Blog | Linux Foundation

Linux Foundation Research Addresses the Value of Open Source in Health Data Management

Written by Anna Hermansen | Dec 5, 2024 9:56:31 PM

My first exposure to the application of emerging technologies in healthcare was during my time at the Blockchain Research Institute as manager of digital projects and research. There, I learned about the work of Dr. Phil Baker, tracking unused medication using an implementation of Hyperledger Besu as a means to redirect and reduce pharmaceutical waste. Writing a case study report on this implementation served two important purposes for me: I tackled the technical details of Besu, and I learned how more effective data sharing and tracking is needed in the healthcare sector. 

In my next role with a research lab at the BC Cancer Research Institute, I was again studying the application of blockchain for data sharing, this time in the context of precision oncology. In this role I learned about the immense gap between data needs and data access. Without access to data, researchers cannot evaluate the clinical effectiveness of an oncology treatment program, hamstringing advancements in care.

How can technology help? Blockchain was applied in these two cases as a way to securely track the sharing of data among authorized institutions. The metadata of these datasets could be read and shared by authorized parties, and their chain of custody and access became transparent. 

Despite the promise of blockchain and other technologies to improve health data accessibility and portability, roadblocks remain. In particular, the institutions hosting this data lack interoperability among their data platforms, and their datasets suffer from non-standardization. This is a foundational issue that we at the Linux Foundation sought to investigate from the lens of open source in our latest research, An Open Architecture for Health Data Interoperability

Open source is about the public good – the idea that we contribute software, hardware, standards, and data to benefit beyond ourselves. Healthcare in and of itself is similarly a public good. So why do we not see more open source adoption in the sector? This study investigated this question. Below are a few key findings from our research. 

Innovation is slow in healthcare. Regulatory risk aversion, systems complexity, and the diversity of stakeholders make it a challenging sector to adopt and iterate on new technologies. In this climate, incumbent electronic health record (EHR) vendors have been able to build a market hegemony that maintains their dominance at the cost of competition and interoperability. 

The interoperability discussion is not new in healthcare. However, data accessibility is becoming increasingly important as our digital data stores exponentially grow, and as the generative AI market sees increasing value in using this data for training LLMs. Without better data access, models risk being ineffective, inaccurate, and potentially harmful. 

How can we address data accessibility? Greater interoperability between and within institutions is in order. As we’ve seen in other sectors, such as in energy, open source encourages collaboration, transparency, and bootstrapping innovation, leading to a more productive environment for standardization and interoperability. Using open source tooling in health infrastructure also reduces costs and vendor lock-in, and provides the building blocks to de-risk innovation. 

One direction forward is to build out a precompetitive architecture that acts as the standard protocol for data sharing. Similar to other industries, health platforms and services would rely on an open “instruction manual” that defines and standardizes the different components, nodes, and technologies within the system and how they all fit together. This would allow solutions to still compete at the application level, without having to reinvent a standard or data sharing protocol. 

As this report shows, the data accessibility problem in health requires a shift in thinking about health data. We face a healthcare system that has all the data to perform life-saving care, research, and innovation, but lacks the structures in place to operationalize this data. Approaching health record-keeping from a collaborative, open science approach, where data is collected using one open standard and is shared using one protocol, will allow research and innovation to flourish. Read the full report to learn more about these recommendations, and if you know of open source projects working to improve healthcare, we'd love to hear about them!