We recently invited three colleagues for a fireside talk on their recent diplomas – not as a celebration of academic titles, but as a deep dive into how structured research transforms the way we build for healthcare. At Kodality, nurturing deep domain expertise is our primary tool for ensuring that our technology survives contact with complex clinical reality.
The Participants:
- Kiira, Junior Software Developer. Developed a web application for physiotherapists to design and manage exercise programs (Bachelor’s, TalTech, 2025)
- Merit, Analyst. Validated TermX* by mapping Estonian radiology procedures to SNOMED CT, proving that the tool can create and handle complex clinical semantics (Master’s, TalTech, 2024).
- Marina, Software Developer. Designed and implemented a CTS2- and FHIR-compatible terminology server architecture, establishing TermX as a standards-based bridge between legacy terminology services and modern interoperability frameworks (Master’s, TalTech, 2024).
*TermX is an open-source platform, developed by Kodality in collaboration with the Estonian Ministry of Social Affairs, that simplifies the design, governance, and publication of clinical terminologies and information models – enabling healthcare organisations to build a consistent, future-proof data layer across projects, teams, and systems.
The Motivation: Why the Deep Dive?
Kodality: You have all moved past your graduation ceremonies, but your research remains at the core of what you and we do. Why was it important for you – and for Kodality – to tackle these specific academic challenges rather than relying solely on learning on the job?
Kiira: For me it was personal. I have been a physiotherapy patient myself. I saw how much work therapists put in and how much of it is still manual. Creating exercise programs often means copying, editing, adjusting the same things again and again: writing repetitions, setting frequencies, reformatting instructions. It takes time. And that time could be spent with patients instead. When I started thinking about my thesis, I didn’t want to build something “innovative”. I wanted to fix something obvious. There was a clear inefficiency. So I decided to analyse it properly and solve it.
Merit: I was told that mapping radiology procedure classifications to SNOMED CT had been attempted before and had failed. That immediately made it a challenge I wanted to take on. Coming from a clinical background I could see why the original classifications were structured the way they were but I also saw how inconsistent and muddy they were from a data perspective. I wanted to prove that it could be done properly and that by applying semantic structure, we could make something confusing, transparent, even richer.
For me, it was about bridging clinicians and informatics and showing what becomes possible when you truly understand both sides.
Marina: My focus was on something most hospitals do not see but constantly depend on – the architectural layer that ensures terminology remains consistent across systems. Healthcare institutions have decades of accumulated classifications, local code lists, and vendor-specific implementations. My thesis explored how to create a standards-based bridge that allows legacy and modern systems to coexist while sharing the same semantic foundation. For Kodality, this was not theoretical. It directly influenced how we designed TermX to operate as a sustainable interoperability layer rather than just another tool.
The Human Element (Kiira’s Research)
Kodality: Kiira, your research focused on a web application for physiotherapists. Hospital management often views rehabilitation as a difficult area to digitise because of its hands-on nature. What did your work reveal about that?
Kiira: The core issue is time and structure. Rehabilitation sessions are hands-on, yes. But a lot of the workload sits around the session such as, creating exercise plans, adjusting them, documenting changes, explaining them clearly, and hoping the patient remembers everything at home. During my research I saw that physiotherapists still do a surprising amount of manual work when creating programs. Even when digital tools exist, they are often rigid, slow to adjust, or not adapted to local workflows. So therapists fall back to Word documents, paper sheets, or rewriting plans repeatedly. That creates two problems: it wastes time, and it reduces consistency.
What I built was not meant to replace the session. It was meant to remove the repetitive parts around it. Structuring exercises. Reusing templates. Adjusting repetitions and frequencies quickly. Giving patients a clear view of what they need to do.
Kodality: And that is the Kodality philosophy – it is not just a tracker, right?
Kiira: Exactly. It is a clinical tool, designed to enhance communication and collaboration between therapists and patients. Beyond organising exercises, it helps therapists track progress over time, identify patterns, and adjust programs based on real patient data, not just assumptions. It also supports personalisation, so each patient’s plan can evolve as they improve, and therapists can see what works best. In the end, it’s about making the entire rehabilitation process smarter, more consistent, and easier to follow, both in the clinic and at home.
The Language of Medicine (Merit’s Research)
Kodality: Merit, you addressed one of the most significant strategic challenges in healthcare: the standardisation of procedure coding in X-ray and Angiography. Why would this be a priority for a modern hospital?
Merit: Because without standardisation radiology procedure codes become inconsistent across systems, which prevents reliable patient safety safeguards, accurate billing and reporting, operational planning and meaningful data reuse. All this makes it impossible for a modern hospital to function as a transparent, interoperable and data-driven organisation.
Kodality: This makes cross-institutional data sharing almost impossible.
Merit: Precisely. Radiology does not operate in isolation. Patients move between institutions, registries collect national data and digital health systems increasingly rely on structured data exchange.
In my thesis, I focused on ensuring that radiology procedure data carries the same meaning beyond the originating hospital, making cross-institutional interoperability technically and semantically possible.
The Digital Backbone (Marina’s Research)
Kodality: Marina, you looked at the architectural backbone – Implementing CTS2 and FHIR compatible Terminology Server. Why does this technical bridge matter for a hospital’s future?
Marina: Every system introduces its own way of storing and interpreting clinical codes. Over time, this creates fragmentation: the same procedure or diagnosis may exist in multiple forms across systems.
The challenge is not only technical integration – it is semantic consistency. FHIR has become the dominant interoperability standard for modern health data exchange, while CTS2 defines robust terminology service capabilities. If these frameworks remain disconnected, hospitals face repeated transformation efforts every time they integrate a new system.
My work demonstrated that a terminology server can act as a stable intermediary layer – exposing terminology through FHIR interfaces while maintaining structured governance and lifecycle management aligned with CTS2 principles. This ensures that meaning remains consistent even as applications change.
Kodality: So, your thesis essentially validated TermX’s role as the architectural layer that allows a hospital to innovate without having to abandon its existing data foundations?
Marina: Yes. I used TermX as the practical implementation context to demonstrate how a standards-compliant terminology service can centralise classification management, reduce redundant mappings, and provide a single governed source of truth.
One Thread Across All Three Diplomas
Kodality: As we hear your stories, a clear theme emerges: meaning has to survive the journey – from clinic to home, from department to dashboard, from one system to another.
Kiira: Yes. If meaning breaks, rehabilitation depends on individual memory and informal routines. Administrative burden grows and consistency declines. But when program creation is structured and reusable, the process remains stable over time. That stability makes rehabilitation efficient and manageable.
Merit: If meaning breaks, data may still move but it no longer carries understanding. And without shared understanding modern healthcare cannot operate safely or strategically.
Marina: If meaning breaks, interoperability becomes fragile and expensive. Every integration requires custom transformations, every reporting project needs manual reconciliation, and every system upgrade risks semantic drift.
But if terminology is governed centrally and exposed through standards-based services, meaning becomes portable. Systems can change, but the semantic backbone remains stable. That stability is what makes true interoperability economically sustainable.
Kodality: And this is where your work connects back to Kodality’s daily principle: we build systems that matter.
A Quiet Point About TermX – and Why the Research of Our Experts Touched It
TermX appeared in the two master theses discussed because, at Kodality, we believe terminology is not a side feature. It is the layer that decides whether clinical information stays consistent as it is reused for care delivery, integration, research, quality, and management reporting.
Our solution is built to simplify the complexity of clinical taxonomy management – enabling hospitals and institutions to model and publish terminology according to their specific needs. In practice, as proven by Merit and Marina’s work, that means:
- Clear and consistent procedure definitions that clinicians and systems interpret the same way.
- Less duplicate maintenance of local code lists and fewer inconsistencies between departments.
- Reduced manual reconciliation when integrating with different systems.
- Reliable alignment with international standards enabling safe data exchange beyond a single institution
What Hospitals Can Take from This
For management looking to future-proof their institutions, these three bodies of research offer a clear roadmap:
Kiira: I would suggest that hospitals adopt digital tools to simplify the creation and management of exercise programs. A web application that structures exercises, reuses templates, and adjusts repetitions quickly can reduce repetitive manual work and ensure patients clearly understand what to do at home.
Management can start by integrating such tools into therapist´s workflows, freeing clinicians to focus on hands-on care while making rehabilitation more consistent and scalable across departments. Over time, this approach improves patient engagement, adherence, and overall outcomes.
Merit: I would start by stabilising and governing clinical terminology as a strategic asset rather than treating it as system configuration. Over time, local procedure lists, departmental variants, and ad-hoc mappings accumulate across systems, creating hidden semantic fragmentation. This makes integrations fragile, reporting inconsistent, cross-institutional exchange difficult, and digital initiatives unnecessarily expensive.
Future resilience comes from consolidating classifications into a governed source of truth, aligning them with international standards and establishing clear ownership and lifecycle management. When terminology is structured, transparent, and consistently maintained, clinical meaning becomes portable and can survive system upgrades, organisational reforms, registry integrations, and cross-institutional data exchange.
Technology will continue to change. Platforms will be replaced. Regulations will evolve. But if the semantics are stable, the organisation remains interoperable, analytically capable, and operationally resilient. Future-proofing is not primarily about acquiring new systems, it is about ensuring that meaning survives every transition.
Marina: I would advise hospital leadership to treat terminology services as core infrastructure rather than as a technical afterthought embedded inside individual systems.
Many interoperability challenges are not caused by incompatible software, but by the absence of a governed semantic layer. Without it, each integration becomes a bespoke engineering project, and each system replacement risks loss of meaning.
Future-proofing requires centralising lifecycle management, version control, and mapping governance, so hospitals can ensure that clinical meaning remains stable across integrations, analytics initiatives, regulatory changes, and vendor transitions.
Closing
Supporting university studies is not a perk at Kodality. It is part of how we protect quality in mission-critical healthcare delivery. These diplomas show the same thing from different angles: healthcare technology works when clinical intent, operational reality, and technical standards stay aligned – in the clinic, in the record, and across systems.
If you are working on interoperability, procedure coding standardisation, or terminology governance, we are always open to a practical conversation – grounded in what hospitals actually need, and informed by the kind of structured research you have just seen.