Standard requirements for GCP-compliant data management in multinational clinical trials

Abstract

BackgroundA recent survey has shown that data management in clinical trials performed by academic trial units still faces many difficulties (e.g. heterogeneity of software products, deficits in quality management, limited human and financial resources and the complexity of running a local computer centre). Unfortunately, no specific, practical and open standard for both GCP-compliant data management and the underlying IT-infrastructure is available to improve the situation. For that reason the "Working Group on Data Centres" of the European Clinical Research Infrastructures Network (ECRIN) has developed a standard specifying the requirements for high quality GCP-compliant data management in multinational clinical trials.MethodsInternational, European and national regulations and guidelines relevant to GCP, data security and IT infrastructures, as well as ECRIN documents produced previously, were evaluated to provide a starting point for the development of standard requirements. The requirements were produced by expert consensus of the ECRIN Working group on Data Centres, using a structured and standardised process. The requirements were divided into two main parts: an IT part covering standards for the underlying IT infrastructure and computer systems in general, and a Data Management (DM) part covering requirements for data management applications in clinical trials.ResultsThe standard developed includes 115 IT requirements, split into 15 separate sections, 107 DM requirements (in 12 sections) and 13 other requirements (2 sections). Sections IT01 to IT05 deal with the basic IT infrastructure while IT06 and IT07 cover validation and local software development. IT08 to IT015 concern the aspects of IT systems that directly support clinical trial management. Sections DM01 to DM03 cover the implementation of a specific clinical data management application, i.e. for a specific trial, whilst DM04 to DM12 address the data management of trials across the unit. Section IN01 is dedicated to international aspects and ST01 to the competence of a trials unit's staff.ConclusionsThe standard is intended to provide an open and widely used set of requirements for GCP-compliant data management, particularly in academic trial units. It is the intention that ECRIN will use these requirements as the basis for the certification of ECRIN data centres.

Open access tools for quality-assured and efficient data entry in a large, state-wide tobacco survey in India

“…Most of the data entry errors were identified and corrected. Double data entry and validation was performed, which is considered the gold standard for reducing data entry errors [3,4]. Auto-recording of time taken to enter each record (this cannot be edited) ensured that actual double data entry happened and the DEO did not ‘copy and paste’ single entered data.…”

“…In this, data are entered independently twice and the two databases are compared for discordances, followed by their resolution by referring to the original data collection forms [3,4]. To achieve this, we replicated a model of combining open-access tools for quality-assured data entry (previously described in an operational research setting) in a large sub-national tobacco survey (TNTS) in a resource-constrained setting [5].…”

How can the Quality of Medical Data in Pharmacovigilance, Pharmacoepidemiology and Clinical Studies be Guaranteed?

“…Many data management procedures exist. [11][12][13] However, on analysis it becomes clear that these guidelines are insufficiently detailed on a critical and essential point: validation of the quality of source data and information in order to guarantee the validity/accuracy in the medical sense (e.g. is what is recorded as myocardial infarction is really myocardial infarction; what criteria were used for this diagnosis?).…”

Comment garantir des données médicales de qualité dans les études cliniques, pharmaco-épidémiologiques et en pharmacovigilance ?

“…Il existe ainsi de nombreuses procédures relatives au data management. [11][12][13] Cependant, l'analyse permet de mettre en évidence que ces recommandations ne sont pas assez détaillées sur un point essentiel et critique : la validation de la qualité des données sources et de l'information qui permet d'en garantir la validité/véracité au sens médical du terme (par exemple, ce qui est déclaré comme un infarctus du myocarde est-il vraiment un infarctus du myocarde ; quels critères ont été utilisés pour ce diagnostic ?). Les recommandations existantes ne mentionnent en effet pas (ou peu) ces aspects, en particulier la validité médicale de la donnée source.…”