EVALUATION OF A LINKED SERVICE MODEL OF CARE FOR HIV-POSITIVE, HOMELESS, AND AT-RISK YOUTHS

Two instruments were used to evaluate an agency's type and availabilit y of services for HIV-positive and at-risk adolescents, and to assess opinions concerning healthcare referral patterns. These instruments we re administered to representatives of 22 agencies from 10 categories o f healthcare services. Nonmetric multidimensional scaling was used to model ratings of interagency knowledge, referral patterns, and general satisfaction with services. We found that no agencies offered youth s ervices for inpatient adolescent-specific mental health treatment or s hort-term residential drug treatment; however, few offered long-term r esidential substance abuse detoxification services (5%), outpatient dr ug maintenance (5%), HIV-specific inpatient services (9%), intensive d ay treatment for substance abusers (9%), HIV home care (14%), HIV hosp ice care (14%), inpatient medical services (14%), short-term shelters (14%), long-term housing (18%), HIV-specific clinical trials (18%), an d dental services (23%). Barriers to expanding care included lack of f unding, transportation, and lack of awareness among youths about servi ces. A multidimensional scaling analysis identified a tight service cl uster of two community health centers and the largest public hospital serving poor communities of color, as well as a relatively tight clust er of three service agencies located on the Boston Common serving home less youths. A third service cluster consisted of two university-affil iated medical centers and one community health center. In conclusion, we found that many critical services for HIV-positive youths are relat ively scarce. Multidimensional scaling provides a visual presentation of the relationships of network sites. This evaluation of services ind icates a need for increased, accessible youth-oriented HIV services an d suggests that linkages across the three distinct clusters of service providers should be solidified. These methodologies can be used to de velop a generic model describing the stages of linkage formation in HI V care service networks.