Executive Summary
Imagine your organization has genuine control over its clinical data assets. Information flows freely between compliant applications via standard APIs that you define and manage. Switching vendors or adding new specialized applications becomes straightforward because your data layer is stable and independent of any single supplier. Your internal informatics expertise thrives, and your strategic decisions are no longer subordinate to vendor roadmaps or locked into proprietary formats. When you evaluate new technologies—including AI and advanced analytics—you do so from a position of strength, with your data accessible and computable on your terms.
Now imagine that your clinicians and researchers can produce evidence findings from their desks in minutes or hours rather than weeks or months—without writing a single line of code. Imagine collaborating with institutions around the world on multi-site studies without ever sharing patient data. Imagine transforming the fragmented data trapped in your legacy systems into standardized, analysis-ready formats with AI-powered tools that are 100 times faster and eight times more precise than traditional methods.
This is not a distant aspiration. It is an achievable architecture built on a proven principle: separating your data from your applications, supported by three open standards, an emerging ecosystem of open terminology services, and purpose-built platforms like Evidentli's Piano that accelerate every step of the journey.
The foundation of your ecosystem rests on leveraging three open standards working in concert in recognition that clinical care, data exchange, and longitudinal analysis represent fundamentally distinct processes, each demanding tailored solutions [1]. Rather than forcing a single standard to serve all purposes, your architecture embraces three complementary open standards, each optimized for its domain.
openEHR provides patient-centric clinical persistence using reusable archetypes and templates—a two-level modeling approach that keeps your longitudinal records vendor-independent and semantically consistent over time [2]. HL7 FHIR handles real-time data exchange through RESTful APIs with rigorous terminology binding and an extension framework that accommodates new use cases without breaking existing interfaces [3]. OMOP normalizes your data for longitudinal analysis, mapping all external terminologies to a unified vocabulary that enables population-level analytics and participation in research networks like OHDSI [4].
Underpinning all three standards is an emerging ecosystem of open terminology services. Publishers of LOINC, SNOMED CT, and RxNorm already provide open access to FHIR-based terminology APIs, enabling your organization to access authoritative vocabularies through standardized operations rather than maintaining fragmented, duplicated terminology infrastructure [5]. Evidentli's Piano platform integrates with these terminology services, providing seamless access to standardized vocabularies throughout the data transformation and analytics workflow.
The vision of open, standards-based data infrastructure confronts a practical reality: preparing data for analysis is traditionally a long and arduous process requiring both data management skills and deep understanding of clinical practice—a rare combination. A 2020 survey by the National Cancer Research Institute and Cancer Research UK found that fewer than half of researchers were successful in accessing the health data they needed, with time and logistics for access cited as the highest barrier, followed closely by data sharing agreement issues [6]. The costs and risks of data extraction, combined with scarcity of specialized skills, create tremendous barriers. Industry surveys confirm the scale of this burden: over 30% of all staff work hours in clinical trials are consumed by data management activities [7], clinical data managers spend an average of 12 hours per week per study on manual data reconciliation and cleaning alone [8], and more than half of biomedical researchers report spending over five person-hours weekly on data handling tasks [9].
Evidentli's Piano platform dramatically accelerates and improves the efficiency of this transformation. Piano is the result of decades of research and development in artificial intelligence and health informatics—an end-to-end solution that streamlines the process of turning healthcare data into actionable evidence. Piano's proprietary AI and natural language processing algorithms, specifically designed for clinical data, transform information from electronic medical records, patient administration systems, laboratory and radiology information systems, clinical registries, and other sources into the OMOP Common Data Model with over 99% accuracy. Critically, Piano maintains a human in the loop throughout the transformation process, and all mappings are fully documented, transparent, and auditable.
Piano's data ingestion workflow includes database and stream connectors that can ingest data from FHIR, relational databases, files, or directly from source systems. An intelligent staging database aggregates and joins data from multiple sources and automatically discovers relationships between tables and columns. Piano supports FHIR ingestion either as an interoperability layer or as bulk upload, with data deliverable through an interoperable HAPI server, JSON files, or SQL databases.
The transformation workflow includes configurable de-identification with options for partial or complete re-identification, AI-driven data transformation tools that clean, aggregate, and standardize staged data, extensive data quality controls for pre- and post-transformation validation, population characteristics reports and alerts, and smart data versioning for precise research reproduction. What traditionally took months of specialized work can now be accomplished in days, enabling your organization to realize value from its data assets far more rapidly.
Managing access to sensitive patient information is a daunting responsibility for any data custodian, especially when the only means to provide access has been copying data into extracts that then proliferate beyond institutional control. Most data custodians prefer not to deal with extracts—they lack resources to manage them effectively and often lack the specialized skills to interrogate them safely.
Piano's governance framework is modeled on ethics permissions as practiced internationally by Institutional Review Boards and Human Research Ethics Committees—the "Helsinki" protocol [10]. With a simple but granular interface, non-technical custodians can grant, revoke, and monitor access by project and by user. Projects are granted access as required to serve their purpose, and members of projects gain access to data according to limits set by the custodian. Users never gain direct, uncontrolled access to underlying data.
Piano's data warehousing is purpose-built for healthcare interoperability. Transformed data is organized, de-identified, and appropriately governed and secured to run applications at scale. The system enforces and audits data access permissions and can be controlled directly and immediately by your administrators.
Evidentli designs every feature with data safety in mind. Piano's security measures include anonymization of identifying and potentially identifying fields, obfuscation of dates while preserving clinically relevant time intervals, query suppression to prevent re-identification through small-cell queries (configurable thresholds ensure queries returning only a handful of patients return zero results), project-based restrictions limiting access to minimum necessary data sets, and detailed audit logs tracking every change by user, project, and site.
Piano's unique security architecture employs a proprietary three-tier system that enforces access at the network gateway, by the application, and by the databases. Data segregation ensures that data and metadata with different levels of sensitivity are stored in different databases, actively preventing data movement between databases whether intentionally or by accident.
Your clinicians are willing and able to produce clinical evidence, but specialized tools are needed to make this feasible. Piano's analytic toolbox provides specialized tools that can interrogate data without a single line of computer code or SQL query. Using concepts every clinician understands, and in accordance with best practices for peer-review and reproducibility, Piano users can produce evidence findings in minutes or hours rather than weeks or months.
Piano's scientific toolbox includes the most commonly used analytical tools in clinical research: defining and refining patient cohorts and sub-groups through inclusion and exclusion criteria, statistical comparison and summary, patient recruitment into trials, and more. Natural language generation algorithms automatically produce reports that accurately describe research results with methodology explained so that research can be exactly reproduced. Intuitive drag-and-drop workflows assemble analytics or machine learning components so that complete projects can be created without specialized programming skills.
Once data has been converted to the OMOP Common Data Model through Piano, evidence can be generated using Piano's scientific toolbox, or through Piano's RESTful API, which provides flexibility by allowing any third-party tool to integrate for visualization, dashboards, and advanced analytics. Native integration with advanced research tools means that data scientists can discover or test predictors, create truly portable AI for decision support, and pursue sophisticated analyses—all while working from the same governed, standardized data infrastructure that serves your clinical researchers.
Whether pursuing quality improvement, patient-centered care, operational management, clinical trials, population health, or precision medicine, co-design and reproducibility are key to reliable evidence. Yet collaboration becomes complicated when data must be shared, especially across jurisdictions with different privacy regulations and governance requirements.
Evidentli's Evidence Hub is a global research collaboration and translation platform that solves this problem by enabling real-time research replication without sharing patient data. Each participating institution maintains its own Piano instance, governed locally according to its own policies. Evidence Hub allows Piano users to collaborate without sharing sensitive data—keeping data local to each institution while combining aggregate data for federated analytics.
Piano's Federation capability uses Piano's REST API to securely connect multiple Piano instances, ensures analysis protocols are identical across all sites, and pulls only aggregate results from each location. Aggregate data can be used for site-to-site comparison and understanding overall health system performance while protocol compatibility is assured and data governance is preserved.
Piano connects to Evidence Hub to share methodology transparently and precisely using open standards. With just a few clicks, anyone with shared values and research interests can collaborate without accessing each other's patient-level data. The primary forms of collaboration in medical research—peer-review of methods, replication, and reproduction—are all enabled through Evidence Hub without requiring data sharing.
Evidence Hub functions as a central repository of research and ingestion methods—essentially GitHub for medical research. Piano automatically generates exact descriptions of research protocols that use OMOP, and Evidence Hub shares these descriptions with other Piano users who can peer-review methods without needing to see original data schemas or extracts. Piano users can download research projects from Evidence Hub and run them on local data sets (subject to data access permissions), facilitating instant reproduction and avoiding the need for replication altogether.
Critically, Piano's architecture ensures that sensitive data never leaves the customer's data center. Piano segregates data into different sections in different databases, with only non-sensitive configuration information—representing the methods used in research projects—transmitted to Evidence Hub. Patient data, aggregate results, and other sensitive information remain entirely within institutional control.
Your Clinical Data Repository stores longitudinal patient records using openEHR's two-level modeling approach. Clinical concepts remain computationally tractable and semantically consistent across time. Your archetypes and templates are reusable, extensible, and independent of any vendor's proprietary data model.
Your Integration and Exchange Layer uses FHIR's RESTful APIs with rigorous terminology binding. The specification requires explicit versioning of terminology systems, ensuring reproducibility and consistency across implementations. When new use cases emerge, FHIR's extension framework accommodates them without breaking existing interfaces. Piano supports FHIR ingestion and connects to open terminology service endpoints for real-time validation and expansion operations.
Your Analytical Data Warehouse implements the OMOP Common Data Model through Piano's AI-powered transformation pipelines that ingest data from FHIR streams, relational databases, and other sources with over 99% reliability. All external terminologies map to standardized vocabulary, enabling studies across disparate data sources and participation in networks like OHDSI without manual reconciliation of coding systems. Piano's RESTful APIs provide flexibility by allowing third-party tools to integrate for visualization, dashboards, and advanced analytics.
Your Governance Infrastructure is managed through Piano's access controls modeled on research ethics protocols, with project-based permissions, configurable de-identification, query suppression for small-cell protection, and comprehensive audit logging. Non-technical data custodians can manage access directly through Piano's intuitive interfaces.
Your Analytics Layer is powered by Piano's scientific toolbox, providing self-serve tools for clinicians and researchers alongside integration points for data scientists using R, Python, and specialized machine learning frameworks. Natural language generation produces reproducible methodology descriptions automatically.
Your Collaboration Infrastructure operates through Evidentli's Evidence Hub, enabling federated analytics across institutional boundaries through Piano Federation, sharing only methodology and aggregate results while keeping patient data local. Evidence Hub serves as a repository for validated research protocols that can be downloaded, peer-reviewed, and reproduced across participating Piano sites worldwide.
Solutions based on Piano are flexible and deployed in your environment. Piano can be deployed on a cluster of one or more virtual machines, hosted on premises or in private clouds offered by all major cloud providers. This ensures your data never leaves your control while still enabling the full range of collaboration capabilities through Evidence Hub.
For organizations requiring complete isolation, Evidentli offers Private Hub deployment—an instance of Evidence Hub hosted in your own data center. While this configuration requires additional maintenance and limits collaboration to within your organization, it provides maximum control for environments with the most stringent security requirements.
Rather than purchasing monolithic "solutions," you mandate platform compliance as the control mechanism for all acquisitions. Vendors and integrators must conform to your specified standards—openEHR for persistence, FHIR for exchange, OMOP-readiness for analytics—along with your terminology binding requirements. This inverts the traditional relationship: instead of investing in a vendor's vision and becoming locked into their proprietary data formats, you define what compliance looks like and select applications that meet your requirements.
Your procurement becomes a roadmap-based process rather than a single large contract. You incrementally purchase or build components according to clinical priority and financial capacity, with each new element connecting to your common platform infrastructure. Evidentli's Piano accelerates this roadmap by providing the transformation, governance, and analytics capabilities needed to realize value from your standards-based architecture immediately.
A new kind of solution integrator emerges to serve your needs: partners who embrace platform standards and perform the technical work of obtaining and connecting specific applications and services to your specifications, rather than selling you their proprietary ecosystem.
This model positions you as the integrator of your own ecosystem. Your data—your most important asset—remains yours, persisted in open formats, exchangeable through standard APIs, analytically accessible through normalized vocabularies, and governed according to policies you control. Your ability to innovate, respond to regulatory change, adopt emerging technologies, and serve your clinical mission is no longer constrained by decisions made years ago in a procurement process that handed control to someone else.
The infrastructure exists. The three open standards have active and growing communities that continue to improve them. Open terminology services demonstrate the feasibility of accessible, API-based semantic infrastructure. Evidentli's Piano platform can convert your legacy data into standardized OMOP formats with unprecedented speed and precision. Evidence Hub enables multi-site research collaboration without the traditional barriers of data sharing. Piano's scientific toolbox puts evidence generation in the hands of your clinicians.
Your clinicians gain the ability to generate evidence from their own data. Your researchers can participate in global networks through Evidence Hub while maintaining local governance through Piano. Your administrators can grant and revoke access with confidence. Your strategic decisions remain your own.
The question is whether you choose to build your future on a foundation you control—leveraging the complementary strengths of openEHR, FHIR, and OMOP, accelerated by Evidentli's Piano and Evidence Hub that democratize access to your data's value—or whether you continue investing in vendor-specific visions that optimize for their interests rather than yours.
1. Tsafnat G, Dunscombe R, Gabriel D, Grieve G, Reich C. Converge or Collide? Making Sense of a Plethora of Open Data Standards in Health Care. J Med Internet Res. 2024;26:e55779. doi:10.2196/55779
2. openEHR Foundation. openEHR Architecture Overview [Internet]. London: openEHR Foundation; 2021 [cited 2024 Dec 1]. Available from: https://specifications.openehr.org/releases/BASE/latest/architecture_overview.html
3. Health Level Seven International. HL7 FHIR Release 4 [Internet]. Ann Arbor: HL7 International; 2019 [cited 2024 Dec 1]. Available from: https://hl7.org/fhir/R4/
4. Observational Health Data Sciences and Informatics. The Book of OHDSI [Internet]. OHDSI; 2021 [cited 2024 Dec 1]. Available from: https://ohdsi.github.io/TheBookOfOhdsi/
5. Gabriel D. Bridging the Terminology Divide: From Community Vision to Technical Implementation in Healthcare (1.0). Zenodo; 2025. doi:10.5281/zenodo.17792077
6. National Cancer Research Institute, Cancer Research UK. Health data and the experience of researchers accessing data for research [Internet]. London: NCRI; 2020 Nov [cited 2024 Dec 1]. Available from: https://www.ncri.org.uk/accessing-health-data-for-research/
7. Madurasinghe VW, Bower P, Eldridge S, Fitzpatrick R, Goldsack J, Knapp P, et al. A good use of time? Providing evidence for how effort is invested in primary and secondary outcome data collection in trials. Trials. 2022;23:1018. doi:10.1186/s13063-022-06951-8. PMCID: PMC9793601
8. Clinical Trials Arena. Revealing the human and business cost of clinical trial inefficiencies for Data Managers and CRAs [Internet]. London: Clinical Trials Arena; 2024 Dec [cited 2024 Dec 15]. Available from: https://www.clinicaltrialsarena.com/
9. Fink JL, Bourne PE. Issues in biomedical research data management and analysis: needs and barriers. AMIA Annu Symp Proc. 2007;2007:234–238. PMCID: PMC2244904
10. World Medical Association. Declaration of Helsinki: ethical principles for medical research involving human subjects. JAMA. 2013;310(20):2191–2194. doi:10.1001/jama.2013.281053