This post is the second in a series on research data management presented by the Records Management Roundtable.
Submitted on behalf of Anita Vannucci, Records Manager at Emory University.
In the fall of 2015, I began a review and update to the research data portion of my institution’s retention schedule. The catalyst for this work came from a campus task force looking at the evolution of research data at academic institutions and anticipating increased interest following the February 2013 White House Office of Science and Technology Policy memo. One recommendation the task force made was to revise the retention schedule.
The 30 series retention schedule was last updated in 2008, long before my arrival. Several challenges became clear when I started my review. Some series were based on federal statutes that had been repealed, or cited code that had no clear connection to the series. Some “series” were actually record types. For clinical records, a records series existed for each stage of research (Phase I, Phase II, Phase III, Phase IV) rather than a single series based on the point the research closed.
I began researching federal and state code, best practices, and the policies of peer institutions. Federal guidance varied from agency to agency, so many of the institutions I surveyed relied on contract language, rather than a centralized retention schedule, to set requirements.
Another challenge was identifying a focus group of subject matter experts. In fiscal year 2015, my institution received 2,923 sponsored awards. While I spoke to people doing research and managing the information created from it, it was impossible to get a representative sample, so I chose to talk to their central monitoring department, the Office of Compliance. It was Compliance who helped me solidify the information I’d gathered into concrete series based on four types of research: behavioral, treatment, FDA-regulated, and non-FDA regulated. Unfortunately, the retention requirements aren’t as straight forward as that finite group of series would lead you to believe.
I began meeting monthly with the deputy chief compliance officer, working through my draft of the schedule. He provided feedback and insight along the way, while I learned about everything from the differences between dangerous drugs and controlled substances to retention requirements around sealed versus non-sealed sponsored study contracts. In the process, we tackled two major challenges:
- Can we create big buckets? Risk analysis is a big part of creating a big bucket schedule. We discussed when keeping information longer than required through a simplified schedule posed a greater risk than the value gained by making the schedule easier to use. For example, alleged misconduct investigations involving federally-funded research must be retained for seven years. There’s more flexibility around privately funded research, so in instances where a complaint is determined to be unfounded, there was a desire to dispose of the information sooner. By keeping misconduct investigations of privately-funded research according to the federally-funded requirements, we’d lose the ability to destroy unfounded complaints and retain spurious or potentially libelous information longer than necessary. We decided to settle this issue on a series-by-series basis after considering the volume of records involved and potential risk for each series.
- How simple can we make the retention requirement? Retention periods set out in code are rarely simple. (Take a look at 21 CFR 58.195 if you don’t believe me.) We discussed the value of using those retention periods word-for-word versus simplifying and risking the loss of nuances. In this instance, making the schedule user friendly won out. We decided to simplify the language but also link to each citation.
The schedule isn’t quite done yet, but it will be ready to go in time for the launch of a new fiscal year and a new round of awards.