Clinical Research Oriented Workshop (CROW) Meeting: July 17, 2014

Present:  Marianne Burke, Juvena Hitt, Kairn Kelley, Rodger Kessler, Ben Littenberg, Connie van Eeghen

Start Up: Ben and Rodger have been invited to a faculty development, 2 day conference at UVM in August.  The focus is big data and there is an opportunity to provide a “flash talk” (5 minutes, 1 slide, and a hook).  Ben may do license data or DARTNet – or both, as an overview, to invite other collaborators to join him.  Rodger has a CCRN data set (similar to DARTNet but not yet merged into it) and/or the SAFETINet data over multiple states.  This is an opportunity to get ideas and meet people.

1.                  Discussion: Rodger’s VIP project

a.       Interest is high in this area.  The Profile could become the assessment tool for the Toolkit that support operationalizing these concepts.

b.      The Profile has 8 domains, each of which is evaluated as an average of the questions that were responded to.  In total, there are 35 questions (elements).

c.       Validation strategies:

                                                  i.      Econometric model: detailed vignettes are developed for standardized scores; the elements are calculated to determine their standards

                                                ii.      A priori declaration that each domain is equally important.  Or, split the four domains that were two domains from the original model.

                                              iii.      Arbitrary ratings for each domain, rating (or ranking) them against each other.

                                              iv.      Gold standard outcome measure, e.g. patient outcome, correlated with VIP scores (this is the  least feasible alternative).

                                                v.      SF36 model: two scores, created from all elements (not domains).  Consider doing a factor analysis. 

1.      Two factors proposed: clinical services and practice characteristics.  Need a couple hundred survey results to conduct the analysis.

d.      Next step: VIP version 1 will not create a single score, but will provide a weighted score at the domain level.  Future versions can move to a weighted domains.

e.       Plan:

                                                  i.      Initial validity testing:

1.      Currently in the field for responses around discriminate validity

2.      Two raters in each site

                                                ii.      Update to version 3 (newly updated 7/17/14)

                                              iii.      New scenarios that are rank ordered and associated with VIP ratings, created around the elements in the eight domains, to be assembled in correspondence with specific VIP outcomes

2.                  Next Workshop Meeting(s): Thursdays, 11:30 a.m. – 12:45 p.m., at Given Courtyard South Level 4.   Remember: the first 15 minutes are for checking in with each other.

a.       July 24: Marianne – research update

b.      July 31: Open

Recorder: Connie van Eeghen