Gordon Kuntz is a health care consultant and strategic advisor with over 30 years’ experience in a multitude of health care settings, working with payers and providers, and in technology and strategy. He began his involvement with oncology care pathways in 2004 as a consultant with US Oncology as they were deploying Level 1 pathways. He subsequently led payer strategy with ION Solutions, a division of AmerisourceBergen, where he gained familiarity with many other pathways vendors, especially in the context of the oncology medical home model. As senior director of strategy for Via Oncology, he deepened his knowledge of the pathway development process, physician adoption, and how both impact cancer center strategies. Mr Kuntz now provides support in strategy and product design as well as assistance in navigating the cancer care ecosystem to established and emerging companies.
Oncology clinical pathways are used by thousands of oncologists to select treatments for hundreds of thousands of patients every year.1 Oncologists may access pathways from a variety of sources, including practice-focused vendors such as Elsevier, McKesson, and Flatiron; vendors contracted by payers such as New Century Health, Oncology Analytics, and Eviti; or even created by a cancer center for their own use.1-6
Pathways have helped reduce unnecessary variations in care, helped countless patients receive the latest in lifesaving cancer treatments, and facilitated the enrollment of thousands of patients in clinical trials.7-9 Arguably, oncology clinical pathways are one of the single most important tools clinicians have in treating cancer patients. However, with all of those benefits and advances, all pathways systems share certain limitations.
These limitations are not necessarily good or bad, however, it is important to understand how pathways are built and what the limits of the technology, processes, and recommendations are so that the results can be better understood. Keeping in mind that pathways are a tool with certain limitations, and that they are only meant to augment but not replace the physician’s professional judgment, will make them even more useful for all stakeholders.
The Classic Three-Step Criteria for Pathway Creation
Since their inception, virtually all pathways systems in the United States have been and still are built using a three-step criteria. Upon evaluating treatment options for a given disease state, if one treatment has evidence that it is more effective than all others, based on published data, it is named as the treatment recommendation. If two or more treatments have comparable effectiveness, toxicity profiles are compared. Again, if there is a single drug or combination that is notably safer, it is recommended. If two or more drugs/combinations have comparable toxicity profiles then, and only then, is cost considered, ie, about 5% of the time.10 The goal for most pathways is that the recommendations would be appropriate for about 80% of all patients with that disease presentation (the 80% including, in most cases, the standard-of-care recommendation as well as clinical trials, palliative care, and any alternative recommendations offered for patient-specific situations not otherwise covered by the pathway).
This approach has been followed for the past 15 years by almost every pathways developer.11 The intent of this article is not to offer an alternative approach—this approach being perhaps too ingrained to allow much of a change anyway—but to create an understanding of the limitations of this approach and the resulting recommendations. Additionally, this article is not intended as a criticism or critique of pathways or any vendor but to help users be more aware of the workings of the systems they rely on.
Limitations in Oncology Clinical Pathways
Pathways, like all systems, are naturally limited in several ways. The data and information that are available as inputs, the way that they are built and managed, and the way they are used in practice all create limitations that are important for stakeholders to understand.
Pathways can only be created for diseases with an identifiable standard of care.
Pathways cannot be developed for disease presentations that lack one or more effective treatments supported by adequate evidence. Major pathways programs cover disease presentations that account for 95% to 97% of all cancer diagnoses in the United States today. Pathways may offer clinical trial options for rare diseases if research is ongoing. Of note, when dealing with rare/orphan diseases, pathways companies may be slow to develop pathways because of the cost of developing and maintaining these algorithms for such rare diseases as well as the difficulty in finding experts to lead the disease teams.
The rate of change in cancer care continues to accelerate.
The pace of change in treatment options is a bit of a double-edged sword for pathways. The proliferation of targeted therapies and new indications increases the necessity of pathways to help clinicians stay informed on the latest and most effective treatments for even more common disease presentations. In fact, the ability to help clinicians sort through the myriad of available treatments is one of the main reasons that pathways are so necessary and effective in oncology compared to many other diseases.
However, with these rapid advances comes the imperative for pathways developers to process more information faster. Considering that when pathways were initially developed for cancer care in the early 2000s,11 there were only a handful of drugs compared to the number of options now. Maintaining pathways is a very laborious process with numerous quality checks, and pathways companies could well have a challenge maintaining the current release schedule that some criticize as already too slow.
Pathways do not automatically reflect the latest drug approvals.
In most pathways systems, Food and Drug Administration (FDA)-approved drugs can be chosen as off-pathway regimens, however, they may not appear as on-pathway recommendations for some time, if ever. The reason for this, initially at least, has to do with the timing of pathways updates and the availability of data. If the FDA approval happens as one release cycle is closing, the review is likely to be delayed for 90 days or more, and if approved for inclusion, would not be released for about 120 days, including the time to load and quality check the recommendation set.
Although most pathways systems allow for priority review in the event of practice-changing data, that is a fairly high bar given the disruption and cost to insert another review and release cycle into their processes. Furthermore, as always, the new drug or new indication must show superiority in efficacy, safety, or cost to the existing standard of care. Although many clinicians or observers of pathways are surprised that they do not always reflect the latest advances in cancer care, this particular limitation is actually a feature, not a bug—the pathways development process is designed to deliver the most effective, least toxic, lowest cost recommendations, not simply the latest.
Pathways rely on published data.
Pathway’s reliance on published data is by design. Most pathways systems rely only on data that can be referenced and has public visibility. Information from internal manufacturers’ studies or physician experience does not rise to this level and is not considered in the decision-making process.
There are several reasons underlying this limitation: (1) the availability of published data sources allows pathways reviewers and clinicians to access the data independently and without bias to create the best recommendations for patients; and (2) most pathways systems provide links or reference citations for research used to support pathways decisions, so having vetted data that is published and available for reference is important.
Pathways are ultimately based on clinical trials.
It may seem obvious, but oncology pathways are ultimately based largely on the results of clinical trials. The inherent challenges of clinical trials are therefore reflected in pathway recommendations. First, head-to-head comparison studies are rare. Pathways developers attempt to mimic these through the pathway evaluation process, but, given the variability in trial design and endpoint selection, this can be difficult. Moreover, clinical trials are tightly controlled environments with significant amounts of support. The treatment given to patients enrolled in clinical trials does not necessarily reflect the situations patients experience in the real world. Patients in clinical trials may have less concerns about financial toxicity and receive more support than patients receiving drugs in a routine clinical setting. Finally, there may be inherent biases in clinical trials12 that could influence the development of pathways. For one, some studies with negative results are simply not published, despite recent regulations requiring it.13
Also, they often have a small sample size due to rarity of disease or specificity to disease presentation.
The presence of these limitations related to clinical trials as the underlying source material for most pathways is far from disqualifying but requires an understanding of the underlying data and the biases that may be introduced.
The strict efficacy/toxicity/cost model of pathways development makes defining value elusive.
The way most organizations develop pathway re-commendations is blind to value. If there is a single regimen in the review that has superior efficacy, then that treatment becomes the standard-of-care recommendation, regardless of the toxicity or cost (to society, payers, or patient). Similarly, if there are two or more with similar efficacies, safety and side effects are considered, but without visibility to cost.
Most systems use Average Sales Price (ASP) or some other industry-wide pricing, generally for the drug alone. Payer-specific reimbursements, costs of administering the drug and managing side effects, and the margin the practice might make are highly variable and difficult for a pathway vendor to specify and compare. Similarly, the other side of the value equation is benefit, a concept that is difficult to characterize independently.
The models that have tried to incorporate value have been difficult to implement.14 One reason may be that, as a society, the United States has resisted the notion of explicitly considering treatment benefit and total cost in a way that allows them to be compared.
Pathways do not reflect patient preferences.
Patient preferences—in the form of a lower toxicity regimen, a lower cost regimen due to insufficient insurance coverage, with the ability to minimize certain side effects, or through therapy designed to maintain quality of life for a period despite a poor prognosis—are not factors that are incorporated into the pathways determination process. Instead, these considerations may occur after the oncologist selects an initial treatment regimen or at best during a shared decision-making process once the standard-of-care recommendation is known. If these are common issues, there may be an alternate recommendation available in the pathway or may result in an off-pathway treatment selection.
Pathways are often limited to chemotherapy or immunotherapy only.
Some systems do include other treatment modalities or supportive care services, but many do not adequately reflect the needs of side-effect management, patient support, workup, and survivorship. The simple fact is that these areas of practice have fewer published studies supporting best practice and standards of care. Each clinic may create their own standards of care or look to one of several industry bodies that have published best practices approaches. As we saw above, creating pathways requires published evidence, which is rare in these areas of care.
Another similar challenge is that pathways may not reflect the technology that is available in the clinic. For example, if a patient in a community practice is eligible for chimeric antigen receptor T-cell therapy, the oncologist may be reticent to refer that patient to a distant cancer center due to the patient’s frailty or an unwillingness to lose the patient. Conversely, a cancer center that has invested millions in advanced radiology treatments may hesitate if the pathways do not include proton therapy, for example. Pathways vendors design their products to be used in a shared environment and are not in a position to reflect the resources available to each clinical site.
The existence of these multiple and possibly conflicting limitations means that pathways, while still a very valuable tool, need to be viewed through the lens of an oncologist’s’ expertise. Every pathways program allows for off-pathway treatment selection, a recognition that, ultimately, the physician is responsible for the patient’s care and has the best perspective on the best treatment for the patient. The notion of a single best recommendation, as is offered in some pathways programs, may imply precision, specificity, and authority instead of recognizing that pathways represent the current state of the art at any moment in time and are an attempt to codify the necessarily limited data and processes inherent in their design. Again, pathways are just like any other tool that comes with certain limitations; they are meant to augment but not replace the physician’s professional judgment. Understanding pathways limitations will make them even more useful for all stakeholders.
1. Daly B, Zon RT, Page RD, et al. Oncology clinical pathways: charting the landscape of pathway providers. J Oncol Pract. 2018;14(3):e194-e200. doi:10.1200/JOP.17.00033
2. Schroeder A. Clinical pathways: a current snapshot, and the journey ahead. J Clin Pathways. 2017;3(2):33-40.
3. Wong W. The evolution of clinical pathways in oncology. J Clin Pathways. 2015;1(1):37-42.
4. An in-depth look at Moffitt Cancer Center’s clinical pathways program and use in payer strategies—part I. J Clin Pathways. https://www.journalofclinicalpathways.com/depth-look-moffitt-cancer-centers-clinical-pathways-program-and-use-payer-strategies-part-i?page=0. Published online March 5, 2019. Accessed February 4, 2020.
5. Michael Hassett: DFCI pathway workflow integration and interoperability [video]. J Clin Pathways. https://www.journalofclinicalpathways.com/multimedia/michael-hassett
-dfci-pathway-workflow-integration-and-interoperability. Published online November 4, 2019. Accessed February 4, 2020.
6. Highlights from the 2019 Oncology Clinical Pathways Congress. J Clin Pathways. 2019;5(9):59-79.
7. 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
8. 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
9. 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.
10. 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
11. Gesme DH, Wiseman M, Strategic Use of Clinical Pathways. J Oncol Pract. 2010;7(1):54-56.
12. Higgins JPT, Savović J, Page MJ, Elbers RG, Sterne JAC. Chapter 8: Assessing risk of bias in a randomized trial. In: Cochrane Handbook for Systematic Reviews of Interventions. Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, eds. Cochrane Handbook for Systematic Reviews of Interventions. Version 6.0. West Sussex, England: John Wiley & Sons Ltd.; 2019. https://training.cochrane.org/handbook/current/chapter-08. Accessed February 4, 2020.
13. P&T Community. Rule on reporting clinical trial results often violated. January 21, 2020. ptcommunity.com website. https://www.ptcommunity.com/news/20200121/rule-reporting-clinical-trial-results-often-violated. Accessed February 4, 2020.
14. Vose J. ASCO Value Framework update [blog]. ASCO Connection. May 31, 2016. https://connection.asco.org/blogs/asco-value-framework-update. Accessed February 4, 2020.