NLP Internal Workshop

Present: Kohane, Szolovits, Churchill, Zeng, Uzuner, Solti

Discussed workflow improvements in NLP required over the next two years.

Discussed specific longer term projects including those that fit best in the upcoming NCBC recompetition.

i2b2 Annual Retreat

Attending: i2b2 External Advisory Committee including Peter Tarczy-Hornoch, Dan Masys, and Piet de Groen.

Crimson and i2b2 cell integration

Bry, Kohane, Murphy, Gainer, Churchill et al.

Argued for the refactoring of the current Crimson xml payload so that it will become fully compliant with the i2b2 Workbench.

Rheumatoid Arthritis

Plenge et al.

Discussed the difference between the training and the test sets. This was again visible on the principal component analysis and Kat went into greater detail about the differences.

Major Depression

Iosifescu, Smoller, et al.

Discussed the filtering of the patients from the MDD data mart to select more patients with MRI's to study the effect of SSRI's and SSRI resistance (in combination with a SNP study).

Jordan reviewed what it takes to get the right specificity and sensitivity and chewed over the issue of how in diseases of prevalence less than 5%, the danger to a GWAS may be more at the level of lack of specificity than lack of sensitivity (if as is the case in i2b2 we have a large enough pool of cases).

Review of national meetings and year 6

Vivian, Zak, Shawn, Diane, Griffin

Reviewed the results of an IOM meeting this week (see photo) in DC where i2b2 was extensively discussed. Griffin gave an update on a very productive and "brass tacks" meeting at Scottsdale, AZ. Distributed querying using i2b2 is clearly becoming a priority for others as well for ourselves.