Clinical Trial Management
Data Management Systems versus Data Warehousing – What Really Matters
Norbert Fritz1 and Charlotte Mauron2
1. Development Leader, Business Intelligence Warehouse, Product Development, F Hoffmann-La Roche Ltd, Basel; 2. Manager, Data Modelling and Acquisition, Biometrics, F Hoffmann-La Roche Ltd, Basel
Abstract
Data integration is the central mechanism that enables simultaneous collection, organisation and usage of data by a set of systems interconnected within a dedicated landscape. This article intends to clarify how different types of system landscapes are suited to support the integration of clinical data. For this purpose, two architectures with different central platforms are compared: data management systems versus data warehouses. In addition to technical features, the data modelling capabilities, given by options and constraints to represent and manage clinical data in a semantically correct way, have to be considered for each landscape architecture. With this in mind, the current industry trend to design clinical system landscapes with data warehouse as the central components is examined.
Keywords
Pharma development, clinical trial data, clinical system landscape, data management systems, data warehouse, data modelling, semantic data integration
Disclosure: The authors have no conflicts of interest to declare. Acknowledgements: This article includes concepts that have been discussed with many colleagues at F Hoffmann-La Roche Ltd in the context of various initiatives. Nevertheless, the content of this paper represents solely the opinion of the authors. The authors have referred to many concepts without citations as they are widely used in debates about data warehousing, even if they originate from specific publications. Part of the conclusions made here have been presented elsewhere. Received: 27 August 2010 Accepted: 17 November 2010 Citation: Drug Development, 2010;5:44–7 Correspondence: Norbert Fritz, Product Development, F Hoffmann-La Roche Ltd, Grenzacherstrasse 124, 4070 Basel, Switzerland. E:
norbert.fritz@
roche.com
During the past decade, the management of clinical data has become increasingly challenging due to the expanding data volume, greater complexity of clinical trials and amount of medical information and the rising diversity of data usage for operational and analytical purposes. Thus, clinical system landscapes have developed that interconnect multiple systems that are dedicated to specific tasks, such as collection, processing, hosting and presentation of clinical data. In general, this raises the challenge to integrate clinical data arising from multiple data sources and systems. This article intends to clarify how different types of landscapes support data integration. For this purpose, clinical system landscapes centred around data management systems are compared with those centred around data warehouses.
Technical aspects will not be the main focus, as methodologies for technical system integration have significantly matured over the last decade and are well developed, such as extraction, transformation, loading (ETL) and service-orientated architecture. Rather, a closer look will be taken into the ability to represent, transmit and share medical information within system landscapes. Such analysis reveals that semantic aspects are the major factors that determine the (semantic) interoperability between interconnected systems.
To compare data management systems with data warehouses as central data integration platforms, data modelling capabilities and semantic data integration will be investigated. It will be pointed out that consistent data modelling across systems is an essential condition for the processing and usage of clinical trial data in any
44
system landscape. Some advantages to the current trend towards designing clinical system landscapes with a data warehouse as the central unit are also highlighted.
Data Sources
The conduct and management of clinical trials entail a multiplicity of tasks, processes and information that generate and consume different classes of data. For example, clinical trial protocols cover:
• clinical data; • safety data; • administrative data; • sampling and kit information; • vendor information; • randomisation data.
These diverse types of data are managed in different systems that are typically designed, customised and implemented for specific business purposes. Such systems are often exclusively used by part of an organisation, for example clinical trial management systems, planning and finance systems, clinical data management systems, safety systems, data analysis and reporting tools, etc.
A fundamental aspect to consider is that each system has its own way of representing data, which is primarily related to a specific purpose, but this also depends on the underlying architecture. As a consequence, the same information stored across systems for different purposes (e.g.
© TOUCH BRIEFINGS 2010
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68