A cloud computing database for data extraction in a Cochrane Review

Article type
Authors
Stordal B1, Kapros E2, Busschots S1, Lawlor D1, Doherty B1, Smith L3, Hennessy B4, O’Mathúna D5
1Department of Histopathology, St James’ Hospital and Trinity College Dublin, Dublin 8, Ireland
2Department of Computer Science, Trinity College Dublin, Dublin 2, Ireland
3Department of Social Work and Public Health, Brookes University Oxford, UK
4Department of Medical Oncology, Beaumont Hospital, Dublin 9, Ireland
5School of Nursing & Human Sciences, Dublin City University, Dublin 9, Ireland
Abstract
Background: An online database was needed for the Cochrane Review ‘Taxanes for the treatment of platinum pre-treated epithelial ovarian cancer’ for several reasons. Firstly, a large amount of clinical parameters were to be extracted from 68 identified studies and the data was to be subdivided into 16 different subgroups. Secondly, there were six authors performing data extraction for this Cochrane Review, working at four different institutions. Therefore data could not be entered and stored centrally at one institution.

Objectives: Development of a cloud computing database accessible online for all authors performing data extraction.

Methods: The database utilises cloud computing, such that the information is stored online and can be accessed from anywhere and is populated through a web-based user interface.

Results: Using an online database-as-a-service and cloud computing to query and compare data allows the elimination of transaction concurrency conflicts. That is, the two authors can simultaneously review the same study without breaking the database. In the data extraction phase the cloud computing model allows authors to all work remotely. Each study in our review must have the data extracted by two review authors. The database will crosscheck information entered by two authors and flag discrepancies that need to be resolved by discussion. In the analysis phase the database allows each relevant paper containing data for each subgroup to be easily identified amongst the identified studies. Data to be presented will include a demonstration of functionality of the database and excerpts of results highlighting the usefulness of the database in complex subgroup categorisation.

Conclusions: The cloud computing database was invaluable for performing this Cochrane Review as the subgroups overlapped in many publications. Many complex Cochrane Reviews could benefit from this approach to data extraction which facilitates collaboration across multiple institutions.