OPTIONS & TOOLS
DOI: 10.1200/JOP.2017.024935 Journal of Oncology Practice - published online before print October 5, 2017
Where Are the Opportunities for Reducing Health Care Spending Within Alternative Payment Models?
The Oncology Care Model (OCM) is a highly controversial specialty care model developed by the Centers for Medicare & Medicaid aimed to provide higher-quality care at lower cost. Because oncologists will be increasingly held accountable for spending as well as quality within new value-based health care models like the OCM, they need to understand the drivers of total spending for their patients.
This retrospective cohort study included patients ≥ 65 years of age with primary fee-for-service Medicare insurance who received antineoplastic therapy at 12 cancer centers in the Southeast from 2012 to 2014. Medicare administrative claims data were used to identify health care spending during the prechemotherapy period (from cancer diagnosis to antineoplastic therapy initiation) and during the OCM episodes of care triggered by antineoplastic treatment. Total health care spending per episode includes all types of services received by a patient, including nononcology services. Spending was further characterized by type of service.
Average total health care spending in the three OCM episodes of care was $33,838 (n = 3,427), $23,811 (n = 1,207), and $19,241 (n = 678). Antineoplastic drugs accounted for 27%, 32%, and 36% of total health care spending in the first, second, and third episodes. Ten drugs, used by 31% of patients, contributed 61% to drug spending ($18.8 million) in the first episode. Inpatient spending also substantially contributed to total costs, representing 17% to 20% ($30.5 million) of total health care spending.
Health care delivery in the United States is rapidly moving toward new payment models, which will fundamentally redefine the reimbursement structure for physicians. The Medicare and CHIP Reauthorization Act mandates paying physicians through the Merit-Based Incentive Program beginning in January 2019 unless they participate in alternative payment models.1,2 Regardless of the pathway chosen, providers will be accountable for total health care spending associated with care delivered to their patients. This will be challenging for physicians, who often lack knowledge of the number and cost of services their patients receive.3 A better understanding of oncology care spending will be essential, because it is expected to exceed $170 billion by 2020.4
Understanding health care spending will be particularly important for oncologists who participate in alternative payment models, such as the Center for Medicare and Medicaid Innovation (CMMI) Oncology Care Model (OCM).5 This model includes a supplemental, per-beneficiary, Monthly Enhanced Oncology Services payment for patients receiving antineoplastic therapy and performance-based payments on the basis of reductions in Medicare spending < 96% of expected spending in a one-sided risk model, and < 97.5% in a two-sided risk model, with payments adjusted based on scores on quality measures.2
Limited data are available to guide physicians on how to reduce health care spending within payment models such as OCM. Granularity regarding service-specific types of spending is needed to identify opportunities where health care spending can be effectively reduced using interventions supported by the up-front payments from Centers for Medicare & Medicaid Services. Thoughtful resource allocation will be essential for success in reaching the triple aim of better patient health, better health care, and lower cost of care.6
Our primary objective was to quantify spending across types of health care services within the OCM framework to identify opportunities for spending reductions. The secondary objective was to assess the types and amounts of spending within the prechemotherapy period (the period from cancer diagnosis to initiation of antineoplastic therapy not considered in the OCM model) to identify additional areas for potential reductions in overall health care spending.
We conducted a retrospective cohort study of Medicare administrative claims data from 2012 to 2014 for all older adults with cancer who received care within the University of Alabama at Birmingham Health System Cancer Community Network (CCN). The CCN included 12 cancer centers of varying size and practice structure located in Alabama, Georgia, Florida, Mississippi, and Tennessee.7 Data were collected for the purpose of evaluating a lay navigation program called the Patient Care Connect Program. Claims data from all patients receiving care at the participating institutions were merged with local tumor registry data. The Institutional Review Boards at UAB and each of the participating CCN sites approved the evaluation of the Patient Care Connect Program.
This study included CCN Medicare beneficiaries ≥ 65 years of age with primary Medicare Part A, B, and D insurance coverage and no health maintenance organization coverage who were diagnosed with a stage I to IV cancer from 2012 to 2014 and received OCM-eligible chemotherapy, hormone therapy, or targeted therapy.
Data on cancer treatment and costs were from 2012 to 2014 Medicare Part A, B, and D claims for inpatient, outpatient, physician (carrier), home health, skilled nursing facility, hospice, durable medical equipment (DME), and prescription drug event files. Data on date of birth, race, and comorbidity were also obtained from Medicare claims data. Comorbid conditions were abstracted from claims data at any time within the database and classified using a weighted score of 0, 1, 2+ on the basis of the Klabunde modification of the Charlson comorbidity index.8-10 Data on cancer type, stage, and date of diagnosis were received from the local cancer registries of the 12 CCN sites.
Claims for antineoplastic therapy were drawn from the physician (carrier), outpatient, DME, and Part D Medicare data files. Health Care Common Procedure Coding System codes, Current Procedural Terminology codes, National Drug Codes, Berenson-Eggers Type of Service codes, or generic drug names were used to identify receipt of antineoplastic therapy (Data Supplement).
Initiation of an OCM episode of care was triggered by receipt of oral or parenteral antineoplastic treatment with an eligible medication. Day 0 is defined as the start of first treatment. Episodes 1, 2, and 3 were defined as 0 to 180, 181 to 360, and 361 to 540 days, respectively. Patients were eligible for an additional episode if alive and continuing antineoplastic therapy. The prechemotherapy period was defined as the time between diagnosis and day 0.
The primary outcome was total health care spending for all patients in each OCM episode, including payments to providers by Medicare, other payers, and patients. The prechemotherapy period spending, defined as the health care spending occurring from time of diagnosis to initial treatment, was also calculated. Health care spending was further characterized by type of service, including inpatient, home health, DME, skilled nursing facility, hospice, chemotherapy drugs (infusion and oral), prescription drugs, growth factor, evaluation and management, radiation, testing/pathology, imaging, emergency department visits, and other uncategorized types. Following the OCM methodology, we evaluated the impact of Winsorization of outliers. Winsorization assigns health care spending below the 5th and above the 95th percentile to be equal to the 5th and 95th percentile, respectively. We calculated the total health care spending for the subset of patients above the 95th percentile, the spending for all patients using Winsorization, and the total spending for all patients excluding above the 95th percentile. Finally, we computed the amount needed to reduce total spending by at least 4%.
Overall sample characteristics were described using means and standard deviations for continuous measures or frequencies and percentages for categorical measures. Total and service-specific spending were quantified by summing all costs of care per patient included during the specified analytical time frame and assessing mean, median, 95% confidence limits, and interquartile range. The service-specific percentage of health care spending was calculated for each time frame. The percentage of health care spending for specific antineoplastic and growth factor drugs was calculated.
The population included 3,427 patients with cancer (Table 1). The mean age at time of diagnosis was 72.9 years (SD, 6.7); 85.5% of the patients were white. The most common cancer types were breast (26.6%), lung (19.7%), and GI (19.4%); 24.9% of patients had stage IV disease. Comorbidities were common; > 50% of patients had two or more comorbidities in addition to their cancer diagnosis (Table 1).
Average spending during the prechemotherapy period was $16,208/patient (N = 3,427) over a median duration of 48 days. When considering service-specific spending, inpatient care accounted for the greatest percentage of the total, representing 41% of health care spending during this period (Fig 1A). The average spending related to laboratory testing and pathology was $1,289/patient (7% of total costs; n = 3,222; Fig 1B). Spending for outpatient services (average cost of $2,394/patient) and physician services (average cost of $1,879/patient) accounted for 12% and 11% of the total.
Thirty-five percent of patients had at least two episodes of care; 20% had at least three episodes. Total health care spending declined substantially for each successive OCM episode (Data Supplement). Average total spending for the first three OCM episodes of care were $33,838 (N = 3,427), $23,811 (n = 1,207), and $19,241 (n = 678). Average total spending per OCM episode overall was $25,630/patient. Within our sample population of 3,427 patients, health care spending over all three OCM episodes totaled > $157 million. Winsorization would reduce that amount to $150 million (a 4.5% reduction). The total spending attributed to the Winsorized patients (above the 95th percentile) was $28 million, so exclusion of all spending on those patients would have reduced the $157 million total to $129 million (an 18% reduction).
The largest category of spending was infusion and oral chemotherapy drug spending, which accounted for $44.9 million and represented the largest single category of health care spending over the first three OCM episodes (first: 27%, average spending of $9,045/patient; second: 32%, average spending of $7,654/patient; third: 36%, average cost of $6,997/patient). Inpatient spending was the second-largest contributor to total spending, representing 20%, 18%, and 17% of the first, second, and third episodes (Fig 1A), for a total of $30.5 million over the three episodes. When considering inpatient spending, 22%, 26%, and 31% of the first, second, and third episode costs were incurred in the last 30 days of life. Reducing total spending for the three episodes by 4% ($6.3 million) could be achieved by either a 14% reduction in chemotherapy spending or a 21% reduction in inpatient spending.
Antineoplastic drug–specific spending in the first OCM episode of care totaled > $30 million. However, drug spending was driven by a small number of commonly administered medications (Table 2). Ten drugs (rituximab [n = 192], pemetrexed [n = 151], lenalidomide [n = 69], ipilimumab [n = 18], bevacizumab [n = 115], trastuzumab [n = 70], bendamustine [n = 49], oxaliplatin [n = 195], capecitabine [n = 137], bortezomib [n = 66]) accounted for 61% of total antineoplastic drug spending at $18.8 million but were used in 31% of patients receiving antineoplastic therapy (Table 2). One hundred seventeen other antineoplastic drugs were administered during the first OCM episode. Among drugs commonly administered in conjunction with chemotherapy, growth factors (pegfilgrastim and filgrastim) accounted for an additional $10.4 million in spending across the three episodes, whereas others, such as bisphosphonates and erythropoiesis-stimulating agents, made up < 1% of total costs within the first OCM episode.
This report characterizes health care spending for cancer care for patients within the framework of the OCM episodes and highlights areas where the largest opportunities for savings may lie. Antineoplastic drug spending accounted for a largest percentage (27% to 36%) of total spending in this population. This percentage is higher than observed in the RAND Corporation’s analysis, on which the OCM was designed, in which chemotherapy drug and administration accounted for 10% to 31% of total spending.11 An additional $10.4 million is spent on growth factor support during chemotherapy. Previous literature has demonstrated a substantial portion of growth factor is used in patients at low risk for neutropenic fever, presenting an opportunity for savings with growth factor omission.12 In contrast, other supportive medications, such as erythropoiesis-stimulating agents and bisphosphonates, are less amenable to intervention in this population. Another potential target for reducing cost is use of biosimilars, estimated by the RAND Corporation to lead to a $44.2 billion savings over the next decade.13 However, excitement for biosimilars has been tempered by modest observed reductions due to regulatory restrictions, prescribing patterns, and pricing environments.14,15 Drug spending on both antineoplastic therapy and growth factor can be influenced, at least in part, by physician choice. In many cancers, physicians have several cost-varying, guideline-based treatment options where pathway programs can guide providers to higher value choices. Pathway programs across the country have demonstrated increased standardization of treatment, reduced health care spending, and maintenance of efficacy end points,16-18 with one pathway program leading to > $10 million reduction in spending.19
Pathway programs, however, can only achieve savings when there is a choice of lower-cost drugs, and total drug spending is driven by the underlying cost of the antineoplastic drugs. This study shows that 10 chemotherapy agents account for > 61% of the spending on chemotherapy in the first episode. As additional high-price immunologic agents are added to the oncology milieu, such as nivolumab and atezolizumab, spending increases will occur beyond the control of oncologists.20,21 In addition, we observed that chemotherapy spending was proportionally higher in the second and third episodes than in the first episode, which may be due to higher proportions of patients with persistent and/or recurrent, metastatic disease. It is important to ensure that oncologists are only held responsible for appropriate treatment selection and not for drug price increases.
Inpatient spending is almost as high as drug spending in this analysis, accounting for 20% of the total spending. This is less than observed in the RAND study, in which inpatient costs were the largest category of spending, accounting for 25% to 45% of total spending.11 Projects focusing on reducing hospitalizations have had some success in reducing spending.22,23 The Patient Care Connect Program demonstrated that emergency department visits, hospitalizations, and intensive care unit admissions decreased by 6.0%, 7.9%, and 10.6% for navigated patients compared with nonnavigated patients, for an estimated $19 million reduction in annual costs to Medicare.23 The Community Oncology Medical Home (COME HOME) project, which included the development of an oncology medical home,24 reported a 12.5% reduction in hospital admissions and a 7.2% reduction in total Medicare spending (ASCO, Editor. 2016). These types of practice transformations promote systematic screening of patient needs and early intervention, which are potential approaches to reduce spending. To achieve the 4% savings goal of the OCM by exclusively focusing on inpatient spending, practices would need to reduce overall inpatient spending by 20%, which is more than achieved by either of these programs alone. CMMI has included requirements for patient navigation and adherence to guideline-concordant care within the OCM, which may aid in achieving reductions in health care spending.
End-of-life care is an important target within the OCM model, where costs can be reduced and quality enhanced. Chastek et al25 reported that cancer-related expenses for Medicare patients in the 6 months before death exceeded $74,000; 55% of cost was attributed to inpatient costs. Notably, both the Patient Care Connect Program and the COME HOME project savings included substantial reductions in spending in the last 90 days of life, contributing to the success of these demonstration projects.26 Therefore, this should be considered in cost reduction efforts.
Other opportunities for cost reductions occur before the initiation of medical therapy for cancer. Analyses of adherence to Choosing Wisely guidelines, which recommend against advanced imaging for several early-stage cancers, have shown that overuse of diagnostic testing contributes to higher spending.12,27-30 Other payment models, such as ASCO’s Patient-Centered Oncology Payment model, better address these opportunities by initiating payment reforms at the time of diagnosis. Our study is the first to highlight the potential of these components to be targets for spending containment. Initial surgical management is another component of cancer care that often occurs before initiation of chemotherapy and may account in part for the higher proportion of inpatient spending observed during the prechemotherapy period in this study. The OCM model does not encourage or reward efforts to reduce spending on surgery; however, it includes adjustments to spending calculations to prevent physicians from withholding neoadjuvant (presurgery) therapy for cancer. Triggering payments only after chemotherapy begins limits the ability to engage surgical oncologists in practice transformation to achieve savings. For example, contralateral prophylactic mastectomy in breast cancer is commonly performed, yet it is associated with higher complication risk and increased spending compared with more conservative surgical approaches.31,32 Discussions between surgical oncologists and their patients about less-aggressive surgical management could benefit patients and result in savings for payers. Currently, there is no support under OCM for the extra time required for optimal treatment planning when patient anxiety can be substantial and coordination between providers is essential. We advocate for continued analysis of practice patterns and comparisons between the OCM model and the ASCO’s Patient-Centered Oncology Payment model, which includes the prechemotherapy period.
Using Winsorization to adjust spending instead of excluding outlier patients altogether could unfairly penalize oncologists. In this analysis, these individuals contribute nearly one fifth of total spending, and there is no obvious reason to believe that physicians can control spending at the 94th percentile better than they can at the 96th. Although excluding outlier patients entirely could potentially create a perverse incentive to increase spending on expensive patients even more, Winsorization alone can significantly penalize a physician for caring for patients with uniquely expensive problems.
Despite the limitations of the OCM, such as the Winsorization methodology, we applaud CMMI for their tremendous effort in moving forward the value proposition in oncology. Undoubtedly, modifications to this model will be needed as both practices and CMMI gain experience within the OCM. We believe that the key to long-term model sustainability will be the continued sharing of claims data by CMMI with OCM participants, which creates a unique opportunity for stakeholders to conduct in-depth analysis of practice patterns and cost. Such analysis should guide implementation and provide a data-driven approach to future modifications of the model.
This study describes the spending for patients with cancer with Medicare, and the results may not be applicable to younger patients with commercial insurance. Younger patients differ in the prevalence of specific cancers,33 and may be more likely to choose expensive aggressive care, but may also be better able to tolerate the adverse effects of toxic regimens and thus avoid hospitalization. Moreover, this study was conducted within the Southeast, and findings may be different in other geographic regions. Data were originally intended for use in the evaluation of cost savings after implementation of a patient navigation program. Although it is possible that reductions in cost due to navigation may result in underestimation of total costs in this analysis, only 33% of the patients included in our sample received navigation; therefore, we expect our results to be generalizable to both navigated and nonnavigated populations. In addition, claims data are designed for billing rather than research, which affects the ability to evaluate relevant clinical factors. For example, we used standard definitions for comorbidity but may have missed mild or noncoded comorbidities. This analysis does not adjust for case mix between episodes or specify spending from preventable hospitalizations. It does not differentiate care received at a primary institution versus care received locally, where there may be fewer opportunities to influence spending. Finally, this analysis does not consider actual costs to health systems, such as personnel and space. Practices and hospitals may incur significant unreimbursed costs in implementing interventions designed to reduce spending, and the upfront payments and performance-based payments they receive may not be adequate to cover those costs.
In conclusion, average total spending per OCM episode was $25,630/patient. Spending was heavily driven by antineoplastic drugs and inpatient care, highlighting the need to consider interventions that target both types of spending and the need for protections in payment models to ensure that physicians are not held accountable for drug price increases beyond their control.
Conception and design: Gabrielle B. Rocque, Courtney P. Williams, Elizabeth A. Kvale, Edward E. Partridge, Maria Pisu
Financial support: Gabrielle B. Rocque
Administrative support: Gabrielle B. Rocque
Collection and assembly of data: Gabrielle B. Rocque, Courtney P. Williams, Karina I. Halilova, Margaret M. Sullivan, Rod P. Rocconi, Edward E. Partridge
Data analysis and interpretation: Gabrielle B. Rocque, Courtney P. Williams, Kelly M. Kenzik, Bradford E. Jackson, Andres Azuero, Elizabeth A. Kvale, Warner K. Huh, Edward E. Partridge
Manuscript writing: All authors
Final approval of manuscript: All authors
Accountable for all aspects of the work: All authors
The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/journal/jop/site/misc/ifc.xhtml.
Consulting or Advisory Role: Genentech, Pfizer
Research Funding: Pack Health, Medscape, Carevive Systems, Genentech, Pfizer
Travel, Accommodations, Expenses: Pfizer, Genentech
No relationship to disclose
No relationship to disclose
No relationship to disclose
No relationship to disclose
Research Funding: Carevive Systems (Inst)
Honoraria: Intuitive Surgical
Consulting or Advisory Role: Genentech
Speakers' Bureau: Genentech, AstraZeneca
No relationship to disclose
Employment: Aspire Health
Stock or Other Ownership: Aspire Health
Consulting or Advisory Role: LI-COR Biosciences, Merck HPV Vaccine, Antiva Biosciences
No relationship to disclose
No relationship to disclose
Supported by a Walter B. Frommeyer, Jr Fellowship in Investigative Medicine (G.B.R.) and by the Centers for Medicare & Medicaid Services Grant No. 1C1CMS331023. Presented at the 2017 ASCO Quality Care Symposium, Orlando, FL, March 3-4, 2017. The Walter B. Frommeyer, Jr Fellowship in Investigative Medicine had no role in the design and conduct of the study; collection, management, analysis, or interpretation of the data; and preparation, review, or approval of the manuscript for publication. The Centers for Medicare & Medicaid Services had no role in any aspect of the study or manuscript preparation and submission.
|1.||Clough JD, McClellan M: Implementing MACRA: Implications for physicians and for physician leadership. JAMA 315:2397-2398, 2016 Crossref, Medline, Google Scholar|
|2.||Kline RM, Bazell C, Smith E, et al: Centers for medicare and medicaid services: Using an episode-based payment model to improve oncology care. J Oncol Pract 11:114-116, 2015 Link, Google Scholar|
|3.||Reichert S, Simon T, Halm EA: Physicians’ attitudes about prescribing and knowledge of the costs of common medications. Arch Intern Med 160:2799-2803, 2000 Crossref, Medline, Google Scholar|
|4.||Mariotto AB, Yabroff KR, Shao Y, et al: Projections of the cost of cancer care in the United States: 2010-2020. J Natl Cancer Inst 103:117-128, 2011 Crossref, Medline, Google Scholar|
|5.||Center for Medicare & Medicaid Services: Oncology Care Model. http://innovation.cms.gov/initiatives/Oncology-Care/ Google Scholar|
|6.||Porter ME: What is value in health care? N Engl J Med 363:2477-2481, 2010 Crossref, Medline, Google Scholar|
|7.||Rocque GB, Partridge EE, Pisu M, et al: The Patient Care Connect Program: Transforming health care through lay navigation. J Oncol Pract 12:e633-e642, 2016 Link, Google Scholar|
|8.||Deyo RA, Cherkin DC, Ciol MA: Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol 45:613-619, 1992 Crossref, Medline, Google Scholar|
|9.||Charlson ME, Pompei P, Ales KL, et al: A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis 40:373-383, 1987 Crossref, Medline, Google Scholar|
|10.||Klabunde CN, Potosky AL, Legler JM, et al: Development of a comorbidity index using physician claims data. J Clin Epidemiol 53:1258-1267, 2000 Crossref, Medline, Google Scholar|
|11.||Huckfeldt PJ, Chan C, Hirshman S, et al: Specialty payment model opportunities and assessment: Oncology model design report. Rand Health Q 5:11, 2015 Medline, Google Scholar|
|12.||Rocque GB, Williams CP, Jackson BE, et al: Choosing Wisely: Opportunities for improving value in cancer care delivery? J Oncol Pract 13:e11-e21, 2017 Link, Google Scholar|
|13.||Mulcahy AW, Predmore Z, Mattke S: The cost savings potential of biosimilar drugs in the United States. http://www.rand.org/pubs/perspectives/PE127.html Google Scholar|
|14.||Yang YT, Chen B, Bennett CL: Biosimilars-curb your enthusiasm. JAMA Oncol 10.1001/jamaoncol.2017.1530 [epub ahead of print on June 8, 2017] Google Scholar|
|15.||Bennett CL, Sartor OA, Armitage JO, et al: Biosimilar and generic cancer drugs unlikely to bend cost curve in the USA. Lancet Oncol 18:22-23, 2017 Crossref, Medline, Google Scholar|
|16.||Feinberg BA, Lang J, Grzegorczyk J, et al: Implementation of cancer clinical care pathways: A successful model of collaboration between payers and providers. J Oncol Pract 8:e38s-e43s, 2012 (suppl 3) Google Scholar|
|17.||Jackman DM, Zhang Y, Dalby C, et al: Cost and survival analysis before and after implementation of Dana-Farber clinical pathways for patients with stage IV non-small-cell lung cancer. J Oncol Pract 13:e346-e352, 2017 Link, Google Scholar|
|18.||Ellis PG, O’Neil BH, Earle MF, et al: Clinical pathways: Management of quality and cost in oncology networks in the metastatic colorectal cancer setting. J Oncol Pract 13:e522-e529, 2017 Link, Google Scholar|
|19.||Kreys ED, Koeller JM: Documenting the benefits and cost savings of a large multistate cancer pathway program from a payer’s perspective. J Oncol Pract 9:e241-e247, 2013 Link, Google Scholar|
|20.||Rittmeyer A, Barlesi F, Waterkamp D, et al: Atezolizumab versus docetaxel in patients with previously treated non-small-cell lung cancer (OAK): A phase 3, open-label, multicentre randomised controlled trial. Lancet 389:255-265, 2017 Crossref, Medline, Google Scholar|
|21.||Brahmer J, Reckamp KL, Baas P, et al: Nivolumab versus docetaxel in advanced squamous-cell non-small-cell lung cancer. N Engl J Med 373:123-135, 2015 Crossref, Medline, Google Scholar|
|22.||Meltzer DO, Ruhnke GW: Redesigning care for patients at increased hospitalization risk: The Comprehensive Care Physician model. Health Aff (Millwood) 33:770-777, 2014 Crossref, Medline, Google Scholar|
|23.||Rocque GB, Pisu M, Jackson BE, et al: Resource use and medicare costs during lay navigation for geriatric patients with cancer. JAMA Oncol 3:817-825, 2017 Crossref, Medline, Google Scholar|
|24.||Waters TM, Webster JA, Stevens LA, et al: Community oncology medical homes: Physician-driven change to improve patient care and reduce costs. J Oncol Pract 11:462-467, 2015 Link, Google Scholar|
|25.||Chastek B, Harley C, Kallich J, et al: Health care costs for patients with cancer at the end of life. J Oncol Pract 8:75s-80s, 2012 Link, Google Scholar|
|26.||Colligan EM, Ewald E, Ruiz S, et al: Innovative oncology care models improve end-of-life quality, reduce utilization and spending. Health Aff (Millwood) 36:433-440, 2017 Crossref, Medline, Google Scholar|
|27.||Rao VM, Levin DC: The overuse of diagnostic imaging and the Choosing Wisely initiative. Ann Intern Med 157:574-576, 2012 Crossref, Medline, Google Scholar|
|28.||Hahn EE, Tang T, Lee JS, et al: Use of imaging for staging of early-stage breast cancer in two integrated health care systems: Adherence with a choosing wisely recommendation. J Oncol Pract 11:e320-e328, 2015 Link, Google Scholar|
|29.||Makarov DV, Soulos PR, Gold HT, et al: Regional-level correlations in inappropriate imaging rates for prostate and breast cancers: Potential implications for the Choosing Wisely campaign. JAMA Oncol 1:185-194, 2015 Crossref, Medline, Google Scholar|
|30.||Ramsey SD, Fedorenko C, Chauhan R, et al: Baseline estimates of adherence to American Society of Clinical Oncology/American Board of Internal Medicine Choosing Wisely initiative among patients with cancer enrolled with a large regional commercial health insurer. J Oncol Pract 11:338-343, 2015 Link, Google Scholar|
|31.||Osman F, Saleh F, Jackson TD, et al: Increased postoperative complications in bilateral mastectomy patients compared to unilateral mastectomy: An analysis of the NSQIP database. Ann Surg Oncol 20:3212-3217, 2013 Crossref, Medline, Google Scholar|
|32.||Albornoz CR, Cordeiro PG, Mehrara BJ, et al: Economic implications of recent trends in U.S. immediate autologous breast reconstruction. Plast Reconstr Surg 133:463-470, 2014 Crossref, Medline, Google Scholar|
|33.||American Cancer Society: Estimated number of new cases for the four major cancers by sex and age group, 2017. https://www.cancer.org/content/dam/cancer-org/research/cancer-facts-and-statistics/annual-cancer-facts-and-figures/2017/estimated-new-cases-for-the-four-major-cancers-by-sex-and-age-group-2017.pdf Google Scholar|