CCTS Director Dr. Robert Kimberly discussed next generation DM techniques with UAB Libraries Assistant Professor Dorothy Ogdon at the Dec. CCTS Forum.The importance of data management (DM) is easily forgotten in the rush to start a research project, as Dr. David Redden, CCTS Biostatistics, Epidemiology, and Research Design (BERD) CoDirector and UAB Biostatistics Vice Chair, knows all too well. At the Dec. CCTS Forum, he shared several stories about data disasters that drew gasps and groans from attendees.
“‘Where did that p-value come from?’ is a question that can make any investigator break into a cold sweat when you can’t find the data or don’t recall which version of a data set you used in your analysis. I know, because it’s happened to me. But with sufficient planning and effort, this is a question you will be able to easily answer about your studies 10 years from now,” he promised.
Redden and UAB Libraries Assistant Professor Dorothy Ogdon led a lively discussion on the many challenges inherent in managing one’s research data to address short and long-term goals, including the need to meet policy requirements for data reproducibility and security. “Don’t wait until you are generating data to start planning how you will manage it,” they urged. Below are several tips and techniques they shared to help investigators avoid data distress.
- At the project planning stage, determine what you aim to show, what you need to measure, and how you will measure it.
- What kind of data will your study generate? Different data types lend themselves to different types of analyses (e.g., EHR data for observational studies, compiled data for a systematic review, etc.)
- Build a data dictionary (aka a metadata file) that contains all variables, how they were measured, how you recorded missing data, and acceptable value ranges. Make sure others will understand what any rating scales mean.
- Consider how you will store, back up, and protect your data, especially if it contains patient data and other personal information protected under federal law.
- Understand your institution’s policies concerning shared data repositories.
- If working with collaborators, agree on a process for version control, including a file naming protocol.
- Worried you will forget an important step? Download a free research DM plan template from the Internet and/or consult with a DM expert at a weekly CCTS BERD Drop-in Clinic.
Redden noted these are just the basic principles and announced CCTS is launching a new hands-on workshop series to provide a “deeper dive” into DM techniques starting in Feb. 2018. In case you missed it, visit our CCTS Forum page for links to the Dec. 6th slide deck and video. For more tips and a list of UAB DM experts, see “Who Has the Data? Strategies for Successful Research Data Management.”
Subscribe to CCTS Digest to stay up-to-date on all of our learning and funding opportunities.