Article type
Year
Abstract
Background: Multiple publications arising from the same trial are common and can lead to bias in systematic reviews if these are not identified. The Cochrane Dementia and Cognitive Improvement Group (CDCIG) trial-based Register was developed to organise and maintain reference and trial records for possible inclusion in CDCIG systematic reviews. Records in the Register are organised primarily by trial rather than bibliographic reference.
Methods: MEDLINE. EMBASE, PsyINFO, CINAHL and LILACS have been searched over all years available using the groups sensitive search strategy. The abstracts of all references which resulted were read. Those that related to randomised or controlled clinical trials were added to the Register. These searches are updated every six months. Ongoing trial databases like NRR and Clinicaltrials.gov are searched quarterly. Electronic subscriptions to all major journals in the field of dementia ensure that in many cases trial reports are added to the Register well before the twice yearly update searches take place. Hard copies were obtained of all 3000+ references in the CDCIG Register. Initially, a trial record was created for each reference. Major characteristics of the trial were then coded into fields such as study design, health condition, number of participants, intervention, blinding, main diagnostic criteria, outcome, etc. Later, each reference was checked against existing trial records to make sure it was linked to the appropriate trial.
Results: The result is a comprehensive, up-to-date register of trials in the field of dementia with detailed coding of all major study characteristics.
Conclusions: The trial-based register has proved to be a most valuable tool. It helps to cut across salami-sliced publishing. For example, information from conference proceedings and websites which does not appear in published reports is linked to the relevant trial. It greatly reduces the work required by authors in producing systematic reviews. It makes possible extremely focussed searches on coded study fields, thus facilitating research on a wide range of questions. For example, performance of specific outcome measures or the response of particular symptoms can be readily examined across trials.
Request for searches are welcome and should be directed to the Coordinator of the Group.
Methods: MEDLINE. EMBASE, PsyINFO, CINAHL and LILACS have been searched over all years available using the groups sensitive search strategy. The abstracts of all references which resulted were read. Those that related to randomised or controlled clinical trials were added to the Register. These searches are updated every six months. Ongoing trial databases like NRR and Clinicaltrials.gov are searched quarterly. Electronic subscriptions to all major journals in the field of dementia ensure that in many cases trial reports are added to the Register well before the twice yearly update searches take place. Hard copies were obtained of all 3000+ references in the CDCIG Register. Initially, a trial record was created for each reference. Major characteristics of the trial were then coded into fields such as study design, health condition, number of participants, intervention, blinding, main diagnostic criteria, outcome, etc. Later, each reference was checked against existing trial records to make sure it was linked to the appropriate trial.
Results: The result is a comprehensive, up-to-date register of trials in the field of dementia with detailed coding of all major study characteristics.
Conclusions: The trial-based register has proved to be a most valuable tool. It helps to cut across salami-sliced publishing. For example, information from conference proceedings and websites which does not appear in published reports is linked to the relevant trial. It greatly reduces the work required by authors in producing systematic reviews. It makes possible extremely focussed searches on coded study fields, thus facilitating research on a wide range of questions. For example, performance of specific outcome measures or the response of particular symptoms can be readily examined across trials.
Request for searches are welcome and should be directed to the Coordinator of the Group.