Abstracts

A Web-Based Relational Database System for Management of an Observational Study: Design and Development

Abstract number : 2.073
Submission category : Clinical Epilepsy-Pediatrics
Year : 2006
Submission ID : 6512
Source : www.aesnet.org
Presentation date : 12/1/2006 12:00:00 AM
Published date : Nov 30, 2006, 06:00 AM

Authors :
1Anthony Marmarou, 1John M. Pellock, 1Lori L. Davis, 1Thomas E. Cason, 1James P. Agnew, 2Shlomo Shinnar, 2Christine O[apos]Dell, 3Darrell V. Lewis, 4Douglas

A Web-based relational database management system (RDBMS) for electronic entry of paper Case Report Forms (CRFs) transmitted by 6 study centers were linked dynamically to a series of databases for prospective and retrospective cohorts (48.6% Female, 52.3% Caucasian, 41.1% Black, and 17.8% Hispanic with a median age of 16.7 months) of children experiencing febrile seizures., The FebSTAT study includes 5500 data elements. Included are demographics, neuropsychological scoring, MRI structural analysis, EEG clinical assessment, virology, genetics, medications, seizure start and stop, continuous intermittent seizure type and follow up data. A relational database allowing entry tracking and data reporting was designed utilizing enterprise database software residing on a Dell PowerEdge 350 server. Over 150 entity- relationship tables address form data and samples including error and submission tracking. All data transmissions were HIPAA compliant. Archiving and backups of the complete database are performed nightly., The system promotes efficiency of personnel via the automation of tasks such as form, sample, and data error tracking. Status of data transmission and receipt can be viewed by the investigators and study coordinators. The system eliminates the need of personnel to manually re-query outstanding data by using programmatic means (SQL) to re-query on a preset schedule. The database reflects the most recent updates of information at all times, rather than obsolete or older data which can result from the manual entry of data updates by personnel. The system is capable of performing analysis to find trends such as persistently missed questions, study center performance and CRF design problems The query system for quality control has resulted in 0.30% error rate and of 2441 queries, only 53 (2.0%) remain unresolved., Despite the large number of variables, the FebSTAT database optimized for this study was achievable and allows easy access for analysis and study by the FebSTAT Investigators., (Supported by NINDS - R01 NS43209 Consequences of Prolonged Febrile Seizures in Childhood.)
Antiepileptic Drugs