At what price — dynamism in EDC
Geoff Low · Pharmaceutical Programming · 2010
In the beginning, there was the case report form (CRF), printed in NCR, bound and sent to site. It was obvious what data the sites needed to complete. Today, Electronic Data Capture (EDC) systems can make this less transparent as some data fields will remain hidden until a key question is completed; this functionality can be called 'dynamic questions'. The dynamic question approach reduces the amount of validation required as, if the dependent field is missing, there is no chance of any unexpected data being recorded. However, should the key variable be queried and the site staff change the result, where do the data from the dynamic pages go? To further complicate life, the idea of dynamic pages or dynamism no longer stops at the CRF. We see more and more complex trial designs, where a given subject can traverse the protocol defined assessments following a number of paths. Not only can a single data point change the appearance of a single CRF, but even direct the subject's progress through the whole trial. This can be conceptually simple to specify but technically complex to build and validate. Practical consideration of dynamism also raises other challenges. First, the database programmer will have to add other layers of complexity during database design. Second, the EDC systems used must be flexible enough to allow the users to correct their key data decisions within the database. Third, statisticians analysing the study data will have to cater for perfectly correct 'holes' in tabular data when analysing dynamically recorded data. This paper will focus on the issues from the perspective of a database programmer relating to dynamism that is available in leading edge EDC systems of today and comment on some of the processes surrounding those systems.