Noise-Canceling Informatics: A New Framework for Personalized Medicine

Chief Executive Officer, R&D Ribitzky, and Distinguished Visiting Professor, Rockefeller Center for e-Health, Kigali Institute of Science and Technology, Rwanda

New discoveries in the diagnosis and treatment of cancer abound, yet how much of it is just noise, regardless of how scientifically promising, intellectually intriguing, or inspirational it might be?

By “noise,” I mean information that does not translate to practice, lead to clinical decision, or provide options for patients.

According to Francis S. Collins, MD, PhD, Director of the National Institutes of Health (NIH), “We have an abundance of scientific advances and new technologies, but we are using old tools for many steps in therapeutic development.” Announcing the National Center for Advancing Translational Sciences in December 2011, Dr Collins stated that, “Patients suffering from debilitating and life-threatening diseases do not have the luxury to wait the 13 years it currently takes to translate new scientific discoveries into treatments that could save or improve the quality of their lives. The entire community must work together to forge a new paradigm.”

In this article, I propose a framework to reduce noise in personalized medicine. The purpose of this framework is to accelerate the transformation of scientific discovery to clinical practice and patients’ choice. I propose to call it a “noise-canceling framework.”

But, first, has anything fundamentally changed in the practice of clinical decision-making (ie, diagnostics) over the past 200 years? My experience in clinical medicine made me realize that being comfortable making decisions in the face of uncertainty is inevitable. “The volume and complexity of the knowledge that we need to master has grown exponentially beyond our capacity as individuals,” stated Atul Gawande, MD, in his Stanford School of Medicine commencement speech in 2010.

Physicians are faced with 13,600 illnesses, more than 4000 medical and surgical procedures, and more than 6000 drugs that may qualify for reimbursement to treat these illnesses. More than 1.5 million medical terms could be used in the treatment of a patient, according to Jim Rossiter of NextGen, a healthcare information technology company.

Yet, referring to John Keats in 1816, Sherwin B. Nuland, MD, stated that “Medical education was brief in those days.…There was little of real usefulness to learn.…Patient care was conducted in a pervasive atmosphere of inexactness.”

He added that the 19th century “was characterized by…a form of practice based on knowledge gained by observation, hypothesis, experiment, and verification.…Uncertainty, if a factor at all, would be quickly dispersed by…laboratory or epidemiological studies.”

Yet, according to Dr Gawande, “half a century ago, medicine was neither costly nor effective.” As reflected by Lynn C. Epstein, MD, “[T]he passage of time has too often proven the espoused remedies of one era to be of limited value or frankly harmful in the next….”

Interactions between 2 or more illnesses that occur in the same patient, simultaneously or sequentially (ie, a comorbidity), make diagnosis and treatment more complicated. As Martine Extermann, MD, pointed out, comorbidity plays a significant factor in risk, survival, disease progression, and treatment of patients with cancer. “Older patients have an average of 3 comorbidities in addition to their cancer.”

Going forward, the cost of genomic sequencing is approaching consumer retail pricing, and “whole-genome sequencing for cancer is entering the clinic.” It appears, therefore, that value can easily be created by knowing what works and what works best, why it works, who it works for, and when it should be used optimally¹—founded not merely on clinical evidence, but also on weighted strength of that evidence.

That value proposition attracted many different organizations in the private and public sectors, as well as in grassroots individuals. Managed care companies began offering web-based information and knowledge services to their insured members; and healthcare provider organizations began offering similar services to their patients.

“Health 2.0 initiatives” became the umbrella term for collaborative user-generated healthcare content.

Established software vendors and content providers, as well as new ventures, began offering a wide range of knowledge-based solutions—from personalized and highly specialized cancer disease management (eg, n-of-One, Proventys, and CollabRx) to action-driven content aggregators (eg, FasterCures and the Combating Cancer of the European Commission’s Program Framework 6), disease-centric clinical documentation (eg, ProVation Medical), and collaboration enabling platforms (eg, BiomedExperts, the National Cancer Institute’s CaBIG, and the NIH-funded National Center for Biomedical Computing i2b2 [Informatics for Integrating Biology and the Bedside]), just to name a few. The Personalized Medicine World Conference 2012 is a microcosmic example of the growing scale and diversity of endeavors in this space.

Yet, making the right patient care decision still feels elusive, despite the ever-increasing signposts that are produced with great intellectual effort and scientific precision at immense cost. In sharp contrast to John Keats’s observation back in 1816, the practice of medicine evolved into an uphill struggle of making sense of too much information.

Our increased insight into illness and wellness, translation and expression, cause and effect, pathways and epigenetics, and preventive care and proactive medicine may not directly lead to a concrete clinical action that would yield a desired outcome, not today nor in the near future, which is why we need a noise-canceling framework.

The purpose of that framework is to guide the development of a new category of information technology that would have the following main capabilities:

• ŸA visual explorer that is capable of easily navigating intelligently and purposefully across multiple and diverse specialized systems, discovering explicit knowledge about the core elements that are at play in a particular situation (eg, patient, breast cancer, gene, drug, pathway), and the specific interrelationships among them (eg, cause and effect, historical association). By “easily” I mean for medical professionals, as well as for patients who may not have previous medical knowledge
• A visual narrator that is capable of organizing and presenting the search results using action-oriented visual semantics, which are contextually meaningful and weighted for value (eg, strength of evidence, outcome measures, trajectory to desired outcomes)
• A knowledge constructor that would retain these results for future reference and incorporation into a perpetual personalized care and discovery that will be capable of linking the clinical events and encounters with the then-current explicit knowledge (over time that information may be used for large-scale population-based discovery, thereby perfecting the personalized care process).

Developing this new category of information technology can be accelerated by:

• Grass-root demand for such capabilities from patients and the physicians taking care of them on the front line. That demand should be directed at the organizations that control which information technology solutions they use and to the information technology vendors that develop them
• New business models to enable low cost of development and massive distribution through social media channels, as well as attract grassroots developers to that cause in a way similar to Apple’s iTune and Apple’s Developer Programs
• ŸInnovative design mindset, which is a fresh design approach free from the conventional legacy drag and which may only be achieved by multidisciplinary, cross-pollinating teams of individuals from diverse disciplines
• Frontier technologies that take advantage of the collective capabilities and opportunities made possible by design philosophies, software development techniques, technologies (eg, Semantic Web), and microapplications design approach (eg, Mashup, service-oriented architecture, web services, data mining, agile software development).

Reference

1. Burrill GS. Biotech 2010 Life Sciences: Adapting for Success. April 14, 2010; MIT Enterprise Forum.