development of knowledge to guide change in practice patterns. The practice-based research networks developed in some medical specialties (and described in Box 4.3) do have this goal. They provide the opportunity for those who must implement the research to be represented in setting the agenda and to participate in the research. The partnership between Arapahoe House and their university research partners demonstrates that intimate collaborations are feasible, as do the collaboration models described in the next chapter. However, failure to develop such relationships is not surprising given the lack of research institutions in many communities and the commitment and investment required on both sides to make such a partnership work.

University-based treatment researchers are obviously familiar with treatment programs, and they are generally engaged in treatment. But many in CBOs feel that these researchers are often not in touch with the realities of delivering services "on the ground." Some workshop participants suggested that the researchers may ignore the "real clinical issues" when they are not relevant to their research interests as illustrated by the vignette that begins the next chapter.

The committee identified a number of variables that appear to interact to affect potential linkages between clinical programs and academically oriented researchers (including those working in nonacademic centers, government, and other applied research settings). These interacting variables—theoretical view of addiction; type of research; research functions and roles,