1) Tell us about the world as you see it. (A description of the status quo and context in which you will be working.) Healthcare systems around the world have major problems with access to care and outcomes. Many of the problems are driven by artificial scarcity of resources due to regulations such as licensing requirements. This results in a shortage of doctors which leads to the inability to provide care. Patients who recognize this problem are largely blocked from taking their healthcare into their own hands as they cannot legally order tests or get access to prescription medication without going through a doctor. Accordingly, the diagnostic process is generally not optimized for good outcomes, but rather for patient throughput. For non-obvious cases, this often results in a wait and see approach with minimal investigation, until the condition becomes catastrophic. This approach causes a lot of harm and suffering. A study published by BMJ found the rate of medical diagnostic errors in outpatient settings to be more than 5%, of which half were estimated to potentially cause harm. There is currently limited accountability for diagnostic errors outside of malpractice lawsuits. Another major issue is the lack of portability and patient control of medical records. It is a tedious and error-prone process to transfer information between doctors. Additionally, if patients have control of their medical records, they can efficiently share information and collaborate. Although not experts, patients have much more time to spend researching than a doctor who must make a judgement in minutes. 2) What change do you want to make in the world? (A description of what you want to change about the status quo, in the world, your personal vision for this area.) By giving patients the power to take their healthcare into their own hands, access and outcomes will be improved. Ownership and sharing of medical records will enable the curation of an extensive crowdsourced database for machine learning algorithms that can improve diagnostics and provide recommendations to those without access to an appropriate doctor. Additionally, diagnosis recommendations can be crowdsourced from other people similar to CrowdMed, except there would be no required fee. It would also be simple to share records with other platforms including CrowdMed or doctors if that was preferred. This would lead to a shorter time to diagnosis for difficult cases. Sharing misdiagnosis information would allow for accountability of poor performing doctors and enable discovery of systemic problems diagnosing specific diseases. These types of analyses have been very limited in mainstream literature and there is a lot of data out there waiting to be analyzed. Misdiagnosis rates can be reduced which will result in less harm to patients. The DIYBio and open hardware spaces have been expanding which will likely lead to open assays and instruments being used more and more to support diagnosis. P2pdx can support results from any of these that people decide to share. It can also support information regarding sources of affordable generic medications and lab testing. P2pdx will be a complete repository for patient empowerment. 3) What do you believe has prevented this change to date? (Describe the innovations or questions you would like to explore during the fellowship year) The healthcare system is slow to change and making changes requires a lot of political capital. It is more cost-effective to take a direct action approach and make the change you want to see. A number of organizations have taken this approach to try and solve pieces of the puzzle, but they all have critical flaws. PatientsLikeMe allows patients to enter and share medical data, but then sells the data to third parties and does not openly allow access to the data for research. The profit motive and business model is stifling innovation and many users do not approve of it. PicnicHealth collects medical records and organizes them so they can be shared. However, a hefty initialization fee and monthly membership are required and only medical records are supported, not dental, vision, or other third party records. Additionally, no open data format exists to allow integration across platforms. Forums are frequently used for discussion between patients but are not efficient for sharing or analyzing data. There is no open platform for entering, sharing, and analyzing health data. The closest platforms are OpenHumans and the forthcoming MIDATA, both of which allow sharing health datasets. However, neither addresses medical record data entry, management, or analysis directly. P2pdx will support all of these capabilities. Additionally, it should be possible to integrate with OpenHumans and MIDATA so that users can easily move their data between platforms to join studies or use apps. 4) What are you going to do to get there? (A description of what you actually plan to do during the year.) The primary efforts for the year will be continuing to develop the software with community feedback, working to grow the user base, and recruiting donating members. A roadmap is laid out on the p2pdx GitHub page that provides an estimate for when various software features are expected to be completed for the rest of 2017. Later in the year, the roadmap will be updated for 2018. The roadmap can change depending on community feedback which can be received via Patreon, GitHub, and other platforms. Summer 2017 is when the software is planned to be minimally usable. At this point, it will be possible to work on growing the number of people contributing data. The plan is to make the case to both healthy people who are proactive with their health and people with chronic illnesses. P2pdx offers a strong value proposition which will continue to be shared with various health and patient communities online to demonstrate the benefit. As the user base increases, more effort will be spent on getting member subscriptions as a funding source for sustaining development, maintenance, and outreach. A membership model using Patreon for recurring monthly donations is currently in use which offers different funding levels with various perks. The top level perk gives supporters the right to vote on presented developmental and operational matters, consistent with a co-op organizational model. 5) What challenges or uncertainties do you expect to face? The primary challenge expected is getting enough users to join and contribute data. I have done some surveys and had discussions in online forums where people have expressed interest in the project. Additionally, I am part of the target demographic so I have a good sense of what is needed. Initially, outreach will focus on highlighting features including an organized repository of all medical records, easy to take full history to new doctors, and data visualization. There is a little bit of a chicken and egg problem as the platform appeal will grow with the amount of data contributed. This is because machine learning algorithms to support features such as anomaly detection, preventative health alerts, and patient similarity require sufficient data in order to be implemented. There is also some uncertainty associated with ZeroNet, the underlying decentralized website platform. It is a relatively new and rapidly evolving technology. Currently it offers all of the features necessary to implement the planned p2pdx functionality. The developer is funded to work full-time on the project so development and maintenance should remain stable. The developer is also open to suggestions and contributions so if any platform changes in the future are needed to support any p2pdx function, they should be facilitated. In the unlikely case that they couldn't be, the project is open source and a fork could be made if absolutely necessary to allow p2pdx to move forward. 6) What part does openness play in your idea? Every aspect of p2pdx follows an open philosophy. The platform itself is organized around the idea of platform cooperativism. The project will be run as a democratically controlled cooperative organization where members will be able to vote on operational and developmental matters. All of the software developed for p2pdx is open source. All of the data is stored in openly defined formats. This is critical as it encourages independent research and development as well as integration with other platforms, for example, OpenHumans and MIDATA. It will also enable the development of a broad ecosystem of apps and studies by professionals and hobbyists alike from all over the world. The data is also decentralized in a peer to peer network. Each peer has a copy of relevant data which ensures that access to the data can never be arbitrarily cut-off or controlled by any one entity. The peer to peer topology also offers higher data availability and resilience. In addition, it allows hosting costs to be shared as bandwidth and data storage is spread across peers, without any central server. 10) Who are your current or potential key partners? I currently don't have any partners supporting my work. I am not actively seeking partners besides the Shuttleworth Foundation at this time. The long-term goal is to have the work directly financed by members who benefit from it. The start-up period has been facilitated so far by donating my free time. Funding would accelerate the work, thus shortening the period before a strong value proposition could be made to users to financially support the work. At that point, it is expected that the community would sustainably fund the work for as long as it continues to provide value.