The IHHP notice included several data collection and reporting requirements for the grantees. In addition to providing program-level HOPWA Annual Performance Reports each year, grantees are required to use HUD’s CoC HMIS or an “equivalent client-level information system” to support the planning and reporting activities of their projects. The notice says that “participation in client-level information systems will support community collaboration in providing services across federal HIV/AIDS programs on outcomes related to housing stability for PLWHA and will foster further federal agency collaborations.”
Housing and HIV care data integration can improve the performance of integrated HIV housing programs in multiple ways. At the service level, housing coordinators and HIV medical case managers can share client-specific information by “seeing” key data elements and case notes in the others’ systems. Having access to integrated client-level housing status, housing service use, and housing and health outcome data can help managers evaluate the effectiveness of their programs. Linking client counts across programs also enables more accurate planning at the community level. For these reasons, all four IHHP sites were eager to improve the integration of their local housing (HMIS) and HIV care (CAREWare) systems.
No federal client-level data systems currently include a comprehensive set of housing and health care variables that can monitor the performance and evaluate the housing and health outcomes of integrated housing services. Neither HUD’s HMIS nor HRSA’s CAREWare systems have the comprehensive set of housing and health care data needed to evaluate the effectiveness of the integrated HIV housing programs operated by the IHHP grantees. While the HMIS data system can track changes in housing status, it does not collect key client-level measures of health care access, use, and changes in clinical outcomes. Although CAREWare collects and reports client-level data on HIV status, health access, utilization and clinical outcomes, the data system does not collect baseline data on clients’ housing status and stability at intake, and thus cannot measure client-level change in housing status. As discussed in the quantitative analysis in Chapter II, it was not possible to combine the data across these two data systems to conduct these analyses.32
Because HMIS and CAREWare were developed separately, they are incompatible. In order to build a comprehensive data system appropriate for their programs, the IHHP grantees must select one system or the other (HMIS or CAREWare) as their base, and then customize that system to add their APR data and other missing data elements (Figure III.2). Three IHHP sites (RRHS, PHB, and CARES) are using locally adapted HMIS data systems to collect and report their IHHP data. The fourth site (FPC) is using their state’s CAREWare system for their IHHP program. Among the four sites, we found no direct linkages between local or state HMIS and CAREWare data systems (indicated by the dotted line in Figure III.2). The IHHP sites’ aggregated APRs were also insufficient to track the client-level data needed to manage their integrated programs. The sites needed to incorporate their aggregate HOPWA data into either of the two larger client-level data systems to manage their programs better. All four IHHP sites in this study are using their grants to improve the integration of the housing, HIV care, employment, and other program-specific data elements in their data systems. However, they are facing different challenges in their efforts (Table III.4). For example, while the RRHS program currently tracks client-level FUSE program data in their CoC HMIS database, the local RWP Part A grantee utilizes CAREWare. The two systems are currently separate, but RRHS is spearheading an effort to link HMIS and CAREWare databases to coordinate their services and minimize duplicate data entry. The linking of these systems will open lines of communication between the RWP case managers and the FUSE service coordinators, allowing each to view the others’ case notes and clients’ lab test results (for example, CD4 counts). This will not involve making changes to CAREWare but rather will create external bridges from HMIS to access CAREWare data. Although the site is working to link the two data systems, the Part A grantee is concerned about making functional changes to CAREWare to foster data interoperability.
Figure III.2. HMIS, CAREWare, and HOPWA APR Data Relationships
FPC case managers collect client-level data from HOPWA TBRA, PHP, and STRMU housing applications and housing plans, and measure clients’ housing stability at intake, at annual assessments, and during tenancy. They also collect qualitative information on clients’ sobriety and other characteristics to track the impact of housing stability on sobriety and other outcomes. Although FPC staff use the state’s CAREWare data system to input their HOPWA data, including their case notes, and report that CAREWare, at least in their state, limits their ability to share client-level data with other agencies. The state data system is set up so that FPC and partner agencies can see only program-level data, limiting the data’s utility. Although some demographic fields are shared, many HOPWA-specific health and housing benchmarks are available only at the service provider level. Both FPC and the state’s RWP Part B grant hope to modify their data systems in the near future, but the timing of these information technology (IT) projects may limit the opportunity for system integration. The state anticipates that its IT development process will take longer to complete than FPC’s.
As the lead agency for Multnomah County CoC, PHB manages the CoC’s HMIS data system and uses the system CoC, HOPWA, and RWP reporting. Using HMIS for collecting and reporting non-CoC programs (HOPWA and RWP) has been a challenge. The primary barrier for using HMIS systems for HOPWA data collection is the incompatibility between HOPWA data requirements and HUD’s formal HMIS data standards. Even though HOPWA and CoC programs share common data elements, there are incompatibilities, including differing eligible activity types and performance outcomes. Further, HOPWA reporting does not track data at the client level, and HMIS does not have many data elements tracking changes in health status.
There is also minimal client-level data sharing between PHB and its partners. PHB requires HMIS for reporting HOPWA services; MCHD clinics collect data through electronic health records but use the TOURS system for reporting (soon to be replaced by CAREWare); and the state’s RWP Part B grantee uses CAREWare. Although some RWP funds have been allocated to improve HMIS/CAREWare communicability, the federal sequestration and related state and local budget cuts have hampered this effort. Ultimately, the RWP grantee hopes to develop a data warehouse that would allow multiple users across systems to view common, real-time data.
Table III.4. Integration of Data Systems
|.||RRHS FUSE||FPC Maine IHHP||PBH S4H||Albany CARES FFL|
|Data Systems Used||RRHS collects FUSE client-level data in the CoC HMIS database. RWP Part A utilizes CAREWare||CAREWare is used to input HOPWA program data, including case notesFigure III.2. HMIS, CAREWare, and HOPWA APR Data Relationships||PHB utilizes the CoC HMIS data system. MCHD clinics use electronic health records but report through TOURS system and will soon move to CAREWare. Oregon uses CAREWare||CARES utilizes HMIS, as implemented by Foothold Technology (FT), which includes a customization module called Form Builder. Both ADC and CCCS use the same FT HMIS system|
|Data Sharing and Integration||Currently, CAREWare and HMIS are separate data systems. There is no data bridge between them. The FUSE program coordinator uses an Excel spreadsheet to record clients’ lab results, which are reported in the IHHP grant’s APR||HOPWA data are uploaded to CAREWare, but data sharing is limited because it is set up only for provider-level viewing. FPC also shares HOPWA data with some medical providers||MCHD and CAP had used a common health care and housing assessment tool in the past but now have only a few data elements in common. Only medical engagement data are shared by MCHD and CAP||CARES does not share its data system with other organizations. Currently, an external evaluation tracks health outcomes|
|Outcomes Measured||HMIS tracks housing placement and stability; access to case management and medical care; access to and maintenance of income; housing planning; and CD4 counts and viral loads||FPC tracks housing type at program entry and exit; homelessness history; housing stability based on HOPWA codes; medical care source information; medical insurance coverage; client income; employment; and some health outcomes||PBH tracks housing placement and stability; access to case management and medical care; access to and maintenance of income; housing planning; employment; and engagement with medical case manager||CARES tracks housing placement and stability; access to case management and medical care; access to and maintenance of income; housing planning; and health, wellness, and employment data. The FFL evaluation also tracks health care status, mental health status, substance use, prevention/risk behaviors, and employment status|
|Challenges||The CoC’s old HMIS system went off-line in December 2012, but the new HMIS system did not become operational at RRHS until May 2013. Over 200 paper notes created in the interim must be entered into HMIS||HOPWA-specific data elements in CAREWare are not maintained when the CAREWare system is upgraded on an annual basis. HOPWA-specific health and housing benchmarks are available only at the provider level||HOPWA data elements are incompatible with HMIS data standards, including differences in eligible activity types and performance outcomes. HOPWA reporting does not track client-level data; HMIS does not track key health status data||CARES opted not to provide evaluation data for FFL’s 2012 APR because of data quality issues, including inconsistencies in defining and recording some data elements due to FFL staff turnover|
|Future Planning||RRHS is spearheading an effort to link the new HMIS and CAREWare databases. This will coordinate services and minimize duplicate data entry. The RWP Part A grantee is not seeking full integration of the two systems||FPC seeks a new data system combining housing, case management, and HIV care data. Maine’s RWP Part B grantee also wants to replace CAREWare with a new system to meet state-specific needs. In the new system, there is potential for integrating housing and health care data||MCHD is transitioning to CAREWare from Tours. PHB and CAP are working together to export HMIS data into the Tours system and, eventually, into CAREWare||CARES is developing a common CoC intake and assessment model for housing, mental health, and other services. A new, integrated data system will be needed to manage this common assessment process|
As the CoC HMIS administrator in 21 counties, CARES has the staff capacity to modify its HMIS data system. CARES staff use an HMIS module called Form Builder to collect additional information on health, wellness, and employment at six-month intervals. An outside evaluation of the FFL program is also using HMIS to measure client progress. At an operational level, employment counselors at ADC and CCCS (the FFL program’s partner agencies) use the data system to track clients’ employment, income documentation, and compliance with their medical appointments. The staff record diagnostic information, such as CD4 counts, in their HMIS progress notes. CoCs will now be required to create and monitor a common client intake and assessment system, which will eventually require a new data system.
32 Though it is theoretically possible to conduct a correlation analysis for a subset of HOPWA and RWP clients enrolled in Medicaid. Such an analysis would exclude the impact of integrated housing services on uninsured clients—the bulk of HOPWA and RWP recipients.