Review by Richard Lehman – 12/18/2016 – 10:49am

Data Request Number: 2016-1176 Is the scientific purpose of the research proposal clearly described?: No If no, please explain:

It is stated that the previous ThioMon 2 algorithms have been patented by the Regents of the University of Michigan. The implication is that any correlations found by using early lab tests to predict later lab tests may be used for commercial rather than scientific purposes.

Will request create or materially enhance generalizable scientific and/or medical knowledge to inform science and public health?: Unsure, further clarification from requestor is needed Can the proposed research be reasonably addressed using the requested data?: Unsure, further clarification from requestor is needed Please use this space to provide any comments, feedback or concerns (optional):

I am concerned that the proposed machine learning study is not primarily addressed at a generalizable scientific question. The combined power of these studies may be inadequate to support the use of a prediction algorithm, though I guess finding that out is one purpose of their enquiry.

Recommendation for this data request: Not Approve

See above. Will the “best fit” model, if they develop one, be open to all? I think we need clarification.