Integration between systems that hold the same patient

The engine maps the interfaces between clinical and administrative systems from what the data means, keeps them working when a system changes, and runs inside the infrastructure you already govern.

The sector

Hospitals and health services run a patient record, lab, radiology, pharmacy, scheduling and finance systems, and a long tail of departmental ones. Data also moves between organisations: hospitals, municipalities, primary care and national registries. Patient data is among the most tightly controlled data any organisation holds.

The problem

Each of those systems describes the same patient, episode and order in its own way, even when they claim to follow the same standard. The interfaces between them are kept alive by a small team that is always behind, and replacing any one system means re-mapping everything it touched.

How we fix it

healthcare · where it breaks and what the engine doeswhere it breaks · the fix swipe →
Where it breaksWhat the engine does
Standards that vary in practice HL7 v2 feeds, FHIR APIs and openEHR repositories run side by side, and two systems that claim the same standard rarely fill the same fields the same way. The engine maps by the content and context of each field, so two systems that fill a standard differently can still be matched. How well it reads a particular feed is tested on your data.
System replacement Replacing a patient record or a lab system means re-mapping every interface that touched it. Discovery maps every interface around the system before it is switched off, and the engine re-maps them onto its replacement.
Across organisations Data shared between hospitals, municipalities and primary care crosses organisational lines, and systems built for the same purpose by different vendors. Each sender’s data is mapped by meaning into the shape the receiver needs, and re-mapped when either side changes.
A wrong mapping is a clinical risk An error in a medication or results interface is not an inconvenience. Set the threshold to 1.00 and nothing applies until a person approves it. The engine still does the work of proposing every mapping.

What your architect will ask

What does our DPO need to know?
The engine has no connection to us or to any other third party. Its graph store holds classified data, including personal data if it is in scope: a second copy inside your estate, on your hosts, under your access controls and retention rules. Whatever governs patient data in your organisation, whether Normen, HIPAA or both, applies to it as to any other internal system.
Does it understand HL7 and FHIR?
It works from what those standards deliver once they reach it: API payloads, database tables or files. How well it interprets a particular feed is something to test on your data, not to take from a web page.
What happens if a component stops?
Modules are connected by queues. Records wait instead of being lost, and failed records can be replayed once the cause is fixed.

Where we would start

With discovery around one system that is due for replacement: every interface that touches it, mapped before anything is switched off.

Bring us one of your integrations

We will run it against your systems in a working session and show you what the engine does with it.

Book a demo