Bench to Bedside Discussion

1
Bench to Bedside -- Discussion

• Jim Herbsleb
• CMU
• jdh_at_cs.cmu.edu

2
Outline

• Kudos
• Pick a few nits
• Concluding thoughts

3
Kudos

• Selection of AIDS research as subject of
  collaboratory
• Inherently multi-disciplinary (virologists,
  pathologists, immunologists, etc.)
• Bridging communities (bench to bedside)
• Speed is important
• Systematic preparation
• Collaboration readiness
• Tool selection based on interviews -- address
  problems that are real to the participants
• Training, infrastructure
• Likely sped up adoption
• Collection of both usage and outcome data

4
Benefits of Collaboratory

• Clearly facilitated some existing collaborations
• Apparently led to increase in cross-site
  publications
• Sped up some tasks (e.g., developing protocol)
• Mentoring junior faculty

5
Some Nits

• Non-human primate researchers??
• I found the presentation of results rather
  confusing
• Did increased number of collaborations come from
  recognizing need/opportunity more than
  availability of tools?
• Some clear summary of before/after would be
  really helpful, e.g.,
• contingency table of sites by sites, counts of
  collaborations before, after
• Statistical modeling to see if introduction of
  collaboratory introduces more cross-site
  collaborations
• (may need to add in conference, workshop papers,
  etc, to get numbers up)
• Relationship of collaboratory use to new
  cross-site collaborations
• Comparison of cross-site collaboration with CFARs
  without Collaboratory?

6
Couple More Nits

• Effect on clinicians?
• How did multi-site collaboration on GLR CFAR
  happen without tools?
• What were the problems, how were they solved or
  worked around?
• How were later collaborations different?
• Collaboration readiness -- what did the reported
  results actually mean?
• There were 4 same-site, 4 cross-site, and many
  anticipated cross-site collaborations
• If competitiveness/complementarity is so
  important, why doesnt collaboration readiness
  focus on this?

7
Issues

• Tool selection -- give participants exactly what
  they want? (fax, desktop video, coercive
  counseling for colleagues)
• Do they know what will be useful? E.g., did they
  ask for web site with the specific functionality,
  e-mail distribution lists? Seem to have been
  used a lot.
• Will they actually use what would actually be
  useful? (e.g., perception of hallway
  conversation, mixed record of chat, IM, MUD use)

8
Hypothesis -- distance matters less over time?
(Early in relationship, early in work)

• Group behavior changes greatly over time (e.g.,
  Time Matters, Joe McGrath)
• Nearly absolute barrier to initiating something
• Early interactions create structure (e.g.,
  identification of mutual interests, plans,
  awareness of style, expertise, etc.) that form
  framework for later communication
• After FTF, other communication works better over
  distance
• Software development -- early activities are the
  most collaborative
• After some critical mass of development, open
  source becomes possible
• Need collaboration tools to support some sort of
  life cycle?
• How, in detail, did new collaborations progress
  in CFAR?

9
Understanding Dependencies in Work

• Van de Ven, organizational assessments
• Org theory -- information processing, certainty,
  stability, etc., influences how much
  communication is needed, how well other
  mechanisms (e.g., bureaucracy) work
• Easy to identify hard dependencies, e.g.,
  producer-consumer, shared resources
• But how do you systematically identify and
  support soft dependencies?
• E.g., A and B should really keep in touch on this
  issue
• Would be potentially useful part of collaboration
  readiness