Warehouses Episode 1: Clinical Data Models

In this first episode of the Clinical Data Warehouses Series we start with fundamentals: the data model.

The motivating reasons for investing in data warehouses, for most businesses, is to collect data from all corners of the organisation so it can be mined for a variety of reasons. It makes sense, therefore, to create a custom data model that fits the organisation exactly. As every organisation is different, so it is common for data models to be tailored for each organisation.

This leads organisations with clinical data to think similarly about clinical data warehouses which is:

Mistake 1: Not using the right clinical data model

The international organisation Observational Health Data Sciences and Informatics has membership from governments, academia and industry in almost all countries. It has developed the Common Data Model (CDM) that is now used by thousands of projects around the world. Using the CDM lets you:

• Not have to re-invent a clinical model
• Use third party tools (including many free ones) to leverage algorithmic power in cleaning, standardising and aggregating data from different data sources.
• Leverage the extensive knowledge and documentation of the model that would then not be the IT department’s responsibility to produce and maintain.