If you were diagnosed with a health condition today, would you want a treatment with average effectiveness for most people — or one based on the specific physical, behavioral, and social drivers of your own health? Precision medicine offers the ability to recommend prevention and treatment that will work best for an individual.
Fully realizing this potential will require advanced analytics operating on massive amounts of data in a privacy-preserving very large scale.
What Is Precision Medicine?
This approach accounts for each person’s genes, health behaviors, and environment to provide the right care at the right time for each person.
Doctors know which treatments and preventions will work for most people as they navigate their health conditions. Finding the best one for a specific person, however, often requires a trial-and-error period with many viable treatment options. Precision medicine could help identify the most effective treatment much more quickly, which is why it is in use for conditions like autism, cancer, and stroke.
Elevance Health recently analyzed data on 5.5 million people with Type 2 diabetes and found more than 385 different treatment paths. Using precision medicine, analysts could predict the safest and most effective treatment choices for each person. The doctor could then work with the person to choose from those top options, instead of from hundreds of paths that may not as specifically address their healthcare needs.
Using precision medicine, a care provider can show how certain treatment options have worked in someone just like the person they are talking to. This personalized approach would help the person understand why their care provider recommends one option specifically for them, out of all the possible treatment options.
“We now have the technology and data to act in a way that is predictive, proactive, and personalized,” said Beau Norgeot, staff vice president of clinical AI at Elevance Health. “The potential benefits of precision medicine are a quicker route to safe and maximally effective treatment for each member, which should result in better health outcomes.”
To predict what will work for a specific person, doctors and health plans need quality and appropriately shared information from a variety of people experiencing similar health conditions. The information also needs to represent the full range of people throughout the population.
In order to predict what will work well for a specific person, doctors and health plans need information on what has worked well for highly similar people. The first step to doing this is research, and the National Institutes of Health (NIH)’s All of Us Research Program is helping with this goal of broad, representative data. It aims to build a large database that accurately captures the diversity of the U.S. population for research purposes. All of Us is a program from the federal Precision Medicine Initiative, an effort to transform how we improve health by personalizing healthcare. The next step will be operationalizing those research findings at scale at the point of care.
In addition to having the appropriate data, the health system needs safe and effective ways to analyze and share data and insights.
Patient registries -- observational data on people with a particular disease or condition — help organize data into a usable form. They can track interventions and outcomes as well as standardizing data collection. This is particularly important when many different data collection methods are possible, such as with genetic testing.
However, having a large and diverse data set is just the starting point. Going from observational data to insights about what intervention will work best for each person requires cutting edge analytics and ways to securely share the insights. This task is made even harder because real-world data, such as those found in registries, are often missing data.
Appropriately Sharing Data
Variations in the way data is collected and stored need to be accounted for — and eliminated where possible. The health system has made progress in standardizing and appropriately sharing some health data.
For clinical health data, electronic health records (EHRs) represent a major opportunity. EHRs have a wealth of data on millions of people over time for every condition and disease. The challenges again include data gaps, especially in behavioral and social drivers of health, historical mis- or under-representation of certain groups, and lack of standardization, especially in unstructured data like doctor’s notes.
An additional challenge with EHRs is interoperability. A 2020 interoperability rule from the Office of the National Coordinator for Health Information Technology (ONC) moved forward data and data transmission standards through secure, standardized, Application Programming Interfaces (APIs). Additional rulemaking and activities are anticipated, which will further promote interoperability between healthcare tools and stakeholders.
While there is a lot of progress to be made toward fully realizing precision medicine, the building blocks are in place to start personalizing healthcare decisions based on each person’s unique factors. When this more personalized approach to treatments is in place, people will be able to get the best treatment more quickly, with fewer doctor visits, lower costs, and less frustration.