The Future of Data Sharing Ronald Krall, MD Adjunct Associate Professor of Neurology and Member, Center for Bioethics and Health Law University of Pittsburgh krallrl@gmail.com
Accomplishments Leadership Increased awareness of risks of data sharing Re-identification Misuse Risk to competitiveness Risk of harm to public health Awareness of the complexity of data sharing Shared your own data Enabled research Brought credit to your company Gained credibility for your research Created an infrastructure for sharing October 2, 2014 Ronald L. Krall 2
Why have you shared data? Others expected it of us GSK (or Medtronic) did it Don t want to be seen as luddites Want to change how our company is seen Want to be seen as doing credible research October 2, 2014 Ronald L. Krall 3
Any scientific benefit from sharing your data? Validation of our results Encouraging secondary research Validation of results Generation of new hypotheses New results Development of new analytical techniques October 2, 2014 Ronald L. Krall 4
What are we trying to do when we share data? Validation of results Examination of subset results Assessment of a specific outcome of interest (eg, cardiovascular) often adverse outcomes Performance of specific comparator treatments across trials October 2, 2014 Ronald L. Krall 5
Depression Is there a more sensitive indicator of efficacy? Has there been a change in suicide ideation in trials over time? Lots of depression trials have high drop out rates; how do the various statistical methods of dealing with dropouts perform? Is there a subgroup of patients who respond early to antidepressants? Is there a group of patients in whom adverse effects compromise efficacy? Is the use of placebos really necessary? October 2, 2014 Ronald L. Krall 6
How would the current data sharing environment need to change to allow these kinds of questions to be answered? Have to be able to find trials, independent of sponsor, that could contribute to the answer Have to be able to access ALL the relevant data Have to be able to bring trials from all sponsors together into an analytical environment Agreement on standards for governance of data sharing October 2, 2014 Ronald L. Krall 7
Reasons Academia hasn t begun to share data Uncertain value No pressure No requirement Fear/Risk Loss of control over results Being scooped Being discredited Work required Preparation of data for sharing Prompt analysis and reporting of all research questions Construction of necessary libraries Cost Lack of a platform that seems friendly to non-industry data October 2, 2014 Ronald L. Krall 8
What needs to happen to bring rest of clinical research community to share data? An environment that better supports the broader and potentially more valuable secondary research questions Finding the important trials Bringing data from trials of all sponsors into common analytical environment Reduce the cost Building data standards into research Building sharing into the routine conduct of research Making standard what is currently unique Data sharing agreements Review processes Access methods Volume Acceptance that not all research requires a closed analytical environment Create a library for storing clinical trial data Take the infrastructure burden off large pharma s hands Mandate to share data October 2, 2014 Ronald L. Krall 9
An Alternative Future
https://osf.io/ October 2, 2014 Ronald L. Krall 11
October 2, 2014 Ronald L. Krall 12
October 2, 2014 Ronald L. Krall 13
October 2, 2014 Ronald L. Krall 14
What are the values we want to create? Trust Confidence Truth Knowledge Efficiency October 2, 2014 Ronald L. Krall 15
A Future of Data Sharing Pursuit of a broad range of valuable secondary research questions All trials, all segments of the clinical research community share data Minimal additional cost Data sharing is built in to clinical research Preparation Conduct Reporting Analytical environments that provide Access to all trials Use of a variety of analytical tools Governance according to an agreed set of principles and practices Protection Patients Primary research Property Integrity of research October 2, 2014 Ronald L. Krall 16
What s in it for YOU? You transition to becoming pure data sharers Your responsibility for maintaining web portal review mechanisms and workflow analytical environments becomes consistent with the amount of data you share You participate in the governance of data sharing You participate in the setting of technical standards that enhance the use of clinical trial data for secondary research Your trial data gain equal standing to all other data October 2, 2014 Ronald L. Krall 17
The Future of Data Sharing Ronald Krall, MD Adjunct Associate Professor of Neurology and Member, Center for Bioethics and Health Law University of Pittsburgh krallrl@gmail.com