Minutes submitted by: Rich Ryan (University of Rochester) and Rosemary Johnson (OHS)

Jim Prochaska introduced the general issues, stressing that having some consensus on outcome measures is critical to scientific progress.

A "straw poll" identified a clear consensus that the BCC should attempt to identify common outcome criteria to facilitate systematic comparisons.

One view of outcome criteria is to identify a standard for "successful" outcome in each behavioral domain. For example, in smoking 6- or 12-month abstinence rates; or in diet interventions using a "5-a-day" behavioral criteria. There was discussion about trade-offs between continuous measures and dichotomous ones. However, it was also noted that most continuous measures are "translatable" into success/failure criterion standards.

The committee thus considered the possibility of evaluating behavioral outcomes across sites with three different approaches:

Criterion based: Was the success criterion met or not? We need to identify criteria for each behavior and assess which study sites are collecting measures for each criterion. This can be done by looking at the matrix of measures across behavior workgroups and sites. If sites aren't using the shared measure, can they convert their assessments to match the groups measure outcomes? We also should consider variables vs. measures.

Progress in treatment population: Can we look at movement towards goals. Is it clinically significant, even if criterion success points are not met?

Continuous measures.

Dan Berch pointed out that there are layered issues. One is identifying common outcome criteria. A second concerns identifying the methodological differences between studies; that is different groups might assess the same criteria using more or less sensitive, verified, or economical measures. Finally, effect sizes will of course vary with different populations, and are affected by the differing inclusion/exclusion criteria used in each BCC project.

Another point raised concerned projects with multiple behavior interventions. How does success in one behavior impact other behavior changes? When evaluating interventions, consider the intensity and dosage of intervention.

Having agreed in principle to move toward identifying viable common outcomes, the question of next step arose. It appeared that Trish Jordan and Jim Prochaska were willing to play an organizational role in gathering a work group that would include Ed Deci, Claudio Nigg , Bob Jeffery, Abby King, Karen Peterson, Rosemary Johnson, Marcia Ory, Candace Young,  Dan Berch, as well as others in the room and elsewhere that have surely been omitted.

A next step considered was having one or more experts on each behavior (i.e., smoking, exercise, nutrition) identify a potential criterion standard for that behavior, and circulate that standard to each PI., asking the questio: Can we translate our outcome measures into this criterion? This would address whether common outcome variables can be assessed by all groups, given the fact of differing measurement approaches (some of which are already "fixed").

 Action items include:

1. Identify criteria for each behavior.

2. Look at matrix of measures across sites/workgroups.

3. Schedule a workgroup conference call.