Collecting systematic review data

Tinazzi A., Floriani I. and Torri V. Istituto di Ricerche Farmacologiche "Mario Negri", Milano, Italy.

Introduction: Although strong efforts have been made to provide guidelines and instruments for better data collection, the specific features of any systematic review, in terms of the type of data to be collected and outcome measures, require the design of "ad hoc" forms and the handling of information within appropriate databases.
Objectives: To build up a database able to match all the aspects to be taken into account when designing a form, with particular emphasis to the information to be collected, formats to be used, data-dictionary to be implemented.
Results: A practical application using relational database has been developed; main features are:
Definition of specific sessions for:

• References and reviewer codes
• Verification of criteria for study inclusion
• Description of study characteristics, with particular emphasis to the following aspects:
  • definition of internal data-dictionary in order to avoid format duplication
  • computing of summary variable and automatic conversion of scales
  • flexibility in collecting outcome data according to different ways of presentation of results
• Conversion of database file compatible with common statistical analysis software

Discussion: The database has been successfully used for collecting data relating to the overviews of the different diseases, in particular in the analysis of the efficacy of vaccines for allergies and in chemotherapy for advanced breast cancer. Further work for making it compatible to Revman is needed

Maryland 1998 O13