practice methods within a particular disease entity (Goldberg et al., 1994). The PORTs, studying practice in such areas as stroke, acute myocardial infarction, low-back pain, and knee replacement did some extraordinary work studying care in the various areas, and an extensive literature is emerging. The output from the PORTs was to fuel AHCPR's guideline development process.

The guideline development process did emerge and AHCPR became the official government agency creating guidelines in many important areas of clinical practice. But there were a variety of problems with the federal approach to guideline creation, including the evidence cited above that governmentally created guidelines was not the most effective way to influence clinical practice. Moreover, there was significant political fall-out from this process, including a move by one group of medical specialists to abolish AHCPR as a result of their unhappiness with the contents of a guideline. Cooler heads prevailed and the threat to the agency dissipated.

With experience came a rethinking of the guidelines/PORT model and AHCPR created a new model. The current thinking follows from the approaches discussed above, that guidelines are best created by sponsors closer to the actual clinical care, including managed care programs, medical specialty groups and the like. But the major impediment to guidelines creating is still the paucity of evidence reviews in many clinical areas. So AHCPR has now named twelve Evidence-Based Practice Centers to produce the evidence-based reviews intended to facilitate improvement in clinical practice. For the current status of this effort, see the AHCPR web site (http://www.ahcpr.gov). Further, AHCPR has created a national nomination process for assisting in determining priorities for the particular evidence-based reviews to be created. It is likely that a similar process would be extremely helpful in the area of substance abuse treatment.

Numerous impediments make it difficult for counselors, program managers, and state agency staff to sift through the research literature, critique it effectively, and select findings to implement in treatment. Techniques such as the consensus conference mechanism and the evidence-based reviews approach might begin to close the gap and to improve treatment, as well as to enhance the potential for broader use of treatment guidelines in drug abuse treatment. The first NIH Consensus Development Statement on drug abuse treatment is included as Appendix F and may also be found on the NIH Consensus Development Statement web site (http://consensus.nih.gov).

Workshop participants described a variety of "top-down" models, most of which could be fairly described as "money with strings" that would