapy-related hospitalization¶ (based on Admitted Patients Data Collection); outpatient medical services costs†† (based on Medicare Benefits Scheme); and emergency department visit costs‡‡ (based on Emergency Department Data Collection). GP, general practitioner; NSCLC, non-small-cell lung cancer.
Total
1502 (61–3549) Chemotherapy medicines† Other prescription 134 (0–343) medicines‡ Chemotherapy 364 (18–998) administration§ Inpatient care¶ 1657 (0–4826) 452 (114–868) Out-of-hospital care†† Radiology and 172 (28–340) pathology Radiation 139 (0–589) oncology Medical, GP and 129 (41–253) specialist visits Allied health and 8 (0–17) Nurse visits Surgery 5 (0–16) Emergency 53 (0–170) department care‡‡
Care element
Table 3 Mean (10%–90% quantiles) monthly (28 day) costs (in 2011 Australian dollars) of cancer care by care element, overall and according to site and stage of cancer
184 RL Ward et al.
Asia-Pac J Clin Oncol 2015; 11: 178–186
185
Cost of cancer chemotherapy in Australia
NSCLC and colorectal cancer, a significant proportion of patients entered this study in or close to their final year of life. Like US studies,15,18,19 the current work showed large inter-patient variations in the costs of chemotherapy independent of tumor site. Subsidy of new cancer drugs such as trastuzumab for adjuvant breast cancer, erlotinib for metastatic NSCLC and bevacizumab for metastatic colorectal cancer in part explained the high mean monthly treatment costs and variability in chemotherapy costs in these groups. However, we propose that new medications, tumor type and stage, and the extent of inpatient care did not account fully for the observed inter-patient variations in treatment costs. Differences in treatment practices, site of treatment, price paid for each element of care, and volume and duration of hospital admissions may also contribute to observed differences and will be the subject of future work. One of the strengths of this study is that by using primary and secondary data, costs were able to be reported by tumor stage and type. Second, enrollment was permitted at any point in the treatment course rather than only at initiation. This ensured a broad capture of the costs of adverse events which may vary in time of onset depending on duration of chemotherapy exposure. The study also has a number of limitations, including the fact that treatment costs were likely to be underestimated because only secondary health data sources were used to determine the cost of the elements of care. Omitted costs include direct patient costs, private scripts and drugs such as trastuzumab which were provided outside the PBS. The primary data showed 32 patients with metastatic breast cancer received trastuzumab, yet this drug cost was not captured at all in the secondary data sources. Another limitation of the study was that a participation bias (resulting largely from increased frequency and duration of clinic visits) led to an overrepresentation of breast cancer patients receiving protocols containing trastuzumab. Finally, it was not possible to reliably distinguish the monthly cancer specific treatments from the background of noncancer healthcare costs for all patients in this study. Despite these limitations, to our knowledge, the study provides the only snapshot of the elements and costs of cancer care for patients receiving chemotherapy for three common cancers in Australia. The data collected in this study will be made available to researchers for future studies of patterns of care, resource use and costs.
Asia-Pac J Clin Oncol 2015; 11: 178–186
ACKNOWLEDGMENT This study was funded by the National Health and Medical Research Council (Health Services Research Grant ID 455366).
REFERENCES 1 Warren JL, Yabroff KR, Meekins A, Topor M, Lamont EB, Brown ML. Evaluation of trends in the cost of initial cancer treatment. J Natl Cancer Inst 2008; 100: 888–97. 2 Smith TJ, Hillner BE. Bending the cost curve in cancer care. N Engl J Med 2011; 364: 2060–5. 3 Delivering High-Quality Cancer Care: Charting a New Course for a System in Crisis. The National Academies Press, 2013. 4 Australian Institute of Health and Welfare Health system expenditure on cancer and other neoplasms in Australia 2008–09. Canberra, 2013. 5 Experts in Chronic Myeloid Leukemia. The price of drugs for chronic myeloid leukemia (CML) is a reflection of the unsustainable prices of cancer drugs: from the perspective of a large group of CML experts. Blood 2013; 121: 4439– 42. 6 Lipscomb J. Estimating the cost of cancer care in the United States: a work very much in progress. J Natl Cancer Inst 2008; 100: 607–10. 7 Cancer Institute NSW. NSW Central Cancer Registry Data Access. 2012. Sydney. [Cited Nov 2013.] Available from: http://www.cancerinstitute.org.au/data-and-statistics/ cancer-registries/nsw-central-cancer-registry-data-access. 8 NSW Health. Health Statistics NSW – Admitted Patient Report. 2012. Sydney. [Cited Aug 2014.] Available from: http://www.health.nsw.gov.au/pubs/2012/admitted_patient _rpt2012.html. 9 The Australian Government Department of Health. Round 13. 2008–2009. Cost Report. 10 The Australian Government Department of Health. Round 14. 2009–2010. National Public Cost Weight Tables. 11 The Australian Department of Health. Hospital Casemix Protocol Annual Report. 2009–2010. Canberra. [Cited Aug 2014.] Available from: http://www.health.gov.au/ internet/main/publishing.nsf/Content/health-casemix-datacollections-publications-HCPAnnualReportsArchived. 12 The Australian Department of Health. Hospital Casemix Protocol Annual Report. 2010–2011. Canberra. [Cited Aug 2014.] Available from: http://www.health.gov.au/ internet/main/publishing.nsf/Content/health-casemix-datacollections-publications-HCPAnnualReports. 13 The Australian Department of Health. Hospital Casemix Protocol Annual Report. 2008–2011. Canberra. [Cited Aug 2014.] Available from: http://www.health.gov.au/ internet/main/publishing.nsf/Content/health-casemix-datacollections-publications-HCPAnnualReportsArchived.
© 2015 Wiley Publishing Asia Pty Ltd
186
14 New South Wales Health Costs of Care Standards 2009/ 10. Sydney. [Cited Aug 2014.] Available from: http:// www0.health.nsw.gov.au/policies/gl/2011/pdf/GL2011_ 007.pdf. 15 Yabroff KR, Lamont EB, Mariotto A et al. Cost of care for elderly cancer patients in the United States. J Natl Cancer Inst 2008; 100: 630–41. 16 Riley GF, Potosky AL, Lubitz JD, Kessler LG. Medicare payments from diagnosis to death for elderly cancer patients by stage at diagnosis. Med Care 1995; 33: 828– 41. 17 Langton JM, Blanch B, Drew AK, Haas M, Ingham JM, Pearson SA. Retrospective studies of end-of-life resource utilization and costs in cancer care using health administrative data: a systematic review. Palliat Med 2014; 28: 1167–96. 18 Meropol NJ, Schulman KA. Cost of cancer care: issues and implications. J Clin Oncol 2007; 25: 180–6.
© 2015 Wiley Publishing Asia Pty Ltd
RL Ward et al.
19 Ferro SA, Myer BS, Wolff DA et al. Variation in the cost of medications for the treatment of colorectal cancer. Am J Manag Care 2008; 14: 717–25.
SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article at the publisher’s web-site: Table S1 Decision rules for the calculation of the costs by funding agent. Table S2 Decision rules for the calculation of costs by care element. Table S3 Socioeconomic characteristics of 2009 and 2010 cohorts obtained from baseline recruitment interview. Table S4 The five most frequently administered chemotherapy protocols according to site and stage of cancer.
Asia-Pac J Clin Oncol 2015; 11: 178–186
Copyright of Asia Pacific Journal of Clinical Oncology is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
