By Kayte Spector-Bagdady JD, MBioethics
Department of Obstetrics & Gynecology; Research Ethics Service, Center for Bioethics & Social Sciences in Medicine, University of Michigan Medical School, Ann Arbor, MI
With recent reports of Google’s data deals with Ascension health and the University of Chicago, there has been a lot of attention paid recently to the sharing and use of health data by unexpected entities.
But we know that patients are uncomfortable when hospitals “commercialize” or sell their health data or biospecimens to industry. In fact, the recent revisions to the human subjects research regulations included a specific biospecimen commercialization disclosure requirement.
Yet hospital executives report being “flooded with requests” by industry to sell or share health data. This is because – for health data research – size matters. Promising scientific advancements like precision medicine and other personalized approaches to healthcare require access to genetic, health, lifestyle, and outcomes data from as many people as possible.
But are academics also using health data generated by industry for their own research as well? And does it matter?
For our article published in Genetics in Medicine, the official journal of the American College of Medical Genetics and Genomics, “Genetic data partnerships: Academic publications with privately owned or generated genetic data,” we explored how often academic researchers are using private genetic data in publications and whether this usage is increasing over time. To do so, we assessed PubMed publications that utilized privately owned or generated human genetic data from 2011 to 2017.
We had four important findings:
- The number of publications utilizing private genetic data continually increased from 4 in 2011 to 57 in 2017 for an overall total of 181 publications. The majority of these publications had at least one academic collaborator (who was most often listed as first or last author or both). (See figure)
- Almost all papers with an academic author performed secondary analysis on data already existing in private databanks (95%) as opposed to collecting the data themselves and using the company as an analysis platform.
- 45% of the articles disclosed at least some National Institutes of Health (NIH) support for this research with private data.
- It was challenging to discern from many of the published articles what type of informed consent was obtained from contributors (43%).
Because of the critical need for health data for life-saving research, the federal government has also made the diversity of and accessibility to these types of databanks a priority. It is currently building the All of Us research data and biobank to fill this need. All of Us currently has ~247,000 participants fully enrolled.
By contrast, 23andMe, the leading direct-to-consumer genetic testing company that we queried in our research, has over 10 million participants (about 80% of whom consent to research). Via its Research Innovation Collaborations Program, 23andMe specifically solicits researchers to propose work with its databank (it does not currently charge for this kind of access to the data).
Privately held genetic databanks can enable academic researchers to accomplish important work they had not been able to before. But the NIH has also put an enormous amount of time and effort into updating and encouraging data sharing and accessibility for federally funded research. It is unclear exactly how this is being managed in individual academic/industry data use agreements across the country.
We look forward to continuing this work assessing the academic use of private genetic datasets and refining soft governance mechanisms to enable best practices moving forward.
Figure: Private Genetic Data Use: Total Publications with Academic vs. Non-Academic Collaborators, by Year, 2011-2017