Upload
miranda-dawson
View
216
Download
2
Tags:
Embed Size (px)
Citation preview
A clinical database for networking in multicenter studies
Ulrike Haase1, Marc Bonin1, Stephan Flemming2, Frank Buttgereit1, Till Sörensen1, Thomas Häupl1
1Department of Rheumatology and Clinical Immunology, Charité University Hospital, Berlin 2Pharmaceutical Bioinformatics, Institute of Pharmaceutical Sciences, University of Freiburg
Motivation
Aim of this project was to develop a clinical database system for sharing of clinical information in multicenter research studies. The system should be flexible to store any kind of data and to enable the user to include any new parameter. It should also conform to data privacy regulations and should be accessible through the internet.
Methods
Programming was based on a web framework in Ruby on Rails. MySQL was used as database system (Figure 1).The complete software package is running on a virtual machine with Windows 7, which acts as a normal server.
Results
For storage of clinical information, two tables were defined. The first table contains all data in structured format, the second table harbors all information about the structure for each type of parameter. In detail, one or more values of a given parameter are linked to two different masterkeys and three different parameter keys (Figure 2).
The masterkeys define the patient identity (patient-ID) and the time point of the visit (Figure 3).The parameter keys define the structure of tables (names of the tables, the rows, and the columns), which can be generated by the user to store clinical information (Figure 3). As an example, a data structure for clinical information from patients with rheumatoid arthritis was established (Figure 3 + 4).
In addition, a system of user and data specific rights was established to own and share data for reading, writing and administrating rights (Figure 5). Graphical user interfaces guide the user through administrative work (create or delete tables, rows and columns, share data) and data collection (enter new patients, visits, data values).
Conclusion
With the new database any kind of information can be stored including images. Every new parameter e.g. new tables, rows or columns can be added without programming knowledge. Many users can access at the same time from their own PC via an internet browser. Access is protected via user specific login and different levels of rights. Thus the database enables to collect data in the clinics, to share these with scientists, to enable biobanking, or sample tracking. The database is currently part of the national research network Arthromark (BMBF-funded), the EU-network NanoDiaRA (EU-funded) and the German-Singaporean collaboration DeSiNet-Rheuma (BMBF-IB-funded) (Figure 6).
Thomas Häupl, MDDepartment of Rheumatology and Clinical ImmunologyBioinformatics GroupCharité University HospitalCharitéplatz 1D-10117 Berlin Germany
phone: +49 30 450 513 291fax: +49 30 450 7 513 291e-mail: [email protected]
Contact: Ulrike HaaseDepartment of Rheumatology and Clinical ImmunologyBioinformatics GroupCharité University HospitalCharitéplatz 1D-10117 Berlin Germany
phone: +49 30 450 513 296fax: +49 30 450 513 968e-mail: [email protected]
Figure 6 Collaborating Networks
Table name • diagnostics
Parameterkey 2 Parameterkey 3
Parameterkey 1
Beginn of symptoms
Date of diagnostics End of disease
rheumatoid arthritis 02/2008 06/2008
diabetes mellitus 06/2009 12/2009
arterial hypertonia 12/2010 12/2010
pneumonia 03/2011 03/2011 04/2011
value (date)
Row names• e.g. rheumatoid arthritis
Column names • e.g. Begin of symptoms
Diagnostics
Figure 2
1.Table in database : saving the structered data sets
2. Table in database : structure information
Figure 1
Database system
Web framework
Ruby on Rails
Clinical Database - Aims
• sharing of clinical information in multicenter research studies• flexible system
• should store any kind of data• should enable the user to include any new parameter
• conform to data privacy regulations• should be accessible through the internet
Aims of the clinical database and used methods (software)
medBioClinicDB
Parameter-key 1
Parameter-key 2 Parameter-key 3
data type values …
laboratory CrP Value Numberlaboratory CrP Unit String mg/l, mg/dl
laboratory CrPReference range
String
… … …
ID Master-key 1
Master-key 2
Master-key 3
Parameter-key 1
Parameter-key 2
Parameter-key 3
Value Numbe
r
Value String
Value Date
User Created at
… … … … … … … … … … … …
Consecutively items Logs „who“ saved the data „when“
define data type for input data
input options for user
Figure 3
Masterkey 1
for storing patient specific data (common data about patient)
for storing patient and visit specific data (visit dependent data about patient)
Parameterkey 2 Parameterkey 3
Parameterkey 1
Value Unit Standard Value
CrP 0.7 mg/dl0.5
ESR 1h 30 mm/1h 15
ESR 2h 50 mm/2h 30
white blood cells (WBC) 7.3 /nl 4..10
value (number,string)
Laboratory
+
Patients (Patient-ID)
Masterkey 2Visits (visit date)
Masterkey 1Patients (Patient-ID)
Figure 4
A table example taken from the clinical database for rheumatoid arthritis. The rectangles with dotted lines show the types of defining parameters(masterkeys, paramaterkeys and values)
Masterkey 1 Masterkey 2 Parameterkey 1
Figure 5 sharing all data from one patient (patient-ID 165)
Value
different levels of user rights
Choose which data should be shared
The main two tables in the database for saving the data sets and its structure
Definition of masterkeys and parameterkeys
Parameterkey 3
values (number,string)
Parameterkey 2