Predicting Homelessness Among Transitioning U.S. Army Soldiers.

INTRODUCTION: This study develops a practical method to triage Army transitioning service members (TSMs) at highest risk of homelessness to target a preventive intervention. METHODS: The sample included 4,790 soldiers from the Study to Assess Risk and Resilience in Servicemembers-Longitudinal Study (STARRS-LS) who participated in 1 of 3 Army STARRS 2011-2014 baseline surveys followed by the third wave of the STARRS-LS online panel surveys (2020-2022). Two machine learning models were trained: a Stage-1 model that used administrative predictors and geospatial data available for all TSMs at discharge to identify high-risk TSMs for initial outreach; and a Stage-2 model estimated in the high-risk subsample that used self-reported survey data to help determine highest risk based on additional information collected from high-risk TSMs once they are contacted. The outcome in both models was homelessness within 12 months after leaving active service. RESULTS: Twelve-month prevalence of post-transition homelessness was 5.0% (SE=0.5). The Stage-1 model identified 30% of high-risk TSMs who accounted for 52% of homelessness. The Stage-2 model identified 10% of all TSMs (i.e., 33% of high-risk TSMs) who accounted for 35% of all homelessness (i.e., 63% of the homeless among high-risk TSMs). CONCLUSIONS: Machine learning can help target outreach and assessment of TSMs for homeless prevention interventions.

Using electronic health record metadata to predict housing instability amongst veterans.

Housing instability is considered a significant life stressor and preemptive screening should be applied to identify those at risk for homelessness as early as possible so that they can be targeted for specialized care. We developed models to classify patient outcomes for an established VA Homelessness Screening Clinical Reminder (HSCR), which identifies housing instability, in the two months prior to its administration. Logistic Regression and Random Forest models were fit to classify responses using the last 18 months of document activity. We measure concentration of risk across stratifications of predicted probability and observe an enriched likelihood of finding confirmed false negative responses from veterans with diagnosed housing instability. Positive responses were 34 times more likely to be detected within the top 1 % of patients predicted at risk than from those randomly selected. There is a 1 in 4 chance of detecting false negatives within the top 1 % of predicted risk. Machine learning methods can classify between episodes of housing instability using a data-driven approach that does not rely on variables curated from domain experts. This method has the potential to improve clinicians' ability to identify veterans who are experiencing housing instability but are not captured by HSCR.

A Research Agenda for Criminal Justice Involvement Among U.S. Veterans.

INTRODUCTION: A substantial proportion of adults in the U.S. criminal justice system are military veterans. Justice-involved veterans are of particular public concern given their service to the country and the high rates of health and social problems in the general veteran population. This article describes the development of a national research agenda for justice-involved veterans. MATERIALS AND METHODS: In the summer of 2022, the VA National Center on Homelessness among Veterans in partnership with the VA Veterans Justice Programs Office convened a national group of subject matter experts and stakeholders across three listening sessions that included 40-63 attendees per session. These sessions were recorded, and transcriptions of all sessions and chats were synthesized to generate a preliminary list of 41 agenda items. The Delphi method involving two rounds of ratings from subject matter experts was used to develop consensus. RESULTS: The final research agenda consists of 22 items covering five domains: Epidemiology and knowledge of the population, treatment and services, systems and systems interface, methodology and research resources, and policies. CONCLUSIONS: The intent of sharing this research agenda is to spur stakeholders to conduct, collaborate, and support further study in these areas.

Food Insecurity and Suicidal Ideation: Results from a National Longitudinal Study of Military Veterans.

OBJECTIVE: Research examining social determinants of suicide risk in veterans suggests a potential link between food insecurity and subsequent suicidal ideation in military veterans. The objective of this study is to investigate, if and how, food insecurity predicts subsequent suicidal ideation in a nationally representative longitudinal survey of veterans. METHODS: A national longitudinal survey was analyzed of participants randomly drawn from over one million U.S. military service members who served after September 11, 2001. N = 1,090 veterans provided two waves of data one year apart (79% retention rate); the final sample was representative of post-9/11 veterans in all 50 states and all military branches. RESULTS: Veterans with food insecurity had nearly four times higher suicidal ideation one year later compared to veterans not reporting food insecurity (39% vs 10%). In multivariable analyses controlling for demographic, military, and clinical covariates, food insecurity (OR = 2.37, p =.0165) predicted suicidal ideation one year later, as did mental health disorders (OR = 2.12, p = .0097). Veterans with both food insecurity and mental health disorders had a more than nine-fold increase in predicted probability of suicidal ideation in the subsequent year compared to veterans with neither food insecurity nor mental health disorders (48.5% vs. 5.5%). CONCLUSION: These findings identify food insecurity as an independent risk marker for suicidal ideation in military veterans in addition to mental disorders. Food insecurity is both an indicator of and an intervention point for subsequent suicide risk. Regularly assessing for food insecurity, and intervening accordingly, can provide upstream opportunities to reduce odds of suicide among veterans.HIGHLIGHTSMilitary veterans with food insecurity were at elevated risk of suicidal ideation.Veterans with mental health disorders had higher odds of suicidal ideation one year later.Food insecurity plus mental health disorders led to a substantial increase in suicidal ideation.

Criminal legal involvement among recently separated veterans: Findings from the LIMBIC study.

OBJECTIVE: This study investigated individual-level and neighborhood-level predictors of criminal legal involvement of veterans during the critical transition period from military to civilian life. HYPOTHESES: We hypothesized that substance use, mental health, and personality disorders will increase the incidence of criminal legal involvement, which will be highest among veterans living in socioeconomically disadvantaged neighborhoods after military discharge. METHOD: We analyzed data from a longitudinal cohort study of 418,624 veterans who entered Department of Veterans Affairs (VA) health care after leaving the military. Department of Defense (DoD) data on clinical diagnoses, demographics, and military history were linked to VA data on neighborhood of residence and criminal legal involvement. RESULTS: Criminal legal involvement in the 2 years following military discharge was most strongly predicted by younger age, substance use disorder, and being male. Other predictors included the military branch in which veterans served, deployment history, traumatic brain injury, serious mental illness, personality disorder, having fewer physical health conditions, and living in socioeconomically disadvantaged neighborhoods. These factors combined in multivariable analysis yielded a very large effect size for predicting criminal legal involvement after military separation (area under the curve = .82). The incidence of criminal legal involvement was 10 times higher among veterans with co-occurring substance use disorder, serious mental illness, and personality disorder than among veterans with none of these diagnoses, and these rates were highest among veterans residing in more socioeconomically disadvantaged neighborhoods. CONCLUSIONS: To our knowledge, this is the largest longitudinal study of risk factors for criminal legal involvement in veterans following military discharge. The findings supported the hypothesis that veterans with co-occurring mental disorders living in socioeconomically disadvantaged neighborhoods were at higher risk of criminal legal involvement, underscoring the complex interplay of individual-level and neighborhood-level risk factors for criminal legal involvement after veterans leave the military. These results can inform policy and programs, such as the DoD Transition Assistance Program (TAP) and the VA Military to Civilian Readiness Pathway program (M2C Ready), to enhance community reintegration and prevent criminal legal involvement among veterans transitioning from military to civilian life. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Establishing a Research Agenda for Suicide Prevention Among Veterans Experiencing Homelessness.

Suicide among Veterans experiencing or at risk for homelessness remains a significant public health concern. Conducting research to understand and meet the needs of this at-risk population remains challenging due to myriad factors (e.g., clinical complexity including multimorbidity, difficulty monitoring risk across systems). To address this challenge, the United States Department of Veterans Affairs (VA) convened the Health Services Research and Development (HSR&D) Suicide Prevention in Veterans Experiencing Homelessness: Research and Practice Development meeting, bringing together subject-matter experts in the fields of homelessness and suicide prevention, both from within and outside of VA. During the meeting, attendees identified 10 potential research priorities at the intersection of suicide prevention and homelessness. After the meeting, Delphi methodology was used to achieve consensus on the relative importance of the identified research domains. Through this iterative Delphi process, agreement was reached regarding the need to increase understanding of barriers and facilitators to suicide risk assessment and emergency intervention for Veterans experiencing homelessness by examining the perspectives of both Veterans and healthcare providers. Elucidating the complex relationships between risk periods, subgroups, suicide means, and drivers of suicide among Veterans experiencing homelessness was also considered a top priority. This article documents the Delphi process and provides a research agenda for researchers, funding agencies, and policymakers to prioritize the most relevant and potentially impactful research domains aimed at preventing suicide among Veterans experiencing or at risk for homelessness.

Association of statewide stay-at-home orders with utilization of case management and supportive services for veterans experiencing housing insecurity.

The US Department of Housing and Urban Development-Department of Veterans Affairs (VA) Supportive Housing (HUD-VASH) program provides Veterans with a subsidy for rent and case management. In response to the Coronavirus 2019 pandemic, many states enacted stay-at-home orders that may have limited access to case managers. Therefore, we examined the association between statewide stay-at-home orders and utilization of HUD-VASH case management. We linked data on whether a state implemented a statewide stay-at-home order between March 1, 2020 and April 30, 2020 with VA medical records. Analysis time was centered on the date of a state's stay-at-home order (exposed states). For Veterans in states without a stay-at home-order (unexposed states), we used the average date exposed states implemented an order (March 27, 2020). We used a difference-in-difference design and adjusted linear regression models to compare total, in-person, telephone, and video case management encounters per Veteran in the 60 days after a stay-at-home order relative to the prior year. There was no significant difference in utilization of case management between Veterans who lived in states that did and did not issue a stay-at-home order. Across all states and in the 60 days after the index date relative to the prior year, Veterans had more total, telephone and video, and fewer in-person encounters. Statewide stay-at-home orders did not differentially affect utilization of case management. Virtual case management in HUD-VASH can increase program reach; however, the effect of virtual case management on outcomes such as quality of life and Veteran satisfaction is unknown.

Prevalence of Alzheimer's disease and related dementias among veterans experiencing housing insecurity.

INTRODUCTION: Housing insecure veterans are aging, but the prevalence of Alzheimer's disease and related dementias (AD/ADRD) in the population is unknown. METHODS: We calculated the prevalence of AD/ADRD diagnoses in 2018 among veterans that experienced homelessness, were at-risk for homelessness, or were stably housed. We determined acute care (emergency department, hospitalizations, psychiatric hospitalizations), and any long-term care (nursing home, and community-based) use by housing status among veterans with an AD/ADRD diagnosis. RESULTS: The overall prevalence of AD/ADRD diagnoses for homeless, at-risk, and stably housed veterans was 3.66%, 13.48%, and 3.04%, respectively. Housing insecure veterans with AD/ADRD used more acute care, and were more likely to have a nursing home admission compared to stably housed veterans. At risk, but not homeless veterans, were more likely to use US Department of Veterans Affairs-paid home and community-based care than stably housed veterans. DISCUSSION: The prevalence of AD/ADRD diagnoses is greater among housing insecure veterans than stably housed veterans.

Incidence of Homelessness among Veterans Newly Diagnosed with Alzheimer's Disease and Related Dementias.

BACKGROUND: To determine the incidence of homelessness among Veterans diagnosed with Alzheimer's disease and related dementias (ADRD). METHODS: We used Veterans Affairs (VA) administrative records to identify Veterans with a new ADRD diagnosis anytime between 2010-2019. Among these Veterans, we calculated the incidence of homelessness, and estimated the association between demographics, comorbidities and hazard of homelessness. RESULTS: The incidence rate of homelessness was highest for Veterans diagnosed with ADRD between 18-49 years of age (14.9 per 1,000 person-years; 95%CI: 13.6, 16.3) and lowest for Veterans diagnosed with ADRD at 90+ years (0.3 per 1,000 person-years; 95%CI: 0.2, 0.4). The adjusted hazard ratio of homelessness was higher for unmarried Veterans, and those with alcohol use disorder, substance use disorder, liver disease, depression, hypertension, lung disease, post-traumatic stress disorder and psychoses. CONCLUSIONS: Younger age and being unmarried at the time of ADRD diagnosis are associated with a greater risk of experiencing homelessness.

Redesign of a Screening Process for VA Homeless Housing.

Standardizing the screening processes for homeless housing among VA facilities can make programs more accessible to veterans experiencing homelessness and improve provider knowledge of existing and available services.

Rapid HIV testing experience at Veterans Affairs North Texas Health Care System's Homeless Stand Downs.

In the USA, 21% of the estimated 1.1 million people living with human immunodeficiency virus/acquired immune deficiency syndrome (HIV/AIDS) are unaware they are HIV-infected. In 2011, Veterans Health Administration (VHA)'s Office of Public Health in conjunction with VHA's Health Care for Homeless Veterans Program funded grants to support rapid HIV testing at homeless outreach events because homeless populations are more likely to obtain emergent rather than preventive care and have a higher HIV seroprevalence as compared to the general population. Because of a Veterans Affairs North Texas Health Care System (VANTHCS)'s laboratory testing requirement, VANTHCS partnered with community agencies to offer rapid HIV testing for the first time at VANTHCS' 2011 Homeless Stand Downs in Dallas, Fort Worth, and Texoma, Texas. Homeless Stand Downs are outreach events that connect Veterans with services. Veterans who declined testing were asked their reasons for declining. Comparisons by Homeless Stand Down site used Pearson χ², substituting Fisher's Exact tests for expected cell sizes <5. Of the 910 Veterans attending the Homeless Stand Downs, 261 Veterans reported reasons for declining HIV testing, and 133 Veterans were tested, where 92% of the tested Veterans obtained their test results at the events - all tested negative. Veterans' reported reasons for declining HIV testing included previous negative result (n=168), no time to test (n=49), no risk factors (n=36), testing is not a priority (n=11), uninterested in knowing serostatus (n=6), and HIV-infected (n=3). Only "no time to test" differed significantly by Homeless Stand Down site. Nonresponse rate was 54%. Offering rapid HIV testing at Homeless Stand Downs is a promising testing venue since 15% of Veterans attending VANTHCS' Homeless Stand Downs were tested for HIV, and majority obtained their HIV test results at point-of-care while further research is needed to determine how to improve these rates.

Time to depression treatment in primary care among HIV-infected and uninfected veterans.

BACKGROUND: Multiple factors, including patient characteristics, competing demands, and clinic type, impact delivery of depression treatment in primary care. OBJECTIVE: Assess whether depression severity and HIV serostatus have a differential effect on time to depression treatment among depressed patients receiving primary care at Infectious Disease or General Medicine clinics. DESIGN: Multicenter prospective cohort, (Veterans Aging Cohort Study), comparing HIV-infected to uninfected patients. PARTICIPANTS AND MEASURES: The total cohort consisted of 3,239 HIV-infected and 3,227 uninfected patients. Study inclusion criteria were untreated depressive symptoms, based on a Patient Health Questionnaire (PHQ-9) score of greater than 9, and no antidepressants or mental health visits in the 90 days prior to PHQ-9 assessment. Treatment was defined as antidepressant receipt or mental health visit within 90 days following PHQ-9 assessment. Depression severity based on PHQ-9 scores was defined as mild-moderate (greater than 9 to 19) and severe (20 or greater). Kaplan-Meier curves were used to estimate time to treatment by depression severity and HIV serostatus. Cox proportional hazards methods adjusted for covariates were used. KEY RESULTS: Overall, 718 (11%) of the cohort met inclusion criteria, 258 (36%) of whom received treatment. Median time to treatment was 7 days [95% confidence interval (CI) = 4, 13] and was shortest for severely depressed HIV-infected patients (0.5 days; 95% CI = 0.5, 6, p = 0.04). Compared to mildly-moderately depressed uninfected patients, severely depressed HIV-infected patients were significantly more likely to receive treatment [adjusted hazard ratio (HR) 1.67, 95% CI = 1.07, 2.60), whereas mildly-moderately depressed HIV-infected patients (adjusted HR 1.10, 95% CI = 0.79, 1.52) and severely depressed uninfected patients (adjusted HR 0.93, 95% CI = 0.60, 1.44) were not. CONCLUSIONS: In this large cohort, time to primary care treatment of depression was shortest among severely depressed HIV-infected patients. Regardless of HIV serostatus, if depression was not treated on the assessment day, then it was unlikely to be treated within a 90-day period, leading to the majority of depression being untreated.

Trends in perimortal conditions and mortality rates among HIV-infected patients.

OBJECTIVES: To describe trends in perimortal conditions (pathological conditions causing death or present at death but not necessarily the reported cause of death) during three periods related to the availability of HAART, pre-HAART (1992-1995), early HAART (1996-1999), and contemporary HAART (2000-2003); annual mortality rates; and antiretroviral therapy (ART) prevalence during 1992-2003. DESIGN: Multicenter observational clinical cohort in the United States (Adult/Adolescent Spectrum of HIV Disease [ASD] project). METHODS: Proportionate mortality for selected perimortal conditions, annual mortality rates, and ART prevalence were standardized by sex, race/ethnicity, age at death, HIV transmission category, and lowest CD4 cell count of ASD decedents. Multivariable generalized linear regression was used to estimate trends in proportionate mortality, as linear trends through all three HAART periods, mortality rates, and ART prevalence. RESULTS: Of 9225 deaths, 58.6% occurred during 1992-1995, 29.5% during 1996-1999, and 11.9% during 2000-2003. Linear trends in proportionate mortality for noninfectious diseases (e.g., liver disease, hypertension, and alcohol abuse) increased significantly; proportionate mortality for AIDS-defining infectious diseases (e.g., pneumocystosis, nontuberculous mycobacterial disease, and cytomegalovirus disease) decreased significantly. Mortality rates decreased from 487.5/1000 person-years in 1995 to 100.6 in 2002. Of 36 256 patients from ASD, 75.7% (standardized average) were prescribed ART annually. CONCLUSIONS: Among HIV-infected patients, the majority of whom were prescribed ART, the increasing trend in common noninfectious perimortal conditions support screening and treatment for these conditions in order to sustain the trend in declining mortality rates.

Effect of hepatitis C coinfection on discontinuation and modification of initial HAART in primary HIV care.

OBJECTIVES: To estimate the effect of hepatitis C (HCV) coinfection on time to first occurrence of either discontinuation or modification of initial HAART among previously antiretroviral therapy-naive HIV-infected patients. METHODS: The analysis included antiretroviral therapy-naive patients who initiated HAART prior to November 2003 and were participating in the University of North Carolina Center for AIDS Research, HIV/AIDS observational clinical cohort. The effect of HCV status on time to first occurrence of either HAART discontinuation or modification was assessed using Kaplan-Meier survival estimates and multivariable proportional hazards regression was used to estimate hazard ratios. RESULTS: Of 296 patients initiating HAART, 22% were coinfected with HCV. During a median follow-up of 473 days [interquartile range (IQR), 167-940] from HAART initiation, 104 (35%) patients discontinued and 91 (31%) modified their first regimen. Reasons for discontinuation and modification were comparable by HCV serostatus and included treatment failure (12%), toxicity (41%), and barriers to adherence (47%). The median time to first occurrence of either discontinuation or modification among HCV-infected patients was 401 days (IQR, 128-821), and among HCV-uninfected patients was 493 days (IQR, 204-952) (P = 0.22). After adjustment for baseline demographic and clinical characteristics, the hazard ratio contrasting HCV-infected with HCV-uninfected patients was 1.39 (95% confidence interval, 0.95-2.03; P = 0.09). CONCLUSION: HCV coinfection was only marginally associated with a shorter duration of an initial HAART regimen, suggesting optimization of a first HAART regimen may not appreciably depend on HCV serostatus.