Government and private organizations use Palantir Foundry to unlock the potential of health data, turning data into insights to improve patient lives and the efficiency of healthcare delivery.
Organizations are facing rapid growth in both the scale and complexity of clinical and biomedical data. Simultaneously, there are paramount public and regulatory concerns around the security and protection of personal information—making the use of health data challenging. Palantir Foundry helps organizations overcome this challenge by providing an infrastructure to integrate health data that accelerates life sciences research and enables data-driven decision making in clinical and operational settings while preserving privacy, reproducibility, and transparency.
Palantir Foundry enables critical work across the health and life sciences landscape:
U.S. researchers are bringing together massive-scale real-world electronic health record (EHR) data from across dozens of research institutions in Palantir Foundry to better understand COVID-19.
Governments around the world are coordinating their COVID-19 pandemic responses using Palantir Foundry.
A top 5 pharma company is integrating clinical outcomes, genotyping, and lab data for patients across thousands of clinical trials.
Palantir Foundry capabilities that are central to health and life sciences organizations include:
Bidirectional connectivity with EHR/EMR systems (including a FHIR-specific API)
Ingestion and harmonization of HL7 (all versions), FHIR, and CDISC data
Interoperability with OMOP, FHIR, TriNetX, PCORnet and ACT/i2b2 common data models and LOINC, SNOMED, mCODE, and MedDRA data standards
Make Data Usable
Health data is becoming more complex: specialized technologies produce a diverse range of data types, sources tend to be error-prone, format standards frequently evolve, and the systems capturing it are disconnected from one another. Deriving insights from and making decisions using health data requires not only connecting disparate systems, but also harmonizing diverse data types and formats for a comprehensive understanding of the data and governing access to this data for a wide range of users. Palantir Foundry offers organizations diverse capabilities for managing the complexity of health data:
- Automated data integration in any format (e.g., HL7, FHIR, CDISC, genomic file formats such as BAM or VCF, high-content image files)
- Open architecture and APIs for bi-directional connectivity to any system type and data formats (e.g., EHRs/EMRs)
- Common ontology that maps data to real-world concepts familiar to users, links data between systems, and is flexible to interoperate with standard health data models (e.g., FHIR, OMOP) and vocabularies (e.g. LOINC, SNOMED, MedDRA) and accommodate changing needs over time
- Data health checks and data quality validation
- Resource versioning, data provenance tracking, and branching of both code and data
- Industry-leading data governance and protections, including granular access controls
Protect Data At All Times
Healthcare institutions are trusted with the stewardship of highly sensitive data—i.e. personally identifiable information (PII)/protected health information (PHI)/patient identifiable data (PID)—and concerns around access, security, and management often prevent full use of this data. Palantir Foundry addresses these concerns by providing organizations data governance capabilities and features that enable oversight and accountability while promoting data usage that is proportionate, transparent, and compliant with other applicable privacy considerations.
With Palantir Foundry, data owners and authorized administrators manage granular data controls that are provenance aware—controls that propagate throughout the entire platform. Palantir Foundry’s multi-tier role- and team-based granular access controls exist at the project, dataset, object, and data row level and govern all operations without exception to protect sensitive data and ensure the data is only accessed on a need-to-know basis. Palantir Foundry can also be configured for justification-based or purpose-based access controls, allowing administrators to implement governance policies based on intended data use, such as enforcing users to provide justification prior to accessing or interacting with sensitive data and providing the ability to audit usage. Purpose-based access controls ensure institutions can conduct critical work with sensitive data in a controlled and responsible way by providing administrators with continuous oversight of who has access to what data, and why, and can dynamically grant and revoke access as necessary.
Our focus on privacy and security extends to our hosting approach: Palantir Technologies develops and operates cloud-hosted SaaS platform environments that meet industry leading regulations. Palantir Foundry is, for example, FedRAMP Moderate authorized by the U.S. Government.
Palantir Foundry also enables organizations to comply with various regulations and standards including GDPR, FISMA, HIPAA, and other security standards and legal and regulatory requirements. All data in Palantir Foundry remains owned by the customer—never by Palantir Technologies.Learn more about our approach Watch Double Click demo
To accelerate scientific research and improve the delivery of care, diverse users from across cooperating institutions must be able to work together on integrated, timely data. Palantir Foundry helps make data accessible in a highly controlled environment and encourages collaboration by providing:
- A data model with real-world concepts. Palantir Foundry’s dynamic data modelingallows data to be mapped onto real-world objects like “patients,” “diagnoses,” and “genes” that allow the presentation and exploration of objects and relationships between them in intuitive ways.
- Analytical tools for users of all technical abilities. Palantir Foundry provides a range of options for all users, including point-and-click tools that can be scaled to billions of data points. The platform also provides a number of ways in which complex workflows can be made into reusable templates so that any user can run them in just a few clicks.
- Advanced analytics allow users to write code as well as test, deploy, and manage models—all on top of live production data (more information on Palantir Foundry’s AI/ML offering). Palantir Foundry integrates with open-source or third-party libraries, models, and technologies, offering full analytical flexibility.
Provide Transparency and Reproducibility
Reproducibility is critical but can be particularly onerous when projects use multiple data sources and involve many collaborators—common characteristics of health and life sciences efforts. By automatically recording rich metadata about all resources in the platform and providing a range of exploratory tools for data lineage and provenance, Palantir Foundry offers full transparency and traceability of data, analyses, and user activity for attribution of work and ease of re-running any analysis. Dataset versioning further ensures that a dataset, analysis, or report can be retrieved in the state that it was at a particular point in time. For AI/ML workflows, these features enable users to inspect data, understand its validity, and identify any biases—which is critical for performant modeling and accurate decision-making.Watch Double Click demo
Delivering at Scale
Palantir Foundry is a mature Software as a Service platform built for scale. It is the result of hundreds of thousands of engineering hours and side-by-side user engagement. Palantir Foundry enables thousands of users to concurrently collaborate on massive-scale data. Palantir Foundry offers a seamless user experience, backed by a rapidly scalable architecture and Palantir’s state-of-the-art continuous delivery system, Apollo.Learn more about Apollo and how it delivers disruption-free updates
Powering Clinical Data Integration for R&D
Randomized controlled trials (RCTs) capture rich, high-quality data about enrolled patients: medical histories, genomic profiles, and of course, clinical responses to therapies. In both the private and public sector, institutions are increasingly looking for ways to use this data to fuel further research and development. Pharmaceutical companies see potential for pooling data to identify patterns in subgroup response or generate hypotheses for indication expansion. Similarly, government agencies want to assess the safety and efficacy of drug classes across hundreds of trials. However, unlocking the potential of this data requires overcoming significant hurdles, including harmonization, security, and accessibility hurdles.
An institution can centralize its RCT data in Palantir Foundry and implement transparent, granular access control policies. With Palantir Foundry, an institution can trace and audit all access and analysis, ensuring complete control over all clinical data use. The Palantir Foundry platform also provides the data engineering tools necessary to integrate and harmonize data into a common, comparable data model. Meanwhile, Palantir Foundry’s transparent data provenance builds trust, encouraging the institution’s researchers to reuse cross-trial patient pools and compounding the institution’s data harmonization efforts.
A government agency used Palantir Foundry to integrate clinical trials from all oncology drug applications submitted for regulatory review, and has harmonized nearly 100 trials for cancer immunotherapies, involving over 25,000 patients. Bringing this data together has enabled research into cardiac-related adverse events and safety signals that arise when patients concurrently receive radiation and immuno-oncology drugs. Similarly, the agency can now conduct meta-analysis on the efficacy of immunotherapies when patients are jointly administered antibiotics for common infections.
Connecting the Commercial Organization at a Medical Device Manufacturer
As elective procedure volumes return to pre-COVID-19 levels, a medical device manufacturer found that the device demand from hospitals didn’t resemble historical patterns.
This raised two questions:
- How can we generate new product demand forecasts that accurately reflect the new reality?
- How do we most effectively prioritize sales and marketing efforts as facilities ramp up?
The organization used Palantir Foundry to create a foundation of live commercial and operational field data, allowing users to map procedure information to specific products and calculate market share on different dimensions for the first time. This market intelligence was used to make decisions across Sales, Marketing, Product, Commercial Operations, and Finance.
- Within 2 weeks, Sales Reps were capturing intelligence on each account (i.e. which procedures were restarting when). By the end of the 3-month pilot over 850 sales users across 25+ countries were using the platform to make decisions to accelerate sales cycles and collaborate on demand forecasting in rapidly changing environments.
- An Account Inbox provided Sales Reps with guidance from their managers, as well as the latest benchmarks and insights for their accounts at the regional and country level. Alerts automatically identified accounts due for reorder, preventing account backlogs.
- Sales Managers and Business Unit Directors surfaced new commercial opportunities in Foundry, and deployed action plans across their lines of business. 1200+ previously unseen opportunities were surfaced by Sales Directors valued at $5.5M. Over 500 of these opportunities were actioned by Sales Reps.
Unlocking Real World Evidence for Commercial Pharma
Pharmaceutical companies are increasingly looking to generate and use real-world evidence (RWE) in order to understand the efficacy, safety, and cost-effectiveness of their products. Working with real-world data introduces challenges of both scale and data type. The data is often extremely high-scale, databases can include millions of patients and billions of total data points, and RWE can have many different data types (e.g. diagnoses, prescriptions, labs). And, unlike when working with 'gold-standard' clinical trial data, there are few established best-practices around the governance of RWE work.
Sanofi, one of the world’s leading pharmaceutical companies, has used Foundry to provide core data infrastructure and an analytical platform as part of their platform for RWE research, receiving a 2020 Gartner Healthcare and Life Sciences Eye on Innovation Award for this work. The platform is used to integrate medical data sources for over 300 million patients into a common foundation totaling tens of billions of medical records and many types of structured and unstructured data (e.g. claims, electronic medical records, patient registries, wearables, etc.). Foundry provides the governance tools required to ensure that this data asset is used appropriately and in accordance with all relevant regulations and usage agreements. Its data management tools ensure that the data is continuously updated as new real-world data becomes available. And, its suite of powerful analytical tools allow Sanofi researchers to implement sophisticated epidemiological studies using statistics and machine learning.
- The platform provides a secure environment for managing access to RWE data and allows Sanofi users of varying technical skills to run sophisticated analyses against huge populations.
- The platform also allows the organization to run studies in collaboration with a range of internal and external Sanofi partners - using the platform’s access control frameworks to ensure that these partners only see data and information related to studies that they are involved with.
- Partnering on studies in a transparent way on a collaborative platform ensures that the full intellectual property of all work done as part of a study is captured- including the code lists, inclusion and exclusion criteria, definitions of outcome variables, and advanced statistical and machine learning algorithms. The platform’s data governance and data lineage features allow all studies to be conducted with full traceability and transparency into all transformations and analyses that happen between raw data and the final results.
- Factors associated with SARS-CoV-2 infection and outbreaks in long-term care facilities in England: a national cross-sectional survey, Feb. 11, 2021, The Lancet
- Heterodimeric IL-15 delays tumor growth and promotes intratumoral CTL and dendritic cell accumulation by a cytokine network involving XCL1, IFN-γ, CXCL9 and CXCL10, Apr. 30, 2020, Europe PMC
- Impact of radiotherapy on risk of adverse events in patients receiving immunotherapy: A U.S. Food and Drug Administration pooled analysis, May 25, 2020, Journal of Clinical Oncology
- The National COVID Cohort Collaborative (N3C): Rationale, design, infrastructure, and deployment, Aug. 17, 2020, Oxford Academic
- Pain medication use in patients with HR+, HER2-neg advanced breast cancer treated with endocrine therapy and a CDK 4/6 inhibitor: A U.S. FDA pooled analysis, May 25, 2020, Journal of Clinical Oncology
- The Atlantic: America’s Most Reliable Pandemic Data Are Now at Risk
- The Verge: The Ambitious Effort to Piece Together America’s Fragmented Health Data
- Washington Post: The Path Forward: Combating COVID-19 with Palantir CEO Alexander Karp
- FDA: Unleashing the Power of FDA Data to Support COVID-19 Vaccine Distribution to Food and Agriculture Workers