in November 1997 and concluded there are promising results in some areas (e.g., dental and postoperative pain and chemotherapy nausea and vomiting); in other situations (e.g., addiction, stroke rehabilitation, and asthma), acupuncture may be useful as an adjunct treatment or included in a comprehensive management program. They cited the emergence of plausible mechanisms for the therapeutic effects of acupuncture as encouraging and concluded that "there is sufficient evidence of its potential value to conventional medicine to encourage further studies. There is [also] sufficient evidence of acupuncture's value to expand its use into correctional medicine and encourage further studies of its physiology and clinical value" (NIH, 1997).
Studies of patient factors, treatment factors, and community factors in treatment outcome research are all needed, as are studies of the effect of payment level and political environment on treatment outcome. Treatment provider professionals have a variety of questions that could be addressed in research but are not receiving sufficient attention. Patient factors have been much more widely studied than have treatment setting or modality, perhaps because there are few measures of treatment setting or treatment services. Treatment providers speaking to the committee recommended directing research attention to such challenging problems as community resistance to the placement of drug treatment facilities, the so-called "NIMBY" (not in my back yard) problem. This is a problem that requires measurement of neighborhood and organizational systems, as well as individuals. Several workshop participants commented on the role that the different perspectives of researchers and treatment programs played in determining what research was done. One participant reported that in her state the treatment and research communities held differing views of addiction, one favoring the disease model and the other a behavioral model, which presented a barrier to research collaboration. There are substantially different views about the desired "outcome" of an addiction treatment. For example, studies using an outcome of "percentage improvement'' in needle use will have little credibility with a clinician who believes that abstinence is necessary for recovery to occur.
Many clinical trials exclude the classes of patients that are most prevalent in community-based agencies, and consequently findings from such research do not seem relevant when viewed by treatment providers. To illustrate, studies of treatment techniques for cocaine abusers commonly screen out potential participants who are also abusing alcohol. This not only limits the generalizability of the research, it also reduces the study's credibility to the provider community, because cocaine abusers normally present for treatment with alcohol abuse and a variety of other problems that would have led to their exclusion from much research. Another difficult but important population needing study is the large and seemingly