Precision Medicine: It's Complicated
By Tim A. Sayed, MD, MBA, FACS
For The Record
Vol. 28 No. 8 P. 8
By now you've probably heard about precision medicine. It's a term that entered the professional and consumer health care vernacular circa 2013, and became popular following President Obama's 2015 State of the Union address in which he announced the formation of a national precision medicine initiative.
Precision medicine is frequently used interchangeably with the more controversial personalized medicine. Many bloggers and organizations, including the National Research Council, the National Institutes of Health, and various Fortune 500 companies, have spent energy parsing whether these terms describe distinct spheres of clinical and research medicine or are synonymous. Whatever the case, labeling something "precision" or "personalized" medicine begs the question: Have we been practicing imprecise and impersonal medicine all along?
The answer is complicated. At its most basic, precision medicine involves the tailoring of medical care to individuals based on some synthesis of unique patient data coupled with data derived from studying populations. This may occur through developing clinical pathways based on well-designed, randomized, double-blind controlled studies, followed by further modification at the point of care informed by patient genomics, proteomics, metabolomics, microbiomics, and other "-omics" signatures.
A true precision medicine paradigm requires statistical and logical tools that can extract trends, associations, and aggregations to yield meaningful information (signal) from background data chatter (noise).
No Clear Winner
Certainly, there has been no shortage of Big Data repositories and data science engines sprouting up in the last several years, most of which promise insightful clinical analytics to help drive actionable intelligence that achieves and manages changes in health care behavior or outcomes.
But, for the most part, these features are now table stakes—the minimum entry for cloud-based, mobile-ready, user experience-oriented, and data-heavy companies to offer a new spin on data analysis.
In health care informatics, like in the delivery of care by providers, no clear "winner" for driving precision care has emerged, and some may wonder, in an era of data security breaches, identity theft, and even willful public oversharing of private information, do we really want one?
Seeking a True Definition
What exactly is precision medicine? Is it simply tailoring drug prescriptions to a patient's unique genetic profile as culled from comparisons to large data warehouses holding meta-analyses of billions of DNA sequence variants obtained at research institutions? What about using old-fashioned evidence-based clinical trial results and applying the "better" therapies identified in treatment arms to only selected members of the cohorts? After all, we now may have additional biochemical and genetic data that would suggest some patients actually do better with the control arm therapy even though, at a population level, the control was statistically inferior to the treatment arm.
Perhaps precision medicine is combining these 20,000-foot surveys of therapeutic outcomes with the molecular view at an individual patient's cellular level to create a comprehensive, bespoke care plan. This care plan is one that evolves as additional data—from wearable fitness device logs and mobile device operating system health app kits to consumer-purchased behavioral analyses—begin to paint a motion picture of the patient's health that is neither static nor episodic, but continuous.
This may be called a "pan-optical" view of a patient's health, bridging telescopic with microscopic perspectives to create a cinemascopic patient record that evolves in real time.
Physicians may argue that they have been doing their best to provide precision medicine all along by crafting individual patient management plans that take into account biopsychosocial characteristics—personal care preferences, cultural norms, communication styles, financial and community resources—and supplementing them with published clinical evidence on a case-by-case basis, often with limited time to put it all together cohesively.
Today, there is even more information available, as well as substantial structured data in EHRs that used to be buried in unstructured, archived handwritten notes. Whatever the sources of these data—some gleaned through empiric observation and some through direct measurement or transfer of data from other stakeholders—physicians pride themselves on their ability to rapidly integrate them to deliver a high-quality and generally efficacious interaction with patients—the so-called art of medicine.
The difference is that physicians now have many novel, interesting, and diverse ways to track and quantify these heretofore seemingly unquantifiable factors. They can count a patient's "likes" or shared posts on certain social media platforms, canvass the opinions of targeted groups (eg, patients with a rare metabolic condition) through online survey tools, estimate a patient's caloric expenditure and intake from wearable devices and smart tools such as refrigerators that detect content movement, calculate with extreme accuracy the location of the nearest health resources (eg, physician offices, hospitals, pharmacies, labs) and find out what everyone else thinks about them, and untold other measurable indices of what makes each person that person.
Perhaps true personalized, precision medicine is the marrying of this swarm of quantified self-data to the simple act of asking patients what they want for their health and wellness. Whether through telemedicine portals or face to face, for these encounters to be effective, physicians cannot be expected to manually incorporate every quantum bit of information that could remotely influence the patient's health outcomes.
What's needed are smart aggregating and interpretive technological tools that can distill the information into simple, humanly readable, aesthetically pleasing interfaces that help users triage the following critical questions:
• Who needs the most immediate attention?
• What patient information is "can't miss" and what is "nice to know" when it comes to preventing harm?
• Can a patient's status be changed through an intervention or education?
• How quickly can changes be realized? Is care immediately actionable or will it take mental, physical, or even financial buy-in from the patient?
• How do the interests of the patient, physician, and health payer align? If they're in agreement, how can intervention be consolidated in order to cover the most ground for all stakeholders?
• Is there a more cost-effective way to get to the same outcome, or are medical savings coming at the expense of increased risk?
Business school marketing classes teach students the differences between functional and experiential brands. Think hardware supplies contrasted with fashion or makeup lines. One is primarily about accomplishing the goal that the product or service seeks to deliver, whereas the other is mostly about how the customer feels when purchasing or using the product.
Recently, an emphasis has been placed on improving the patient experience, including the introduction of software designed with the user in mind. More importantly, the movement has extended to providers, hospitals, payers, and other constituents hoping to boost patient interaction. These efforts are driven by online reviews, youthful startup companies challenging the status quo, and the widespread availability of application programming interfaces that allow users to plug-and-play a variety of powerful products into unified visual platforms.
At the end of the day, achieving precision and personalization in medicine will require not only tools that can collect, quantify, organize, parse, analyze, integrate, associate, and prioritize but also promote collaboration among patients, providers, and payers in an ecosystem of shared goals and incentives.
— Tim A. Sayed, MD, MBA, FACS, is vice president of patient engagement at Interpreta.