Background: Bundled payment programs play an increasingly important role in transforming reimbursement for oncologic care. We assessed determinants of oncologists’ willingness to participate in bundled payment programs for breast cancer. We hypothesized that providers would be more likely to participate in bundled payment programs if offered higher levels of reimbursement for each episode of care. Methods: Oncologists from Florida, New Jersey, New York, and Pennsylvania were identified in the AMA database or by patients listed in state cancer registries. Providers were randomized to receive one of four versions of a survey describing bundled payment programs offering different levels of compensation for the first year of localized breast cancer treatment ($5000, $10,000, $15,000, or $20,000). Physicians rated their likelihood of participation in a bundled program on a Likert scale. Logistic regression was used to analyze determinants of likelihood of participation in bundling. Results: Among 460 respondents, only 17% of oncologists were highly likely to participate in a bundled program paying $5000 for the first year of care, rising to 41% for the $15,000 program, but falling to 34% for the $20,000 program. Likelihood of participation was higher among oncologists who were male, older, and believed that cancer patients should not be offered high-cost drugs with minimal survival benefit. Conclusion: Our results suggest that medical oncologists have limited enthusiasm for bundled payments, and higher payments may not overcome resistance to bundling among a substantial proportion of physicians. Keywords: Bundled payment, Physician compensation, Breast cancer, Oncology, Payment reform Background physicians’ willingness to participate . Physicians’ sup- Rising medical costs have motivated initiatives to re- port is especially critical for the Center for Medicare and design physician compensation. Bundled payment pro- Medicaid Services (CMS) Oncology Care Model, which grams, in which providers receive fixed fees to care for reimburses oncologists through bundled payments for patients during illness “episodes,” have received increas- episodes of cancer care, while still covering certain ser- ing national attention [1, 2]. While oncology practices vices through traditional fee-for-service payment models have traditionally operated using fee-for-service payment alongside pay-for-performance incentives . It is the models [3, 4], in 2013 Medicare launched the Bundled first of Medicare’s large-scale bundling initiatives to offer Payments for Care Improvement initiative, with plans to payment programs directly to solo-practitioners rather transition 50% of payments from volume-based reim- than exclusively to hospitals or group practices . The bursement to alternative models by 2018 [5, 6]. Early American Society of Clinical Oncology has also detailed uptake of bundled payment has predominantly been potential reimbursement plans that make use of both bun- confined to large hospitals, but the long-term success of dled payment and pay-for-performance models [3, 4, 8]. Medicare’s initiatives will depend upon individual Prior studies have raised concerns about physician op- position to bundling [1, 2]. At least 10% of health care costs are generated by patients with cancer [8, 9], yet lit- * Correspondence: firstname.lastname@example.org tle is known about medical oncologists’ attitudes towards Department of Medicine, Massachusetts General Hospital, 55 Fruit Street, bundled payments. In this study, we report on the GRB 740, Boston, MA 02114, USA Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Murciano-Goroff et al. BMC Health Services Research (2018) 18:391 Page 2 of 5 results of a survey asking oncologists about their willing- estimates . Subjects recorded their likelihood of par- ness to participate in bundled programs for breast ticipation in bundled programs on a 5-point Likert scale. cancer treatment. In accordance with previous studies [1, 2], the primary analysis dichotomized responses into extremely/very Methods likely versus less likely to participate in bundling. We surveyed medical oncologists from Florida, New Items regarding bundled payment programs were Jersey, New York, and Pennsylvania. embedded in a larger 17-item survey as part of a study The physicians included were listed in the American of disparities in genetic testing [10–12]. The survey Medical Association Masterfile or were identified by collected information on provider demographics, as well cancer patients who were surveyed as part of a larger as the characteristics of providers’ patient panels, includ- study of disparities in genetic testing [10–12]. The states ing the percentage of patients who are black, as well as included were chosen for the diversity of their popula- the percentage who have no health insurance or are in- tions as well as the ability to recruit patients directly sured by Medicaid. Items asking respondents to rate from the Pennsylvania and Florida state cancer regis- their agreement with a series of statements about the tries of the State Departments of Health. Physicians were costs of care were modeled after previous studies, contacted by email and/or postal mail. Institutional review including: “patients should have access to all effective board approval was obtained. The provider survey treatments for their cancer regardless of cost,”  response rate was 29.2% using American Association for “oncologists have a responsibility to balance the poten- Public Opinion Research rate 4 . Respondents and tial benefit of a drug with the potential cost of the drug,” non-respondents did not differ significantly by age “it is only important to consider the costs of treatment if (p = 0.69) or sex (p = 0.10). The University of Pennsylvania they are not covered by insurance” , and “high cost and Massachusetts General Hospital Institutional Review drugs should not be offered to patients when they have Boards approved the study, and considered completion of minimal effect on survival.” a questionnaire as implicit informed consent. Logistic regression was used to examine whether level of Four hundred and sixty medical oncologists confirmed compensation, physician characteristics, views about costs that they see breast cancer patients and responded to of care, and patient-panel demographics predicted physi- questions about bundling. Breast cancer care was felt to cians’ likelihood of participation in bundling. We addition- be an important focus for a study of bundled payments ally carried out ordered logistic regressions, with likelihood owing to the prevalence of the disease . A desire to of participation analyzed as a 5-level dependent variable. understand the utility of bundling for breast cancer has also led to the development of a pilot bundling program Results focused on this cancer . Providers were randomized Sixty-eight percent of oncologists were male. Mean age was to receive one of four survey versions each describing a 50. Twenty-two percent of providers were extremely or bundled program paying a specific amount ($5000, considerably involved in insurance contracting (Table 1). $10,000, $15,000, and $20,000) for medical oncology and The majority of surveyed providers supported ensuring infusion costs for the first year of localized breast cancer patient access to effective treatments regardless of cost treatment. For example, for providers randomized to re- (75%), but a similar number of respondents felt that on- ceive the version of the survey describing the $5000 cologists have a responsibility to balance the costs and bundled payment program, the survey asked “how likely benefits of drugs (78%) (Table 2). would you be to participate in bundled payment for lo- The proportion who were extremely or very likely to calized breast cancer if you received a single payment of participate in bundled programs was lowest among $5,000 for the first year of treatment of a patient with lo- providers who received the survey describing a $5000 calized breast cancer? The payment would include all program (17%), but was higher for the $15,000 (41%) medical oncologist and infusion costs but not drug, im- than for the $20,000 program (34%) (Table 2). In the re- aging or other costs.” Similarly, providers randomized to gression and unadjusted ordered logistic regression receive the $10,000, $15,000, and $20,000 versions of the model, likelihood of participation increased from the survey were asked how likely they would be to partici- $10,000 to the $15,000 program, but not from the pate in a bundled payment program offering $10,000, $15,000 to the $20,000 program (Table 3, Table 4). The $15,000, or $20,000, respectively, for the same list of ser- adjusted ordered logistic regression model yielded simi- vices for patients with localized breast cancer. There lar results (OR 2.3, p < 0.001 for $15,000 vs. $5000; OR were no differences in the patient scenarios or excluded 1.9, p = 0.02 for $20,000 vs. $5000). treatments described between the different versions of Likelihood of participation was higher among older the survey. Payment levels were selected based upon oncologists (OR 1.03 for each year, p = 0.005) and those expert opinion and were in keeping with published believing cancer patients should not be offered high-cost Murciano-Goroff et al. BMC Health Services Research (2018) 18:391 Page 3 of 5 Table 1 Characteristics of study participants Table 2 Medical oncologists’ willingness to participate in bundling and agreement statements about costs of care Medical oncologists (n = 460) Willingness to participate in bundled Extremely likely/very payment programs paying likely to participate Sex Female 145 (31.5%) $5000 (n = 123) 17% Male 315 (68.5%) $10,000 (n = 129) 25% Age 30–39 y/o 105 (22.8%) $15,000 (n = 116) 41% 40–49 y/o 115 (25.0%) $20,000 (n = 92) 34% 50–59 y/o 124 (26.9%) Costs of care: Strongly/somewhat agree 60–69 y/o 97 (21.1%) Patients should have access to all 75% 70–79 y/o 16 (3.5%) effective treatments for their cancer 80–89 y/o 3 (0.7%) regardless of cost (n = 459). State Florida 100 (21.7%) Oncologists have a responsibility to 78% balance the potential benefit of a New Jersey 81 (17.6%) drug with the potential cost of the New York 154 (33.5%) drug (n = 460). Pennsylvania 125 (27.2%) It is only important to consider the 23% costs of treatment if they are not Race or ethnicity Asian 100 (21.7%) covered by insurance (n = 460). Black 13 (2.8%) High cost drugs should not be 60% offered to patients when they have Hispanic 20 (4.4%) minimal effect on survival (n = 457). White 305 (66.3%) Other 29 (6.3%) from $5000 to $15,000, but did not increase further with Percentage of provider’s patient 0–5% 145 (31.6%) payment above $15,000. This threshold effect raises the panel who is black 6–100% 314 (68.4%) possibility that increases in price may not overcome re- luctance to participate among a substantial proportion Percentage of provider’s patient 0–5% 210 (45.8%) panel on Medicaid of oncologists. 6–100% 249 (54.3%) Participation in bundled payment requires assuming a Percentage of provider’s patient 0–5% 360 (78.4%) level of risk as patients’ clinical courses may be unpredict- panel with no health insurance 6–100% 99 (21.6%) able leading to uncertainty regarding costs [2, 17, 18]. Degree of provider’s Extremely or 101 (22.2%) Even among patients with localized breast cancer, like involvement in insurance considerably those described in this survey, a variety of disparate treat- contracting Somewhat, 354 (77.8%) ments may be required and complications may arise. Our slightly, or not survey focused exclusively on medical oncology and at all infusion costs, which may differ according to a variety of factors including a patient’s tumor size, hormone receptor drugs with minimal survival benefit (OR 2.1, p = 0.002), and Her2/neu status, tumor gene expression sta- but was lower among females (OR 0.48, p = 0.009). tus, whether the patient is pre- or post-menopausal, as well as individual patient variability . Bundled payment Discussion programs thus require physicians to accept a degree Understanding physicians’ attitudes toward bundling is of uncertainty. critical to the success of current compensation reform Tolerance of risk varies substantially across individual efforts. To our knowledge, this study is the first to focus physicians [20–22]. Furthermore, some providers have on oncologists’ willingness to participate in bundled relatively little understanding of the costs of care [23, 24] payment programs at different reimbursement levels and Table 3 Unadjusted ordered logistic regression on likelihood of offers two major findings. participation in a bundled payment program First, a minority of oncologists were interested in par- Category Odds ratio p-value 95% confidence interval ticipating in bundled payments for breast cancer care. Bundled payment amount Previous studies estimated that 6–17% of physicians sup- port or are “very enthusiastic” about bundling [1, 2], $5000 Ref similar to the percentage of our respondents who would $10,000 1.78 0.10 1.14–2.76 participate in a $5000 program. Second, the proportion $15,000 2.62 < 0.001 1.65–4.17 of respondents interested in participating in bundled $20,000 2.22 0.002 1.34–3.62 programs increased as the compensation level increased Murciano-Goroff et al. BMC Health Services Research (2018) 18:391 Page 4 of 5 Table 4 Logistic regression and ordered logistic regression modeling likelihood of participating in bundled payment program Logistic regression Ordered logistic regression Category Odds ratio p-value 95% confidence Odds ratio p-value 95% confidence interval interval Bundled payment amount $5000 Ref Ref $10,000 1.60 0.16 0.83–3.07 1.82 0.01 1.16–2.86 $15,000 3.19 < 0.001 1.69–6.03 2.35 < 0.001 1.45–3.79 $20,000 1.94 0.07 0.96–3.93 1.87 0.021 1.01–3.17 Age 1.03 0.005 1.01–1.05 1.02 0.01 1.00–1.04 Gender Male Ref Ref Female 0.48 0.009 0.28–0.83 0.60 0.009 0.41–0.88 Level of involvement in insurance contracting decisions for practice Somewhat/slightly/not at all involved Ref Ref Extremely/considerably involved 1.55 0.10 0.92–2.62 1.04 0.86 0.67–1.61 Views about the costs of cancer care: High cost drugs should not be offered to patients when they have minimal effect on survival. Somewhat or strongly disagree/neither agree nor disagree Ref Ref Somewhat or strongly agree 2.09 0.002 1.30–3.38 1.61 0.009 1.12–2.29 Patients should have access to all effective treatments for their cancer regardless of cost. Somewhat or strongly disagree/neither agree nor disagree Ref Ref Somewhat or strongly agree 1.02 0.93 0.61–1.71 0.96 0.84 0.65–1.42 Oncologists have a responsibility to balance the potential benefit of a drug with the potential cost of the drug. Somewhat or strongly disagree/neither agree nor disagree Ref Ref Somewhat or strongly agree 1.01 0.98 0.59–1.73 1.34 0.17 0.88–2.04 It is only important to consider the costs of treatment if they are not covered by insurance. Somewhat or strongly disagree/neither agree nor disagree Ref Ref Somewhat or strongly agree 1.31 0.32 0.77–2.22 1.27 0.26 0.84–1.95 The regression model was adjusted for physicians’ geographic location by US state as well as for characteristics of physicians’ patient panels, including the percentage of patients who are black, have no health insurance, and are covered by Medicaid. Responses to questions about physicians’ patient panels were collected on a 5-point scale (< 1%, 1–5%, 6–20%, 21–50%, 51–100%), and dichotomized as < 5% of patients versus 6–100% and may not feel responsible for managing costs . Our affiliation, which correlate with attitudes regarding bund- results accord with studies in which providers affirmed ling. Our study is further limited by the fact that our sur- the importance of cost-conscious care but opposed vey did not include items to explicitly assess providers’ restricting access to effective therapies [1, 16, 25]. Also in level of understanding of payment systems. In practice, keeping with previous investigations showing that older physicians’ with different levels of understanding of reim- physicians were more likely to support withholding costly bursement systems are likely to be affected by changes to therapies with little clinical efficacy , older oncologists physician reimbursement schemes, and further study is in our study were more supportive of bundling. Unexpect- needed to elucidate how education about payment reform edly, women were less likely to be interested in bundled impacts physicians’ willingness to accept bundled pay- payment for breast cancer care, an association that has ments. Our analysis did control for self-reported involve- not previously been described to our knowledge. Further ment in insurance contracting and did not find level of research will be needed to understand these patterns. involvement in contracting to be a significant determinant This study is limited by small sample size and response of willingness to participate in bundling. As an increasing rate. While our model controlled for such demographic number of initiatives are launched to reform physician factors as the percentage of patients in physicians’ panels payment, oncologists’ may be further exposed to bundled who are black or are uninsured, there may be other differ- payment programs and their attitudes may change . ences in practice characteristics, such as academic Our study focused on providers’ willingness to participate Murciano-Goroff et al. BMC Health Services Research (2018) 18:391 Page 5 of 5 in bundled payment programs for localized breast cancer, 5. Burwell SM. Setting value-based payment goals–HHS efforts to improve U. S health care N Engl J Med. 2015;372(10):897–9. and it is unknown whether these results are applicable to 6. Press MJ, Rajkumar R, Conway PH. Medicare's new bundled payments: other tumor types. design, strategy, and evolution. Jama. 2016;315(2):131–2. 7. Tsai TC, Joynt KE, Wild RC, et al. Medicare's bundled payment initiative: most hospitals are focused on a few high-volume conditions. Health Aff Conclusion (Millwood). 2015;34(3):371–80. In summary, this study of oncologists supports the 8. Oncology ASoC. Potential approaches to sustainable, long-lasting payment reform in oncology. J Oncol Pract. 2014;10(4):254–8. growing body of evidence that physicians have limited 9. Newcomer LN, Gould B, Page RD, et al. Changing physician incentives for enthusiasm for bundled payments and raises the possi- affordable, quality cancer care: results of an episode payment model. bility that higher compensation may not overcome re- J Oncol Pract. 2014;10(5):322–6. 10. McCarthy AM, Bristol M, Domchek SM, et al. Health care segregation, sistance to bundled programs among a substantial physician recommendation, and racial disparities in BRCA1/2 testing among proportion of oncologists. women with breast Cancer. J Clin Oncol. 2016;34(22):2610–8. 11. Dean LT, Moss SL, McCarthy AM, et al. Healthcare system distrust, physician Acknowledgements trust, and patient discordance with adjuvant breast Cancer treatment The authors thank Younji Kim for her administrative assistance in the recommendations. Cancer Epidemiol Biomark Prev. 2017;26(12):1745–52. manuscript submission. The Departments of Health of Florida and 12. Kim Y, McCarthy AM, Bristol M, et al. Disparities in contralateral prophylactic and Pennsylvania provided State Cancer Registry data. Published findings mastectomy use among women with early-stage breast cancer. NPJ Breast and conclusions are those of the authors and do not necessarily represent Cancer. 2017;3:2. the official position of the Florida or Pennsylvania Departments of Health. 13. (AAPOR) AAfPOR. Standard definitions: final dispositions of case codes and outcome rates for surveys. Lenexa, Kansas: AAPOR; 2008. Funding 14. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2017. CA Cancer J Clin. This work was supported by a National Institutes of Health/National Cancer 2017;67(1):7–30. Institute grant (5-R01-CA133004). 15. Warren JL, Yabroff KR, Meekins A, et al. Evaluation of trends in the cost of initial cancer treatment. J Natl Cancer Inst. 2008;100(12):888–97. 16. Berry SR, Bell CM, Ubel PA, et al. Continental divide? The attitudes of US and Availability of data and materials Canadian oncologists on the costs, cost-effectiveness, and health policies The research database is maintained by author K.A. associated with new cancer drugs. J Clin Oncol. 2010;28(27):4149–53. 17. Hassett MJ, Elkin EB. What does breast cancer treatment cost and what is it Authors’ contributions worth? Hematol Oncol Clin North Am. 2013;27(4):829–41. ix KA obtained funding and designed the study. KA, PWG, and SMD were 18. Frakt AB, Mayes R. Beyond capitation: how new payment experiments seek involved in acquisition of the data. YRMG, KA, and AMM carried out the data to find the 'sweet spot' in amount of risk providers and payers bear. Health analysis. MB and UNM offered administrative support. YRMG and KA drafted Aff (Millwood). 2012;31(9):1951–8. the manuscript, which was approved by all authors. 19. Gradishar WJ, Anderson BO, Balassanian R, et al. Breast Cancer Version 2. 2015. J Natl Compr Canc Netw. 2015;13(4):448–75. Ethics approval and consent to participate 20. Hancock J, Roberts M, Monrouxe L, et al. Medical student and junior The University of Pennsylvania and Massachusetts General Hospital doctors' tolerance of ambiguity: development of a new scale. Adv Health Institutional Review Boards approved the study, and considered completion Sci Educ Theory Pract. 2015;20(1):113–30. of a questionnaire as implicit informed consent. 21. Pines JM, Hollander JE, Isserman JA, et al. The association between physician risk tolerance and imaging use in abdominal pain. Am J Emerg Competing interests Med. 2009;27(5):552–7. Dr. Armstrong has a consulting role with GlaxoSmithKline. The remaining 22. Tubbs EP, Elrod JA, Flum DR. Risk taking and tolerance of uncertainty: authors have no conflicts of interest to disclose. implications for surgeons. J Surg Res. 2006;131(1):1–6. 23. Allan GM, Lexchin J, Wiebe N. Physician awareness of drug cost: a systematic review. PLoS Med. 2007;4(9):e283. Publisher’sNote 24. Reichert S, Simon T, Halm EA. Physicians' attitudes about prescribing and Springer Nature remains neutral with regard to jurisdictional claims in knowledge of the costs of common medications. Arch Intern Med. published maps and institutional affiliations. 2000;160(18):2799–803. 25. Schrag D, Hanger M. Medical oncologists' views on communicating with Author details patients about chemotherapy costs: a pilot survey. J Clin Oncol. Department of Medicine, Massachusetts General Hospital, 55 Fruit Street, 2007;25(2):233–7. GRB 740, Boston, MA 02114, USA. Department of Medicine, University of Pennsylvania Perelman School of Medicine, Philadelphia, PA, USA. University of Pennsylvania Abramson Cancer Center, Philadelphia, PA, USA. Received: 16 June 2016 Accepted: 13 May 2018 References 1. Tilburt JC, Wynia MK, Sheeler RD, et al. Views of US physicians about controlling health care costs. Jama. 2013;310(4):380–8. 2. Federman AD, Woodward M, Keyhani S. Physicians’ opinions about reforming reimbursement: results of a national survey. Arch Intern Med. 2010;170(19):1735–42. 3. Clough JD, Kamal AH. Oncology care model: short- and long-term considerations in the context of broader payment reform. J Oncol Pract. 2015;11(4):319–21. 4. Narayanan S, Hautamaki E. Oncologist support for consolidated payments for Cancer Care Management in the United States. Am Health Drug Benefits. 2016;9(5):280–9.
BMC Health Services Research – Springer Journals
Published: May 31, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
All the latest content is available, no embargo periods.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud