The Role of Data in the Push for Precision Medicine
By Tariq Dastagir, MD, Lead Medical Director
In many ways, the care experience has always been personalized for an individual’s particular medical situation. But as we defined in our recent infographic “Deep Dive into Population Health: 5 Forces that Shape a Person’s Health,” there are many factors, beyond traditional medical variables, that impact a person’s disease state, including their lifestyle, social standing and genomics — factors that will increasingly come to bear on clinical decision making.
Precision medicine — defined by the National Institutes of Health (NIH) as “an emerging approach for disease treatment and prevention that takes into account individual variability in genes, environment and lifestyle for each person” — is a way to incorporate this more comprehensive set of patient data and use it to tailor treatments for the individual.
Such an approach yields numerous potential benefits. For example, a physician may avoid administering a treatment that a patient is unlikely to respond well to due to his/her unique attributes, thereby helping to diminish adverse drug events. Moreover, physicians may be able to better predict how chronic diseases might progress in individuals, as well as determine the likelihood that they will remain compliant with a particular treatment course. All of these scenarios could result in improved care and reduced costs — part of the Institute for Healthcare Improvement (IHI) Triple Aim.
Evidence-Based Decision Making
In order for precision medicine to be truly predictive and personalized, health systems have to analyze data from a wide variety of sources, such as electronic health records (EHRs), claims data, lab results, behavioral information, genomic results and more — both for individual patients and for the broader patient population. This enables technology to assist care teams to make clinical decisions that are not only tailored to a particular individual — using drugs that target the genetic mutations of a specific cancer, for instance — but are grounded in evidence for a large population.
For example, if a large number of blood cancer patients of a certain age, medical history and lifestyle tend to react poorly to a particular treatment path, a physician could leverage this evidence to avoid an intervention to a patient with similar characteristics.
In many areas, physicians already use similar kinds of data analysis to aid in clinical decision making. This can include for example, the use of algorithms to determine the risk of bleeding with the use of blood thinners for patients with atrial fibrillation (irregular heart rhythm), or the Framingham Risk Score, a gender-specific algorithm that can help to assess an individual’s cardiovascular risk over ten years.
However, such tools still largely depend on manual data entry and calculation by physicians and do not benefit from a broader wealth of population data. Ideally, an interoperable data analytics platform would automatically collect and analyze such data from a number of healthcare stakeholders. Without such capabilities, health systems will be limited in their ability to leverage patient data, including genomic and behavioral data, for the purposes of precision medicine.
Barriers to Precision Medicine
While the technology for precision medicine currently exists (at least in its initial stages), three obstacles, in general, prevent it from taking hold on a larger scale:
- The Lack of Structured Data. Much of the data contained within EHRs is unstructured (such as a physician’s notes) and therefore unusable for healthcare analytics. As Becker’s Health IT & CIO observes, however, natural language processing could reduce some of the effort needed to translate this data into mineable, analyzable formats. Otherwise, such efforts are largely manual.
- Aligning Incentives. Many healthcare systems are not yet able to justify the internal investment of adopting the analytic systems needed for precision medicine. However, as payments become increasingly aligned with outcomes and value, these systems may become more appealing to health system executives, as they can help to reduce costs and improve quality. Many health systems are already using risk prediction scores to reduce 30-day readmissions, for example, as they help to accomplish the same goal.
- A Uniquely Capable Workforce.To implement analytic systems which can enable precision medicine, healthcare organizations need a sufficiently large informatics workforce that not only understands healthcare analytics technology but its possible clinical applications and the highly specific considerations involved. Thus far, a lack of sufficiently trained professionals has presented a barrier.
While precision medicine shows incredible promise, there is still a long road ahead before health systems will be able to leverage its full potential. That said, as organizations increasingly adopt data and analytics technologies, their ability to wield precision medicine techniques and improve outcomes will only continue to advance.
How will your organization use precision medicine to enhance the quality of care?