Abstract: Clinical pathways are a critical structural element in the effort to reduce variation in health care delivery, allowing for cost reduction as well as more predictable, measurable, and accountable outcomes. At the 2nd Annual Clinical Pathways Congress (September 8-10, 2017; Washington, DC), a mock clinical pathway was conducted using a modified Delphi methodology. The goal was to make the process of developing consensus-driven, evidence-based clinical pathways in oncology transparent and educational. Nine medical oncologists and 1 surgical oncologist of diverse geography and practice type served on a steering committee to simulate the development of a clinical pathway for the treatment of metastatic breast cancer. The simulation provided an opportunity for meeting attendees to observe the process by which oncologists determine initial diagnostic tests critical to treatment selection, the most appropriate treatment for each line of therapy, and criteria for subsequent treatment selection.
Consensus-driven, evidence-based clinical pathways represent the natural evolution of treatment guidelines such as those developed by the National Comprehensive Cancer Network® (NCCN), which categorize the universe of evidence-based care options that might be included within pathways. Value-based care requires comparisons of various care options to establish consensus standards. It might be concluded that clinical pathways are the end products of those consensus standards and thereby critical to the transition to value-based health care delivery.
A host of oncology clinical pathways programs have been developed and have demonstrated an ability to modify physician prescribing behavior thereby improving clinical and financial outcomes.1-7 Like the NCCN guidelines, these pilot pathways programs were created by physician representatives of sponsoring institutions, eg, practices, payer networks, academic institutions. These representatives performed in-depth evidence reviews and achieved consensus in an iterative process to reduce the number of treatment options recommended for common clinical scenarios based on efficacy, toxicity, and cost. The resulting pathways were then disseminated to the broader physician group where adherence was encouraged, monitored, and usually rewarded financially.
The first pathways program of its kind was reported in 1998, when a group of private practice medical oncologists designed and implemented a novel cancer care delivery model that included the use of clinical pathways and treatment protocols.1 In the 3 years following the implementation of that model, resource use, specifically hospitalization, was reduced by 50%.1 Nearly 2 decades later, results from other clinical pathway programs have affirmed that cancer care cost savings and resource utilization, specifically reductions in cancer-related emergency room visits and inpatient admissions, can be achieved with provider participation in payer-supported or practice-initiated oncology pathway programs.2-7 These analyses also affirmed that as much as a 15% savings can be achieved in the first year of a pathways program on aggregated breast cancer, colon cancer, and lung cancer spending, with as much as a 7% reduction in hospital admissions.3,4 One analysis suggested an annual savings in excess of $30,000,000 for a mid-Atlantic commercial insurer with 3 million covered lives if the pathway program was expanded beyond the pilot to the entire provider network.4
It seems increasingly more apparent that consensus-driven, evidence-based clinical pathways in oncology will become a foundational platform in the transformation to value-based cancer care. Whether that care is delivered through an accountable care organization or an oncology medical home, reimbursed via episode of care or bundled payment, or configured as an Oncology Care Model (OCM), clinical pathways are a critical architectural element that reduces variance while also making costs and outcomes more predictable, measurable, and accountable.2 Previous assertions that pathway success requires appropriately aligned incentives and must be instituted as a collaboration between a payer, the majority of community providers, and other network providers may no longer apply in an era in which providers hold accountability for treatment cost, as is now the case for OCM-participating practices.
Increased attention to the utility of clinical pathways in value-based care has also heightened interest in the development process of clinical pathways. In 2016, the American Society of Clinical Oncology (ASCO) published a policy statement on oncology clinical pathways recommending that, in order for collaboration and consensus to occur, transparency is an absolute requirement.8 Cardinal Health Specialty Solutions (CHSS), an innovator in the pathway movement,2,3 has developed a mock clinical pathway program to simulate a process undertaken by payers and provider groups, allowing health care stakeholders to witness first-hand the process by which physicians determine rules, review evidence, and reach consensus in the creation of disease-specific oncology pathways. To remove any bias from the mock pathway steering committee’s decision-making, the programs have historically been double-blinded; that is, the steering committee members are blinded to the interested stakeholder, and the stakeholder is blinded to the participants (ie, physician name, practice name, city, state).
At the second annual Clinical Pathways Congress (CPC) in September 2017, a live mock clinical pathway was conducted, simulating the process of developing a clinical pathway for managing metastatic breast cancer (MBC) patients. The goal of the mock pathway simulation was to examine the willingness of community providers to participate in pathway development and to provide an opportunity for interested stakeholders—payers, providers, pharmaceutical companies, and patients—to observe the processes by which oncologists determine initial and subsequent diagnostic tests critical to treatment selection, timing and extent of such testing, the most appropriate treatment for each line of therapy, definition of treatment failure, and criteria for subsequent treatment selection.
Background
The Journal of Clinical Pathways was established in July 2015 to provide a peer-reviewed journal for the presentation of pathway-related research. CPC, the official meeting of Journal of Clinical Pathways, first convened August 2016 in order to provide an opportunity for stakeholders to further discuss clinical pathways and their impact on value-based care. Based on journal editorial interest and feedback from the first conference, the CPC program committee engaged CHSS to provide a mock demonstration of how a clinical pathway is developed to be presented live at the second annual CPC meeting, September 8-10, 2017, in Washington, DC. Subsequent discussion resulted in the decision to simulate the development of a clinical pathway in MBC.
CHSS developed a feasibility screener to identify 12 oncology specialists agnostic to group provider organization affiliation or institution from diverse geographic locations and practice type to serve as a steering committee. Clinicians were randomly selected from a proprietary master recruitment list. Clinicians were screened to identify oncology specialists (surgical, radiation, and mostly medical) who had at least 10 patients currently with breast cancer under active treatment per week and who have previously been involved in the development of practice- or payer-sponsored cancer treatment pathways.
A total of 12 oncologists met the screening criteria, of which 10 (9 medical, 1 surgical) accepted the invitation to serve on the steering committee for this initiative. A summary of the backgrounds of the 10 committee members is presented in Table 1. Physicians were compensated for their time at established fair market value rates, and their room, board, and travel were reimbursed.
The NCCN Clinical Practice Guidelines in Oncology for Breast Cancer with Evidence Blocks (version 2.2017)9 were distributed via email to the steering committee members with request to review them before the live event in preparation for mock development. The steering committee was only brought together live the day of the mock pathway program without prior interaction.
The Delphi method, a structured communication technique, developed as a systematic, interactive forecasting method which relies on a panel of experts, was described as the process by which pathway consensus would be achieved. The Delphi exercise consists of several iterative phases, beginning with broad concepts that become increasingly specific, gaining eventual consensus to all items in the pathway.10 Consensus was determined when at least 80% (8/10) of the experts agreed on a given item. Items not reaching consensus were summarized by the moderator and re-presented to the panel for further discussion with the aim to achieve consensus. New items for discussion needed to be voiced by at least 2 panelists in order to be considered for further refinement and achievement of consensus. The Delphi method for this exercise was slightly modified as a pre-meeting survey was not obtained.
Clinical Pathways Development Process
The process was conducted/moderated by a CHSS chief medical officer who is a board-certified medical oncologist with over 25 years of experience. The moderator was responsible for guiding the discussion of the committee as it attempted to reach a consensus decision.
Participants were instructed that the purpose of the exercise was to simulate a national payer-sponsored pathway development process for the treatment of MBC, taking into account the following criteria when selecting treatments: first efficacy; if efficacy is equal, then toxicity; if toxicity is comparable, then cost, where the less costly treatment is selected. Participants were instructed to keep in mind that the goal of the pathway was to reduce variability in treatment, attain meaningful clinical outcomes, and deliver the right treatment to the right patient at the right time. The preferred outcome was to achieve consensus on a MBC systemic treatment pathway with the fewest treatment option(s) needed for each line of therapy.
Ground Rules for Pathway Development
Setting the ground rules is critical to the transparency of pathway development, which is a key objective of the exercise. Many believe that the rules by which treatments are selected as on- or off-pathway are more critical than the treatment selections themselves. The experience and qualifications of the steering committee members; the type, consistency, and quantity of evidence to be used in consideration; the criteria upon which treatments will be selected for inclusion; and the adherence threshold by which an extended network of providers will be judged, are often the least understood but most important aspects of a pathway development process. Only once such ground rules are established can the actual work of pathway development begin.
Consistent with the process of being a payer-sponsored pathway, the moderator provided the panelists with the following general rules for deciding which treatments to include on the pathway for MBC: (1) a clinical trial is always compliant and the preferred therapy when available; and (2) palliative care and hospice are reasonable at any time for the appropriate patient.
The next discussed rule related to pathway adherence—the rate at which providers adhere to or deviate from the pathway selections. The concept was familiar to the participants as all had had some familiarity with clinical pathway development, but, for the audience, an explanation was provided. Adherence thresholds are critical to pathway development and success and are usually tied to reimbursement. Most of the published pathways have used an 80% adherence threshold: treatment selection above 80% is rewarded while below 80% is penalized. In the payer-sponsored environment, such penalties were usually financial. The more restrictive the pathway, the lower the adherence threshold, as more clinical scenarios arise in which the patient’s best interest would not be served by the pathway. Essentially, an 80% adherence rule states that it is expected and considered good clinical medicine that up to 20% of patients be treated off-pathway. Given the objective to develop a pathway with the fewest options at each branch point, an 80% threshold was unanimously selected.
The next set of ground rules related to the parameters by which treatment options would be evaluated and the type, extent, and consistency of evidence required. When discussing efficacy parameters by which to compare treatment options, the panelists considered overall survival (OS) to be the most important measure of efficacy, though they noted that this represents a challenge, as long-term data is required to draw proper conclusions. Panelists also noted that multiple lines of therapy can impact OS of patients without knowing which line of therapy contributed to the outcomes. They cited ASCO’s Net Benefit Score described in the recent ASCO Value Framework11 as a helpful approach for determining the overall clinical benefit. They also discussed the value of the NCCN Evidence Blocks9 in this regard. The discussion of the most important toxicity consideration drew a distinction between toxicity that is most important to the provider (in this case, neutropenia) and those that are most important to the patient (eg, hair loss, neuropathy, and gastrointestinal disruption). Although this discussion did not alter the efficacy, toxicity, or cost paradigm of comparative effectiveness, it was helpful to both participants and audience to level set on these critical rules of engagement.
Diagnosis, Patient Selection, and First-Line Treatment
The panel began by discussing approaches to patient selection. Breast cancer classification by pathological subtype: luminal A (HR+, HER2-); luminal B (HR+, HER2+/-); HER2 type (HR- and HER2+); or triple-negative/basal-like (HR- and HER2-) was rejected in favor of the simpler: hormone sensitive (HR+), HER2 sensitive (HER+), and triple-negative breast cancer (TNBC). The panel felt that this classification is more practical and reflects their daily clinical practice. Based on the distinctly inferior survival of TNBC, the participants agreed to begin the pathway there.
NCCN Guidelines for metastatic TNBC simply recommend “systemic chemotherapy.” The specific options for chemotherapy regimens within these guidelines are many, further categorized as preferred single agents, other single agents, and combinations. Varying scores on measures of efficacy, safety, quality of evidence, consistency of evidence, and affordability for these options are assigned and presented in NCCN Evidence Blocks.9 The steering committee members felt that the NCCN guidelines did not represent the current practice of TNBC management, with many of the listed options not used at this time in the United States. The panelists offered their own list of single-drug and combination-drug regimens (Table 2).
When comparing the efficacy of combination therapy and sequential single agents, the panelists were evenly split in their preference, resulting in a healthy debate. A few members felt strongly that combination therapy was needed for patients with high-volume, symptomatic, rapidly progressive disease, but 80% believed this to be a small minority of patients and well within the 20% off-pathway allowance.
The participants began their discussion of treatment for metastatic TNBC patients with a robust discussion of treatment-related selection factors that included the following considerations:
- Prior therapy (neoadjuvant/adjuvant)-type, response, adverse events
- Patient comorbidities
- Site(s) and volume of metastases
- Nature and extent of symptoms
- Patient choice issues: alopecia, oral vs intravenous
- Programmed death ligand 1 expression
- BRCA gene mutation status
- Psycho-social support and transportation
However, when the participants were challenged to identify which of the above treatment selection criteria were relevant to more than 20% of patients and thereby not adequately addressed by the 80% adherence threshold, only one emerged as critical to the pathway: BRCA status. Their explanation was that TNBC is more common in women with BRCA1 gene mutations and that BRCA may herald differential treatment response. Thus, it was determined that the clinical pathway for patients with TNBC should have 2 initial branches: one for BRCA-positive patients, and one for BRCA-negative patients.
Placing a branch point early into the treatment algorithm that required genomic testing begged the question of practicality for broad adoption. When these same participants were asked about the frequency with which they perform genetic testing, answers ranged from a low of 60% to 70% of those eligible for testing to universal testing of all patients meeting current guidelines (with the group suggesting that, in their experience, 20%-25% of patients meet these criteria).
Despite the realities of BRCA testing frequency, everyone on the panel agreed that, based on the strength of the clinical data, patients should first be stratified by BRCA mutation and, if positive, patients should first receive platinum-based therapy. There was consensus that BRCA testing would likely improve if spelled out in the pathway. When the panelists were asked to make a generalization as to what treatment they used for most BRCA-negative patients, they were first reluctant to choose a broad first-line therapy. However, when the question was phrased, “Looking at your last 10 charts, what would be the therapy used most often?” the panel evenly split between preferring Abraxane (protein-bound paclitaxel) or Taxol (paclitaxel). There was healthy debate among all panelists about the benefits of those 2 options. Discussion centered around toxicity of paclitaxel, with a focus on risk of developing neuropathy. Cost of therapy for Abraxane was also considered. Although the panel began evenly split between the 2 options, consensus was eventually reached to choose Taxol with certain patient exclusion factors (eg, preexisting neuropathy, prior adjuvant therapy) incorporated into the pathway.
Disease Progression
The panel agreed that second-line treatment need not take into consideration BRCA mutational status. An initial vote on a preferred second-line agent revealed preferences for each of the single agents: doxorubicin (if not used in adjuvant treatment), capecitabine, gemcitabine, and eribulin. The debate over capecitabine focused on the quality-of-life benefit of an oral therapy vs toxicity concerns, but eventually it was decided this option should be excluded from the pathway as a second-line treatment. Ultimately, doxorubicin (if not used adjuvantly) and eribulin were given parity status on the pathway in the second-line setting. The panel determined that gemcitabine, if not used in combination with platinum for a BRCA-positive patient, was the preferred treatment in the third-line setting, otherwise capecitabine.
The final clinical pathway developed by the panel for the treatment of patients with metastatic TNBC is shown in Figure 1.
It should be noted that mock pathways as conducted by Cardinal Health represent but one approach to pathway development and are not necessarily represntative of all pathways development. Cardinal Health brings unique experience to the process as it was previously the largest vendor for payer-sponsored clinical pathways in oncology and rheumatology between 2008-2014. But, as a result of MACRA and OCM shifting pathway development sponsorship from payers to providers, that business unit is no longer operational. Nonetheless, the observations from that work and the experience gained from moderating steering committees comprised of specialty physician representatives in oncology and rheumatology from large regional Blue Cross and national payer provider networks resulted in unique insights into the process. Structural elements like steering committee selection criteria and Delphi methodology consensus-building are not common in the published literature. Despite these differences the available evidence of pathways development processes suggest there is more that is common among them than is different. In this regard, we believe the process described herein provides valuable insights to all interested stakeholder.
Conclusion
The mock pathways experience validates the willingness and ability of community practitioners to actively participate in clinical pathway development. Mock pathways conducted in the described fashion provide an opportunity for stakeholders to witness and better understand the factors, considerations, and discussions that contribute to a successfully developed pathway.
References
1. Feinberg B, Feinberg I. Overall survival of the medical oncologist: a new outcome measurement in cancer medicine. Cancer. 1998;82(suppl 10):2047-2056.
2. Feinberg B, Milligan S, Olson T, et al. Physician behavior impact when revenue shifted from drugs to services. Am J Manag Care. 2014;20(4):303-310.
3. 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. 2012;18(5):e194-e199.
4. 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. 2013;9(5):e241-e247.
5. Nabhan C, Mato AR, Feinberg BA. Clinical pathways in chronic lymphocytic leukemia: Challenges and solutions. Am J Hematol. 2017;92(1):5-6.
6. 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.
7. 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.
8. Zon RT, Frame JN, Neuss MN, et al. American Society of Clinical Oncology policy statement on clinical pathways in oncology. J Oncol Pract. 2016;12(3):261-266.
9. National Comprehensive Cancer Network (NCCN). NCCN Clinical Practice Guidelines in Oncology for Breast Cancer with Evidence Blocks (Version 2; 2017). Updated April 26, 2017. Accessed February 3, 2018.
10. Fink A, Kosecoff J, Chassin M, Brook RH. Consensus methods: characteristics and guidelines for use. Am J Public Health. 1984;74(9):979-983.
11. Schnipper LE, Davidson NE, Wollins DS, et al. Updating the American Society of Clinical Oncology value framework: revisions and reflections in response to comments received. J Clin Oncol. 2016;34(24):2925-2934.
