2. Ruby implementation of openEHR
● Done
– RM, AM
implementatio
n
– ADL 1.4
parser
– Archetype to
Rails app
● Doing
– OPT parser
– Rails Engine
for openEHR
– Sample app
for our next
generation
EHR
● To do
– Terminology
(ontology)
flattener
– ADL 2.0
parser
● RR
grammar
3. Dolphin project
● Union of regional health care system
– Kyoto, Miyazaki, Kumamoto, Tokyo
– Web based system for hospitals and patients.
● Standardised by MML
– XML standard for sharing clinical information
– 22 XML formalism
● 12 years history
– Mainly coded by Perl and M(Mumps).
– DB Engine is Cache(TM).
4. rDolphin Project
● Revised Dolphin
– Modernise architecture. → MVC Framework
– Object oriented programming, Cache to RDBMS(or
NoSQL)
● Research Dolphin
– For implementation research of openEHR
– For clinical research
● Ruby Dolphin
– Ruby and Ruby on Rails implementation
5. Implementation overview
● Ruby on Rails(Ruby 2.2.3, Rails 4.2)
● Re-construction of MML by archetype concept
models
– Hand generated forms and classes
● PostgreSQL database engine
● Open Source Software
6. Logical model of Ruby classes
Person
Party
Organization EHR Composition
ContentItem
ContentItem
ContentItem
ContentItem
9. Performance bench mark(2)
Document
size
Concurrent
access: 1
Concurrent
access: 10
Concurrent
access: 100
HTML 8,500bytes 29.99 30.35 30.15
XML 812bytes 104.62 104.48 106.22
JSON 172bytes 108.83 109.08 107.01
CSV 16bytes 140.74 146.85 143.93
HTTP server: unicorn 4.8.3,
CPU Core i7 4770, 3.4GHz, RAM 32GB, SSD 255GB
Measured by Apache bench to localhost. Unit: request/second.
Each result was the avarage of 1,000 trials
10. Further works
● Rails Engine
– Auto-generation from OPT(Operational Template) to
● Web forms
● DB schema
● Ruby codes for models, controllers, views on Rails
● Javascripts and CSSs
● DB Engine
– RDBMS: PostgreSQL, MySQL
– NoSQL: Google Big Table, Hadoop