People tend to say that data is ‘the new oil’ and yet, due to privacy and regulatory hurdles, that oil is deeply buried underground in many life science companies instead of flowing vastly. While the benefits of sharing large research and clinical data sets have shown tremendous benefit in the development of drugs against SARS-CoV-2, shortening the estimated drug development timeline from 4-5 years to about 1-2 years in total, there is currently no standardized infrastructure in place that allows sensitive research and clinical data to be shared effectively and securely at scale. This highlights the need for privacy-enhancing technologies to enable secure and scalable sharing of sensitive research and clinical data.
Sharing insights, not data:
The common denominator and major hurdle to achieving data sharing is the reticence of large companies to share data sets with third parties. The main reason is that these data sets are proprietary and/or contain sensitive clinical or patient information.
Now, let’s run a thought experiment to ask ourselves: what if we didn’t have to share clinical data sets directly with third parties, but could instead give it to a middleman that encrypts the data in a way where it can’t be seen or accessed by the third party?
And what if we could then analyze the data within that encrypted environment using machine learning or AI, and obtain a report on the outcome of the data analysis to provide meaningful insights, without the data itself ever changing hands? Would that make large corporations comfortable?
Enter Privacy-Enhancing Technologies (PETs); a rapidly evolving group of technologies that have the potential to revolutionize our global data-sharing infrastructure, and an essential force to drastically change the efficiency of the current drug development model used by pharma and biotech. PETs represent several interesting business and investment opportunities that pharma and biotech companies, big and small, can use to their advantage.
Clinical data is becoming increasingly rich and complex… And increasingly hidden:
Pharma and biotech companies are sitting on large data sets of proprietary information from in vitro, preclinical and clinical studies. While companies can use this data for their own development pipelines, the majority of this data gets swept into a central server and will never see the light of day. Not only does the number of data sets increase, but the resolution and amount of parameters in the data are also increasing rapidly.
The Tufts Center for the Study of Drug Development published a report in 2021 outlining several stats on the increasing complexity of clinical trial data. Their team found that the average Phase III clinical trial generates an average of 3.6 million data points, an increase of 3-fold from the decade before.
Data from a Tufts report shows that clinical trial data is becoming richer and more complex every year and, when shared, would yield an enormous potential to improve global healthcare.
“Pharma companies have already started to explore data marketplaces, and this is driving the uptick in startup companies that offer PET. I expect that PET will be used 5-10 years from now, driving new forms of revenue and new ways for big pharma and startups to collaborate” ~ Innovation Lead at PET Technology Company
The pandemic caused a (temporary) paradigm shift in global data sharing:
The COVID-19 pandemic has taught us that, if we choose to do so, it is possible to share data at scale with those barriers in place. Within a matter of months, after the pandemic started, several global consortia were formed that shared real-world clinical data to advance drug development and vaccine development efforts.
A notable example includes the International Consortium for Clinical Characterization of COVID-19 by EHR (4CE). This consortium gathered data from over 27,584 Covid-19 patients between Jan – April 2020, including 187,802 laboratory tests from 96 hospitals across five countries within a matter of 3 weeks. Another example is the International COVID-19 Data Alliance, which was built as an open international research partnership making data accessible to health researchers and scientists worldwide.
Because time was of the essence during the pandemic, the willingness and urgency to collaborate were driving the inception of these collaborations, supported by the use of privacy-enhancing technologies. Now that this pandemic is fading out like a candle, this urgency is no longer there, taking with it funding from governments.
Based on this, it seems tempting to conclude that ample funding contributed to an increased willingness of pharma/biotech to participate. Therefore, it seems essential that any new effort for safe data sharing needs a clear funding or business model for those involved, to minimize the risk of investment and to increase return on investment as much as possible.
Privacy-enhancing technologies enable encrypted data analysis for business and product development insights across healthcare:
Sparked by the pandemic, there have been several initiatives that look at ways to share data sets between stakeholders in a secure and compliant way, while opening up potential licensing or business models for those who participate. Broadly, there are 3 types of PET buckets to consider:
To give color to these technology buckets, let’s look at two examples of PET companies: Roseman Labs and Linksight.
Roseman Labs:
Roseman Labs offers a product that enables stakeholders to collaborate without exposing sensitive data to each other. Roseman Labs offers a relational database, which allows for applying common operations such as joining, filtering, and data analytics, with a range of machine learning models.
The virtual data lake is set up so that the entering data is never concentrated in one place. It is instead encrypted and distributed over servers that run the Roseman Lab’s product. The encrypted data is accessed through a secure multi-party computation (MPC) protocol, where involved parties can authorize beforehand certain analyses that may be run on the combined data. This means that data owners control how their data is processed. Roseman Labs is currently partnered with a top 10 global pharma company to set up a system for their biologics research program.
Linksight:
Another example of private computation is Linksight, which offers privacy-by-design data collaboration. Instead of a cloud-based approach, organizations can install a local Linksight “data station” and connect to other organizations’ stations. MPC protocols are run between the data stations, allowing for computations on data while the data itself stays fully encrypted and is thus unreadable.
These two companies are examples of a booming, nascent field in the process of overcoming technical hurdles while trying to resolve (perhaps a healthy portion of) the skepticism from industrial stakeholders through the implementation of privacy-enhancing technologies.
While in their early stages of development, both Roseman Labs and Linksight prove that there are workarounds allowing computation or analysis of data sets in a way that ensures privacy and data ownership.
Roseman Labs and Linksight are not the only players in the field and PET companies have boomed over the last 5 years (>50 startups globally, since 2020), many of them working as part of a public-private partnership (PPPs) or in close collaboration with industry and government bodies.
PET can revolutionize data-sharing for drug discovery and clinical trial design:
As technical hurdles and trust issues are slowly overcome, it’s obvious that privacy-enhancing technologies have a huge potential to improve global healthcare and address public health challenges. To make effective use of these technologies, big pharma, biotech, academia, and governments will have to smith collaborations to rethink how we can use and reuse research and clinical data sets to make data-driven decisions, in a regulatory-compliant manner, while benefiting the greater good.
By sharing insights, and not the data itself, privacy and data ownership by industry and/or academic stakeholders can be protected. A major challenge lies in formulating clear incentive strategies to get academic groups, governments, and (especially) industry stakeholders on board. It seems logical to assume that public-private partnerships (PPPs) can play a crucial role in driving this process forward, by de-risking investment and incentivizing stakeholder participation.
