Data Management Considerations for Adaptive Trial Designs
With the added weight of an FDA guidance document and the rapid advancement of the technology available for use, adaptive clinical trial designs are becoming much more appealing. In a nutshell, an adaptive trial involves the ongoing analysis and monitoring of unblinded data by a selected group of independent clinicians and statisticians at defined points during the trial, with the potential for study design changes. Utilising this approach, answers to questions not typically realised before the end of a study can be obtained. For instance; whether certain features of the trial need to be changed, whether the current enrolment target is sufficient to meet the defined statistical endpoints or, possibly, which dose amongst multiple treatment arms will be the most beneficial to the patient. The study team will, therefore, be in a position to propose and take counteractive steps much sooner. The decisions made based on these interim analyses can save on the potential failure of a study due to, for example, incorrect dose choices or miscalculated sample sizing and ensure that the chances of a successful study are optimised.
Adaptive trial designs are invariably complex and, as a result, often riddled with logistical challenges. Firstly, in providing the data to be reviewed for these key decisions to be made and then to execute any modifications to the trial emanating from them. Each, as ever, having to be done within a restricted time.
Data management sits at the fulcrum of this process; ensuring that data are captured correctly and in a timely manner, plus cleaned and released to programming/stats teams to produce the necessary outputs. Therefore, from the outset, this group needs to be consulted and kept informed of all that is going on.
Due to the nature of adaptive trials, it is imperative that whatever data capture tool(s) is employed it is flexible enough to cope with potential mid-study changes and, if various technologies are to be used, thought should be given during the planning/start-up stage as to how the data from these different sources will be integrated and reconciled. It will be role of the data management group to co-ordinate this data integration and reconciliation effort throughout the course of the trial, but will also be able to place focus on particular sub-groups of patients and data, as directed by the study team, for the purposes of the interim analyses. To do this, data management will need to work closely with multiple clinical colleagues and external data providers.
The impact of any resulting modifications to the conduct of the study on the data capture tools will need to be assessed and, along with input from the study team, the data management team will initiate, build in and test any updates to the EDC system and review/test updates made by external data providers. All testing will need to be approved before any changes are implemented. The impact of any updates on existing data will also need to be considered. For instance, there may be a modification to an existing edit check, or addition of a new one, which could cause discrepancies in data already deemed clean. The data management team will also ensure that the development instance of the database is kept accessible throughout the duration of the study to conduct any testing required so that use of the live database is not affected by this process.
An extra measure that can be taken to further reduce the time taken to implement database changes is for the data management team to create different versions of eCRF pages that are likely to require modification as a result of the data review meetings within the development database during the start-up stage. Each version will represent a potential scenario and can be utilised as and when required.
It is likely there will be an increase in the number of trials embracing this approach as the advantages to the sponsor and the benefits to patient are clear. From a data management perspective careful planning, tracking and effective communication with all stakeholders are key to ensuring the associated challenges are met head on.