i2b2 Japan

The gap this fills#

Hospitals hold enormous amounts of clinical data, but a researcher cannot ask questions of it without a system built for the purpose — one that can count how many patients match a set of criteria before a study is designed around them.

i2b2↗ is the open-source standard for this. It was developed in the United States and is used by research hospitals worldwide, and it is free to install. Uptake in Japan has been thin all the same, for two reasons that have nothing to do with the software: the documentation is in English, and it does not explain how to connect i2b2 to the systems Japanese hospitals actually run.

Why this is worth doing#

The software is free, already written, and running in research hospitals around the world. What decides whether a hospital in Japan gets to use it is something much smaller: whether someone here can read the schema and start a server. That is where it has been stopping.

The repository that runs it#

To make i2b2 easier to stand up, there is a second piece: i2b2-jp-docker↗.

It builds on the Docker images i2b2 publishes officially, adapted for the way hospitals here will run it. Clone it, run docker compose up, and a server comes up ready to log into from a browser.

The implementation guide on the site follows this repository step by step.

What the site covers#

i2b2 Japan is a Japanese-language documentation site for people evaluating or deploying i2b2 at a Japanese institution.

Architecture

The cells and the hive, and what each one is responsible for.

Data repository (CRC)

The star schema, table by table — observation facts, the dimension tables, and the mapping tables.

Ontology

How the ontology cell organises the terms a query is written against.

Access control

User roles and what each level of data access permits.

Implementation

Standing a server up with Docker, and the parts that catch people out.

There is also a full table reference, and a set of links to the i2b2 resources worth reading.

Read the documentation↗.