Curating and analyzing health-care data at scale has spurred innovation in personalized medicine. When we combine a range of data inputs, we improve our ability to deliver personalized care. Real-world data captures the patient experience and provides insights beyond what we can learn from clinical trials. The curation and use of real-world data is happening today, but it’s too limited.
As we continue to make data-informed care a more routine aspect of health care, we must also strive to create a more inclusive and representative data infrastructure. To do this, we must reach beyond the traditional avenues of working with multimodal data to further diversify the inputs that can inform personalized care.
Partnering with large academic medical centers is a strong start to delivering on the promise of precision medicine and establishing a robust multimodal data framework. But to better understand the profiles of patients from many different backgrounds, who may seek care in a variety of settings, we need to do more. We must broaden our approaches to collecting real-word data to fully capture the multitude of patient experiences.
The data we rely on should be representative and inclusive to truly further innovation in health care.
Focusing on community hospitals allows us to deliver advanced care to patients across the country, with different racial and ethnic backgrounds, living in a range of rural and urban settings. It also helps us understand their biological characteristics and clinical experiences to improve our ability to identify more tailored treatments. At Tempus, we’ve established integrations with both academic medical centers and community health-care networks to help make our genomic testing data more accessible and actionable for patients in different care settings. With these integrations, our genomic testing platform can connect with patients’ electronic medical records to make ordering tests and delivering results seamless and smarter throughout their treatment journey. By tracking their progress over time, we also have the ability to use longitudinal data to more fully understand a patient’s experience in a way that can help better manage their care.
We also can do more to help patients contribute their own health-care data in a structured way, to obtain truly personalized recommendations for treatment. We’re used to relying on apps to help manage our lives. Patient-reported outcomes tools deployed in apps provide evidence-based methods for tracking patients’ health and quality of life. They can be an effective tool for monitoring response to medications and helping patients get on the right therapy.
When we combine measurement-based care with powerful sequencing technologies and structured clinical data, we can help physicians find more targeted treatments for their patients. For example, our measurement-based care app is available to physicians providing mental health care in a variety of outpatient settings. This approach helps us support an even broader range of patients and empowers them to evaluate their own treatment, according to their unique profile.
As we continue to aggregate data to help individualize medicine, we can also use artificial intelligence to develop and deploy new algorithms. AI allows us to sift through vast amounts of data collected during routine medical care and, in turn, generate insights into conditions that were previously undetectable. These algorithms may be trained on internal datasets as we seek to make their outputs clinically actionable. However, to improve their real-world performance, we can conduct an external validation on an orthogonal dataset from one or more different patient population.
Working with academic partners, research institutions, and large health systems across the country to increase the diversity of validation datasets helps build confidence in AI-based tools. For AI to have the greatest impact in health care, it must be understandable and accessible. We need to explain our approach to training and validation to make clear that our work to validate on diverse datasets will help deliver AI-based precision medicine solutions to patients of different racial and ethnic backgrounds.
These are just some of the ways that we are working to expand the reach and increase the impact of real-world data to help individual patients get on an optimized treatment path. When we aggregate data from diverse sources, we also make it possible to discover new targeted therapies and design more efficient clinical trials, which can make our regulatory processes more efficient and effective. The data that we rely on to achieve all of these goals should be representative and inclusive to truly further innovation in health care.