“One day, all cancers will be rare cancers.”
I remember oncologist and Flatiron vice president Neal Meropol saying this at a national oncology conference a few years ago. His point was that in the past, we typically applied ‘one-size-fits-all’ therapies for metastatic colon cancer to treat 30,000 patients a year. Now, however, we can apply a series of molecular modifiers, each of which adds predictive or prognostic data to the diagnosis and allows us to deliver more precise treatment. For example, only 4% of patients with metastatic colorectal cancer - those with mismatch repair-deficient (MMRd) colon cancer - are likely to benefit from new immunotherapy drugs. With molecular diagnostics, we can avoid treating tens of thousands of patients for whom these drugs are unlikely to work and instead prescribe them for the relatively small number of patients most likely to benefit from them.
But here’s the problem with precision medicine: little attention is paid to what happens after the treatment decision is made. In other words, once the mutation is revealed and the care is delivered, the focus on precision ends. One of the biggest hurdles to controlling costs is a lack of innovation in precision delivery of healthcare.
What is ‘precision healthcare delivery’? I first heard this term from physician and healthcare consultant Blake Long about the challenges faced by famous physician and author Atul Gawande as he attempts to redesign healthcare delivery for a new enterprise backed by Amazon, Berkshire Hathaway, and JP Morgan. Blake explained that Gawande is charged with making care more efficient and less costly, while optimizing the individual patient experience – in short, he needs to innovate in precision healthcare delivery.
The ultimate aim of precision healthcare delivery is to account for each patient’s unique situation, preferences, and quality of life. In short, we need to find better ways to engage with our patients.
What this boils down to is understanding how to get the right drug to the right patient, at the right time, and at the right price. At Duke Forge, we’re seeking ways to use data to improve care. How then can we apply the principles of data science to innovation in precision healthcare delivery? We see several key areas for focused efforts:
- Incorporate patient-reported outcomes (PROs) into clinical decision-making. Recent evidence suggests that collecting and acting on PROs in patients receiving cancer treatment can increase survival by months, the same range of survival improvement seen with some new immunotherapies.
- Enable continuity of care. Even in this era of supposedly connected EHRs, I regularly see patients in clinic who have records faxed to our office, scanned into hazy PDFs, or even hand-carry reams of paper records to their appointments. Evidence suggests that better continuity in care can reduce healthcare utilization, and while EHR interoperability is only a start in terms of better care transitions, but it is a key component of care delivery that is inadequately addressed.
- Provide actionable data about cost of therapy. Out-of-pocket costs for patients often remain on the periphery of treatment decisions despite strong evidence that they impact quality of life and adherence to care. This happens partly because clinicians and staff lack the kinds of data that would empower them to address costs. However, information on how much a patient would owe for any given medical intervention is out there—it’s just often not available at the point of care. If price transparency is integrated into the clinical decision, patients can at least plan and account for upcoming costs.
- Document goals of care. Unsurprisingly, patients differ in terms of what’s important to them, particularly in the context of a life-limiting illness. In oncology, we’re seeing a trend toward prioritizing discussions about goals of care between clinicians and patients. Yet little is being done to ensure those discussions are documented and readily available. To complicate matters further, if the patient travels to another facility, the undocumented components of those discussions are lost.
Broad efforts are underway to incorporate data science into the precision healthcare delivery. For example, the Center for Medicare and Medicaid Innovation’s Oncology Care Model (OCM) is built on the points described above. However, barriers still remain, as OCM practices are struggling financially to meet care delivery standards. Payers should look to value-based insurance design, considering patient cost in relation to expected treatment outcomes. Clinicians should collect patient-reported outcomes as a part of standard care, track them over time, and use them as an ‘early warning system’ to identify patients at risk for adverse events.
The ultimate aim of precision healthcare delivery is to account for each patient’s unique situation, preferences, and quality of life. In short, we need to find better ways to engage with our patients. Before the term ‘precision medicine’ came into vogue, we used to talk to about ‘personalized medicine.' A mentor once told me, “First and foremost, personalized medicine means talking to the patient, personally.” It’s still true today.