The written component is a series of answers to questions posed by the Shuttleworth Foundation. Responses are limited to 1500 characters.
A description of the status quo and context in which you will be working.
How do we study ourselves?
Science is a fundamental tool for understanding the world, and it benefits from many open values. Unfortunately, when it comes to studying ourselves – human subjects research – we’ve failed to translate many principles of openness. I see pervasive challenges and missed opportunities.
Data sharing is one example. An increasing swath of research data is identifiable. For instance, genetic data is a core aspect of biomedical research, but a 2013 study re-identified nearly 50 “anonymous” genomes. Our current approach treats data sharing and privacy as fundamentally at odds – and as a result, we often give up on sharing.
Moreover, studies rarely share data they generate with their own participants – even when it’s potentially sensitive or meaningful. This can feel unfair, even dystopian: the US government’s dbGaP database holds genetic data from 1.4 million research subjects, almost none of whom have had access to it.
In addition, each study is fundamentally isolated. Participants rarely meet, and they can’t connect accounts between studies. Follow-up and iterations on research can ONLY occur through the original study: it is the sole conduit for engaging the participant. When that study ends, numerous opportunities end with it.
Finally, regulatory aspects of human subjects research make it difficult territory for citizen science. Non-scientists have valuable technical skills to share, but a chasm divides researchers and citizen scientists.
A description of what you want to change about the status quo, in the world, your personal vision for this area.
I want a new approach to human subjects research that embraces open values – extending these ideals to participants and citizen scientists. I believe this can enable new types of research and innovation.
I want a new approach for data sharing – an alternative to the status quo, where the study is the sole conduit for data. We can unlock new opportunities if we allow participants to manage and share data generated by a study. Participants can choose to make some of their data a public resource, and we can make it easy for participants to share their data with new projects.
I also want individuals to be empowered to access and explore their own data. And I want study participants to be empowered as a community: to be able to meet each other and explore data together.
I want to enable a new type of “longitudinal research”, using data that can be aggregated across studies, over time, with the individuals themselves.
I want to give citizen scientists the opportunity to make significant contributions, to bridge the gap between participant and researcher. Citizen-led projects can share new ways to explore data – or add new data sources that researchers can use, e.g. by building apps or “hacking” their own medical equipment.
To achieve this, I believe a community and platform is needed that brings together technical solutions, advocacy, contributing participants, and researchers. To this end, I co-founded Open Humans in 2015.
Describe the innovations or questions you would like to explore during the fellowship year
The way we currently perform human subjects research is a product of the pre-digital era. Current methods and regulations were developed in the 80s, without the context of modern informatics and automation. “Informed consent” was a verbal process, and involved paper forms. Data was hard to store, copy, and analyze.
As a result, management of cohorts has been centralized and top-down. Participant decision-making has been limited to a binary choice: agree to the study’s “informed consent” document, or do not participate. Data sharing and follow-up research could only occur through the decision and resources of the study itself. In an era of paper forms and face-to-face transactions, it was technically infeasible to empower participants to have individual, ongoing decision-making or control.
The internet and web-based engagement make these things possible! We can imagine participants having accounts within a website or app – where they can access data, meet other participants, authorize sharing, and contribute to new research.
But even in the digital era, the solutions are nontrivial. It is difficult for any single study to create an alternative approach: their scope is limited to investigating a specific research question. Achieving a substantial change requires something larger than a study: an ongoing platform that studies can use, enabling individuals to move themselves and their data between projects.
A description of what you actually plan to do during the year.
In Open Humans, there is now a strong technical foundation enabling my vision. Individuals can manage and share their data – publicly or directly with projects they join. It’s also an “open garden”, where anyone can create a project, and all features work immediately. This technology is key, but it’s only a first step: people need to use it.
Focusing on growing the Open Humans community is my plan for the coming year, with a particular focus on “projects”. Projects can be research studies that invite and/or return data, or citizen science apps that add or explore data. Projects give members things to do, ways to contribute, and solidify our role as a place to conduct and contribute to research.
I plan to take a threefold approach to growing projects in Open Humans: outreach, assistance, and improvements. With a team of advisors to guide strategy, I want to expand our research study partnerships and target citizen scientists through direct communication, conferences, online forums, and/or academic centers. Once we have leads, I’ll assist projects and help them launch – at this early stage, each study builds a precedent giving other researchers confidence in our platform. As more projects use Open Humans, I’ll receive feedback which I can use to guide improvements. This combined strategy builds awareness, confidence, and usability. Also, as we have more projects, I want to use these as opportunities to grow our participant community – a key draw for researchers.
The participant-centered approach enabled by Open Humans is innovative but unfamiliar. Adoption by the academic community will take time and leadership – studies can take years to plan and conduct. We also need to educate: to describe why this matters, why it’s better, and to be advocates for changing what society expects from research.
I also worry that society may be at a pivotal moment. Which platform will connect people and their data to research? Centralization often occurs as attention is focused on a few solutions – examples include Facebook (closed) and Wikipedia (open). If centralization happens, we’ll be locked in: the technologies and formats used will have a strong influence on the future of human subjects research.
There are new entities similar to Open Humans that are commercial and far less “open”, e.g. 23andMe’s “23andWe” program and PatientsLikeMe’s “Data For Good” campaign. Both invite contributions of data and effort to research – and both are closed models that sell access to their data and/or cohort (largely to pharmaceutical companies). For-profit goals encourage these companies to guard access to their cohorts and data, reinforcing the closed values of the status quo.
Will society adopt an approach that is open or closed? Existing closed commercial efforts pose a large challenge, but there are reasons for optimism. Open approaches are valued by researchers and participants, especially in the context of altruism and science.
I think there are many ways to pursue openness in human subjects research. Reflecting this, Open Humans is open on many levels – from our open source code to our “open garden” for projects on the site.
Initially, we focused on data sharing by individuals – publicly, or privately with projects. We continue to support this: members decide when to share their data, creating a participant-led data resource.
We also support studies returning data to participants. This makes it possible for individuals to aggregate data, and choose to share it.
Through this, we have something new: a sort of open cohort. An indefinite number of studies can invite the same participants to join, and studies can return data to member accounts where it can be shared with other studies.
The “projects” aspect of Open Humans is also open: we require ethics board certification to be listed as a study, but the platform is open to non-study projects. Citizen scientists can create projects, apps, and tools. For example…
I am thrilled and inspired by these examples of how citizen-led efforts can contribute to “real” science, made possible by a combination of technology platform and community.