Homelessness

Contents lists available at ScienceDirect

Mental Health & Prevention

journal homepage: www.elsevier.com/locate/mhp

Impact of homelessness and unstable housing on adolescent health

Michael J. Cima, R. David Parker⁎

West Virginia University, School of Public Health, 1 Medical Center Drive, HSC N G104, Morgantown, WV 26506, USA

A R T I C L E I N F O

Keywords: Homeless adolescents Service utilization Service costs Mental health

A B S T R A C T

The purpose of this study was to describe service utilization and associated costs of a statewide, Medicaid- insured population of homeless adolescents (HA). This was a secondary data analysis of statewide individual service-level data. The total costs for this sample (n=5259) was $84,077,009 over the study period. Mental health related services accounted for 40% of all costs. A small subpopulation (n=65) of the sample accounted for 28% of the total costs, all of which had mental health related services. In order to reduce costs and improve health outcomes, there is a need to expand mental health services for HA.

1. Introduction

Homelessness persists as an international public health concern, with more than 500,000 homeless individuals in the United States and 400,000 in the European Union on any given night ("The 2015 Annual Homeless Assessment Report to Congress," 2016; Fazel, Geddes, & Kushel, 2014). However, the prevalence of homelessness may not be comparable between countries due to the lack of a unifying definition of what constitutes being homeless. For example, the European Union considers those who are ‘unsheltered’ to be homeless, but there is variation between countries with other components of the definition of homelessness (Fazel et al., 2014). Comparatively, the United States defines chronic homelessness as one year or more of homelessness or four occurrences in the last three years (Defining chronic homelessness: a technical guide for HUD programs, 2007). Despite these differences in definitions, the United States in particular struggles with homelessness.

Homelessness in the United States diminishes public health, the financial wellbeing of health care institutions, and has negative impacts on communities beyond these factors (Fitzpatrick, La Gory, & Ritchey, 2003; Larimer et al., 2009). Healthcare access among persons experi- encing homelessness (PEH) is compounded by the lack of housing itself, as PEH are less likely to be connected to regular medical care due to lack of transportation, no insurance, substance abuse issues, and mental illness among others (Parker & Dykema, 2013). As a result, there is a high utilization of acute care services among PEH, and increasing costs often absorbed by health care institutions (Han & Wells, 2003; Kushel, Vittinghoff, & Haas, 2001; Zur, Mojtabai, & Li, 2014).

Health care costs for PEH in the United States have been studied extensively with many studies using macro level data, such as national

or state indicators, which do not necessarily translate to the individual service level (Peterson et al., 2015). Although macro scale studies using administrative or billing data have utility, individual level data may be more important to practitioners who wish to translate findings into practice. Practice-based research could improve services for PEH and delivery by agencies.

Characterized as a population with high rates of mental illness, substance abuse, sexually transmitted infections, and victimization, homeless adolescents (HA) are especially vulnerable, and experience unique barriers to care (Ammerman SD et al., 2004). In particular, HA tend to be under- or uninsured, consequently reducing primary healthcare utilization, and increasing emergency department use (Christiani, Hudson, Nyamathi, Mutere, & Sweat, 2008; Kushel, Perry, Bangsberg, Clark, & Moss, 2002; Winetrobe, Rice, Rhoades, & Milburn, 2016). In addition, homeless and precariously housed adolescents have less knowledge about their eligibility for government health insurance programs, like Medicaid, compared to their housed counterparts (Fryling, Mazanec, & Rodriguez, 2015). Adolescents with low-income, pre-existing disabilities, or certain familial situations are automatically eligible for Medicaid coverage, and as of 2015, over 36 million adolescents were insured by Medicaid (“2015 Number of Children Ever Enrolled Report", 2016). However, homelessness is a costly challenge to insurers. A large proportion of Medicaid enrollees who are homeless receive services reimbursed through fee-for-service ar- rangements. (USDHHS, 2014) While constituting a small number of Medicaid beneficiaries, fee-for-service enrollees account for the major- ity of Medicaid spending due to frequent hospitalizations (Billings & Mijanovich, 2007).

According to the 2015 Annual Homeless Assessment Report (AHAR), on a single night there were 180,760 HAs (under 25 years

http://dx.doi.org/10.1016/j.mhp.2017.05.002 Received 13 March 2017; Received in revised form 17 April 2017; Accepted 19 May 2017

⁎ Correspondence to: West Virginia University, 1 Medical Center Drive, HSC Annex G103, Morgantown, WV 26506, USA. E-mail address: [email protected] (R.D. Parker).

Mental Health & Prevention 7 (2017) 8–11

Available online 22 May 2017 2212-6570/ © 2017 Elsevier GmbH. All rights reserved.

MARK

old), representing almost one third of PEH (“The 2015 Annual Homeless Assessment Report to Congress,", 2016). This estimation is derived from a bi-annual point in time count in which volunteers and service providers seek out PEH. Although useful, point in time counts are methodologically limited and likely underestimate the actual number of PEH. Strict definitions of homelessness, as well as barriers specific to rural areas, including remote locations, lack of resources and isolation social networks, contribute to the underestimation of the true number of homeless individuals (Link et al., 1995; Tompsett, Toro, Guzicki, Manrique, & Zatakia, 2006).

This study builds on current research methodologies but focuses on the financial implications and healthcare utilization of HA. Information obtained may elucidate areas for interventions to reduce healthcare costs associated with providing care and services to HA.

2. Methods

This secondary data analysis was conducted on a linked, individual service level data system including persons receiving at least one housing or homeless service, one Medicaid service, reside in the state of West Virginia (WV), and who were 24 years of age or younger. We used Housing and Urban Development's (HUD) definition of chronically homeless, which is one year or more of homelessness or four occur- rences in the last three years (Defining chronic homelessness: a technical guide for HUD programs, 2007). The service dates were 01 January 2012 through 31 March 2015. The study was acknowledged by the university IRB as an exempt protocol. The state of West Virginia's Homeless Management Information System (HMIS) contains records of persons receiving services related to housing and homelessness and is estimated to cover 90% of the PEH population. These persons were then linked to their records from the statewide Medicaid dataset using social security number and date of birth. Once the linkages were complete, the data were deidentified for the purposes of this study.

Data included demographics and service data, such as: date and type of each health service accessed; service provider; location of provider; type of service; date of birth; race; payor/insurer; ICD 9 codes of diagnoses at visit; prescriptions; and all total visit charges. Frequencies and percentages were used to describe categorical data, while medians and ranges were used to summarize continuous data. Medians were used to reduce the influence of extreme values. Since this dataset was compiled at the service and individual visit level, subsetting was used to de-duplicate data for descriptive purposes. Analyses involving cost data were conducted on the full, duplicated dataset. R was used for all analyses (Team, 2016).

3. Results

A total of 5259 HAs received at least one homeless service and one Medicaid service between 2012 and 2015. This sample was comprised of approximately equal numbers of men (n=2,608, 49%) and women (n=2,598, 49%), was predominantly white (n=3,754, 74%); with most persons aged 5–18 years (n=2,571, 54%). Disabling conditions, those expected to be of a long-term duration were reported among 12.4% (n=545); 8.7% (n=365) were chronically homeless, and 2.4% (n=123) had been continuously homeless for at least one year. The most common diagnoses were routine infant or child health check (n=697, 13%), ADHD (n=150, 3%), and acute upper respiratory infection (n=150, 3%).

The total healthcare costs across all visits and service providers was $84,077,009 for the 2012–2015 period. These costs were categorized as: practitioner costs, prescription costs, and other costs. Other costs included incentives, capitation, dental, facility encounter, inpatient, inpatient crossover, long term care, non-facility encounter, outpatient, outpatient crossover, Part B crossover, and voids. Medians and ranges for each of these types of costs are reported in Table 1. Data from the ten costliest patients are presented in Table 2. These ten individuals

accounted 2% of the total cost of the sample (n=5259) at nearly $1.4 million over three years. In this subgroup, women ($836,859.34) and HAs between 18 and 24 years ($869,588.46) had the highest expense totals.

The most frequent service types provided by practitioners or other care providers and associated costs are outlined in Table 3. Mental health related services were the most frequent and costly type of service provided for this sample with 72,662 patient visits for 1085 people, costing $32,870,450. A small percentage (n=66) of those 1085 individuals received a ‘Mental Health Inpatient’ service from a provider other than a practitioner costing $23,649,773. In other words, 6% of those receiving a mental health service accounted for almost three quarters of the total cost for all mental health services rendered. Physician non-specialty inpatient services also accounted for a large portion of the total cost at over $18 million.

4. Discussion

A very limited number of published manuscripts have examined the service utilization and associated costs for HA at the individual service level (Bharel et al., 2013; Larimer et al., 2009; Lin, Bharel, Zhang, O'Connell, & Clark, 2015; Parker, Regier, Brown, & Davis, 2015). These data suggest an overrepresentation of minority groups compared to the corresponding proportion of the states total population. Although white HA represented the majority (74%), black HA comprised 20% of the total sample. Comparatively, according to census data, black race was identified by 3.6% of the total state population (StatesCensusBureau: Statistical Abstract of the United States, 2011). Generally, men tend to experience homelessness at a higher rate than women, however this sample had nearly equal numbers of men and women (~49%) ("The 2015 Annual Homeless Assessment Report to Congress,", 2016). Also in contrast to national statistics, which estimates 89% of HA were between

Table 1 Characteristics of homeless adolescents, 24 years and younger, receiving one homeless service and one Medicaid service, 2012–2015 (n=5259).

Characteristic Frequency (%)a

Gender Male 2608 (49.6) Female 2598 (49.4)

Race White 3886 (73.9) Black 1131 (21.5) Other 242 (4.6)

Age Category < 5 853 (16.8) 5–17 2751 (54.1) 18–24 1482 (29.1)

Disabling Condition? Yes 545 (12.4) No 3839 (87.1)

Chronically Homeless? Yes 365 (8.7) No 3848 (91.3)

Continuously Homeless? (1 year) Yes 123 (2.4) No 799 (15.7) Costs Median (min - max) Practitioner costs $70.51 ($0.00 - $6969.00) Prescription costs $9.80 ($0.00 - $24,990.00) Other total costsb $73.82 ($0.00 - $53,600.00)

TOTAL 5259

a Variable totals may not equal 5259 due to missing values. b Other costs included incentive pays, capitation, dental, facility encounter, inpatient,

inpatient crossover, long term care, non-facility encounter, outpatient, outpatient cross- over, Part B crossover, and voids.

M.J. Cima, R.D. Parker Mental Health & Prevention 7 (2017) 8–11

9

18 and 25, the majority of this sample (70%) were under 18 years of age. This is more than likely due to our inclusion of ‘precariously housed’ persons but may yield a more relevant number of children and adolescents facing homelessness and many children who are removed from the families once they become homeless (“The 2015 Annual Homeless Assessment Report to Congress,", 2016). Considering regional nuances, including barriers to care and availability of services, the data presented here may only be representative of HA within the central Appalachian region of the United States.

In this study, costs were categorized as related to a practitioner, prescription, or “other”. “Other”, in this case, included incentives, capitation, dental, facility encounter, inpatient, inpatient crossover, long term care, non-facility encounter, outpatient, outpatient crossover, Part B crossover, and voids. In terms of median costs, the “other” category was the largest ($73.82), followed by practitioner ($70.51), and then prescription ($9.80). “Other” costs also had a much wider range compared to the other two categories ($0.00-$53,600.00), indicating a possible need for intervention to reduce costs.

HA were found to experience high rates of mental illness, another finding similar to existing research (Ammerman SD et al., 2004). Mental health visits were the most prevalently billed service with 72,662 patient visits. Consequently, mental health services accounted for nearly 40% of the total costs for this sample. In addition, there is evidence suggesting that few individuals cost a disproportionate amount compared to the rest of the sample. For example, 65 HA received a mental health inpatient service from a provider other than a practitioner costing a total of $23,649,773. This corresponds to 1% of the sample accounting for 28% of all costs. Further, ten individuals accounted for nearly 2% of the total cost for the entire sample (Table 2). Of these individuals, females costed a disproportionate amount com- pared to males, despite equal numbers of males to females. While the majority of the total sample was under 18 years old, five of the most costly individuals were between 18 and 24, and cost 1.7 times the remaining five of ten high cost patients under 18 years.

Table 3 outlines the 10 most frequent services provided by practi-

tioners and providers other than a practitioner. Congruent with current literature, emergency room and mental health related services were the most frequent and costly services (Ammerman SD et al., 2004; Christiani et al., 2008; Kushel et al., 2002; Parker & Dykema, 2013; Winetrobe et al., 2016). While consistently noted as areas for future interventions among PEH, few, if any, studies present the financial implications for an insured, statewide population. It is apparent from these data there exists a need for expanding preventative services, as well as providing more cost effective mental health services. For instance, compare HA who received a mental health inpatient service from a practitioner to those receiving the same from “other” service providers. Approximately equal numbers presented to each, but HA seeing an “other” service providers presented for nearly double the amount of visits and accounted for almost 80 times the cost. These cost- related data have implications for healthcare policy regarding HA. Enhanced surveillance for mental illnesses among HA through state- wide, individual service-level approaches may serve to increase linkage to regular mental health practitioners, which may significantly reduce costs and the number of patient visits. Additionally, these data only relate to Medicaid insured HA. Uninsured HA may have different patterns of mental health and related costs. Consequently, targeted approaches aimed to increase linkage to mental health related services should consider both insured and uninsured HA.

Multiple studies, including this one, have found a high rate of mental illness among homeless populations (Ammerman SD et al., 2004; Rhoades et al., 2014). However, there is a gap between the high rates of mental illness and utilization of services. Some studies report less than a third of homeless individuals needing mental health services received them (Small, 2010). Other studies speculate that this gap may be attributable to prioritization of survival needs over mental health (Gelberg, Andersen, & Leake, 2000). Therefore, alleviating basic survi- val necessities through housing may help to better identify and treat mental illness within homeless populations.

The information provided by this study support the need for preventative and mental health services among PEH, as well as the

Table 2 Ten costliest patients who received at least one Medicaid service and one homeless service, 2012–2015.

Characteristic Total (n=10) Sex Age Categories

Female (n=5) Male (n=5) < 5 (n=1) 5- < 18 (n=4) 18- < 24 (n=5)

Costs Practitioner $161,725.78 $85,118.89 $76,606.89 $4797.30 $60,821.49 $96,106.99 Prescription $84,507.87 $15,642.25 $68,865.62 $1618.46 $19,752.14 $63,137.27 Othera $1,137,939.00 $736,098.20 $401,840.80 $60,658.92 $366,935.90 $710,344.20 Total $1,384,172.67 $836,859.34 $547,313.31 $67,074.68 $447,509.53 $869,588.46

a Other costs included adpays, capitation, dental, facility encounter, inpatient, inpatient crossover, long term care, non-facility encounter, outpatient, outpatient crossover, Part B crossover, and voids.

Table 3 Most frequent service types provided by practitioners or other care providers and associated total costs across all visits.

Practitioner Service Category Visits Persons Total Costa Other Service Category Visits Persons Total Costa

Mental Health Other Outpatient 58,488 954 $8,927,850 Facility Outpatient ER 25,157 684 $3,794,065 Physician Non-Specialty Office Visits 17,021 455 $1,158,092 Facility Outpatient Other 24,367 669 $2,528,340 Physician Non-Specialty ER 13,933 460 $887,522 Physician Non-Specialty Office Visits 19,847 497 $1,400,089 Physician Non-Specialty Inpatient 6963 103 $18,441,429 Dental 15,929 396 $3,745,694 Professional Services Other 5816 122 $802,187 Laboratory Outpatient Other 15,189 300 $412,302 Professional Transportation 5530 124 $724,362 Physician Specialty ER 10,086 304 $537,642 Mental Health Inpatient 4995 65 $292,827 Mental Health Inpatient 9179 66 $23,649,773 Substance Abuse Other Outpatient 4607 68 $531,927 Physician Specialty Office Visits 5903 167 $419,143 Professional Diagnostic Services 2814 72 $108,923 Laboratory Outpatient Chemistry Tests 5280 107 $213,659 Physician Non-Specialty Outpatient Other 2654 52 $156,907 Physician Non-Specialty Outpatient Other 4355 85 $220,530

**Other services could have been provided through incentives, capitation, dental, facility encounter, inpatient, inpatient crossover, long term care, non-facility encounter, outpatient, outpatient crossover, Part B crossover, and voids.

a Total cost rounded to the nearest dollar.

M.J. Cima, R.D. Parker Mental Health & Prevention 7 (2017) 8–11

10

financial implications of not linking HA to appropriate and cost- effective care. However, this study is limited in that information was only available for those who received at least one homeless service and one Medicaid service. Additionally, costs and service utilization may vary among HA that are not insured by Medicaid. Some self-reported information, such as demographic information, disabling conditions, and factors related to a person's homeless status, was subject to reporting bias.

Acknowledgements and declarations

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Conflicts of interest

The authors declare no conflicts of interest.

References

The 2015 Annual Homeless Assessment Report to Congress. (2016). In U. S. D. o. H. a. U. Development (Ed.). Washington D.C. Housing and Urban Development.

2015 Number of Children Ever Enrolled Report. (2016). Centers for Medicaid and Medicare Services.

Ammerman SD, E. J., Kirzner, R., Meininger, E. T., Tornabene, M., Warf, C. W., Zerger, S., & Post, P. (2004). Homeless Young Adults Ages18–24: Examining Service Delivery Adaptations. National Health Care for the Homeless Council, Inc.

Bharel, M., Lin, W. C., Zhang, J., O'Connell, E., Taube, R., & Clark, R. E. (2013). Health care utilization patterns of homeless individuals in Boston: preparing for Medicaid expansion under the Affordable Care Act. The American Journal of Public Health, 103(Suppl 2), http://dx.doi.org/10.2105/AJPH.2013.301421.

Billings, J., & Mijanovich, T. (2007). Improving the management of care for high-cost Medicaid patients. Health Affairs, 26(6), 1643–1654. http://dx.doi.org/10.1377/ hlthaff.26.6.1643.

Christiani, A., Hudson, A. L., Nyamathi, A., Mutere, M., & Sweat, J. (2008). Attitudes of homeless and drug-using youth regarding barriers and facilitators in delivery of quality and culturally sensitive health care. Journal of Child and Adolescent Psychiatric Nursing, 21(3), 154–163. http://dx.doi.org/10.1111/j.1744-6171.2008.00139.x.

Defining chronic homelessness: a technical guide for HUD programs. (2007). Fazel, S., Geddes, J. R., & Kushel, M. (2014). The health of homeless people in high-

income countries: Descriptive epidemiology, health consequences, and clinical and policy recommendations. Lancet, 384(9953), 1529–1540. http://dx.doi.org/10.1016/ S0140-6736(14)61132-6.

Fitzpatrick, K. M., La Gory, M. E., & Ritchey, F. J. (2003). Factors associated with health- compromising behavior among the homeless. Journal of Health Care for the Poor and Underserved, 14(1), 70–86.

Fryling, L. R., Mazanec, P., & Rodriguez, R. M. (2015). Barriers to homeless persons acquiring health insurance through the affordable care Act. The Journal of Emergency Medicine, 49(5), http://dx.doi.org/10.1016/j.jemermed.2015.06.005.

Gelberg, L., Andersen, R. M., & Leake, B. D. (2000). The behavioral model for vulnerable

populations: Application to medical care use and outcomes for homeless people. Health Services Research, 34(6), 1273–1302.

Han, B., & Wells, B. L. (2003). Inappropriate emergency department visits and use of the health care for the homeless program services by homeless adults in the Northeastern United States. Journal of Public Health Management and Practice, 9(6), 530–537.

Kushel, M. B., Perry, S., Bangsberg, D., Clark, R., & Moss, A. R. (2002). Emergency department use among the homeless and marginally housed: results from a community-based study. American Journal of Public Health, 92(5), 778–784. http:// dx.doi.org/10.2105/ajph.92.5.778.

Kushel, M. B., Vittinghoff, E., & Haas, J. S. (2001). Factors associated with the health care utilization of homeless persons. Journal of American Medical Association, 285(2), 200–206 (doi: jcu00007 )(pii).

Larimer, M. E., Malone, D. K., Garner, M. D., Atkins, D. C., Burlingham, B., Lonczak, H. S., & Marlatt, G. A. (2009). Health care and public service use and costs before and after provision of housing for chronically homeless persons with severe alcohol problems. Journal of American Medical Association, 301(13), 1349–1357 (doi: 301/13/1349 ) (pii) (10.1001/jama.2009.414).

Lin, W. C., Bharel, M., Zhang, J., O'Connell, E., & Clark, R. E. (2015). Frequent emergency department visits and hospitalizations among homeless people with medicaid: implications for medicaid expansion. The American Journal of Public Health, 105(Suppl 5), http://dx.doi.org/10.2105/AJPH.2015.302693.

Link, B., Phelan, J., Bresnahan, M., Stueve, A., Moore, R., & Susser, E. (1995). Lifetime and five-year prevalence of homelessness in the United States. American Journal of Orthopsychiatry, 65(3), 347–354. http://dx.doi.org/10.1037/h0079653.

Parker, R. D., & Dykema, S. (2013). The reality of homeless mobility and implications for improving care. Journal of Community Health. http://dx.doi.org/10.1007/s10900- 013-9664-2.

Parker, R. D., Regier, M., Brown, Z., & Davis, S. (2015). An inexpensive, interdisciplinary, methodology to conduct an impact study of homeless persons on hospital based services. Journal of Community Health, 40(1), 41–46. http://dx.doi.org/10.1007/ s10900-014-9892-0.

Peterson, R., Gundlapalli, A. V., Metraux, S., Carter, M. E., Palmer, M., Redd, A., & Fargo, J. D. (2015). Identifying homelessness among veterans using VA administrative data: Opportunities to expand detection criteria. PLoS One, 10(7) (doi: ARTN e0132664. 10.1371/journal.pone.0132664).

Rhoades, H., Wenzel, S. L., Golinelli, D., Tucker, J. S., Kennedy, D. P., & Ewing, B. (2014). Predisposing, enabling and need correlates of mental health treatment utilization among homeless men. Community Mental Health Journal, 50(8), 943–952. http://dx. doi.org/10.1007/s10597-014-9718-7.

Small, L. (2010). Use of mental health services among people with co-occurring disorders and other mental health co-morbidities: Employing the behavioral model of vulnerable populations. Mental Health and Substance Use: Dual Diagnosis, 3(2), 81–93.

Team, R. (2016). R: A Language and Environment for Statistical Computing. from ⟨http:// www.R-project.org/⟩.

Tompsett, C., Toro, P., Guzicki, M., Manrique, M., & Zatakia, J. (2006). Homelessness in the United States: Assessing changes in prevalence and public opinion, 1993–2001. American Journal of Community Psychology, 37(1), 47–61. http://dx.doi.org/10.1007/ s10464-005-9007-2.

United States Census Bureau: Statistical Abstract of the United States: 2011 (130th Edition). (2011, 2010). from ⟨http://www.census.gov/statab/www/⟩.

USDHHS (2014). A primer on using Medicaid for people experiencing chronic homelessness and tenants in Permanent Supportive Housing.

Winetrobe, H., Rice, E., Rhoades, H., & Milburn, N. (2016). Health insurance coverage and healthcare utilization among homeless young adults in Venice, CA. Journal of Public Health, 38(1), 147–155. http://dx.doi.org/10.1093/pubmed/fdv001 (Oxford).

Zur, J., Mojtabai, R., & Li, S. (2014). The cost savings of expanding Medicaid eligibility to include currently uninsured homeless adults with substance use disorders. The Journal of Behavioral Health Services and Research, 41(2), 110–124. http://dx.doi.org/ 10.1007/s11414-013-9366-7.

M.J. Cima, R.D. Parker Mental Health & Prevention 7 (2017) 8–11

11