Abstract: Clinical pathways represent an effort to optimize value-based clinical decision making by standardizing (when appropriate) evidence-based care to offer patients the best outcome while remaining mindful of cost. Pathways developers uniformly consider efficacy and toxicity first. This paradigm is challenged in clinical areas where evidence generation is so rapid that such data is incomplete or does not exist. We describe our collective experience in developing a pathway for chronic lymphocytic leukemia, a clinical area where options have exploded but evidence is immature, to focus discussion on how to best proceed in similar clinical circumstances. We provide a summary of lessons learned as they apply to the broader context of pathways development.
The process of clinical treatment pathways development in oncology is, at first glance, fairly straightforward. The mantra of pathways developers has always been to look at all available published evidence and consider efficacy, toxicity, and cost (in that order).1,2 This paradigm has not been without criticism. For example, patients have generally not had a voice in pathways construction, and pathways do not consider real-world evidence. In addition, clinical trial data is often used to design pathways, but there is little doubt that patients treated on clinical trials are younger and healthier than patients seen in practice, ie, than those being treated on a pathway. In some cases, the development process for pathways is deemed proprietary and thus not transparent to other health care stakeholders.3
Despite the above limitations, particularly regarding the issue of transparency of pathway development processes, there has been remarkable consistency in final pathways content, and payers and many providers have adopted such pathways to improve care and control cost.4,5 Specifically, it has been found that all pathways programs embrace clinical trial enrollment, aiming to increase trial accrual; all recognize the need for an “exception” process that allows individualization of therapy when the pathways-endorsed regimen is inappropriate; and all pathways programs espouse regular and timely updating as evidence evolves. But what happens when the evidence evolves too fast? What happens when the evidence is incomplete, as it often is for drugs approved via accelerated Food and Drug Administration (FDA) programs like the Breakthrough Therapy Designation?
We recently convened a panel discussion of the leading pathways organizations to explore the current state of pathways development and maintenance in light of rapid advancements in treatment options for chronic lymphocytic leukemia (CLL). Participants were chosen in an effort to bring varied experiences together from provider-led pathways groups, payer-facing pathways groups, and commercial vendors. Attendees included representatives from the US Oncology Network, Eviti, New Century Health, ClinicalPath (formerly Via Pathways), the American Society of Clinical Oncology Clinical Pathways Taskforce, and The Community Oncology Medical Home (COME HOME) program. In what follows, we examine the pathways development paradigm (efficacy, toxicity, cost) in light of the evolving treatment landscape for CLL, drawing from the collective learnings of the panel discussion. We conclude with what we propose are the broader learnings beyond this case for pathways developers. These learnings should inform efforts in other diseases with rapidly evolving clinical landscapes.
Challenges With the Traditional Pathways Development Paradigm
The Evidence for Efficacy Is Problematic
The majority of new drug approvals by the FDA over the last few years have been for cancer drugs. Also, the majority of drugs being considered for and approved via accelerated FDA programs have been cancer drugs.6 Although the agency has maintained that these accelerated approvals must pass vigorous regulatory muster, recent evidence suggests that post-approval studies are frequently not completed and, when they are, the results often do not replicate results as initially presented.7-9 Let us consider the challenges faced by pathways developers in light of these accelerated approvals.
First, the regulatory endpoint may or may not have clinical significance. These endpoints are invariably surrogate endpoints and are meant to predict real benefit, but this may or may not prove to be true with further study.10,11 In addition, that benefit may not stand up over time, as these surrogates are really designed to speed the approval process and allow for early access. Progression-free survival (PFS) is, by far, the most common endpoint in CLL trials. But how should PFS be interpreted when one therapy is given continuously and another is given for a fixed number of cycles? This does not even begin to address the question of the significance of the endpoint to patients.
Second, increasingly, the accelerated approval studies either have no control arm or the control arm is of little or no relevance. This is particularly true in CLL trials where chlorambucil (due to its legacy FDA approval) is the control therapy, though it is rarely, if ever, used in practice.12 There is little or no chance that a randomized trial will compare the novel therapy with the true standard of care, and, given the rapidly changing landscape, prior therapies and subsequent therapies available to patients are impossible to control for. Sequence is definitely important, but how do pathways authors develop appropriate sequences with this incomplete data set? It is almost impossible to determine the appropriateness of therapy in relapsed and refractory disease given these limitations.
Third, the control population in the registration trials is not at all representative of patients seen in the real world.13 It has been known for some time that clinical trial enrollees are younger, fitter, and usually Caucasian. This problem is magnified in CLL where the average age at diagnosis is 71 years.14 Interestingly, this also contributes to an apparent discrepancy between how academic physicians view CLL as opposed to community physicians, since academicians are more likely to see clinical trial-like patients. In the case of CLL, academic physicians favor ibrutinib in the first-line setting whereas in the community, depending on the general health of the patient, single-agent rituximab or bendamustine rituximab (BR) is typically preferred, based on panelist opinions. Of course there is a randomized trial comparing ibrutinib and BR, but PFS was the study endpoint (favoring ibrutinib), and whether these patients were “typical” is arguable.15 In addition, the PFS benefit is not difficult to anticipate, since the ibrutinib therapy was continuous.
This academic-community divergence was also observed by panelists when considering biomarker testing. The most frequently discussed biomarker in CLL is 17p del. This clearly has prognostic significance, but whether it is routinely used to make treatment decisions is less clear (ie, whether it has predictive value). Panel discussion asserted that academicians feel this test is of significant value, but community physicians and pathway developers are less convinced. In general, if a biomarker is included in a pathway, the absence of testing leads to a hard stop, ie, the physician cannot complete the order without that data. In CLL pathways, this hard stop does not exist. Hard stops are tough to implement in pathways. They upset physician workflow, and unless there is universal agreement among clinicians, they generate conflict and resistance.
Toxicity Reported May Not Be Representative of Real-World Toxicity
The difficulties in assessing efficacy spill over to evaluating toxicity. This is of obvious significance in patients with comorbidities of varying severity. But there are two other important and frequently ignored toxicity issues. The first is the issue of cumulative toxicity. In most trials, toxicity during the trial period is reported. But with ibrutinib, many patients are on the therapy for years. This leads to two significant cumulative toxicities. The first is cost, or financial toxicity, which will be addressed below. The second is fatigue or asthenia, or, as one pathways advisor observed, “the patient feels like they are carrying a backpack full of bricks uphill.” In fact, analysis of the data of ibrutinib in CLL shows that true resistance is relatively unusual and that patients discontinue therapy for other reasons.16,17
The second toxicity issue that may be underappreciated are those unfamiliar and often serious toxicities listed in black box warnings. Though these toxicities are captured as grade 3 or 4 in the pivotal trials, the ability to recognize and manage them appropriately is key to safe and effective treatment. In the case of idelalisib (and other PI3 kinase [PI3K] inhibitors), these toxicities include autoimmune colitis, pneumonitis, and severe opportunistic infections.18 Many of these should be managed more like autoimmune disorders as opposed to standard chemotherapy toxicity. Recent PI3K FDA approvals like that of duvelisib include recommended actions in the label to reduce the occurrence and appropriately manage these toxicities.19 Venetoclax-associated tumor lysis was a similar intimidating and potentially serious toxicity,20 but most physicians are now well aware of this problem, and, as a consequence, there may not be the same degree of unease. How do pathways capture the essence of these toxicities? How do standardized approaches mitigate these risks, and should they be part of pathways?
The Cost Picture Is Incomplete
Currently, pathways developers only consider cost after efficacy and toxicity. But, because of incomplete data regarding efficacy and toxicity (particularly a lack of long-term data), many of these novel treatments are indistinguishable. In these circumstances, where everything seems more or less the same, cost becomes more of a consideration. Unfortunately, historically, the only available cost metric was the cost of the drug (generally expressed as average sales price, or in the case of oral drugs wholesale acquisition cost or some derivative). Most manufacturers loudly protest this approach to cost analysis; they argue total cost of care is much closer to the truth. In addition, this approach is obviously totally blind to patient out-of-pocket costs, which are definitely relevant. Looking at de-identified Oncology Care Model (OCM) cost data, which includes total cost of care, is instructive.
But what does total-cost-of-care analysis really show? Total cost-of-care data, as currently available, does not provide much insight. For one, this information in claims data is devoid of clinical context, because critical data, including line of therapy and molecular markers, are missing. The attribution methodology also probably does not accurately define the start and stop dates for a given episode. In light of these circumstances, several facts are true in our case study discussion: patients with CLL are not admitted to the hospital very often, and the cost of the drugs are overwhelmingly the largest cost category. But, because oral drug costs are only captured if the Medicare beneficiary has Part D coverage, this is a highly selected patient population. When one considers the numerous commercial pharmacy benefit designs, the challenges with the “average patient out-of-pocket costs” becomes obvious.
The other issue incompletely considered in pathway development is the issue of cumulative cost. In a universe of arbitrary 6-month episodes of care (eg, as defined in the OCM21) this can easily be forgotten. But the community oncology perspective iterated by panelists is that, irrespective of the monthly cost of BR vs ibrutinib, the fact that BR is given for a fixed number of cycles automatically makes it more cost effective—this does not even consider the fact that the patient out-of-pocket cost is invariably lower for drugs paid as a medical benefit compared to drugs paid on the pharmacy benefit.22 There is also ample evidence of the efficacy of ibrutinib in the second-line setting23 as well as that patients have generally good quality of life during their treatment-free interval,24 which almost certainly has a positive impact on the “cost” of therapy.
Do Pathways Continue to Bring Value and How Do We Make Them Better?
Despite all of the limitations noted above, pathways, even in diseases like CLL, are worthwhile.1,4,25-28 Standardization in health care is widely believed to be of value in and of itself, as espoused by the general view articulated by Brent James, MD, of Intermountain Healthcare.29,30 Certainly pathways provide a construct for systematic evaluation of outcomes including cost, and pathways provide a mechanism to control rogue prescribing behavior. But CLL may represent the new normal. Multiple myeloma suffers from the same set of challenges and, arguably, so does renal cell carcinoma. These diseases call for a new, nonbinary pathways approach. These new and improved pathways could involve a “branched logic” paradigm, offering pathways choices based on an if-then model. Although this is not a complete solution, it allows incorporation of as much evidence as is currently available to optimize pathways content.
We posit several take-home messages from the case study provided and in consideration of the collective learnings gathered from the panel discussion at Box 1. These learnings can inform pathways efforts in similar diseases with rapidly evolving clinical landscapes.
Conclusion
Pathways have become an important tool in the physician’s arsenal to provide the best possible care with an eye toward value. The rapidly changing clinical landscape will benefit from the standardization offered by pathways, but pathways must adapt to remain relevant.
References
1. Neubauer MA, Hoverman JR, Kolodziej M, et al. Cost effectiveness of evidence-based treatment guidelines for the treatment of non-small cell lung cancer in the community setting. J Oncol Pract. 2010;6(1):12-18. doi:10.1200/JOP.091058
2. McCutcheon S, Ellis PG, Hess R, Krebs M, Lokay K. Frequency of efficacy, toxicity and cost as the deciding factor when determining clinical pathways. J Clin Oncol. 2017;34(suppl 15):e18169. doi:10.1200/JCO.2016.34.15_suppl.e18169
3. Kuntz G. Recognizing the limitations surrounding clinical pathways to ensure appropriate use. J Clin Pathways. 2019;6(1):22-24. doi:10.25270/jcp.2020.2.00115
4. Zon RT, Frame JN, NeussMN, et al. American Society of Clinical Oncology policy statement on clinical pathways in oncology. J Oncol Pract. 2016;12(3):261-266. doi:10.1200/JOP.2015.009134
5. Zon RT, Edge SB, Page RD, et al. American Society of Clinical Oncology criteria for high quality clinical pathways in oncology. J Oncol Pract. 2017;13(3):207-210. doi:10.1200/JOP.2016.019836
6. Beaver JA, Howie LJ, Pelosoff L, et al. A 25-year experience of US Food and Drug Administration accelerated approval of malignant hematology and oncology drugs and biologics. JAMA Oncol. 2018;4(6):849-856. doi:10.1001/jamaoncol.2017.5618
7. Woloshin S, Schwartz LM, White B, Moore TJ. The fate of FDA postapproval studies. N Engl J Med. 2017;377(12):1114-1117. doi:10.1056/NEJMp1705800
8. Naci H, Smalley KR, Kesselheim AS. Characteristics of preapproval and postapproval studies for drugs granted accelerated approval by the US Food and Drug Administration. JAMA. 2017;318(7):626-636. doi:10.1001/jama.2017.9415
9. Pease AM, Krumholz HM, Downing NS, Shah ND, Ross JS. Postapproval studies of drugs initially approved by the FDA on the basis of limited evidence: systematic review. BMJ. 2017;357:j1680. doi:10.1136/bmj.j1680
10. Chen EY, Raghunathan V, Prasad V. An overview of cancer drugs approved by the US Food and Drug Administration based on the surrogate endpoint of response rate. JAMA Intern Med. 2019;179(7):915-921. doi:10.1001/jamainternmed.2019.0583
11. Gyawali B, Hey SP, Kesselheim AS. Assessment of the clinical benefit of cancer drugs receiving accelerated approval. JAMA Intern Med. 2019;179(7):906-913. doi:10.1001/jamainternmed.2019.0462
12. Burger JA, Tedeschi A, Barr PM, et al. Ibrutinib as initial therapy for patients with chronic lymphocytic leukemia. N Engl J Med. 2015;373(25):2425-2437. doi:10.1056/NEJMoa1509388
13. Elting LS, Cooksley C, Bekele BN, et al. Generalizability of cancer clinical trial results: prognostic differences between participants and nonparticipants. Cancer. 2006;106(11):2452-2458. doi:10.1002/cncr.21907
14. Plute D, Castro FA, Jansen L, et al. Trends in survival of chronic lymphocytic leukemia patients in Germany and the USA in the first decade of the twenty-first century. J Hematol Oncol. 2016;9(28). doi:10.1186/s13045-016-0257-2
15. Woyach JA, Ruppert AS, Heerema N, et al. Ibrutinib regimens versus chemoimmunotherapy in older patients with untreated CLL. N Engl J Med. 2018;379(26):2517-2528. doi:10.1056/NEJMoa1812836
16. Maddocks KJ, Ruppert AS, Lozanski G, et al. Etiology of ibrutinib therapy discontinuation and outcomes in patients with chronic lymphocytic leukemia. JAMA Oncol. 2015;1(1):80-87. doi:10.1001/jamaoncol.2014.218
17. Mato AR, Nabhan C, Thompson MC, et al. Toxicities and outcomes of 616 ibrutinib-treated patients in the United States: a real-world analysis. Haematologica. 2018;103(5):874-879. doi:10.3324/haematol.2017.182907
18. Zydelig [package insert]. Foster City, CA: Gilead Sciences, Inc; 2014.
19. Copiktra [package insert]. Needham, MA: Verastem Inc.; 2018.
20. Venclexta [package insert]. North Chicago, IL: AbbVie Inc. South San Francisco, CA: Genentech USA; 2016.
21. Centers for Medicare & Medicaid Services (CMS). Oncology Care Model Overview. https://innovation.cms.gov/Files/slides/ocm-overview-slides.pdf. Published July 2019. Accessed March 1, 2020.
22. Dusetzina SB, Huskamp HA, Winn AN, Basch E, Keating NL. Out-of-pocket and health care spending changes for patients using orally administered anticancer therapy after adoption of state parity laws. JAMA Oncol. 2018;4(6):e173598. doi:10.1001/jamaoncol.2017.359
23. Byrd JC, Furman RR, Coutre SE, et al. Targeting BTK with ibrutinib in relapsed chronic lymphocytic leukemia. N Engl J Med. 2013;369(1):32-42. doi:10.1056/NEJMoa1215637
24. Osorio MJM, Pavlovsky C, Pavlovsky A, et al. Impact of ibrutinib in quality of life (QoL) in patients with chronic lymphocytic leukemia (CLL): preliminary results of real-world experience. Clin Lymphoma Myeloma Leuk. 2018;18(suppl 1):s218-s219. doi:10.1016/j.clml.2018.07.085
25. Newcomer LN, Malin JL. Payer view of high-quality clinical pathways for cancer.
J Oncol Pract. 2017;13(3):148-150. doi:10.1200/JOP.2016.020503.
26. Hoverman JR, Cartwright TH, Patt DA, et al. Pathways, outcomes, and costs in colon cancer: Retrospective evaluations in two distinct databases. J Oncol Pract. 2011;7(suppl 3):52s-59s. doi:10.1200/JOP.2011.000318
27. 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. 2017;13(5):e522-e529. doi:10.1200/JOP.2016.019232
28. Rotter T, Kinsman L, James E, et al. Clinical pathways: effects on professional practice, patient outcomes, length of stay and hospital costs. Cochrane Database Syst Rev. 2017;17(3):CD006632. doi:10.1002/14651858.CD006632.pub2
29. Bohmer RMJ, Edmondson AC, Feldman L. Intermountain Health Care. Harvard Business School Case 603-066. Published October 2002. Revised March 2013. Accessed March 1, 2020. https://www.hbs.edu/faculty/Pages/item.aspx?num=29326
30. Pearson SD, Goulart-Fisher D, Lee TH. Critical pathways as a strategy for improving care: problems and potential. Ann Intern Med. 1995;123(12):941-948. doi:10.7326/0003-4819-123-12-199512150-00008
31. Blau S, Ellis A, Frownfelter J, et al. Artificial intelligence applications in community oncology. Presented at: The 2019 Community Oncology Alliance Conference; April 4, 2019; Orlando, FL. https://communityoncology.org/wp-content/uploads/sites/20/2019/03/2019-conference-agenda-final.pdf. Accessed March 2, 2020.
