- Email: [email protected]
Attributable costs of differentiated thyroid cancer in the elderly Medicare population
Attributable costs of differentiated thyroid cancer in the elderly Medicare population Melissa M. Boltz, DO, MBA,a,b Christopher S. Hollenbeak, PhD,a,b,c Eric Schaefer, MS,c David Goldenberg, MD,b,d and Brian D. Saunders, MD,b,e Hershey, PA
Background. Little is known about costs associated with differentiated thyroid cancer (DTC) and followup care. This study used data from the Surveillance Epidemiology and End Results (SEER) database to examine cumulative costs attributable to disease stage and treatment options of DTC in elderly patients over 5 years. Methods. We identified 2,823 patients aged >65 years with DTC and 5,646 noncancer comparison cases from SEER Medicare data between 1995 and 2005. Cumulative costs were obtained by estimating average costs/patient in each month up to 60 months after diagnosis. We performed multivariate analyses of costs by fitting each monthly cost to linear models, controlling for demographics and comorbidities. Marginal effects of covariates were obtained by summing coefficients over 60 months. Results. Cumulative costs were $17,669/patient the first year and $48,989/patient 5 years after diagnosis. Regional disease was associated with higher costs at 1 year ($9,578) and 5 years ($8,902). Distant disease was associated with 1-year costs of $28,447 and 5-year costs of $20,103. Patients undergoing surgery and radiation had a decrease in cost of $722 at 5 years. Conclusion. DTC in the elderly is associated with significant economic burden largely attributable to patient demographics, stage of disease, and treatment modalities. (Surgery 2013;154:1363-70.) From the Division of Outcomes Research and Quality, Department of Surgery,a Department of Surgery,b Division of Otolaryngology,d and Division of General Surgery Specialties and Surgical Oncology,e Penn State Milton S. Hershey Medical Center, Hershey; and Department of Public Health Sciences,c Penn State College of Medicine, Hershey, PA
THE INCIDENCE OF THYROID CANCER in the United States has more than doubled since the early 1970s. The yearly incidence has increased from 3.6 per 100,000 in 1973 to 8.7 per 100,000 in 2002.1 Although the incidence is increasing
Supported by a grant obtained from the Department of Health Commonwealth Universal Research Enhancement (CURE) program (Tobacco Settlement Act, Act 2001-77). This study used the linked SEER-Medicare database. The interpretation and reporting of these data are the sole responsibility of the authors. The authors acknowledge the efforts of the Applied Research Program, NCI; the Office of Research, Development and Information, CMS; Information Management Services (IMS), Inc; and the Surveillance, Epidemiology, and End Results (SEER) Program tumor registries in the creation of the SEER-Medicare database. Accepted for publication June 26, 2013. Reprint requests: Christopher S. Hollenbeak, PhD, Department of Surgery, The Pennsylvania State University, College of Medicine, 500 University Drive, H151, Hershey, PA 17033. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2013 Mosby, Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2013.06.042
steadily, thyroid cancer is still relatively uncommon because there were only 56,460 cases diagnosed in 2012.2 The majority of disease is differentiated thyroid cancer (DTC), which includes papillary and follicular cancers, and comprises approximately 90% of all thyroid cancers.1 In addition, the incidence of thyroid cancer also increases with age. Elderly patients often present with more aggressive forms of thyroid cancer, larger tumors, more extensive local growth, or distant metastases.3-5 However, the literature regarding outcomes in this age group remains scarce. The treatment of low-risk DTC is controversial despite the management guidelines recommended by the American Thyroid Association (ATA), which were first developed in 1996 and most recently updated in 2009.1 Given the long-term survival of patients with DTC, the increasing cost of health care, and the need for cost-containment, treatment strategies have become increasingly conservative. Follow-up and treatment of well-differentiated thyroid cancer extends for years beyond the initial operative procedure. SURGERY 1363
1364 Boltz et al
Although long-term survival of patients with DTC is a common occurrence, little is known about the cost attributable to the disease and follow-up care. The objective of this study was to use data from the Surveillance Epidemiology and End Results (SEER) and Medicare linked databases6 to study the costs attributed to the stage of disease and treatment options of DTC in elderly patients, as well as the cumulative cost of the disease over 5 years. METHODS Data. Data for this study were from the SEERMedicare linked data from the National Cancer Institute.7 The SEER program is a national tumor registry that collects data on approximately 26% of the U.S. population and reports the incidence and survival of cancer in the United States.8 SEERMedicare linked data contains the subset of Medicare enrollees in the SEER registry. It then links these patients to all Medicare claims records. Thus, by the use of Medicare billing claims data it is possible to identify specific types of resource utilization in the combined SEER-Medicare data.9 The SEERMedicare database also includes a group of patients without cancer that can be used for comparison purposes.7 Study population. Our analysis was limited to elderly adult Medicare enrollees (66 years and older) diagnosed with a single primary thyroid cancer (papillary or follicular). Claims data were only available for certain types of services since 1994. Therefore, we limited our sample to patients who were first diagnosed between 1995 and 2005. In addition, patients who were enrolled in a health maintenance organization at or after the cancer diagnosis were excluded because these patients do not have any Medicare billing records. The final analysis data set contained 2,823 patients. A comparison group sample of Medicare patients without cancer was selected from the Medicare 5% sample. To control for potentially different distributions of patient characteristics, propensity score matching was used to select a comparison sample that had a similar distribution of background characteristics as the thyroid cancer patients. We chose a propensity score matching approach because there was a large number of characteristics on which to match, including age, sex, race/ethnicity, and several comorbidities. An additional advantage of this technique is that the statistical analysis does not require adjustment for pairing as do exact matching techniques. The propensity score was the estimated probability in a logistic regression of case status (N = 2,823) versus
Surgery December 2013
comparison group status (N = 422,259). Covariates included year of birth (<1900, 5-year increments 1900 1934, $1935), sex, and race/ethnicity (white, black, Asian, other). An interaction between sex and race was also included because it was the only interaction that was clinically important. The predicted probability of being a case was computed for each subject and used as a propensity score. Two controls were matched to each case with the exact propensity score. Hence, each case was matched to two controls with the same sex and race, and born within the same 5year window. With these criteria the final sample included 2,823 patients with papillary or follicular thyroid cancer, and 5,646 matched noncancer comparison group cases. Costs. The cost analyses took the perspective of Medicare as payer and represented actual payments made by Medicare for all-cause treatment of patients. All costs were adjusted for inflation to year 2009 dollars. Billing files were used to create monthly costs for 5 years after the primary thyroid cancer diagnosis. All billing records, regardless of whether the bills were directly for treating cancer, were accumulated in each of the 60 months. For the month of diagnosis, bills from 1 month before diagnosis and the diagnosis month were included. For the month of death, bills from the month of death and 2 months after death were included in order to account for lags in reporting. Covariates. We studied the impact of demographic and disease characteristics, comorbidities, and treatment choice on 1-year and 5-year costs. Data from the SEER registry included the demographic variables age, sex, race/ethnicity, and urbanicity. Disease variables included stage at diagnosis (in situ/local, regional, distant, unstaged). Treatment indicators were identified as no treatment, operative intervention, surgery with radiation, and other treatment. An indicator was also included for chemotherapy, which may have been used in combination with the other treatments. Statistical analysis. Patients in the SEERMedicare data have variable follow-up. Therefore, the approach to estimating costs at 1 year and 5 years was a method proposed by Bang and Tsiatis,10 which accounts for variable follow-up and patient censoring. This approach estimates costs for each patient at regular time intervals (monthly costs) and weights those estimates by the inverse of the probability that the patient is not censored. Bang and Tsiatis recommend a nonparametric estimator of the inverse probability. This was obtained from a Kaplan-Meier estimator of time to censoring. Using 60 monthly observation periods, we
Boltz et al 1365
Surgery Volume 154, Number 6
weighted the monthly costs for each patient that was uncensored by the inverse probability and then summed the weighted costs up to 12 months to estimate 1-year average cumulative costs, and up to 60 months to estimate 5-year average cumulative costs. This was done separately for thyroid cancer cases and comparison group patients. Multivariate analysis was performed using the approach suggested by Lin.11 Using this approach, each of the 60 monthly costs were fit to a linear model, where covariates included patient, disease, and treatment variables. Coefficients for months 1 through 12 were summed to give marginal effects of covariates on 1-year costs; coefficients for months 1 through 60 were summed to give marginal effects on 5-year costs. Ninety-five percent confidence intervals were obtained from 1,000 bootstrap replicates. All analyses were performed using SAS statistical software, version 9.2 (SAS Institute Inc, Cary, NC) and R (2.13.0), open source software available at: http://www.rproject.org/. RESULTS Characteristics of the 2,823 patients with papillary or follicular thyroid cancer are presented in Table I. The majority of patients (30.3%) were first diagnosed between the ages of 66 years and 69 years, were female (72.7%), and white (78.8%). The comorbidity burden was relatively low, with most patients (84.3%) having no comorbidities. Most patients with thyroid cancer lived in a metropolitan area (91%) and had local disease (54.1%) as opposed to regional (32.2%) or distant (10.4%) disease. In addition, 55.9% of thyroid cancer patients did not receive radiation, but of those who did, 61% underwent I-131 therapy. Operative treatment alone was the primary treatment modality (49.5%), followed by a combination of operation and radiation therapy (34%). The average inverse probability-weighted costs through 60 months are presented in Fig 1. A great portion of the costs attributed to thyroid cancer over 5 years occurred in the first 7 months of diagnosis. During those first 7 months the average cost attributed to the disease was $7,000 to $11,000. At 10 months after diagnosis, the costs leveled out over the remaining 60-month follow-up. Fig 2 presents the average cumulative cost curves from diagnosis up to 5 years after the diagnosis for DTC patients and comparison group patients. The costs were calculated under the assumption that the patient survived up to the time period specified on the x-axis. At 1 year after diagnosis, patients with DTC had accumulated
Table I. Demographic and disease characteristics of SEER-Medicare patients with well-differentiated thyroid cancer Total (N = 2823) Age, y 66 69 70 74 75 79 80+ Sex Male Female Race White Nonwhite Urban/rural Big metro Metro/urban Less urban/rural Stage Localized Regional Distant Unstaged Number of comorbidities 0 1 2 $3 Radiation None Beam radiation Radioactive implants Radioisotopes Combination of beam with implants/isotopes Unknown Treatment None Operation only Operation + radiation only Chemo (± operation/radiation)
856 785 636 546
(30.3) (27.8) (22.5) (19.3)
772 (27.3) 2051 (72.7) 2225 (78.8) 598 (21.2) 1675 (59.3) 896 (31.7) 252 (8.9) 1527 910 294 92
(54.1) (32.2) (10.4) (3.3)
2380 (84.3%) 257 (9.1%) 186 (6.1%) 1460 141 38 985 50
(52.9) (5.1) (1.4) (35.7) (1.8)
87 (3.1) 182 1398 961 282
(6.4) (49.5) (34) (10)
Values are n (%). SEER, Surveillance Epidemiology and End Results.
approximately $25,000 whereas comparison group patients had accumulated an average of $7,000. By 5 years, accumulated costs were approximately $66,000 for DTC patients and $43,500 for comparison patients. The difference between the curves is $22,500, meaning approximately $22,500 per patient is attributable to thyroid cancer over the 60 months after diagnosis. In the multivariable models of 1-year costs, several patient, disease, and treatment factors were significantly associated with accumulated costs at 1 year (Table II). At 1 year, patients ages
1366 Boltz et al
Surgery December 2013
Table II. Multivariate regression results for 1-year costs
Fig 1. Average inverse through 60 months.
probability-weighted
costs
Fig 2. Cumulative inverse probability-weighted unadjusted costs for patients with DTC (black line) and matched disease-free comparison group cases (gray line).
70 74 had $4,015 (95% confidence interval [95% CI] $64 7,677) greater costs than younger patients age 66 69. The greater costs incurred by male patients and nonwhite patients were not substantial. Disease stage was an important determinant of cost. Patients with regional disease incurred $9,578 (95% CI $6,539 12,735), and patients with distant disease incurred $28,447 (95% CI $21,554 36,692) in additional costs. Comorbidities were the most substantial driver of costs, where one to two comorbidities added $13,972 (95% CI: $7,420 21,288) and greater than 3
Variable
Parameter estimate, $
Intercept Age 66 69 Age 70 74 Age 75 79 Age 80+ Female Male White Nonwhite Big metro Metro Urban Less urban/rural Localized Regional Distant Unstaged No comorbidities 1–2 comorbidities 3+ comorbidities Operation only No treatment Operation and radiation Other treatment
17,669 Reference 4,015 3,517 3,863 Reference 2,093 Reference 1,491 Reference 6,695 5,146 7,926 Reference 9,578 28,447 8668 Reference 13,972 37,350 Reference 4,552 690 14,151
95% Confidence interval Lower
Upper
14,717
20,965
64 423 676
7,677 7,484 8,097
1362
5,603
2,472
5,364
9,585 10,802 12,422
3,922 816 2,906
6,539 21,554 15,978
12,735 36,692 1,545
7,420 27,131
21,288 48,994
4,591 2,707 8,349
14,535 2,787 20,259
comorbidities added $37,350 (95% CI $27,131 48,994) in costs. Compared with operation alone, patients undergoing both operation and radiation incurred $690 (95% CI $2,707 to 2,787) greater costs, which was not substantially different; however, patients who underwent other treatment including chemotherapy incurred $14,151 (95% CI $8,349 20,259) greater costs. Similar effects were observed for costs accumulated at 5 years after diagnosis (Table III). Patients ages 70 74 years incurred on average $7,362 (95% CI $470 14,815) in greater costs, patients age 75 79 years incurred on average $10,180 (95% CI $2,412 18,016), and patients age 80 or older incurred on average $12,649 (95% CI $4,290 22,065) compared with younger patients (reference age 66 69). Like 1-year cumulative costs, the greater costs of approximately $5,407 (95% CI $956 to 12,434) in male patients was not substantially different. Disease stage was also an important determinant of costs as patients with regional disease accumulated $8,902 (95% CI $2,241 $15,336) and patients with distant metastases incurred on average $20,103 (95% CI
Boltz et al 1367
Surgery Volume 154, Number 6
Table III. Multivariate regression results for 5-year costs
Variable
Parameter estimate, $
Intercept Age 66–69 Age 70–74 Age 75–79 Age 80+ Female Male White Nonwhite Big metro Metro Urban Less urban/rural Localized Regional Distant Unstaged No comorbidities 1–2 comorbidities 3+ comorbidities Operation alone No treatment Operation and radiation Other treatment
48,989 Reference 7,362 10,180 12,649 Reference 5,407 Reference 4,222 Reference 11,261 6,147 13,411 Reference 8,902 21,264 9,831 Reference 27,648 62,234 Reference 714 722 17,492
95% Confidence interval Lower
Upper
42,630
55,617
470 2,412 4,290
14,815 18,016 22,065
956
12,434
4,376
13,291
17,287 17,682 22,916
5,458 5,876 2,020
2,241 7,638 24,380
15,336 32,333 6,230
17,986 43,579
37,954 84,399
16,266 6,917 6,970
16,068 5,183 28,385
$7,638 32,333) in additional costs, which was similar to 1-year costs. Comorbidities were a substantial driver of 5-year costs, where patients having one or two comorbidities incurred $27,648 (95% CI $17,986 37,954) in additional costs at 5 years and greater than three comorbidities resulted in $62,234 (95% CI $43,579 84,399) in additional cumulative costs. Treatment choice was not associated with 5-year costs. Compared with patients who only underwent operation, patients who underwent operation and radiation therapy had a decrease in costs of $722 (95% CI $6,917 to 5,183), which was not clinically significant, and patients who underwent other treatments including chemotherapy had substantially greater costs of $17,492 (95% CI $6,970 28,385). DISCUSSION The rapid and dramatic increase in the incidence of DTC carries with it the corresponding increasing economic burden of thyroid cancer care. The long-term disease-specific survival from this relatively indolent malignancy is excellent. However, close and ongoing surveillance is
necessary to detect and treat potential cancer recurrence for years after the initial diagnosis of a DTC. In this era of increasing public and private health care expenditures, as well as the recognized need for cost-containment, it is vital to have valid data on the true monetary cost attributable to disease management. The costs associated with the increasing incidence of thyroid cancer are not well-understood. There is currently no literature on the overall costs related to nonmetastatic thyroid cancer, the longterm management, or repeat I-131 therapy for refractory disease. Given that most thyroid cancer is nonmetastatic and the number of these cases is increasing, the costs related to therapy will almost inevitably increase.12 Shrime et al13 sought to compare the 20-year cost-effectiveness of an initial hemithyroidectomy versus total thyroidectomy in the management of small papillary thyroid cancer. They used pooled data of 31 studies from the published literature to determine key statistics for decision analysis such as rates of recurrence, rates of complications for all interventions undertaken, and rates of death. The authors found that during 20 years the cost estimates including initial surgery, followup, and treatment of recurrence were between $13,897 and $14,242 for total thyroidectomy and between $15,038 and $15,064 for hemithyroidectomy. Berger et al14 performed a retrospective longitudinal cohort study using a U.S. health insurance claims database and analyzed the costs of healthrelated interventions in 183 patients with newly metastatic thyroid cancer between 2003 and 2005. Inpatient care was the main driver of the total health care expenditure, representing 43% of all costs. I-131 therapy was used in 19%, thyroid surgery in 13%, and chemotherapy in 11% of patients. The costs were substantial and totaled $60,200 per patient during the first year and $35,200 during the second year of follow-up. In the present study we sought to identify the total cumulative costs of medical care for DTC in a well-defined, at-risk population of patients. The Medicare-linked SEER database is a powerful, wellvalidated, and large cross-sectional sample of elderly cancer patients in the United States. The billing files associated with this database represent actual payments by Medicare such that a meaningful measure of costs for DTC care at 1 and 5 years from diagnosis could be obtained. We have identified that the 1-year incremental increase in cost for a patient with DTC is in excess of $15,000. Much of this cost is incurred within the
1368 Boltz et al
first 7 months of a new diagnosis and initial treatment of the thyroid malignancy. This time frame appears logical in that the up-front costs associated with initial surgical extirpation, hospitalization, and, if used, radioactive iodine remnant ablation, is anticipated to be clustered in the several months surrounding a diagnosis. From month 10 after initial diagnosis through to 60 months after the diagnosis, the incremental cost is present over a control patient without malignancy, but remains relatively stable. The sum total excess expenditure attributable to DTC during a 5-year period nears $25,000 per patient. Multivariate analysis of the 1- and 5-year costs revealed numerous factors that significantly influence the costs attributable to the care of patients with DTC. The advancing age of the individual patient plays the largest demographic role in the cost of care, despite treatment and surveillance guidelines that do not differentiate between age groupings in their recommendations. Furthermore, all patients included in this Medicarebased study would be included in the high-risk category on the basis of any number of the thyroid cancer risk scores. Male sex increased the cost of care for this disease process, notably at the 5-year mark; however, this was not clinically important. Finally, more medically fragile patients as measured by their recorded number of comorbid diagnoses had a great impact on and greater costs for care of DTC. Interestingly, of the patients with DTC, approximately 84% of the cohort had no comorbidities. However, this finding is consistent with the current literature. Kuijpens et al15 examined comorbidities in older newly diagnosed thyroid cancer patients and found that only 30% of their patient population had one or more comorbidities, with the number of comorbidities increasing with age. As might be expected, because of a greater disease burden, patients with more advanced disease had a stepwise increment in costs associated with thyroid cancer care. DTC metastatic to regional lymph nodes added more than $8,500 of cost at 1 and 5 years. Disease classed as metastatic added in excess of $28,000 and $20,000 at 1 and 5 years, respectively. The addition of multimodal treatment for DTC is often the standard pathway for the care of these patients. Patients with DTC greater than or equal to 1 cm in size should undergo near-total or total thyroidectomy with adjuvant radioiodine ablation for metastatic disease or a functional thyroid remnant according to ATA guidelines. The current study found that approximately 55% of patients
Surgery December 2013
did not receive radioiodine therapy and 6.5% had additional types of radiation therapies. In other words, they received less-aggressive care that was not consistent with ATA guidelines. Similar findings were observed in a study by Park et al,16 who used SEER data to describe treatment patterns of elderly Americans with DTC and found that elderly patients were more likely to undergo lessaggressive treatment and less likely to receive radioiodine therapy. Although following the current ATA guidelines have been shown to offer decreases in disease recurrence, it does come with an associated economic cost. However, the use of radioisotope therapy in an adjuvant setting only increases the 1-year costs by $700 and interestingly, decreases 5year costs by an average of $700; both findings, however, were not clinically important. This study is limited by the granularity of the data available in the SEER-Medicare database. There was a clear clinical limitation to the broad descriptors of disease stage. Furthermore, the details of operation and adjuvant therapies were limited as well. The strength of these data is the capture of actual payments made by Medicare for the care of these thyroid cancer patients. The excess cost encumbered as a result of DTC is a real contributor to health care costs and is likely to increase as the rate of this diagnosis rises and patients survive longer. The costs were largely accounted for via immutable variables such as necessary upfront treatments and disease stage-specific variables. However, these monetary data can be included in cost effectiveness analyses as future iterations of treatment and surveillance guidelines are researched. Although, not all clinically relevant, these data show too that there were apparent disparities in the cost of care by sex, race/ethnicity, and urbanicity. Although there is likely a complex interplay amongst all these factors, and with disease-specific factors, further investigation into this will be necessary to ensure that costs are contained as much as possible. These data may ultimately suggest that regionalization of care yields the most cost effective care for a diverse and growing population of patients. Furthermore, elderly patients diagnosed with DTC at an older age, with multiple comorbidities, and a more advanced disease process incurred greater costs. In addition, patients who underwent chemotherapy had greater costs at 1 and 5 years, whereas the use of radioactive iodine did not significantly impact these costs. These data suggest
Surgery Volume 154, Number 6
that it may be prudent to treat DTC when detected in the younger patient population and at an earlier stage of disease. REFERENCES 1. American Thyroid Association (ATA) Guidelines Taskforce on Thyroid Nodules and Differentiated Thyroid CancerCooper DS, Doherty GM, Haugen BR, Kloos RT, Lee SL, et al. Revised American Thyroid Association management guidelines for patients with thyroid nodules and differentiated thyroid cancer. Thyroid 2009;19: 1167-214. 2. Surveillance Epidemiology and End Results program. US Mortality Data 1969-2007. Bethesda, MD: National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch; 2010. Released June 2010. 3. Coburn MC, Wanebo HJ. Age correlates with increased frequency of high risk factors in elderly patients with thyroid cancer. Am J Surg 1995;170:471-5. 4. Miccoli P, Iacconi P, Cecchini GM, Caldarelli F, Ricci E, Berti P, et al. Thyroid surgery in patients aged over 80 years. Acta Chir Belg 1994;94:222-3. 5. Whitman ED, Norton JA. Endocrine surgical diseases of elderly patients. Surg Clin North Am 1994;74:127-44. 6. Surveillance Epidemiology and End Results program. SEER Medicare Research Data 1986-2007. Bethesda, MD: National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch; 2010. Released June 2010. 7. NCI. SEER-Medicare Linked Database. 2008; Available at: http://healthservices.cancer.gov/seermedicare/aboutdata/ cases.html. Accessed May 23, 2011. 8. National Cancer Institute. Surveillance Epidemiology and End Results. 2011. Availabe at: http://seer.cancer.gov/. Accessed April 29, 2011. 9. Hollenbeak CS, Boltz MM, Schaefer EW, Saunders BD, Goldenberg D. Recurrence of differentiated thyroid cancer in the elderly. Eur J Endocrinol 2013;168:54956. 10. Bang H, Tsiatis A. Estimating medical costs with censored data. Biometrika 2000;87:329-43. 11. Lin D. Linear regression analysis of censored medical costs. Biostatistics 2000;1:35-47. 12. Brown RL, deSouza JA, Cohen EEW. Thyroid cancer: burden of illness and management of disease. J Cancer 2011;2:193-9. 13. Shrime MG, Goldstein DP, Seaberg RM, Sawka AM, Rotstein L, Freeman JL, et al. Cost-effective management of low-risk papillary thyroid carcinoma. Arch Otolaryngol Head Neck Surg 2007;133:1245-53. 14. Berger A, Edelsberg J, Chung K, Ngyuyen A, Stepan D, Oster G. Healthcare (HC) utilization and costs in patients (pts) with newly diagnosed metastatic thyroid cancer (mTC). J Clin Oncol 2007;25:170-82. 15. Kuijpens JL, Janssen-Heijnen ML, Lemmens VE, Haak HR, Heijckmann AC, Coebergh JW. Comorbidity in newly diagnosed thyroid cancer patients: a population-based study on prevalence and the impact on treatment. Clin Endocrinol 2006;64:450-5. 16. Park HS, Roman SA, Sosa JA. Treatment patterns of aging Americans with differentiated thyroid cancer. Cancer 2010;116:20-30.
Boltz et al 1369
DISCUSSION Dr Ralph P. Tufano (Baltimore, MD): On your graph, when you looked at the patients with thyroid cancer and then the others, there was a parallel over that time. You would think that, potentially, the slope would continue to stay relatively the same for the thyroid cancer patients, but for the others it would kind of level out. Can you account for that parallel increase? It almost looked like there was the same slope as it continued out 60 months or so. Dr Melissa M. Boltz: The costs were all costs for care. So it makes sense that the slope would increase more in the first year because of the up-front costs of treatment and then level out over time. Dr Ralph P. Tufano (Baltimore, MD): Right. But for the patients with a benign condition, all they would need essentially is a follow-up to see whether their thyroid hormone was adequate, whereas the other patients theoretically should be getting ultrasound, thyroid gland assessment, and further workup for potential recurrence. It just looked like there was a parallel. Dr Melissa M. Boltz: It may be that either patient age or comorbidities are more significant than those followup procedures or care. Dr Carrie C. Lubitz (Boston, MA): I have looked at this myself. And there is a considerable cost after treatment for thyroid cancer for complications. I’m wondering if you factored that into your analysis, postoperative complications, radioactive iodine complications. Dr Melissa M. Boltz: Unfortunately, the SEER Medicare database does not look at postoperative complications for this subset of patients. Dr Cord Sturgeon (Chicago, IL): It looks like the area in which you could affect cost containment would be in that first 12 months after the diagnosis of thyroid cancer. So part one, is that true? Dr Melissa M. Boltz: I think that would be a logical conclusion, yes. Dr Cord Sturgeon (Chicago, IL): And it’s not possible, I assume, through this data set to compare the cost in a Medicare patient population, age 65 and older, versus what it would cost in people who are younger than the Medicare age? Dr Melissa M. Boltz: Right. That is a limitation to the database. And I know of no database that we would have cost data for the younger patient population. Dr Cord Sturgeon (Chicago, IL): Is it possible through this data set to compare the cumulative 60month cost for patients with a diagnosis of thyroid cancer versus those who had a thyroidectomy for a benign diagnosis, and show where the impact is? Dr Melissa M. Boltz: I don’t believe the codes differentiate. I looked at papillary and follicular cancer. I suppose there may be a code that differentiates benign thyroid disease as well. I don’t know for certain. Dr Rebecca S. Sippel (Madison, WI): There is dramatic cost to treating thyroid cancer in the elderly that thyroid cancer has a hold of very well, and the life expectancy of some of those patients might be not as long. So looking at your data, patients only get older and sicker,
1370 Boltz et al
and tumors only get bigger over time. But it seems like the one factor that decreased your cost was earlier intervention. So is that one of the take-home points, that is, we should be intervening sooner than later in this population to try to decrease costs?
Surgery December 2013
Dr Melissa M. Boltz: Yes, another logical conclusion. As most patients age, their comorbidities increase, and that is a significant determinant of cost. So I think intervening earlier would prove beneficial for the patient.
Background. Little is known about costs associated with differentiated thyroid cancer (DTC) and followup care. This study used data from the Surveillance Epidemiology and End Results (SEER) database to examine cumulative costs attributable to disease stage and treatment options of DTC in elderly patients over 5 years. Methods. We identified 2,823 patients aged >65 years with DTC and 5,646 noncancer comparison cases from SEER Medicare data between 1995 and 2005. Cumulative costs were obtained by estimating average costs/patient in each month up to 60 months after diagnosis. We performed multivariate analyses of costs by fitting each monthly cost to linear models, controlling for demographics and comorbidities. Marginal effects of covariates were obtained by summing coefficients over 60 months. Results. Cumulative costs were $17,669/patient the first year and $48,989/patient 5 years after diagnosis. Regional disease was associated with higher costs at 1 year ($9,578) and 5 years ($8,902). Distant disease was associated with 1-year costs of $28,447 and 5-year costs of $20,103. Patients undergoing surgery and radiation had a decrease in cost of $722 at 5 years. Conclusion. DTC in the elderly is associated with significant economic burden largely attributable to patient demographics, stage of disease, and treatment modalities. (Surgery 2013;154:1363-70.) From the Division of Outcomes Research and Quality, Department of Surgery,a Department of Surgery,b Division of Otolaryngology,d and Division of General Surgery Specialties and Surgical Oncology,e Penn State Milton S. Hershey Medical Center, Hershey; and Department of Public Health Sciences,c Penn State College of Medicine, Hershey, PA
THE INCIDENCE OF THYROID CANCER in the United States has more than doubled since the early 1970s. The yearly incidence has increased from 3.6 per 100,000 in 1973 to 8.7 per 100,000 in 2002.1 Although the incidence is increasing
Supported by a grant obtained from the Department of Health Commonwealth Universal Research Enhancement (CURE) program (Tobacco Settlement Act, Act 2001-77). This study used the linked SEER-Medicare database. The interpretation and reporting of these data are the sole responsibility of the authors. The authors acknowledge the efforts of the Applied Research Program, NCI; the Office of Research, Development and Information, CMS; Information Management Services (IMS), Inc; and the Surveillance, Epidemiology, and End Results (SEER) Program tumor registries in the creation of the SEER-Medicare database. Accepted for publication June 26, 2013. Reprint requests: Christopher S. Hollenbeak, PhD, Department of Surgery, The Pennsylvania State University, College of Medicine, 500 University Drive, H151, Hershey, PA 17033. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2013 Mosby, Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2013.06.042
steadily, thyroid cancer is still relatively uncommon because there were only 56,460 cases diagnosed in 2012.2 The majority of disease is differentiated thyroid cancer (DTC), which includes papillary and follicular cancers, and comprises approximately 90% of all thyroid cancers.1 In addition, the incidence of thyroid cancer also increases with age. Elderly patients often present with more aggressive forms of thyroid cancer, larger tumors, more extensive local growth, or distant metastases.3-5 However, the literature regarding outcomes in this age group remains scarce. The treatment of low-risk DTC is controversial despite the management guidelines recommended by the American Thyroid Association (ATA), which were first developed in 1996 and most recently updated in 2009.1 Given the long-term survival of patients with DTC, the increasing cost of health care, and the need for cost-containment, treatment strategies have become increasingly conservative. Follow-up and treatment of well-differentiated thyroid cancer extends for years beyond the initial operative procedure. SURGERY 1363
1364 Boltz et al
Although long-term survival of patients with DTC is a common occurrence, little is known about the cost attributable to the disease and follow-up care. The objective of this study was to use data from the Surveillance Epidemiology and End Results (SEER) and Medicare linked databases6 to study the costs attributed to the stage of disease and treatment options of DTC in elderly patients, as well as the cumulative cost of the disease over 5 years. METHODS Data. Data for this study were from the SEERMedicare linked data from the National Cancer Institute.7 The SEER program is a national tumor registry that collects data on approximately 26% of the U.S. population and reports the incidence and survival of cancer in the United States.8 SEERMedicare linked data contains the subset of Medicare enrollees in the SEER registry. It then links these patients to all Medicare claims records. Thus, by the use of Medicare billing claims data it is possible to identify specific types of resource utilization in the combined SEER-Medicare data.9 The SEERMedicare database also includes a group of patients without cancer that can be used for comparison purposes.7 Study population. Our analysis was limited to elderly adult Medicare enrollees (66 years and older) diagnosed with a single primary thyroid cancer (papillary or follicular). Claims data were only available for certain types of services since 1994. Therefore, we limited our sample to patients who were first diagnosed between 1995 and 2005. In addition, patients who were enrolled in a health maintenance organization at or after the cancer diagnosis were excluded because these patients do not have any Medicare billing records. The final analysis data set contained 2,823 patients. A comparison group sample of Medicare patients without cancer was selected from the Medicare 5% sample. To control for potentially different distributions of patient characteristics, propensity score matching was used to select a comparison sample that had a similar distribution of background characteristics as the thyroid cancer patients. We chose a propensity score matching approach because there was a large number of characteristics on which to match, including age, sex, race/ethnicity, and several comorbidities. An additional advantage of this technique is that the statistical analysis does not require adjustment for pairing as do exact matching techniques. The propensity score was the estimated probability in a logistic regression of case status (N = 2,823) versus
Surgery December 2013
comparison group status (N = 422,259). Covariates included year of birth (<1900, 5-year increments 1900 1934, $1935), sex, and race/ethnicity (white, black, Asian, other). An interaction between sex and race was also included because it was the only interaction that was clinically important. The predicted probability of being a case was computed for each subject and used as a propensity score. Two controls were matched to each case with the exact propensity score. Hence, each case was matched to two controls with the same sex and race, and born within the same 5year window. With these criteria the final sample included 2,823 patients with papillary or follicular thyroid cancer, and 5,646 matched noncancer comparison group cases. Costs. The cost analyses took the perspective of Medicare as payer and represented actual payments made by Medicare for all-cause treatment of patients. All costs were adjusted for inflation to year 2009 dollars. Billing files were used to create monthly costs for 5 years after the primary thyroid cancer diagnosis. All billing records, regardless of whether the bills were directly for treating cancer, were accumulated in each of the 60 months. For the month of diagnosis, bills from 1 month before diagnosis and the diagnosis month were included. For the month of death, bills from the month of death and 2 months after death were included in order to account for lags in reporting. Covariates. We studied the impact of demographic and disease characteristics, comorbidities, and treatment choice on 1-year and 5-year costs. Data from the SEER registry included the demographic variables age, sex, race/ethnicity, and urbanicity. Disease variables included stage at diagnosis (in situ/local, regional, distant, unstaged). Treatment indicators were identified as no treatment, operative intervention, surgery with radiation, and other treatment. An indicator was also included for chemotherapy, which may have been used in combination with the other treatments. Statistical analysis. Patients in the SEERMedicare data have variable follow-up. Therefore, the approach to estimating costs at 1 year and 5 years was a method proposed by Bang and Tsiatis,10 which accounts for variable follow-up and patient censoring. This approach estimates costs for each patient at regular time intervals (monthly costs) and weights those estimates by the inverse of the probability that the patient is not censored. Bang and Tsiatis recommend a nonparametric estimator of the inverse probability. This was obtained from a Kaplan-Meier estimator of time to censoring. Using 60 monthly observation periods, we
Boltz et al 1365
Surgery Volume 154, Number 6
weighted the monthly costs for each patient that was uncensored by the inverse probability and then summed the weighted costs up to 12 months to estimate 1-year average cumulative costs, and up to 60 months to estimate 5-year average cumulative costs. This was done separately for thyroid cancer cases and comparison group patients. Multivariate analysis was performed using the approach suggested by Lin.11 Using this approach, each of the 60 monthly costs were fit to a linear model, where covariates included patient, disease, and treatment variables. Coefficients for months 1 through 12 were summed to give marginal effects of covariates on 1-year costs; coefficients for months 1 through 60 were summed to give marginal effects on 5-year costs. Ninety-five percent confidence intervals were obtained from 1,000 bootstrap replicates. All analyses were performed using SAS statistical software, version 9.2 (SAS Institute Inc, Cary, NC) and R (2.13.0), open source software available at: http://www.rproject.org/. RESULTS Characteristics of the 2,823 patients with papillary or follicular thyroid cancer are presented in Table I. The majority of patients (30.3%) were first diagnosed between the ages of 66 years and 69 years, were female (72.7%), and white (78.8%). The comorbidity burden was relatively low, with most patients (84.3%) having no comorbidities. Most patients with thyroid cancer lived in a metropolitan area (91%) and had local disease (54.1%) as opposed to regional (32.2%) or distant (10.4%) disease. In addition, 55.9% of thyroid cancer patients did not receive radiation, but of those who did, 61% underwent I-131 therapy. Operative treatment alone was the primary treatment modality (49.5%), followed by a combination of operation and radiation therapy (34%). The average inverse probability-weighted costs through 60 months are presented in Fig 1. A great portion of the costs attributed to thyroid cancer over 5 years occurred in the first 7 months of diagnosis. During those first 7 months the average cost attributed to the disease was $7,000 to $11,000. At 10 months after diagnosis, the costs leveled out over the remaining 60-month follow-up. Fig 2 presents the average cumulative cost curves from diagnosis up to 5 years after the diagnosis for DTC patients and comparison group patients. The costs were calculated under the assumption that the patient survived up to the time period specified on the x-axis. At 1 year after diagnosis, patients with DTC had accumulated
Table I. Demographic and disease characteristics of SEER-Medicare patients with well-differentiated thyroid cancer Total (N = 2823) Age, y 66 69 70 74 75 79 80+ Sex Male Female Race White Nonwhite Urban/rural Big metro Metro/urban Less urban/rural Stage Localized Regional Distant Unstaged Number of comorbidities 0 1 2 $3 Radiation None Beam radiation Radioactive implants Radioisotopes Combination of beam with implants/isotopes Unknown Treatment None Operation only Operation + radiation only Chemo (± operation/radiation)
856 785 636 546
(30.3) (27.8) (22.5) (19.3)
772 (27.3) 2051 (72.7) 2225 (78.8) 598 (21.2) 1675 (59.3) 896 (31.7) 252 (8.9) 1527 910 294 92
(54.1) (32.2) (10.4) (3.3)
2380 (84.3%) 257 (9.1%) 186 (6.1%) 1460 141 38 985 50
(52.9) (5.1) (1.4) (35.7) (1.8)
87 (3.1) 182 1398 961 282
(6.4) (49.5) (34) (10)
Values are n (%). SEER, Surveillance Epidemiology and End Results.
approximately $25,000 whereas comparison group patients had accumulated an average of $7,000. By 5 years, accumulated costs were approximately $66,000 for DTC patients and $43,500 for comparison patients. The difference between the curves is $22,500, meaning approximately $22,500 per patient is attributable to thyroid cancer over the 60 months after diagnosis. In the multivariable models of 1-year costs, several patient, disease, and treatment factors were significantly associated with accumulated costs at 1 year (Table II). At 1 year, patients ages
1366 Boltz et al
Surgery December 2013
Table II. Multivariate regression results for 1-year costs
Fig 1. Average inverse through 60 months.
probability-weighted
costs
Fig 2. Cumulative inverse probability-weighted unadjusted costs for patients with DTC (black line) and matched disease-free comparison group cases (gray line).
70 74 had $4,015 (95% confidence interval [95% CI] $64 7,677) greater costs than younger patients age 66 69. The greater costs incurred by male patients and nonwhite patients were not substantial. Disease stage was an important determinant of cost. Patients with regional disease incurred $9,578 (95% CI $6,539 12,735), and patients with distant disease incurred $28,447 (95% CI $21,554 36,692) in additional costs. Comorbidities were the most substantial driver of costs, where one to two comorbidities added $13,972 (95% CI: $7,420 21,288) and greater than 3
Variable
Parameter estimate, $
Intercept Age 66 69 Age 70 74 Age 75 79 Age 80+ Female Male White Nonwhite Big metro Metro Urban Less urban/rural Localized Regional Distant Unstaged No comorbidities 1–2 comorbidities 3+ comorbidities Operation only No treatment Operation and radiation Other treatment
17,669 Reference 4,015 3,517 3,863 Reference 2,093 Reference 1,491 Reference 6,695 5,146 7,926 Reference 9,578 28,447 8668 Reference 13,972 37,350 Reference 4,552 690 14,151
95% Confidence interval Lower
Upper
14,717
20,965
64 423 676
7,677 7,484 8,097
1362
5,603
2,472
5,364
9,585 10,802 12,422
3,922 816 2,906
6,539 21,554 15,978
12,735 36,692 1,545
7,420 27,131
21,288 48,994
4,591 2,707 8,349
14,535 2,787 20,259
comorbidities added $37,350 (95% CI $27,131 48,994) in costs. Compared with operation alone, patients undergoing both operation and radiation incurred $690 (95% CI $2,707 to 2,787) greater costs, which was not substantially different; however, patients who underwent other treatment including chemotherapy incurred $14,151 (95% CI $8,349 20,259) greater costs. Similar effects were observed for costs accumulated at 5 years after diagnosis (Table III). Patients ages 70 74 years incurred on average $7,362 (95% CI $470 14,815) in greater costs, patients age 75 79 years incurred on average $10,180 (95% CI $2,412 18,016), and patients age 80 or older incurred on average $12,649 (95% CI $4,290 22,065) compared with younger patients (reference age 66 69). Like 1-year cumulative costs, the greater costs of approximately $5,407 (95% CI $956 to 12,434) in male patients was not substantially different. Disease stage was also an important determinant of costs as patients with regional disease accumulated $8,902 (95% CI $2,241 $15,336) and patients with distant metastases incurred on average $20,103 (95% CI
Boltz et al 1367
Surgery Volume 154, Number 6
Table III. Multivariate regression results for 5-year costs
Variable
Parameter estimate, $
Intercept Age 66–69 Age 70–74 Age 75–79 Age 80+ Female Male White Nonwhite Big metro Metro Urban Less urban/rural Localized Regional Distant Unstaged No comorbidities 1–2 comorbidities 3+ comorbidities Operation alone No treatment Operation and radiation Other treatment
48,989 Reference 7,362 10,180 12,649 Reference 5,407 Reference 4,222 Reference 11,261 6,147 13,411 Reference 8,902 21,264 9,831 Reference 27,648 62,234 Reference 714 722 17,492
95% Confidence interval Lower
Upper
42,630
55,617
470 2,412 4,290
14,815 18,016 22,065
956
12,434
4,376
13,291
17,287 17,682 22,916
5,458 5,876 2,020
2,241 7,638 24,380
15,336 32,333 6,230
17,986 43,579
37,954 84,399
16,266 6,917 6,970
16,068 5,183 28,385
$7,638 32,333) in additional costs, which was similar to 1-year costs. Comorbidities were a substantial driver of 5-year costs, where patients having one or two comorbidities incurred $27,648 (95% CI $17,986 37,954) in additional costs at 5 years and greater than three comorbidities resulted in $62,234 (95% CI $43,579 84,399) in additional cumulative costs. Treatment choice was not associated with 5-year costs. Compared with patients who only underwent operation, patients who underwent operation and radiation therapy had a decrease in costs of $722 (95% CI $6,917 to 5,183), which was not clinically significant, and patients who underwent other treatments including chemotherapy had substantially greater costs of $17,492 (95% CI $6,970 28,385). DISCUSSION The rapid and dramatic increase in the incidence of DTC carries with it the corresponding increasing economic burden of thyroid cancer care. The long-term disease-specific survival from this relatively indolent malignancy is excellent. However, close and ongoing surveillance is
necessary to detect and treat potential cancer recurrence for years after the initial diagnosis of a DTC. In this era of increasing public and private health care expenditures, as well as the recognized need for cost-containment, it is vital to have valid data on the true monetary cost attributable to disease management. The costs associated with the increasing incidence of thyroid cancer are not well-understood. There is currently no literature on the overall costs related to nonmetastatic thyroid cancer, the longterm management, or repeat I-131 therapy for refractory disease. Given that most thyroid cancer is nonmetastatic and the number of these cases is increasing, the costs related to therapy will almost inevitably increase.12 Shrime et al13 sought to compare the 20-year cost-effectiveness of an initial hemithyroidectomy versus total thyroidectomy in the management of small papillary thyroid cancer. They used pooled data of 31 studies from the published literature to determine key statistics for decision analysis such as rates of recurrence, rates of complications for all interventions undertaken, and rates of death. The authors found that during 20 years the cost estimates including initial surgery, followup, and treatment of recurrence were between $13,897 and $14,242 for total thyroidectomy and between $15,038 and $15,064 for hemithyroidectomy. Berger et al14 performed a retrospective longitudinal cohort study using a U.S. health insurance claims database and analyzed the costs of healthrelated interventions in 183 patients with newly metastatic thyroid cancer between 2003 and 2005. Inpatient care was the main driver of the total health care expenditure, representing 43% of all costs. I-131 therapy was used in 19%, thyroid surgery in 13%, and chemotherapy in 11% of patients. The costs were substantial and totaled $60,200 per patient during the first year and $35,200 during the second year of follow-up. In the present study we sought to identify the total cumulative costs of medical care for DTC in a well-defined, at-risk population of patients. The Medicare-linked SEER database is a powerful, wellvalidated, and large cross-sectional sample of elderly cancer patients in the United States. The billing files associated with this database represent actual payments by Medicare such that a meaningful measure of costs for DTC care at 1 and 5 years from diagnosis could be obtained. We have identified that the 1-year incremental increase in cost for a patient with DTC is in excess of $15,000. Much of this cost is incurred within the
1368 Boltz et al
first 7 months of a new diagnosis and initial treatment of the thyroid malignancy. This time frame appears logical in that the up-front costs associated with initial surgical extirpation, hospitalization, and, if used, radioactive iodine remnant ablation, is anticipated to be clustered in the several months surrounding a diagnosis. From month 10 after initial diagnosis through to 60 months after the diagnosis, the incremental cost is present over a control patient without malignancy, but remains relatively stable. The sum total excess expenditure attributable to DTC during a 5-year period nears $25,000 per patient. Multivariate analysis of the 1- and 5-year costs revealed numerous factors that significantly influence the costs attributable to the care of patients with DTC. The advancing age of the individual patient plays the largest demographic role in the cost of care, despite treatment and surveillance guidelines that do not differentiate between age groupings in their recommendations. Furthermore, all patients included in this Medicarebased study would be included in the high-risk category on the basis of any number of the thyroid cancer risk scores. Male sex increased the cost of care for this disease process, notably at the 5-year mark; however, this was not clinically important. Finally, more medically fragile patients as measured by their recorded number of comorbid diagnoses had a great impact on and greater costs for care of DTC. Interestingly, of the patients with DTC, approximately 84% of the cohort had no comorbidities. However, this finding is consistent with the current literature. Kuijpens et al15 examined comorbidities in older newly diagnosed thyroid cancer patients and found that only 30% of their patient population had one or more comorbidities, with the number of comorbidities increasing with age. As might be expected, because of a greater disease burden, patients with more advanced disease had a stepwise increment in costs associated with thyroid cancer care. DTC metastatic to regional lymph nodes added more than $8,500 of cost at 1 and 5 years. Disease classed as metastatic added in excess of $28,000 and $20,000 at 1 and 5 years, respectively. The addition of multimodal treatment for DTC is often the standard pathway for the care of these patients. Patients with DTC greater than or equal to 1 cm in size should undergo near-total or total thyroidectomy with adjuvant radioiodine ablation for metastatic disease or a functional thyroid remnant according to ATA guidelines. The current study found that approximately 55% of patients
Surgery December 2013
did not receive radioiodine therapy and 6.5% had additional types of radiation therapies. In other words, they received less-aggressive care that was not consistent with ATA guidelines. Similar findings were observed in a study by Park et al,16 who used SEER data to describe treatment patterns of elderly Americans with DTC and found that elderly patients were more likely to undergo lessaggressive treatment and less likely to receive radioiodine therapy. Although following the current ATA guidelines have been shown to offer decreases in disease recurrence, it does come with an associated economic cost. However, the use of radioisotope therapy in an adjuvant setting only increases the 1-year costs by $700 and interestingly, decreases 5year costs by an average of $700; both findings, however, were not clinically important. This study is limited by the granularity of the data available in the SEER-Medicare database. There was a clear clinical limitation to the broad descriptors of disease stage. Furthermore, the details of operation and adjuvant therapies were limited as well. The strength of these data is the capture of actual payments made by Medicare for the care of these thyroid cancer patients. The excess cost encumbered as a result of DTC is a real contributor to health care costs and is likely to increase as the rate of this diagnosis rises and patients survive longer. The costs were largely accounted for via immutable variables such as necessary upfront treatments and disease stage-specific variables. However, these monetary data can be included in cost effectiveness analyses as future iterations of treatment and surveillance guidelines are researched. Although, not all clinically relevant, these data show too that there were apparent disparities in the cost of care by sex, race/ethnicity, and urbanicity. Although there is likely a complex interplay amongst all these factors, and with disease-specific factors, further investigation into this will be necessary to ensure that costs are contained as much as possible. These data may ultimately suggest that regionalization of care yields the most cost effective care for a diverse and growing population of patients. Furthermore, elderly patients diagnosed with DTC at an older age, with multiple comorbidities, and a more advanced disease process incurred greater costs. In addition, patients who underwent chemotherapy had greater costs at 1 and 5 years, whereas the use of radioactive iodine did not significantly impact these costs. These data suggest
Surgery Volume 154, Number 6
that it may be prudent to treat DTC when detected in the younger patient population and at an earlier stage of disease. REFERENCES 1. American Thyroid Association (ATA) Guidelines Taskforce on Thyroid Nodules and Differentiated Thyroid CancerCooper DS, Doherty GM, Haugen BR, Kloos RT, Lee SL, et al. Revised American Thyroid Association management guidelines for patients with thyroid nodules and differentiated thyroid cancer. Thyroid 2009;19: 1167-214. 2. Surveillance Epidemiology and End Results program. US Mortality Data 1969-2007. Bethesda, MD: National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch; 2010. Released June 2010. 3. Coburn MC, Wanebo HJ. Age correlates with increased frequency of high risk factors in elderly patients with thyroid cancer. Am J Surg 1995;170:471-5. 4. Miccoli P, Iacconi P, Cecchini GM, Caldarelli F, Ricci E, Berti P, et al. Thyroid surgery in patients aged over 80 years. Acta Chir Belg 1994;94:222-3. 5. Whitman ED, Norton JA. Endocrine surgical diseases of elderly patients. Surg Clin North Am 1994;74:127-44. 6. Surveillance Epidemiology and End Results program. SEER Medicare Research Data 1986-2007. Bethesda, MD: National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch; 2010. Released June 2010. 7. NCI. SEER-Medicare Linked Database. 2008; Available at: http://healthservices.cancer.gov/seermedicare/aboutdata/ cases.html. Accessed May 23, 2011. 8. National Cancer Institute. Surveillance Epidemiology and End Results. 2011. Availabe at: http://seer.cancer.gov/. Accessed April 29, 2011. 9. Hollenbeak CS, Boltz MM, Schaefer EW, Saunders BD, Goldenberg D. Recurrence of differentiated thyroid cancer in the elderly. Eur J Endocrinol 2013;168:54956. 10. Bang H, Tsiatis A. Estimating medical costs with censored data. Biometrika 2000;87:329-43. 11. Lin D. Linear regression analysis of censored medical costs. Biostatistics 2000;1:35-47. 12. Brown RL, deSouza JA, Cohen EEW. Thyroid cancer: burden of illness and management of disease. J Cancer 2011;2:193-9. 13. Shrime MG, Goldstein DP, Seaberg RM, Sawka AM, Rotstein L, Freeman JL, et al. Cost-effective management of low-risk papillary thyroid carcinoma. Arch Otolaryngol Head Neck Surg 2007;133:1245-53. 14. Berger A, Edelsberg J, Chung K, Ngyuyen A, Stepan D, Oster G. Healthcare (HC) utilization and costs in patients (pts) with newly diagnosed metastatic thyroid cancer (mTC). J Clin Oncol 2007;25:170-82. 15. Kuijpens JL, Janssen-Heijnen ML, Lemmens VE, Haak HR, Heijckmann AC, Coebergh JW. Comorbidity in newly diagnosed thyroid cancer patients: a population-based study on prevalence and the impact on treatment. Clin Endocrinol 2006;64:450-5. 16. Park HS, Roman SA, Sosa JA. Treatment patterns of aging Americans with differentiated thyroid cancer. Cancer 2010;116:20-30.
Boltz et al 1369
DISCUSSION Dr Ralph P. Tufano (Baltimore, MD): On your graph, when you looked at the patients with thyroid cancer and then the others, there was a parallel over that time. You would think that, potentially, the slope would continue to stay relatively the same for the thyroid cancer patients, but for the others it would kind of level out. Can you account for that parallel increase? It almost looked like there was the same slope as it continued out 60 months or so. Dr Melissa M. Boltz: The costs were all costs for care. So it makes sense that the slope would increase more in the first year because of the up-front costs of treatment and then level out over time. Dr Ralph P. Tufano (Baltimore, MD): Right. But for the patients with a benign condition, all they would need essentially is a follow-up to see whether their thyroid hormone was adequate, whereas the other patients theoretically should be getting ultrasound, thyroid gland assessment, and further workup for potential recurrence. It just looked like there was a parallel. Dr Melissa M. Boltz: It may be that either patient age or comorbidities are more significant than those followup procedures or care. Dr Carrie C. Lubitz (Boston, MA): I have looked at this myself. And there is a considerable cost after treatment for thyroid cancer for complications. I’m wondering if you factored that into your analysis, postoperative complications, radioactive iodine complications. Dr Melissa M. Boltz: Unfortunately, the SEER Medicare database does not look at postoperative complications for this subset of patients. Dr Cord Sturgeon (Chicago, IL): It looks like the area in which you could affect cost containment would be in that first 12 months after the diagnosis of thyroid cancer. So part one, is that true? Dr Melissa M. Boltz: I think that would be a logical conclusion, yes. Dr Cord Sturgeon (Chicago, IL): And it’s not possible, I assume, through this data set to compare the cost in a Medicare patient population, age 65 and older, versus what it would cost in people who are younger than the Medicare age? Dr Melissa M. Boltz: Right. That is a limitation to the database. And I know of no database that we would have cost data for the younger patient population. Dr Cord Sturgeon (Chicago, IL): Is it possible through this data set to compare the cumulative 60month cost for patients with a diagnosis of thyroid cancer versus those who had a thyroidectomy for a benign diagnosis, and show where the impact is? Dr Melissa M. Boltz: I don’t believe the codes differentiate. I looked at papillary and follicular cancer. I suppose there may be a code that differentiates benign thyroid disease as well. I don’t know for certain. Dr Rebecca S. Sippel (Madison, WI): There is dramatic cost to treating thyroid cancer in the elderly that thyroid cancer has a hold of very well, and the life expectancy of some of those patients might be not as long. So looking at your data, patients only get older and sicker,
1370 Boltz et al
and tumors only get bigger over time. But it seems like the one factor that decreased your cost was earlier intervention. So is that one of the take-home points, that is, we should be intervening sooner than later in this population to try to decrease costs?
Surgery December 2013
Dr Melissa M. Boltz: Yes, another logical conclusion. As most patients age, their comorbidities increase, and that is a significant determinant of cost. So I think intervening earlier would prove beneficial for the patient.