The quantitative study analyzed HUD and HRSA administrative data from HOPWA and RWP. In addition, the team obtained publicly available HIV surveillance data from the CDC and homeless services needs assessment and service data from the HUD Office of Special Needs Assistance Programs. The study team finalized a statistical analysis plan based on the study objectives and on the aggregate nature of the data available, and conducted descriptive analyses of each of the variables of interest and compiled state and national summaries. The team also compared HOPWA and RWP services (such as relative percentages of participants served by type of housing assistance, income level, or demographics). For HOPWA data, the average cost per person served by specific type of housing was also calculated. The analysis drew on information from multiple sources (See Appendix A for details):
- HOPWA Consolidated Annual Performance Evaluation Reports (CAPER) (2007-2010)
- HOPWA Annual Performance Reports (APR) (2007-2010)
- 2010 Annual Homeless Assessment Report (AHAR)
- Homeless Management Information System (HMIS) (2010)
- Ryan White Program Expenditure Reports (2007-2010)
- Ryan White Program Data Reports (RDR) (2007-2010)
- Ryan White Services Report (RSR) (2010)
The qualitative study selected program models drawn from HUD’s seven IHHP grantees in 2011. In May 2013, Mathematica and Cloudburst conducted in-person site visits to four of the seven sites: (1) the River Region Human Services (RRHS) FUSE project in Jacksonville, Florida; (2) the Frannie Peabody Center (FPC) statewide IHHP project; (3) the Portland (Oregon) Housing Bureau (PHB) S4H project; and (4) the Albany Corporation for AIDS Research, Education, and Services (CARES) Foundations for Living (FFL) project in upstate New York. The one-day in-person visits at each site included discussions with IHHP grantees, community partners, and RWP grantees and service providers. Interviews were conducted with IHHP grant administrators, local housing program directors, IHHP housing coordinators, RWP-funded medical case managers, local IHHP grant evaluators and data systems managers, and other local housing agencies involved in planning community-level housing. Appendix L contains the study’s site visit interview protocols.
The study collected information from the four IHHP grantees about three levels of activity: (1) the integration of program policies and practices at the client level; (2) efforts to integrate housing and health care data systems; and (3) community-level efforts to coordinate and align HIV housing, health care, and other support services. Site profiles were created using information gathered during the discussions and based on review of program documents, including grant applications, site reports, and APR data. The site profiles were systematically coded and analyzed to address the study’s research questions. The project also used a mixed-methods approach to integrate the quantitative and qualitative studies and findings. Early findings from the quantitative analysis were used to inform the selection of the four sites for the IHHP case studies and to inform site visit topics and questions. For example, in the site visits, we asked questions about the compatibility of the HOPWA and RWP data sets, based on challenges encountered in the quantitative analysis.