The NCCN guidelines have been described as a multi-lane superhighway. What we need of we want to drive value in oncology is a two-lane road. When I was chairman of the Pharmacy and Therapeutics Committee at US Oncology, our practice in Kansas City, led by Dr. Marcus Neubauer and John Hennessy had an idea. They hypothesized that a careful analysis of evidence supported treatment options (all NCCN endorsed) which considered efficacy, toxicity, and cost could lead to a “ranking” of therapies that truly reflected cost effectiveness. Further, they felt that this approach would benefit patients by standardizing care and that payers would be interested because it considered cost. They recognized that such a pathway would not be appropriate for everyone (for one reason or another), but that this would be a vast improvement over a doctor’s standard practice. The goal was 80% compliance. It was successful in Kansas City, and we decided to implement the clinical pathways network wide.
Over the next several years, we rolled the USON clinical pathways out across the practice network. We established committees of volunteer physicians and hired Pharm D’s to perform the clinical and economic reviews, covering the most common cancers we saw in our community practices. We sought and obtained approval from the corporate team and buy-in from our physicians. We embedded these pathways into the electronic medical record used in our network, providing point-of-care decision support. We measured physician compliance, as well as reasons for exceptions to the pathways in an effort to optimize pathways content. And of course we updated the pathways as new evidence emerged. We published our results, 35% cost savings in lung cancer (https://pubmed.ncbi.nlm.nih.gov/20539725/). But we were unable to help our practices in their payer relationships.
We did get a lot of criticism. We were accused of building the pathways to maximize profit; this criticism was easily dispatched when the pathways content was shown to heavily favor generic drugs. We were accused of “cook-book” medicine. Specifically we were accused of impeding the physician’s ability to personalize care. This criticism was driven by the pharmaceutical industry. In point of fact, since they had no voice in how the pathways were constructed, the pathways drove them nuts. All of their sales efforts were for naught. One legitimate concern was that there was no patient voice in our committees. However, the intent of the pathways was to create a basis for decision making that would be used in shared decision making. In fact, when I discussed with my patients “what I was up to” when I was absent from clinic and I told them about our pathways program, they were genuinely impressed and excited.
If imitation is the sincerest form of flattery, we were on the right track. As the USON clinical pathways program gained steam, several similar programs (some provider based and some commercial) were launched. Payers began experimenting with pathways programs as cost-saving measures. The economic results were mixed. But the IDEA of using a formal assessment of efficacy, toxicity and cost has taken hold. Both the American Society of Clinical Oncology (ASCO) and the European Society of Medical Oncology (ESMO) have published “value frameworks” that embrace this general framework. This “value-based”, evidence driven approach makes sense to oncologists.
When I discussed the patient centered medical home, I argued that although it was an excellent care model it wasn’t such a great payment model, and that drugs should be taken out of the equation. Pathways are an approach to taking drugs out of the equation. Or conversely, pathways should be part of the medical home model. In fact, when Ira Klein and I built the Aetna oncology medical home, we mandated that practices use pathways. And in our experience, Texas Oncology saved almost 20% of the cost of care predominantly through their use of pathways (https://ascopubs.org/doi/full/10.1200/JOP.17.00091?role=tab).). We did not use “Aetna” pathways, Texas Oncology used the USON pathways. Because I was familiar with how USON built their pathways I knew the process had structure, rigor, and integrity. To be fair, the “authors” of the pathway do not matter as much as the process by which the pathways are developed. Several of the other pathways programs have content that is similar to the USON pathways. Pathways are not created by a bunch of docs sitting around a table telling everyone what they think or do when faced with a particular clinical case. That does not constitute evidence based medicine.
Does this approach have legs? Can pathways remain clinically relevant and generate cost savings? Should we hold physicians accountable to care that measures up to an agreed upon evidentiary standard? Is this standardization actually good for patients? I think so. But there are challenges.
The first challenge is the rapidly changing state of medical knowledge. Can pathways adapt quickly enough? I actually do not think that “big” advances are likely to be missed. A new mutation with a targeted therapy is very easy to add to a pathway. I am concerned more about the disconnect between clinical trial results and real world efficacy/toxicity. Given the standardization introduced by pathways as well as the near universal adoption of electronic medical records, we should be able to learn quickly about what works and what doesn’t, and this should impact our pathways. But this is not happening. It needs to. Real world evidence and artificial intelligence are a topic for an upcoming post.
Can pathways still save money? That is a much harder question, but I think the answer is yes. Let’s discuss non-small cell lung cancer (NSCLC), the first disease we built a pathway for at USON. Our pathway was successful because: 1. None of the treatments available were very good and 2. One of the treatments used generic drugs. The result was that the generic regimen was just as good and a lot cheaper. Today, almost every patient with lung cancer gets a check-point inhibitor at some point, and appropriately so. But if you examine the various treatment options for advanced NSCLC there is still significant variability in cost without a ton of evidence that one is better than another. Even the various check-point inhibitors have variable costs without significant clinical differentiation.
Pathways also have a big impact on “cowboy” behavior. Some doctors really think they know more than everybody else, that they have the ability to synthesize an optimal treatment plan for each individual patient. Maybe these special “master clinicians” really exist but I am not sure how you identify them. I prefer to consider them (in the absence of evidence) “cowboys”. Cowboys are a pain in the neck for their practices, because they do things that the health plans do not recognize as evidence based. These denials at the minimum create a lot of work and at the maximum put the practice at financial risk. When we adopted pathways in the USON network, denials disappeared overnight.
Interestingly, pathways might also be a way to learn about novel therapies that enter the clinical space through accelerated FDA approval. If we plan to pay an “introductory price” for a new treatment until the post marketing data requirements have been fulfilled, pathways are an excellent vehicle to ensure the right patients are getting the new treatment. Pathways also provide an excellent deterrent to use these new treatments outside the clinical indications upon which the FDA granted approval. If the new therapies go through the pathways process and are deemed not good enough to be on pathway that tells us something, too.
Finally, pathways may be an excellent quality measure. A physician that is 80% compliant with a recognized pathway is delivering evidence-based care. And because pathways are “standardized” there is every reason to believe that medical errors could be reduced (for a superb presentation of why standardization is good for medicine, I refer the reader to Atul Gawande’s “The Checklist Manifesto”). If we move towards capitation, pathways ensure the patient’s care is not being compromised by omission of services in the pursuit of profit. I strongly believe all physicians should have pathways compliance reported as a quality measure (though that is rarely done today, even at NCCN institutions).
Pathways may not be a good solution forever. Pathway recommendations are designed to give the best possible results in a population similar to that originally studied. That is why exceptions are appropriate for patients that are not well represented in clinical trials. As artificial intelligence helps us better understand why some patients may have different responses to the same therapy we will need to evolve. I think that is a long way off.
Evidence based medicine is a cornerstone of good quality care, even if the evidence is imperfect and constantly changing. Pathways are only one example of how evidence-based medicine might contribute to defining quality. We will next discuss the broader challenges of what constitutes quality in medicine (and oncology in particular).