Local Bidirectional Data-Sharing Collaboration to End Veteran Homelessness: The Erie Model
*The views presented here are those of the authors and do not represent those of any state or federal agency or the US government.
Homelessness is a major issue in the United States. Among adults, military veterans are at increased risk (Donovan and Shinseki, 2013; Fargo et al., 2012; Tsai and Rosenheck, 2015). Studies continue to highlight the greater demand that homeless veterans place on emergency rooms and their increased risk for all-cause mortality, psychological disorders, substance use disorder, and other conditions, including chronic obstructive pulmonary disease, hepatitis, HIV, and traumatic brain injury (Balshem et al., 2011; Tsai et al., 2013; LePage et al., 2014). Obtaining stable housing and finding other supportive services to maintain housing are critical to achieving better health. Brown and colleagues (2015) found lower depressive symptoms in older adults who had obtained housing compared to those who had not.
In 2010 the US government pledged to end veteran homelessness by the end of 2015 (USICH, 2015). Progress toward this goal is assessed through the Annual Homeless Assessment Report (AHAR) submitted to Congress by the US Department of Housing and Urban Development (HUD). This report is also used to allocate resources to the Department of Veterans Affairs (VA) and local communities. The two main data sources used to count homeless individuals, including veterans, are the point-in-time (PIT) counts and the congressionally mandated Homeless Management Information System (HMIS; Balshem et al., 2011; Poulin et al., 2008).
The PIT counts, conducted by HUD’s Continuums of Care (CoC), occur in January of each year. (For more information on the CoC program: https://www.hudexchange.info/programs/coc/) These counts are used to establish the number of individuals who are homeless on a given night. Approximately 50,000 veterans were identified as homeless in 2014, and nearly one-tenth were women (HUD, 2014). In comparison to counts in 2010, veteran homelessness in 2014 had declined by 33 percent. Counts are also available in HMIS, which tracks services received from HUD-funded programs. Each CoC is required to maintain data on all homeless individuals, including veterans, through this data system. Continuums of Care are made up of HUD-funded homeless service providers in more than 400 communities across the United States. Specific to the VA is the Homeless Operations Management and Evaluation System (HOMES), which is a proprietary database to track the services provided to homeless veterans at the medical centers. Each veteran is assessed at the time of intake and at the discretion of the clinician.
Obtaining an accurate count of veteran homelessness is challenging, given the different data sources available to local communities and the VA. For example, PIT counts do not capture homeless individuals who are not available to be counted on the night of the count, and HMIS likely undercounts those receiving services from non-HUD-funded providers. In addition, HMIS does not independently verify whether an identified individual is a veteran or whether he or she is eligible for VA services. Anyone who self-identifies as a homeless veteran is considered to be a homeless veteran regardless of that individual’s VA eligibility. At the same time, HOMES does not independently verify the veteran’s chronic homelessness status, resulting in a veteran who might not qualify as being chronically homeless to HUD but is accepted as being chronically homeless to VA. Reliance on unverified self-reports of veteran status might also lead to inaccurate estimates of the local homeless veteran population (Metraux et al., 2014).
In 2015, the Veterans Health Administration’s (VHA) Office of Health Equity (OHE) initiated an agency-wide environmental scan to establish a baseline of activities that advance health equity. As a part of its mission, OHE aims to position the VHA as a national leader in achieving equity, to champion efforts to address health disparities, and to capitalize on the existing networks in order to advance health equity and achieve equitable health care for all veterans. Medical centers and program offices were asked to identify programs and other activities related to the five focal areas (leadership; awareness; health outcomes; diversity and cultural competency of the workforce; and data, research, and evaluation) of the VHA Health Equity Action Plan (Uchendu, 2014), which is the agency’s strategic framework for achieving health equity. The purpose of this discussion paper is to describe a data-sharing initiative included in the environmental scan that seeks to cure the aforementioned challenges and now serves as an emerging best practice in the VHA for consistently identifying homeless veterans among the VA and local communities.
Description of the Initiative
The Erie model data-sharing initiative to end homelessness among veterans is a bidirectional data-sharing collaboration between the local Veterans Affairs Medical Centers (VAMC) in Erie, Pennsylvania, and the PA-605 CoC. (PA-605 is the Continuum of Care (CoC) program identification for Erie City and County in Pennsylvania). The model benefits from the use of existing VA and HUD data-sharing policies without the need to duplicate data entry. The burden on veterans is reduced by not having to repeatedly provide the same information to the VA and community providers. In addition, the model facilitates enrollment in non-VA and VA services as needed. Veterans are asked during intake if they would like to share their HOMES assessment with the local CoC. The HOMES assessment contains similar data elements found in HMIS. The veterans then sign a release of information, providing their consent and understanding that they may also choose to receive community homeless resources in addition to those offered at the VA. Subsequently, the VA securely faxes the signed release of information and the HOMES assessment to the HMIS administrator, who manually enters the data into HMIS. Through this data-sharing process, the HMIS administrator also provides periodic reports to VA that include names of identified veterans.
The impetus of the data-sharing collaboration was spurred by a concern for missing homeless veterans in need of care. In the fall of 2013, the Erie homeless program supervisor and the HMIS administrator began ongoing monthly meetings to develop an understanding of the local homeless veteran population. These meetings revealed an incomplete picture of the needs of homeless veterans between the VA and the community. The CoC understood the needs of homeless veterans engaged in HUD-funded community services, while the VA understood the needs of homeless veterans presenting for homeless services at the VAMC. In addition, the supervisor of the homeless program was provided a list of homeless individuals who were identified as veterans and had provided consent to share data according to the CoC’s data-sharing policy in the HMIS database. These names were then compared to VA administrative databases and electronic health records, revealing a substantial proportion of individuals who were either not considered to be homeless according to the VA or ineligible for VA services. Together, the supervisor of the homeless program and the HMIS administrator identified an opportunity to share data that relied on VA’s existing release-of-information process and HMIS policies for data sharing. Meetings were then held with VA and CoC leadership to explain the steps of the proposed data-sharing process and to get agreement. These meetings resulted in the CoC bringing the VA into HMIS as a participating agency. Additional meetings were held with VA information security and privacy officers to obtain their approval.
In 2012, HMIS showed 147 veterans being served in Erie County. The number of identified veterans in HMIS increased by 34 percent (n = 197) after the implementation of data sharing in 2013. In 2014, 212 veterans were being served in Erie County. This represents a 44 percent increase in homeless veterans who were previously unaccounted for in HMIS and therefore were not being included in official AHAR reports. The data sharing also revealed that more than 80 percent of the veterans in HMIS who received VA services were discharged to permanent housing, compared to 25 percent of the veterans who did not participate in or were not eligible for VA services (Figure 1). The following benefits of the data-sharing collaboration have been identified locally: increased communication between CoC and VA, decreased barriers for veterans receiving services in the community, increases in VA and community understanding of the homeless veteran population, increased VA and HUD confidence in homeless veteran data validity, and an increase in data-driven interventions.
Donovan and Shinseki (2013; former US secretaries of HUD and VA, respectively) have noted homelessness as a public health problem. More specifically, homelessness is also a health equity issue, given the vital role of stable housing from the perspective of the social determinants of health. Lack of stable housing impacts almost all the domains of the social determinants of health as defined by Healthy People 2020 (Koh et al., 2011). The VA emphasizes the need to first remove barriers to affordable permanent housing and then address issues that lead to homelessness, such as medical and mental health concerns, substance use disorders, and lack of social supports. Because of the successful implementation of the data-sharing model, veterans are now being served without duplication by the VA or community resource best suited to the veteran’s need. Moreover, AHAR data, including PIT counts of the homeless, are now able to be consistently verified in Erie, Pennsylvania. There is no longer a discrepancy between the VA and the local CoC on how many veterans are homeless.
Communities across the country have begun similar data-sharing processes. Currently more than 30 VAMC and community partners have contacted the Erie VAMC and indicated interest. In addition, the VHA Homeless Program National Office (2014) endorsed this data-sharing process as a best practice and produced a white paper. The identification of this program in the VHA health equity environmental scan and subsequent exposure through other mediums will, we hope, spur similar initiatives and partnerships across the country to better serve veterans and other homeless populations. Stakeholders in housing and health care can benefit from applying this approach to tracking and caring for the homeless and those at risk of homelessness. Such efforts will illuminate and underscore the unique connection between housing, health, and overall well-being.
Several issues should be considered when scaling up data-sharing processes. Some HMIS systems may be larger or more extensive than the Erie model and therefore require additional staff or a scheduled data upload system, which is currently being explored at the VA. Large communities may have multiple CoCs/VAMCs serviced by HMIS and may use different software to manage HMIS data. Some homeless men and women, community homeless service providers, HMIS administrators, and other stakeholders might resist sharing homeless data. Education is crucial in each case. In all instances, individual veteran’s choice, privacy, and informed consent are paramount.
The current initiative did not track health outcomes of identified and housed veterans; however, studies targeting the impact of VA engagement and housing on veterans indicate that getting connected and housed lead to positive health (Tsai and Rosenheck, 2015). Fortunately, VA’s infrastructure actively provides care and case management to homeless veterans while linking medical and social services necessary to restore stability in homeless veterans’ lives. For example, low income is routinely addressed by connecting veterans to financial benefits and other supportive services. Outreach programs are available for justice-involved veterans. Also, substance use disorders and mental health issues are addressed in tailored primary medical care (O’Toole et al., 2015). Finally, from a health equity standpoint, adding other vulnerability variables such as race/ethnicity, gender, and military era of service to routine data collection, analysis, and reporting will inform policies and operations for targeted actions toward addressing disparities among vulnerable veterans, including those who are homeless. Future studies should assess potential health improvements in stably housed veterans with a history of homelessness.
Assisting in the identification and dissemination of initiatives to advance health equity is one key role of OHE. This collaborative paper submission among the Erie VAMC, local HMIS, and OHE was made possible by the recent VHA health equity environmental scan. The current work chronicles the necessary synergy to address the needs of homeless veterans in support of the VA mission and a presidential mandate to end veteran homelessness. Current VA secretary Robert A. McDonald recently highlighted the significant strides to end veteran homelessness and the need for local efforts and partnerships with the VA to reach this goal (Shane, 2015). We also stress the importance of all sectors of society identifying veterans in their efforts. One percent of the US population has served in the US military. In addition to being a numerical minority, veterans from various military eras are a vulnerable group with unique experiences and exposures that impact their ability to reach the highest level of health. All public and private sectors are vital in the task of identifying homeless veterans in order to sustain efforts to end homelessness. It will take all hands on deck to end veteran homelessness. Health care organizations can benefit from applying this approach to underscore the connection between health and housing.
Michael Wehrer and Michael Tomlinson were key players in the genesis and development of the local data-sharing project in this submission. Uchenna Uchendu conceived the idea of the health equity environmental scan by the Office of Health Equity and guided the approach for packaging the applicable project for this submission. Kenneth Jones coordinated this effort and worked with all authors toward the writing of the manuscript. All authors contributed significantly to the final product.
Download the graphic below and share it on social media!
- Balshem, H., V. Christensen, A. Tuepker, and D. Kansagara. 2011. A critical review of the literature regarding homelessness among veterans. VA-ESP Project #05-225. Washington, DC: Department of Veterans Affairs.
- Brown, R. T., Y. Miao, S. L. Mitchell, M. Bharel, M. Patel, K. A. Ard, L. J. Grande, D. BlazeyMartin, D. Floru, and M. A. Steinman. 2015. Health outcomes of obtaining housing among older homeless adults. American Journal of Public Health 105:1482–1488.
- Donovan, S., and E. K. Shinseki. 2013. Homeless is a public health issue. American Journal of Public Health 103:S180.
- Fargo, J., S. Metraux, T. Bryne, E. Munley, A. E. Montgomery, H. Jones, G. Sheldon, V. Kane, and D. Culhane. 2012. Prevalence and risk of homelessness among US veterans. Preventing Chronic Disease 9:110–112.
- HUD (Housing and Urban Development). 2014. The 2014 annual homeless assessment report to Congress. https://www.hudexchange.info/resources/documents/2014-AHAR-Part1.pdf (accessed July 24, 2015).
- Koh, H. K., J. J. Piotrowski, S. Kumanyika, and J. E. Fielding. 2011. Healthy people: A 2020 vision for the social determinants approach. Health Education and Behavior 38:551–557.
- LePage, J. P., L. D. Bradshaw, D. J. Cipher, A. M. Crawford, and D. Hoosyhar. 2014. The effects of homelessness on veterans’ health care service use: An evaluation of independence from comorbidities. Public Health 128:985–992.
- Metraux, S., M. Stino, and D. P. Culhane. 2014. Validation of self-reported veteran status among two sheltered homeless populations. Public Health Reports 129:73–77.
- O’Toole, T. P., E. E. Johnson, M. L. Borgia, and J. Rose. 2015. Tailoring outreach efforts to increase primary care use among homeless veterans: Results of a randomized controlled trial. Journal of General Internal Medicine 30:886–898.
- Poulin, S. R., S. Metraux, and D. P. Culhane. 2008. The history and future of homeless management information systems. In Homelessness in America, edited by R. H. McNamara. Westport, CT: Praeger. Pp. 171–179,
http://works.bepress.com/dennis_culhane/101 (accessed July 24, 2015).
- Shane, L. 2015. VA secretary: “We can get to zero” homeless veterans. Military Times. May 15, http://www.militarytimes.com/story/military/benefits/veterans/2015/05/27/va-secretaryhomeless-goals/28011769/ (accessed July 27, 2015).
- Tsai, J., K. M. Doran, and R. A. Rosenheck. 2013. When health insurance is not a factor: National comparison of homeless and nonhomeless US veterans who use Veterans Affairs emergency departments. American Journal of Public Health 103:S225–S231.
- Tsai, J., and R. A. Rosenheck. 2015. Risk factors for homelessness among US veterans. Epidemiologic Reviews 37:177–195.
- Uchendu, U. S. 2014. Institutional journey in pursuit of health equity: Veterans Health Administration’s Office of Health Equity. American Journal of Public Health 104:S511–S513.
- United States Interagency Council on Homelessness (USICH). 2015. Opening Doors: Federal Strategic Plan to Prevent and End Homelessness. https://www.usich.gov/resources/uploads/asset_library/USICH_OpeningDoors_Amendment2015_FINAL.pdf (accessed August 24, 2016).
- VHA Homeless Program National Office. 2014. Erie VAMC homeless programs data sharing and coordination with the Erie continuum of care: An emerging practice in VHA homeless programs operations. http://strat-hub.appspot.com/vh/docs/Erie-Data-Sharingand-Coordination-White-Paper.pdf (accessed July 28, 2015).