2. Model fidelity refers to the degree to which a program is implemented in conformance with an evidence-based treatment model (such as, Assertive Community Treatment), i.e. has elements and levels of intensity that experts have found to be part of effective models and does not have elements that are not thought to be part of effective models.
3. The PROC MIXED procedure of the SAS® software system Version 6.12 (SAS Institute, Cary, NC) was used for this analysis. Random effects were modeled for site in models in which the dependent measures were considered at the agency level and with respect to both site and agency for those models in which the dependent measure was at the dyad level.
4. The decline in the number of respondents by 10 from wave one to wave three at five sites, a 16.7% decline overall, primarily reflects the integration of mental health and substance abuse programs (60%) but also weakening involvement of general health programs (30%), and a consolidation of VA facilities at one site (10%).
5. An alpha level of less than 0.15 was used to determine statistical significance for these analyses because of the relatively small number of observations for testing the significant interactions between site/agency and time.
6. The existence of a fiscal relationship between two agencies was indicated by either the transfer of funds between the two agencies or the ability of one agency to influence the flow of funds to the other agency from third parties.
"report.pdf" (pdf, 296.36Kb)