¶ Mandate
Maximize average healthy human lifespan and minimize net suffering by quantifying the effects of every food, additive, supplement, and medical intervention.
¶ Overview
The Wikipedia model demonstrates the power of crowdsourcing and open collaboration. Despite Microsoft spending billions of dollars hiring thousands of PhDs to create Encarta, Wikipedia generated over 50 times more content in just a few years with the same accuracy levels at a fraction of the cost.
Our goal is a 50X acceleration in clinical discovery by replicating this model for clinical research. We can use real-world evidence from patients, clinicians, and researchers to enable orders of magnitude more insights and discovery.
¶ Problems We're Trying To Solve
- Slow and Expensive: It takes over a decade and $2.6 billion on average to bring a drug to market, with Phase III clinical trials costing around $41k per subject.
- No Data on Unpatentable Molecules: There is a significant lack of data on the long-term effects of the majority of synthetic or natural compounds, as there is insufficient incentive to research non-patentable molecules.
- Neglected Off-Patent Treatments: There is a lack of financial incentive to approve drugs for additional conditions after their patent expires, leaving potential treatments for rare diseases unexplored.
- No Long-Term Outcome Data: Due to financial constraints, long-term effects of drugs are often unknown.
- Publication Bias: Negative results are frequently unreported, leading to redundant research efforts and wasted resources.
- Exclusion of Diverse Patient Populations: Clinical trials often exclude a vast majority of the patient population, limiting the generalizability of findings.
- Vast Amount of Unresearched Combinations: The immense number of possible molecule-disease combinations leaves 99.9999998% of potential knowledge undiscovered.
- Bias Towards Rejecting Effective Treatments: Cognitive biases and personal risk aversion among regulators result in a tendency to reject potentially life-saving treatments.
To overcome these perverse incentives and biases, implementing a Decentralized Autonomous Organization (DAO) regulatory body is proposed. This would distribute responsibility among a large group of experts, mitigating individual risk and bias, and fostering a more balanced, efficient, and innovative drug approval process. 👉 Learn More
¶ Ingredients of the dFDA
¶ Digital Twin Safes / PersonalFDA Nodes
A tool for self-sovereign storage of personal data that enables effortless data sharing with clinical safety and efficacy studies.
Features
- Data Import: Create seamless mechanisms for importing existing health data while ensuring privacy and security. Import data from all your apps and wearables, so you can centrally own, control, and share all your digital exhaust.
- Data Encryption: Implement robust encryption protocols to safeguard sensitive health data.
- Sync to Trusted Instances: Establish secure channels for data synchronization, ensuring integrity and reliability.
- Federated Learning with Homomorphic Encryption: Innovate in secure data analysis, allowing for meaningful insights without compromising data privacy.
- Data Gems NFTs - Data sets can be encrypted and stored in a decentralized manner generating a Data Gem NFT that can be sold on data exchanges granting the possessor access to the data set.
- Digital Twin Skeleton Key NFT - This key gives you the ability to mint Data Gem Data Access NFTs using your imported data.
- The Human File System Protocol SDK - A Simple API for Patient-Controlled Health Data Aggregation, Sharing, and Monetization. Also standard protocol for personal data exchange between studies, apps, and devices.
Potential Implementations, Components or Inspiration
¶ dFDA Wiki
- Knowledge Base: Inspiration could be taken from the Psychonaut Wiki. It's a modified version of MediaWiki with additional quantitative metadata storage regarding the pharmacokinetics of various substances. This could be expanded to document the quantitative effects of every factor on specific health outcomes.
- Data Silos Directory: Compile a comprehensive directory of existing data sources, facilitating integration with the Digital Twin Safe.
- Reputation Scoring: Develop a transparent and reliable reputation-weighted voting system for intervention approval.
- Comparative Policy Analysis - Aggregate existing approval and certification data from existing national regulatory bodies
- Food and Drug Outcome Labels - Ultimately, the most useful output of a decentralized FDA would be Outcomes Label which lists the degree to which the product is likely to improve or worsen specific health outcomes or symptoms. These are derived from real-world data (RWD) and subject to Futarchical-weighted review by the board members of the dFDA.
- Publish Meta-Analyses - Generate meta-analyses from all completed studies at ClinicalTrials.gov
- Certification of Intervention Manufacturers/Sources via a Decentralized Web of Trust derived from end-user data and reviews traced back through an NFT-tracked supply chain
- Intervention Ranking - Elevate the most promising yet little/known or researched treatments
- Decentralized Clinical Trials - This would increase knowledge and access and availability of new and innovative treatments to those who need them urgently.
Related Projects
¶ Benefits
¶ For Individuals
- eBay for health data - You can earn magic internet money by selling your data regarding symptoms, treatments, and
factors to pharmaceutical companies, insurance companies, and other data buyers - Control access and use of your data through fine-grained permissions
- Continuously monitor and audit the data you provide to other organizations
¶ For Health Apps
- Connected real-world data yields better insights for your users
- Apps that embed the exchange in their app earn a 0.5% transaction fee for each data sale
- Connect to third-party sources to enrich your data, or easily connect to a user's existing data
¶ For Pharmaceutical Companies
- Conduct long-term safety and effectiveness studies by linking their clinical trial data to medical claims and
electronic health record data - Refine models for finding rare disease patients by linking diagnostic lab, genomic, and imaging data
- Discover new therapeutic candidates with connected data
¶ For Insurers
- Improve value-based care analytics and sharpen total cost of care estimates by linking to EHR and clinical data
- Connect to the nation's largest ecosystem of health data
- Hone risk adjustment factor calculations by linking claims to social determinant's data, to properly estimate the true
cost of patient care
¶ The dFDA Tech Stack
¶ 1. Personal FDA Nodes
A safe for your digital twin that provides
- Storage
- Security
- Access Control
- De-identified Data Sharing
- API with Advanced Querying Capabilities
¶ 2. The 3rd Party Plugin Framework
- Data import from any source
- Data Format Transformation
- Data visualizations
- Machine learning algorithms
- Data analysis
- Personalized Health Dashboards
Our novel incentive structure overcomes the traditional collaboration and data sharing barriers by encoding contributions through non-fungible tokens (NFTs).
Using smart contracts, the platform will compensate all contributors with royalties.
¶ Problem: You and Everyone You Love Will Suffer and Die
Over 2 billion people are suffering and 150,000 people die every single day from preventable diseases.
For perspective, this is equivalent to:
¶ 💡 Solution: A Collaborative Framework for Personalized, Precision Health
The solution is to use the oceans of real-world evidence to accelerate the discovery of new cures and reveal hidden causes of disease.
The human body can be viewed as a black box with inputs (like diet, treatments, etc.) and outputs (like symptom severity). We're creating a mathematical model of human biology to determine the input factors and values that produce optimal health outcomes.
¶ The Potential of the Oceans of Real-World Evidence
- Discovering Hidden Causes of Illness - Data mining and analysis to identify hidden factors in our daily life that are making us sicker
- Preventative medicine - Predictive analytics and data analysis of genetic, lifestyle, and social circumstances to prevent disease
- Precision medicine - Leveraging aggregate data to determine the precise treatments and dosages for your unique biology
- Accelerated Treatment Discovery - Data-driven medical and pharmacological research to discover new treatments and medicines
- Reduction of adverse medication events - Harnessing big data to spot medication errors and flag potential adverse reactions
- Cost reduction - Driving better patient outcomes for long-term savings through prevention and avoidance of expensive and ineffective treatments
- Population health - Identify health strategies based on demographic, geographic, and socioeconomic trends
¶ Why Clinical Research Has Failed
It takes over 10 years and 2.6 billion dollars to bring a drug to market (including failed attempts).
It costs $41k per subject in Phase III clinical trials.
The high costs lead to:
1. No Data on Unpatentable Molecules
We still know next to nothing about the long-term effects of 99.9% of the 4 pounds of over 7,000 different synthetic or natural compounds. This is because there's only sufficient incentive to research patentable molecules.
2. Lack of Incentive to Discover Every Application of Off-Patent Treatments
Most of the known diseases (approximately 95%) are classified as rare diseases. Currently, a pharmaceutical company must predict particular conditions to treat before running a clinical trial. Suppose a drug is effective for other diseases after the patent expires. In that case, there isn't a financial incentive to get it approved for the different conditions.
3. No Long-Term Outcome Data
It's not financially feasible to collect a participant's data for years or decades. Thus, we don't know if the long-term effects of a drug are worse than the initial benefits.
4. Negative Results Aren't Published
Pharmaceutical companies tend to only report "positive" results. That leads to other companies wasting money repeating research on the same dead ends.
5. Trials Exclude a Vast Majority of The Population
One investigation found that only 14.5% of patients with major depressive disorder fulfilled eligibility requirements for enrollment in an antidepressant trial. Furthermore, most patient sample sizes are very small and sometimes include only 20 people.
6. We Only Know 0.000000002% of What is Left to be Researched
The more research studies we read, the more we realize we don't know. Nearly every study ends with the phrase "more research is needed".
If you multiply the 166 billion molecules with drug-like properties by the 10,000 known diseases, that's 1,162,000,000,000,000 combinations. So far, we've studied 21,000 compounds. That means we only know 0.000000002% of the effects left to be discovered.
¶ Why Digital Health Innovation Has Failed
¶ Duplication of Effort
Despite this massive growth in health data and innovation, we've seen increased costs and disease burden and decreased life expectancy.
The reason is awful incentives. There are more than 350,000 health apps.
Each costs an average of $425,000 to develop.
Most have significant overlap in functionality, representing a cost of $157,500,000,000 on duplication of effort.
Isolated streams of health data can only tell us about the past. For example, dashboards filled with descriptive statistics such as our daily steps or sleep.
If this data and innovation efforts were combined, this could increase the rate of progress by 350,000 times.
The obstacle has been the free-rider problem. Software developers that open source their code give their closed-source competitors an unfair advantage, increasing their likelihood of bankruptcy.
¶ 🏭 Platform
A global open-source platform and plugin framework will enable the transformation of data into clinical discoveries.
¶ Functional Scope
The functional scope of the platform includes:
- Aggregation
- Managing
- Processing
- Storage
of health data from different sources.
¶ Primary goal
Create a basic foundational technology layer suitable for any digital health application that provides better interoperability, portability, availability, analysis, and data security.
¶ Use Cases
- EHR Systems for healthcare providers
- User-centered dashboards for personal health management
- Data sharing with doctors, health coaches, or family members
- Decentralized clinical trial platforms
- Patient recruitment services for clinical trials
- Citizen science platforms
- Health data marketplaces
- Open health databases for research
- Algorithm and scores development (e.g., in-silico trials)
- Niche health applications with specific requirements or custom integrations
The platform consists of two primary components:
- Core Open-Source Platform - The core platform is open-source and includes only universally necessary features. This primarily consists of user authentication, data owner access controls, data storage, data validation, and an API for storage and retrieval.
- Plugin Framework - Plugins will provide additional functionality like data import from specific sources, data mapping to various formats, data analysis, data visualization, notifications. These may be free or monetized by their creator.
¶ Core Components
Data Ingestion and Access API
The Unified Health application programming interface (API) includes an OpenAPI specification for receiving and sharing data with the core database. Software development kits (SDKs) will enable developers to implement easy automatic data access and sharing options in their applications.
Data Mapping and Validation
Data from files or API requests can be mapped from many different proprietary formats into a standard schema.
Data Ownership
Data should be owned by the individual who generated it. It should remain under their control throughout the entire data life-cycle from generation to deletion.
Data Compensation
Value stream management allows the exchange of data for tokens.
¶ Plugin Framework
3rd party plugins can interact with the core and provide additional functionality. They may be free or monetized by their creator. These include:
- Data Import Plugins
- Data Visualization Plugins
- Machine Learning Plugins
- Electronic Health Record System Plugins
- Clinical Trial Management Plugins
Data Analysis Plugins
Data Analysis Plugins will apply statistical and machine learning methods to the ocean of high-frequency longitudinal individual and population-level data. The resulting value will include:
- Personalized Effectiveness Quantification - Determination of the precise effectiveness of treatments for specific individuals
- Root Cause Analyses - Revelation of hidden factors and root causes of diseases
- Precision Medicine - Determination of the personalized optimal values or dosages based on biomarkers, phenotype, and demographics
- Combinatorial Medicine - Discover relationships between variables or combinations of interventions
- Optimal Daily Values - Determination of the personalized optimal dosages of nutrients or medications
- Cost-Benefit Analysis of interventions by weighing clinical benefit against costs in terms of side effects and financial impact
Example Data Presentation Plugins
- Outcome Labels
- Predictor Search Engines
- Root Cause Analysis Reports
- Observational Studies
- Real-Time Decision Support Notifications
¶ ❤ Benefits
¶ Incentive Alignment
This illustrates the flow of value between different stakeholders. Unlike traditional zero-sum games, we can provide everyone with more value from participation than they have to put into it.
¶ 1. For Patients
Incentives for Patients to share their de-identified data will include:
- Actionable ways to prevent and mitigate chronic illnesses.
- The ability to license and earn a share of income for the use of their data for research and development by pharmaceutical companies and other businesses.
¶ 2. For Businesses Housing Data Silos
Businesses housing data silos include health insurers, pharmacies, grocery delivery services, digital health apps, hospitals, etc. These will be incentivized to allow individuals to easily share their data via a well-documented OAuth2 API by:
- A share of income for using their data for research and development.
- An on-site instance of the OAuth2 server to retrieve required data from their on-premise databases.
- Reduction in their employee healthcare costs (one of their most significant expenses)
¶ 4. For Digital Health Businesses
- Reduced costs of software development.
- Massive free marketing exposure through company branded plugins in the Plugin Marketplace.
- Revenue derived from their plugins in the Plugin Marketplace.
¶ 5. For Disease Advocacy Non-Profit Organizations
Disease advocacy nonprofits will benefit from promoting studies to their members by:
- Furtherance of their mission to reduce the incidence of chronic illnesses.
- Member engagement more productive than the traditional charity walk.
¶ 6. For Government Health Agencies
- A reduction in healthcare costs due to discovering new ways to prevent and mitigate chronic illnesses.
- Furtherance of their stated reason for existence to protect and promote the general welfare.
- Their duty to protect the rights of individuals' data. To fulfill this, they must require businesses in possession of it to give them the ability to access and share their data via a well-documented OAuth2 API
- Cost-savings from international cost-sharing by using global open-source software.
- Epidemiological discoveries on the effectiveness of different public health regulations between nations.
¶ 7. For Contributors to the Development of the Platform
- Gitcoin bounties for specific tasks
- Entitling the developer to ongoing royalties in proportion to their contributions.
- Dework bounties for tasks
- Other benefits - depending on the involvement
¶ Conclusion
The dFDA could provide innovative tools, resources, and frameworks, the dFDA empowers Contract Research Organizations, Certifying Agencies, and Regulatory Agencies to perform their roles more effectively, accelerating safety and innovation.
¶ 📚 Resources
- dFDA Governance Protocol - By implementing a continuous improvement process and a Futarchical Voting mechanism, the dFDA aims to enhance the quality of drug approvals and health policies.
- Regulatory Enhancements - The ideal framework for drug approvals aims to harness the collective intelligence of a diverse and knowledgeable crowd, ensuring rigorous, transparent, and inclusive decision-making processes and algorithmic optimization of outcomes using quantitative cost-benefit analysis of every intervention for each condition
- Support of Real World Evidence - Large-scale efficacy trials based on real-world evidence, utilized before the pharmaceutical industry-driven randomized controlled trials mandated post-1962, were more effective in improving health outcomes, as evidenced by a consistent increase in human life expectancy during their implementation.
- dFDA Legal Structure - In considering the optimal legal structure for the dFDA, it is crucial to evaluate the advantages, disadvantages, and suitability of various options to ensure that the organization can fulfill its mission effectively while remaining compliant with legal and regulatory requirements.
- DeSci Exchange - A free market for personal data in the form of an embeddable SDK with a configurable transaction fee for the apps embedding it.