– In a society where nearly everything can be done digitally, customization is almost always expected. Whether it’s including special instructions in a takeout order or curating the perfect playlist, personalization is a major part of any industry – and it makes sense that patients would demand the same of healthcare.
With new sources of data and innovative technologies setting the stage for individualized care and precision medicine, providers can increasingly develop treatments that fulfill patients’ specific needs.
In this emerging, data-driven environment, physicians will have the option of tailoring therapies for every type of patient, including medically complex individuals and those suffering from multiple chronic diseases.
However, personalization in healthcare is easier said than done. Designing an effective treatment plan for a medically complex patient is far more difficult and delicate than choosing the right song. There is a lot of data to consider, and providers don’t always have the resources to assess every piece of valuable information.
Source: Xtelligent Healthcare Media
To accelerate the shift to precision medicine and personalized treatment, federal agencies have emphasized the use of real-world data to modernize clinical trials, an advancement made possible by the 21st Century Cures Act. With real-world data, researchers will be able to go beyond the scope of traditional trials, gaining insights from information collected in routine clinical care.
READ MORE: FDA Launches App to Support Clinical Trials, Real-World Evidence Collection
“A study using real-world data is entirely different from a randomized clinical trial,” said Dan Riskin, MD, CEO of Verantos and adjunct professor of surgery and biomedical informatics research at Stanford University.
“A randomized clinical trial is trying to show efficacy. It asks, how well does this work under optimal circumstances? In contrast, a real-world evidence study is trying to show how well something works in a real-world scenario.”
With real-world evidence, researchers can test which medications are most appropriate for certain patients, leading to individualized care and improved outcomes.
“Real-world evidence studies are commonly used for what’s called comparative effectiveness,” explained Riskin.
“There may be a lot of medications on the market, and all of them are proven to be effective and safe. A doctor may have a choice of 10 or more medications they can use for a patient with a certain condition. The only rational way to choose between these drugs is comparative effectiveness, meaning in the real world, which drug works better for this type of patient? Those real-world evidence studies are not readily available.”
READ MORE: How Precision Medicine Could Boost Chronic Disease Management
Currently, real-world evidence isn’t a major part of the medical industry, Riskin noted.
“To date, we haven’t implemented a lot of real-world evidence in this country, or in healthcare in general,” he said.
“We’ve tried to figure out what is and isn’t working by analyzing claims data, but there hasn’t been a rigorous approach to validate the information, and that creates challenges. We don’t know for sure whether the data is right.”
In a study recently published in the Journal of the American Medical Informatics Association (JAMIA), Riskin and a team of researchers evaluated the benefits and challenges of using real-world evidence in studies focused on cardiovascular medicine.
After examining over 10,000 clinical notes from an academic medical center, the team found that structured data in the EHR did not meet regulatory grade criteria. These results suggest that to achieve valid real-world evidence, researchers must ensure data generalizability and accuracy – the latter of which is usually overlooked.
READ MORE: Multi-“omic” Studies Could Accelerate Precision Medicine for Asthma
“One part of data validity is generalizability, which considers whether a dataset is relevant for a specific patient population,” said Riskin. “For example, if you’re treating a patient in a community hospital in Ohio, does a study conducted in an academic medical center in New York apply to that patient?”
“The other part of data validity is accuracy. If we’re using information that is inaccurate, then the results are questionable,” Riskin added.
“While the generalizability issue has been evaluated quite a bit, the accuracy issue is almost never looked at. That’s starting to change now that we’re using real world evidence to make more clinical assertions.”
Establishing a study protocol will be a critical step in making sure trial data is accurate and generalizable, Riskin stated.
“When we create a randomized clinical trial, we always have a protocol, which explains why the information is believable. Since real-world evidence studies are now influencing the standard of care, we need to have a protocol that includes generalizability as well as accuracy,” he said.
“Have we tested the accuracy of the information, or at least some subset of it? Is the level of accuracy high enough to make the clinical assertion we’re making? For example, if you use a dataset, and you look for, ‘did the patients ever have a heart attack’, and you find that only 30 percent of the people that had a heart attack were documented as having a heart attack, would that dataset be valid?”
Achieving data validity may require the use of advanced tools like artificial intelligence, machine learning, and other analytics technologies, researchers found.
“Data enrichment should not be done by humans. It’s also not financially feasible to have a person go through a million records. That’s where AI is incredibly helpful, because it can go through records over and over again and pull out the pertinent information,” Riskin said.
Policymakers have increasingly stressed the important role these tools will play in modernizing the clinical trial process. In a 2018 speech to the Bipartisan Policy Center, former FDA Commissioner Scott Gottlieb stated that real-world data gathered from EHRs, combined with advances in machine learning, will be critical for structuring the next generation of clinical trials.
And according to experts, that next generation is coming sooner rather than later.
“I think we’re in a transformational period for healthcare,” Riskin said.
“In the 2020s, we’re going to see data-driven healthcare, precision medicine, and personalized treatments. These trends are all based on the same enabling technologies from a decade earlier – high computing power, artificial intelligence, availability of electronic data. It’s natural that we’re going to use these tools to create tailored therapies.”
These advanced technologies, combined with real-world data, will accelerate the integration of precision medicine with routine clinical care.
“In the past with evidence-based medicine, we would just say, ‘All people with hypertension have these 10 options, and all of them are effective, and we don’t know which is best for this person,’” Riskin said. “The next version of healthcare will say, ‘This person is a 73-year-old woman who has diabetes as well as hypertension, and after examining numerous people like her, we feel this particular treatment is the best.’”
“We want to get to a point where there is enough real-world data to compare the outcomes for subgroups of patients. Maybe it’s men over 60, or maybe it’s women who have hypertension and hyperlipidemia. This may help advance personalized medicine or precision medicine, which covers everything from genomics to clinical characteristics of the subgroups.”
As real-world evidence starts to play a key role in care delivery, researchers will need to ensure the data they use is accurate and valid.
“It’s an exciting area. We are going to create data-driven healthcare, the question right now is whether we do it the right or wrong way,” Riskin concluded.
“If we’re using real-world evidence to change the standard of care, it’s our obligation to do these studies properly. The goal is to develop high-quality science so that we can help establish the next generation of medicine.”