Bringing comprehensive health data interoperability to life science and pharmaceutical research

In 2020, catalyzed by a global pandemic, the healthcare segment began moving toward a solution that is now poised to spread across the entire health industry and revolutionize how we manage data in life sciences and pharmaceutical research: FHIR.

Heather Jordan Cartwright | November 19, 2020

If you work in the health and life science industry, you have had an endpoint in health care. From drug discovery to treatment delivery, our health care systems are the frontline of the journey between drug discovery to treatment. This intersection of human connection and clinical intervention creates a unique data estate of discovery, trials, validation, and real-world evidence data, to name a few. But serving this function also creates unique challenges for interoperability, global access, collaboration, security, and compliance. Input from the frontlines of health needs to flow in and out of multiple data platforms and organizational environments that can include enterprise, academic and or government systems. Providers and payors in health care have been grappling with this challenge for decades, but in 2020, the urgency of the global pandemic catalyzed new approaches to data exchange. The challenge of interoperability became poignant not just in health care, but across the entirety of the health ecosystem. How do we create an environment for the future of health industry that is both agile and standardized so that the data can be leveraged in a learning health data continuum?

The emerging solution: cloud technology and the open data frameworks of FHIR (Fast Healthcare Interoperability Resource). Developed to normalize clinical data exchange, the embrace of FHIR truly took off in 2020, to bring speed, interoperability, and improved security in the way health care providers exchange and manage electronic health care information. As we work collectively to fight a global pandemic, examples surface daily where FHIR is being leveraged to reduce the time and development costs of data exchange across systems. But even more exciting, we are beginning to see examples where FHIR accelerates data normalization for machine learning development and enables us to deliver better care.

See also: Microsoft’s Dr. Greg Moore on 5 major trends that will define the future of health care

In certain disciplines of basic research, FHIR can be regarded as ‘just a framework for clinical data’, not applicable to the complexities of imaging, pharmaceutical, or biotechnology workloads. And while FHIR is still early in its development, the adoption of precision medicine endpoints as surrogates for health outcomes is pulling it across the health data continuum.  In 2020, the Azure IOT Connector for FHIR was released to harness data from the ever-expanding Internet of Medical Things (IoMT) and map that data to FHIR. And as provider ecosystems were faced with overnight virtual health care and remote patient monitoring demands during COVID, they suddenly had a tool that could ingest normalized data from medical devices in a secure pipeline and persist that data alongside the clinical record. In October, Microsoft pushed the limits of FHIR, introducing new open-source technology to bring FHIR together with DICOM (Digital Imaging and Communications in Medicine). As data moves to the cloud, DICOM Cast technology can now extract the metadata from medical images, map it to FHIR, and create a new lens for patients that bridges clinical and imaging data. These evolutions have opened the door to seamless interoperability across research, clinical, device, and medical imaging data.

Lighting life sciences and pharma data on FHIR
Much like healthcare, pharmaceutical and life sciences companies grapple with the cost and complexity of clinical research and trials data. The high bar for extensive biomarker criteria, recruitment strategies, regulatory requirements, patient safety, and scientific rigor add unique challenges, but at the foundational level, the same problem exists. Data is locked away in disconnected systems and schemas, driving cost and complexity. So what if we ask the provocative question.  Can pharma and life science speak FHIR? While the FHIR framework isn’t fully designed today for life sciences – it’s foundation as open source enables it to expand as the community embraces and contributes to it, and the potential for it is compelling.

Challengers will submit that every research project is unique and requires special data formats.  There are validated elements of truth here, but rather than continue the development of new data schemas, consider how much faster we could achieve outcomes if projects were able to leverage consistent data exchange with clinical data sets? How much faster we could deploy a virtual clinical trial with the ability to anonymize and leverage data generated in previous efforts? Could we leap-frog epochs of discovery or early phase trials because the scientific community is able to expose previously siloed, disparate data?  Perhaps most importantly, what innovation could we fund simply by speaking the same open-source language health care system are now embracing speak for data exchange?

The ideation around FHIR and clinical data pipelines is no longer just an academic thesis.  In the US, the 21st Century Cures Act was passed in 2016 with a goal to “accelerate the discovery, development, and delivery of 21st century cures.” In 2020, it influenced new regulatory legislation that will require the use of FHIR for data exchange of clinical data. But the developers of the Cures Act are not only focused on the delivery of health care, but they are also strategically looking forward and considering how we can leverage precision medicine workloads, population health data sets and respective technologies to accelerate the future pipelines of artificial intelligence and machine learning.  The Cures Act introduced the with the potential use of real world evidence (RWE). Data such as lab results, prescription records, observational studies, and patient centric data can now be leveraged to help expedite drug discovery. The use of RWE opens a new horizon for the pharmaceutical and life sciences communities to think differently about how we bring frontline patient data back into the ecosystem. 

The ability to collect, annotate, and apply the power of machine learning to health data is the future of better outcomes and health transformation. We are doing it today – but we are doing it slowly because we aren’t speaking the same language for data exchange. The emerging use of the cloud for health workloads and FHIR technologies are creating a paradigm shift in the health industry. How much faster could we innovate when we normalize data exchange across the entire health care ecosystem?  How much faster can we apply science?  How much simpler can we make global collaboration on these datasets?  And how many people can we positively impact? The frontlines of healthcare are the endpoint for all pathways in health, but in 2020, its clear that ‘endpoint’ is just the beginning of the future of health data on FHIR.

About the Author

Heather Jordan Cartwright, Author

Heather Jordan Cartwright is GM of Microsoft Healthcare, the incubation team focused on delivering new experiences in health care technology within Microsoft’s Artificial Intelligence and Research division. In 2019, Heather was recognized in the industry as one of the 50 Most Powerful Women in Health care IT. Cartwright specializes in consumer-focused innovation and is widely recognized for her successful track record in multiple industries. Prior to Microsoft, she spent 10 years at Amazon. She began her career at Ford Motor Company, spearheading multiple projects including early development on the Ford Fusion.