Content uploaded by Lisa M Lines
Author content
All content in this area was uploaded by Lisa M Lines on May 03, 2018
Content may be subject to copyright.
Lifetime and Treatment-Phase Costs Associated With Colorectal Cancer:
Evidence from SEER-Medicare Data
KATHLEEN LANG,* LISA M. LINES,* DAVID W. LEE,
‡
JONATHAN R. KORN,* CRAIG C. EARLE,
§
and JOSEPH MENZIN*
*Boston Health Economics, Inc., Waltham, Massachusetts;
‡
GE Healthcare, Waukesha, Wisconsin; and
§
Harvard Medical School, Boston, Massachusetts
Background & Aims: This study provides detailed esti-
mates of lifetime and phase-specific colorectal cancer (CRC)
treatment costs. Methods: This retrospective cohort study
included patients aged 66 years and older, newly diagnosed
with CRC in a Surveillance Epidemiology and End Results
(SEER) registry (1996–2002), matched 1:1 (by age, sex, and
geographic region) to patients without cancer from a 5% sam-
ple of Medicare beneficiaries. The Kaplan–Meier sample av-
erage estimator was used to estimate observed 10-year costs,
which then were extrapolated to 25 years. A secondary analysis
computed costs on a per-survival-year basis to adjust for dif-
ferences in mortality by stage and age. Costs were expressed in
2006 US$, with future costs discounted 3% per year. Results:
Our sample included 56,838 CRC patients (41,256 colon can-
cer [CC] patients and 15,582 rectal cancer [RC] patients;
mean ⴞSD age, 77.7 ⴞ7.1 y; 55% women; and 86% white).
Lifetime excess costs were $29,500 for CC and $26,500 for RC
patients. Per survival year, stage IV CRC patients incurred
$31,000 in excess costs compared with $3000 for stage 0 pa-
tients. CRC patients incurred excess costs of $33,500 in the
initial phase, $4500/y in the continuing phase, and $14,500 in
the terminal phase. RC patients had lower costs than CC
patients in the initial phase, but higher costs in both the
continuing and terminal phases. Conclusions: Excess costs
associated with CRC are striking and vary considerably by treat-
ment phase, cancer subsite, and stage at diagnosis. Interventions
aimed at earlier diagnosis and prevention have the potential to
reduce cancer-related health care costs.
Currently available estimates of the cost of colorectal cancer
(CRC) vary widely and are outdated. This study extends
previous research by estimating lifetime and phase-specific CRC-
related costs by cancer subsite, age, and stage at diagnosis using
the most current data available. Representative CRC cost data are
needed because they are used to evaluate technologies aimed at
early detection and prevention.1Further, age-specific CRC cost
data are important because the aggressiveness of treatment and
screening guidelines are age-dependent. Finally, understanding
costs by stage at diagnosis is important from a public health
perspective because it affects economic evaluations of new colo-
rectal screening technologies.2,3 The goal of screening is to
affect a positive shift in stage at detection from late to early
stage. The economic benefit of this stage shift is measured in
the incremental cost reduction in detecting CRC earlier.
Methods
Data Sources
This study used 3 data sources: (1) the linked Surveil-
lance Epidemiology and End Results (SEER)–Medicare database
(a collaborative effort of the National Cancer Institute, the
SEER registries, and the Centers for Medicare and Medicaid
Services); (2) the SEER*Stat database, containing clinical and
survival data from the SEER registries; and (3) survival data for
the general population from US life-tables.
SEER is a US cancer surveillance system consisting of pop-
ulation-based tumor registries designed to track incidence and
survival. The registries routinely collect information from mul-
tiple reporting sources about newly diagnosed cancer patients
in geographically defined areas representing approximately 25%
of the US population.4Complete details of the linkage of the
SEER and Medicare data have been described elsewhere.5
Patient Selection
Colorectal cancer cohort. All patients aged 66 years
and older with a new diagnosis of malignant adenocarcinoma of
the colon or rectum (ie, presence of a SEER cancer site recode value
between 15 and 27 and one of the following International Classi-
fication of Diseases for Oncology, 3rd Edition (ICD-O-3) histology
codes: 8140, 8210-11, 8220-21, 8260-63, 8470, 8480-81, and 8490)
reported to a SEER registry between January 1, 1996, and Decem-
ber 31, 2002, were identified for possible inclusion in the CRC
cohort. The index date for each patient was defined as the date of
CRC diagnosis. Patients were required to have a full 12 months of
data available pre-index.
To ensure complete expenditure information for our sample,
patients were excluded if at any time in the period 12 months
before, or anytime after the index date, they were enrolled in a
Medicare HMO, not eligible for both Medicare Part A and B
benefits, or eligible for benefits under the end-stage renal dis-
ease program. We also excluded patients who had claims in the
12-month pre-index period indicating any other cancer, were
diagnosed with CRC at the time of death or autopsy, or could
not be matched to an appropriate comparator.
Comparison cohort. Comparison cohort patients
were selected from Medicare enrollment files using a 5% random
sample of Medicare beneficiaries residing in SEER areas who did
not have cancer. One comparison patient of identical age, sex, and
census region was matched randomly to each CRC patient and
assigned the same index date (so that both patients were followed
up over the same time period). As with CRC patients, comparators
were not eligible for inclusion if they were enrolled in an HMO or
were not eligible for Medicare Part A and B benefits at any point
from 12 months before index through follow-up evaluation. These
Abbreviations used in this paper: CC, colon cancer; CRC, colorectal
cancer; RC, rectal cancer.
©2009 by the AGA Institute
1542-3565/09/$36.00
doi:10.1016/j.cgh.2008.08.034
CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2009;7:198 –204
patients were not required to have used services to be selected for
inclusion, and they were allowed to develop cancers other than
CRC after their index date. Patients in both cohorts were followed
up from their index date until death or the end of the Medicare
claims data (December 31, 2005).
Statistical Analyses
Baseline demographic and clinical characteristics of
both study cohorts were evaluated, including Deyo–Charlson
comorbidity scores.6Health care costs included all Medicare
payments, private insurer payments, and patient copayments
and deductibles for covered services. Covered services included
inpatient hospital and skilled nursing facility stays, outpatient
hospital services, physician and laboratory services, home
health, and hospice care. Prescription drugs (including chemo-
therapy) administered in hospitals were included in inpatient
expenditures. Outpatient chemotherapy costs were included in
outpatient expenditures; oral prescription drugs administered
on an outpatient basis were excluded because they were not
covered by Medicare at the time of the study.
Excess costs attributable to CRC were defined as the difference
in costs between CRC and matched comparison patients. We used
this definition to minimize bias caused by coding inconsistencies
and omissions associated with relying on a sum of expenditures
for medical events with a cancer diagnosis code.7For example,
treatment for the side effects of cancer therapy should be included
in the cost of cancer treatment but may not be coded with a cancer
diagnosis. In addition, patients may be treated for cancer-related
conditions and other conditions in the same medical encounter, in
which case it is impossible to determine from the claims the cost
of the cancer-related portion of the encounter.
Estimation of lifetime costs. Patients in our SEER–
Medicare analysis were accrued between 1996 and 2002 and
were followed up through 2005 (the end of the SEER–Medicare
database), for a maximum of 10 years of follow-up evaluation,
a time horizon unlikely to be adequate for evaluating lifetime
costs among patients diagnosed with less-advanced cancer or
for comparison patients. Thus, modeling techniques for the
analysis of censored cost data were used to estimate lifetime
health care costs attributable to CRC.8–11
Costs for years 1 through 10 were estimated directly from
observable SEER–Medicare data for CRC patients and controls
using the nonparametric Kaplan–Meier Sample Average estima-
tor.12 The Kaplan–Meier Sample Average computes costs by sum-
ming expected costs incurred per time interval, calculated as the
product of the probability of surviving to that time interval, and
the sample average cost among survivors to the start of that
interval.
The Kaplan–Meier Sample Average estimator is calculated by
using the following formula:
C1⫽兺
t⫽1
120
PtCt
Where tis the post–index-date month, Ptis the survival proba-
bility, and Ctis the mean actual costs in period tamong
beneficiaries surviving to month t. The Kaplan–Meier Sample
Average estimator minimizes the bias associated with censored
data by dividing the time period into short intervals. This
calculation provides a nonparametric estimate of the average
costs for patients with variable lengths of follow-up evaluation.
Costs for years 11 through 25 were extrapolated for both
cohorts using the assumption that cohort-specific average an-
nual medical care costs were constant for years beyond the
available data until the year before the final year of life (ie,
continuing costs). These continuing-phase costs were estimated
from the subset of patients in each cohort who lived at least 3
years as the average annual cost for years between year 1 and the
final year of life, exclusive. In addition, we assumed that medical
care costs in the final year of life (ie, terminal costs) for patients
who lived beyond the 10-year study period were the same,
regardless of time from diagnosis; terminal costs were thus
estimated as the average final-year cost among cases in the
relevant cohort who died at least 2 years after the index date.
We used the following formula to estimate costs in years 11
to 25:
C2⫽兺
y⫽11
25
P
ˆyCy
ˆ
Here, Py
ˆis the probability of surviving to year y. For CRC
patients, this survival probability was estimated for years 11 to
16 using SEER*Stat data and for years 17 to 25 by fitting SEER
data from 1988 to 2004 to a Weibull model. US life tables from
the National Center for Health Statistics13 provided survival
probabilities for controls. Cy
ˆis the expected cost in year y,
computed as a weighted average of the annual cost for cases
dying in year y(ie, terminal cost) and those surviving through
year y(ie, continuing cost). Total lifetime costs in each
cohort are calculated by summing cohort-specific estimates
of C1and C2.
Lifetime excess costs among CRC patients were reported
overall and per year of survival, by stage, age at diagnosis, and
cancer subsite, in 2006 US dollars, with future costs discounted
at 3% per year. Estimates of lifetime costs per survival year were
calculated by dividing the estimate of expected lifetime costs for
each age- and stage-specific group by the expected years of
survival for that group.
Phase-specific cost estimates for the entire colo-
rectal cancer population. Our lifetime cost estimation re-
quired estimates of continuing and terminal costs for CRC
patients who survived at least 10 years. For completeness, we
also estimated phase-specific costs for the entire CRC popula-
tion (ie, short-term and long-term survivors). These costs were
estimated as follows: (1) terminal costs were assigned first and
were calculated as the average cost in the final year of life, with
all costs considered terminal for patients living fewer than 13
months after diagnosis; (2) initial costs were calculated as the
average costs in the initial (up to 1 year) period after diagnosis
and before the last year of life and were calculated among those
who lived at least 13 months after diagnosis; and (3) continuing
costs included the period between the first and last year of life
for patients with at least 25 months of survival, and were
reported on a per-year basis.
Results
Patient Characteristics
We identified 56,838 CRC patients (41,256 colon cancer
[CC], 15,582 rectal cancer [RC]) who met our selection criteria.
Demographic and clinical characteristics for CC patients, RC pa-
February 2009 CRC COSTS 199
tients, the combined CRC cohort, and the comparison cohort are
presented in Table 1. The mean ⫾SD age was 77.7 ⫾7.1 years;
about 55% of patients in both cohorts were women and 86% were
white.
Lifetime Cost Estimates
Total lifetime cancer-related costs were $28,500, with
an inverted U-shaped pattern by stage and a U-shaped pat-
tern by age (Table 2). Excess costs for RC patients were
somewhat higher than costs for CC patients for stages I to III
and substantially lower for stage 0. CC costs were $42,000 for
stage 0, $45,000 for stage I, $43,000 for stage II, and $41,000 for
stage III.
Among RC patients, excess costs for stages I to III were
approximately 50% greater than for stage 0 ($47,000 for stage I,
$49,000 for stage 2, and $46,500 for stage III vs $30,000 for
Table 1. Baseline and Demographic Characteristics of Patients With CRC and Matched Comparison Patients
Variable CC patients RC patients Combined CRC cohort Comparison cohort
N 41,256 15,582 56,838 56,838
Age, ya
Mean (⫾SD) 77.9 (7.1) 77.1 (7.1) 77.7 (7.1) 77.7 (7.1)
Median 77.0 76.0 77.0 77.0
Interquartile range 72–83 71–82 72–83 72–83
Femalea57.0% 50.5% 55.2% 55.2%
Race/ethnicity
White 85.6% 86.3% 85.8% 86.2%
African American 8.2% 6.6% 7.7% 6.7%
Hispanic 1.2% 1.4% 1.3% 2.1%
Other 5.0% 5.7% 5.2% 5.0%
Geographic regiona
Northeast 21.9% 22.4% 22.0% 22.0%
Midwest 24.0% 23.3% 23.8% 23.8%
West 40.1% 40.8% 40.3% 40.3%
South 14.0% 13.5% 13.9% 13.9%
Location of residence
Metropolitan county 82.6% 81.8% 82.4% 82.6%
Nonmetropolitan county 17.4% 18.2% 17.6% 17.3%
Missing 0.0% 0.0% 0.0% 0.1%
Year of CRC diagnosis
1996 10.0% 10.3% 10.1% b
1997 10.3% 10.6% 10.4% b
1998 10.2% 10.5% 10.3% b
1999 9.9% 10.0% 9.9% b
2000 19.8% 20.4% 20.0% b
2001 19.7% 19.4% 19.6% b
2002 20.0% 18.7% 19.6% b
Stage at diagnosis
Stage 0 6.7% 8.1% 7.1% b
Stage I 22.2% 27.6% 23.7% b
Stage II 30.5% 20.7% 27.8% b
Stage III 22.0% 18.2% 21.0% b
Stage IV 14.5% 16.5% 15.1% b
Unknown 4.1% 8.9% 5.4% b
Charlson scorec
Mean (⫾SD) 1.9 (1.8) 1.7 (1.8) 1.8 (1.8) 1.7 (1.9)
Median 1.0 1.0 1.0 1.0
Interquartile range 0–3 0–3 0–3 0–3
Selected Charlson comorbidities (%)
Chronic pulmonary/respiratory disease 33.2% 32.1% 32.9% 28.8%
Congestive heart failure 32.7% 28.4% 31.5% 28.3%
Diabetes without complications 27.9% 24.9% 27.1% 24.3%
Cerebrovascular disease 21.6% 19.0% 20.9% 24.2%
Myocardial infarction 15.3% 14.5% 15.1% 14.3%
Peptic ulcer 9.4% 7.4% 8.9% 6.8%
Other major conditionsd33.3% 29.2% 32.2% 33.9%
Data from SEER–Medicare database, 1996 –2005.
aVariables used in matching cohorts.
bCharacteristics do not apply.
cModified Charlson comorbidity index6excluding cancer-related comorbidities.
dOther major conditions include rheumatologic disease, mild liver disease, diabetes with complications, major liver disease peripheral vascular
disease, dementia, renal disease, hemiplegia or paraplegia, and acquired immune deficiency syndrome.
200 LANG ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 7, No. 2
stage 0). Excess costs among stage IV patients were negative,
reflecting patients’ shorter life expectancy and the incorpora-
tion of future medical costs of comparison patients who outlive
CRC patients.
On a per-survival-year basis, excess CRC costs were approx-
imately 9 times greater for patients diagnosed at stage IV versus
stage 0 (Figure 1). Excess CC costs per survival year were
approximately 3 times greater for stage IV than for stage III.
Excess RC costs per survival year were approximately 2 times
greater for stage IV than for stage II. Compared with costs for
CC, costs for RC were similar for stages 0 and I, somewhat
higher for stages II to III, and substantially lower for stage IV.
Across all stages, excess costs per year of survival were highest
among the oldest patients.
Table 3 reports lifetime cancer-related health care costs per
year of survival for combinations of age and stage at diagnosis.
Within stage, costs increase monotonically with age up to stage
III, but the difference is more pronounced in early stage disease.
Within age, costs increase monotonically with stage, with the
increase most pronounced in younger patients.
For the combined CRC cohort, costs for the 85⫹group were
approximately 50% and 30% higher than costs for the 66 to 74
Table 2. Excess Cancer-Related Lifetime Health Care Costs
(2006 US$) by Cancer Subsite, Stage, and Age at
Diagnosis
CC patients RC patients
Combined
CRC cohort
All stages $29,420 $26,544 $28,626
Stage 0 $42,127 $29,983 $38,155
Stage I $45,094 $46,703 $45,435
Stage II $42,847 $49,020 $44,311
Stage III $41,050 $46,614 $42,437
Stage IV ⫺$7428 ⫺$18,770 ⫺$10,864
Unknown/unstaged $25 ⫺$510 ⫺$474
All ages $29,420 $26,544 $28,626
Age 66–74 $36,226 $36,790 $36,401
Age 75–84 $22,815 $16,726 $21,167
Age 85⫹$27,309 $12,960 $23,799
NOTE. Survival probabilities beyond year 10 were estimated with
SEER*Stat data (stage-specific survival data projected beyond year
16 with Weibull models). Future costs were discounted at 3% per year
Data from SEER-Medicare database, 1996-2005.
Figure 1. Excess lifetime cancer-related health care costs (2006 US$)
per year of survival, by cancer subsite, and stage at diagnosis.
Table 3. Excess Lifetime Cancer-Related Health Care Costs (2006 US$) per Year of Survival and by Age and Stage at Diagnosis
All ages Age 66–74 Age 75–84 Age 85⫹
Stage at
diagnosis CC patients RC patients
Combined CRC
cohort CC patients RC patients
Combined CRC
cohort CC patients RC patients
Combined CRC
cohort CC patients RC patients
Combined CRC
cohort
All stages $8909 $8759 $8853 $8521 $9195 $8713 $9397 $9923 $9516 $12,917 $11,602 $12,639
Stage 0 $3348 $2975 $3210 $2629 $2641 $2551 $4267 $4674 $4308 $9350 $5038 $8027
Stage I $4197 $4363 $4238 $3744 $4554 $3954 $4875 $5206 $4988 $6667 $5848 $6419
Stage II $6337 $8401 $6723 $5849 $8412 $6498 $6283 $9136 $6804 $9333 $11,359 $9586
Stage III $10,192 $11,517 $10,516 $10,232 $12,578 $10,867 $10,428 $12,225 $10,926 $13,082 $14,178 $13,221
Stage IV $33,033 $25,455 $30,794 $36,117 $28,790 $34,622 $30,913 $25,335 $29,861 $28,224 $19,992 $25,888
Unknown/
unstaged
$7670 $4804 $6225 $7954 $7004 $7650 $7255 $4750 $6431 $12,941 $4239 $9250
NOTE. Survival probabilities beyond year 10 were estimated with SEER*Stat data (stage-specific survival data projected beyond year 16 with Weibull models). Future costs were discounted at
3% per year.
Data from SEER–Medicare database, 1996 –2005.
February 2009 CRC COSTS 201
and 75 to 84 age groups, respectively. For the CC cohort, costs
for the 85⫹age group were approximately 50% and 40% higher
than costs for the 66 to 74 and 75 to 84 age groups, respectively.
For the RC cohort, costs for the 85⫹age group were approxi-
mately 20% and 10% higher than costs for the 66 to 74 and 75
to 84 age groups, respectively.
Across age groups, RC patients had less variation in costs
than CC patients. Compared with CC patients, RC patients had
higher costs for the 66 to 74 and 75 to 84 age groups and lower
costs for the 85⫹age group.
Phase-Specific Cost Estimates for the Entire
Colorectal Cancer Population
Examining both short- and long-term survivors, CRC-
related initial phase costs were approximately $33,000, excess
continuing phase costs were about $4500 per year, and excess
terminal costs were more than $14,000 (Table 4). Initial-phase
costs were slightly higher for CC patients, whereas continuing-
phase costs were roughly one-third higher for RC versus CC and
terminal costs were higher for RC patients by a small margin.
Phase-specific costs were much higher for stages III and IV
compared with stages 0 to II for all phases. However, little can
be inferred from the difference in stage IV because these pa-
tients are treated continually for active disease. In all phases,
costs were highest for the 66 to 74 age group and lowest for the
85⫹age group. The biggest difference in costs by age occurred
in the continuing phase, in which the 66 to 74 age group had
costs approximately two-thirds higher than that of the 75 to 84
age group.
Discussion
This study evaluated lifetime and phase-specific excess
costs among elderly patients with CRC in the United States. We
found that lifetime CRC-related costs are substantial and vary
by cancer subsite, stage at diagnosis, age at diagnosis, and
treatment phase. Excess lifetime costs show an inverted U-
shaped pattern by stage at diagnosis, and a U-shaped pattern by
age at diagnosis for both CC and RC. Costs for RC patients are
lower than costs for CC patients in stage 0, higher in stages I to
III, and lower in stage IV. On a per-survival-year basis, costs are
substantially higher for both RC and CC patients diagnosed in
stage IV versus any other stage. RC patients had lower costs
than CC patients in the initial phase, but higher costs than CC
patients in both the continuing and terminal phases.
Our findings regarding excess costs associated with CRC are
broadly consistent with 2 previous studies. Etzioni et al9ana-
lyzed excess lifetime costs of care for CRC patients using data
from SEER–Medicare (1986–1994) using a methodology simi-
lar to ours (ie, subtracting total lifetime costs for a noncancer
control group from lifetime costs for CRC patients). Brown et
al11 used 5 years of SEER–Medicare data (1990–1994) to esti-
mate phase-specific (initial, continuing, and terminal) costs and
to project total lifetime costs associated with CRC, not account-
ing for future medical costs that would have been incurred in
the absence of cancer. Because their study did not account for
future medical costs, our estimates of excess costs are somewhat
lower. Total costs in our study for both cancer patients and
comparison patients are higher for all stages and cancer sub-
sites, which we would expect given our use of data that includes
more recent advances in CRC treatment (eg, irinotecan) and the
inflation in the cost of health care services that has occurred
during the decade between data sources.
One previous study found much higher estimates of CRC-
related direct medical costs.14 This study used an administrative
claims database, which included patients insured by private or
Medicare supplemental health plans. The investigators found
that Medicare beneficiaries had considerably lower monthly
expenditures than patients with commercial insurance, likely
explaining our difference in findings. An additional study
found that 41% to 55% of patients who were diagnosed with
CRC more than 5 years ago were still receiving treatment, well
past the time when CRC patients are traditionally thought of as
Table 4. Initial, Continuing, and Terminal Cancer-Related Health Care Costs (2006 US$) by Cancer Subsite, Stage, and Age
InitialaContinuing (per year)bTerminalc
CC RC All patients CC RC All patients CC RC All patients
All patients $33,520 $32,683 $33,294 $3927 $5254 $4280 $14,410 $14,878 $14,538
Stage
Stage 0 $18,052 $13,954 $16,762 $2374 $1744 $2175 $7103 $7161 $7121
Stage I $27,783 $25,659 $27,099 $2347 $3341 $2665 $7774 $9641 $8371
Stage II $35,055 $40,217 $36,092 $2750 $5126 $3216 $11,731 $16,741 $12,755
Stage III $41,222 $43,518 $41,796 $5768 $7142 $6109 $18,417 $20,255 $18,854
Stage IV $42,401 $39,436 $41,562 $19,987 $22,039 $20,582 $27,898 $20,625 $25,714
Unknown $27,841 $28,500 $28,132 $3737 $6179 $4788 $12,785 $12,147 $12,496
Age, y
66–74 $33,980 $35,002 $34,282 $5334 $6828 $5775 $16,788 $16,491 $16,699
75–84 $33,401 $31,741 $32,967 $3277 $4166 $3503 $13,457 $14,273 $13,675
85⫹$32,966 $27,982 $31,869 $2261 $3534 $2502 $12,273 $12,571 $12,346
Data from SEER–Medicare database, 1996 –2005.
aInitial costs were defined as average costs in the initial (up to 1 year) period after diagnosis and before the last year of life and were calculated
among only those who lived at least 13 months after diagnosis.
bContinuing costs were defined as average annual costs in years beyond the initial year and before the last year of life and were calculated among
only those patients who lived at least 25 months after diagnosis.
cTerminal costs were defined as average costs in the final year of life (all costs are considered terminal for patients living ⬍13 months after
diagnosis).
202 LANG ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 7, No. 2
cured.15 This reinforces our findings that continuing costs are
nearly $3500 higher for CRC patients than for the comparison
cohort.
One recently published study of the phase-specific costs of
cancer care using SEER–Medicare data from 1999 to 2003
found similar costs for the initial year of life ($32,101 for male
CRC patients vs our estimate of $33,500 for all CRC patients).16
However, because of the methods used in matching terminal-
phase cancer patients to continuing-phase controls, their esti-
mates of terminal-phase costs were much higher than ours
($39,544 vs $14,500). Our continuing-phase cost estimate
($4500) was slightly higher than theirs ($2444 for men), possi-
bly because of a different approach to annualizing the data.
Our results are consistent with previous studies17–19 showing
that the costs of the last year of life are lower for older patients
than for younger patients, reflecting the fact that older patients
receive less aggressive treatment. This fact is important to
remember when calculating cost-effectiveness ratios, especially
for preventive care. When considering that terminal costs de-
crease with age, costs per quality-adjusted life years for preven-
tive interventions may be lowered by a substantial amount for
some patient groups.17
Our study contributes to the existing literature on costs
associated with CRC by evaluating the latest data, thus reflect-
ing some of the recent changes in treatment patterns for CRC,
such as the introduction of irinotecan and more use of multi-
modality therapy. This study reports colorectal cancer costs per
survival year.
The substantially higher costs for later-stage cancers may
indicate that these cancers receive more drastic or expensive
treatments. If this is the case, it may be important to focus on
earlier detection of CRC as a way to reduce medical expendi-
tures. In addition, per-survival-year costs are lower for younger
patients than for older patients. However, phase-specific costs
are higher for younger patients, perhaps reflecting more aggres-
sive treatments used in younger patients.
Although lifetime costs for RC are slightly lower than costs
for CC patients for all stages combined, costs vary considerably
by stage at diagnosis. For the later-stage cancers, RC patients
have substantially higher costs than CC patients. Our results
suggest that any discussion involving CRC patients would be
improved by focusing on RC and CC patients separately.
Our findings will be useful in shaping policy discussions
regarding CRC treatment, costs, and insurance coverage. In
addition, our findings regarding stage- and phase-specific costs
for the entire CC and RC populations can be used in economic
models such as cost-effectiveness models evaluating the poten-
tial impact of new technologies to detect and treat CRC.
This study is subject to certain limitations that are common
to all studies that rely on retrospective claims data, such as
potential coding errors and incomplete data.20 Our use of the
SEER–Medicare database, which includes complete claims only
for Medicare-eligible patients aged 65 years and older, intro-
duces additional limitations.10,21 Although the elderly comprise
the majority of patients with CRC, this sample is not represen-
tative of all US patients, particularly those with other forms of
health insurance (eg, managed care, private pay). Despite its
limitations, SEER–Medicare data have been used in numerous
published studies of colon cancer, as well as cancers of the
breast, prostate, and lung, among others.22
Only services covered by Medicare were included in this
analysis, and at the time of this study, coverage for most oral
prescription medications was not included. Because claims are
available only through 2005, the most recent changes in CRC
treatment patterns were not captured in the data. Until re-
cently, a combination of fluorouracil and leucovorin was the
standard of care for CRC; numerous other drugs have been
approved for use in CRC patients since 2004, including oxali-
platin, bevacizumab, cetuximab, and panitumumab.23 These
new cytotoxic and biologic agents used in cancer treatment are
among the most expensive technologies in medical care, and the
costs of these agents are substantially greater than costs for
older drugs. The impact of more expensive drugs on the cost of
CRC remains to be seen. If treatment is getting more expensive,
our estimates of CRC-related costs may be lower than the
current costs associated with the disease.
In conclusion, this study showed that excess costs associated
with CRC are striking and vary considerably by treatment
phase, cancer subsite, and stage at diagnosis. Within each treat-
ment phase and on a per-survival-year basis, costs increase
substantially for later-stage diagnoses. Interventions aimed at
prevention and earlier detection of CRC have the potential to
yield sizable economic benefits.
References
1. Winawer SJ, Zauber AG, O’Brien MJ, et al. Randomized compar-
ison of surveillance intervals after colonoscopic removal of newly
diagnosed adenomatous polyps. The National Polyp Study Work-
group. 1993. Available at: http://content.nejm.org/cgi/content/
full/328/13/901. Accessed: February 21, 2008.
2. Frazier AL, Colditz GA, Fuchs CS, et al. Cost-effectiveness of
screening for colorectal cancer in the general population. JAMA
2000;284:1954 –1961.
3. Song K, Fendrick AM, Ladabaum U. Fecal DNA testing compared
with conventional colorectal cancer screening methods: a deci-
sion analysis. Gastroenterology 2004;126:1270 –1279.
4. National Cancer Institute. About SEER. 2008. Available at:
http://seer.cancer.gov/about/. Accessed: January 9, 2008.
5. Warren JL, Klabunde CN, Schrag D, et al. Overview of the SEER-
Medicare data: content, research applications, and generalizabil-
ity to the United States elderly population. Med Care 2002;
40(Suppl):3–18.
6. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity
index for use with ICD-9-CM administrative databases. J Clin
Epidemiol 1992;45:613– 619.
7. Howard DH, Molinari N, Thorpe KE. National estimates of medical
costs incurred by nonelderly cancer patients. Cancer 2004;100:
891.
8. Ramsey SD, Berry K, Etzioni R. Lifetime cancer-attributable cost
of care for long term survivors of colorectal cancer. Am J Gastro-
enterol 2002;97:440 – 445.
9. Etzioni R, Ramsey SD, Berry K, et al. The impact of including
future medical care costs when estimating the costs attributable
to a disease: a colorectal cancer case study. Health Econ 2001;
10:245–256.
10. Brown ML, Riley GF, Schussler N, et al. Estimating health care
costs related to cancer treatment from SEER-Medicare data. Med
Care 2002;40(Suppl):104 –117.
11. Brown ML, Riley GF, Potosky AL, et al. Obtaining long-term dis-
ease specific costs of care: application to Medicare enrollees
diagnosed with colorectal cancer. Med Care 1999;37:1249 –
1259.
12. Etzioni R, Urban N, Baker M. Estimating the costs attributable to
February 2009 CRC COSTS 203
a disease with application to ovarian cancer. J Clin Epidemiol
1996;49:95–103.
13. Arias E. United States life tables, 2002. Natl Vital Stat Rep
2004;53:1–38.
14. Chang S, Long SR, Kutikova L, et al. Estimating the cost of
cancer: results on the basis of claims data analyses for cancer
patients diagnosed with seven types of cancer during 1999 to
2000. J Clin Oncol 2004;22:3524 –3530.
15. Mariotto A, Warren JL, Knopf KB, et al. The prevalence of patients
with colorectal carcinoma under care in the US. Cancer 2003;
98:1253–1261.
16. 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 – 641.
17. Gandjour A, Lauterbach KW. Does prevention save costs? Con-
sidering deferral of the expensive last year of life. J Health Econ
2005;24:715–724.
18. Levinsky NG, Yu W, Ash A, et al. Influence of age on Medicare
expenditures and medical care in the last year of life. JAMA
2001;286:1349 –1355.
19. Yang Z, Norton EC, Stearns SC. Longevity and health care expen-
ditures: the real reasons older people spend more. J Gerontol B
Psychol Sci Soc Sci 2003;58:S2–S10.
20. Schneeweiss S, Avorn J. A review of uses of health care utiliza-
tion databases for epidemiologic research on therapeutics. J Clin
Epidemiol 2005;53:337.
21. Izquierdo JN, Schoenbach VJ. The potential and limitations of
data from population-based state cancer registries. Am J Public
Health 2000;90:695– 698.
22. Surveillance Epidemiology and End Results. SEER bibliography.
Available at: http://seer.cancer.gov/cgi-bin-pubsearch/pubsearch/
index.pl#results. Accessed: February 6, 2008.
23. National Cancer Institute. Colon cancer treatment. 2008. Avail-
able at: http://www.cancer.gov/cancertopics/pdq/treatment/
colon/healthprofessional. Accessed: January 29, 2008.
Address requests for reprints to: Joseph Menzin, PhD, Boston Health
Economics, Inc., 20 Fox Road, Waltham, Massachusetts 02451.
e-mail: jmenzin@bhei.com; fax: (781) 290-0029.
The authors disclose the following: This study was sponsored by a
grant from GE Healthcare, Waukesha, WI. D.W.L. is an employee of GE
Healthcare; K.L., L.M.L., J.R.K., and J.M. received research funding
from GE Healthcare; and C.C.E. is a consultant for Boston Health
Economics.
Portions of this study were presented in preliminary form at the 13th
Annual Meeting of the International Society for Pharmacoeconomics
and Outcomes Research, May 5, 2008, Toronto, Ontario, Canada; and
the 44th Annual Meeting of the American Society of Clinical Oncology,
June 3, 2008, Chicago, IL.
The authors gratefully acknowledge Rick deFriesse, MEd, for assis-
tance with SAS programming, and David J. Vanness, PhD, for helpful
comments on earlier versions of this work.
204 LANG ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 7, No. 2
