Applying SCRUM Methodology to Database Lock

Session chaired by Hemant Gawande

Description:

The industry standard timelines for Last Patient Last Visit(LPLV) to Database lock(DBL) are observed to be 6 to 8 weeks. This average time has been consistently same over past two decades. It is surprising to see less efforts towards improving this metrics even though it is most crucial one. Multiple Sponsors fighting to get a drug or therapy for same disease sometimes observe great results during conduct of their trials and who locks the database and submits the data first becomes pivotal in getting the patent. During pandemics, the equivalence of pulling this time by even one week was equivalent to a million lifes. Can bringing agility to the overall process of getting 'Lock Ready' help reduce the timelines ? We need to approach the database locks in iterative fashion in the form of splitting milestones over sprints starting 90 days before Last Patient Last Visit. This could definitely help brining visibility to what is missing early on. Applying most of the other SCRUM principles while dividing period from [LPLV - 90 days] to DBL over several sprints could bring a lot of efficiency. Can our Study Data Managers or Lead Data Managers or Clinical Data Scientists be upskilled to be SCRUM Masters for Database Locks ? With CDS initiatives being ON across organisations, database lock aspects needs be to given special attention. Coordinating with cross functional stake holders in the crucial phases of DBL becomes even more critical. With increasing sources and formats of data, managing increasing number of external vendors and timely receiving clean data becomes key to success. For the process to become agile and people to become more efficient, technology will have a bigger role to play in shortening the distance between LPLV and DBL. With real-time cross functional clean patient trackers and other visualization dashboards our CDSs can be enriched to become SCRUM masters in project DBL. Better ways of communicating with external vendors for the discrepancy management are still waited, we are still reaching them out via emails. With APIs making applications belonging to different system vendors interact with each other data managers would soon be able to send queries to external vendors right from their workbenches. With process, people and technological improvements, we can aim to be "Lock Ready" on as we approach LPLV and lock the database in days after LPLV instead of weeks.