J Community Health (2013) 38:685–689
DOI 10.1007/s10900-013-9664-2
ORIGINAL PAPER
The Reality of Homeless Mobility and Implications
for Improving Care
R. David Parker • Shana Dykema
Published online: 15 March 2013
 Springer Science+Business Media New York 2013
Abstract Homeless persons are perceived as a highly
mobile population, and have high rates of co-morbid con-
ditions, including mental health and substance use issues.
This study sought to determine the characteristics of the
mobility and reported health conditions of homeless per-
sons. The sample for this cross sectional study (n = 674)
accounted for 88 % of the homeless population in a medium
sized southern city in the United States. Participants were
recruited from a homeless shelter operating during the
winter season. Homeless persons were less mobile than the
general state population (46.11 % were born in-state vs.
40.7 % of the general population) and less transient than the
general state population (78 % reported an in-state zip code
for the last permanent residence). 31.9 % reported a dis-
abling condition of a serious and long term nature. These
findings challenge the concept that homeless persons are
primarily a mobile population. Furthermore, homeless per-
sons in this sample were more likely to remain in the state
where they lived after becoming homeless. Thus, provider
perceptions that homeless persons would not benefit from
referral to a regular source of outpatient care may be mis-
informed. As homeless persons often seek care in emer-
gency departments for conditions that could be addressed
through outpatient care, if a medical care system imple-
mented standard practices specifically for homeless patients,
this could decrease recidivism. Such interventions represent
significant opportunities to reduce costs, conserve resources,
R. D. Parker (&)
S. Dykema
Department of Medicine, University of South Carolina
School of Medicine, 2 Medical Park, Suite 502,
Columbia, SC 29203, USA
e-mail: david.parker@uscmed.sc.edu
S. Dykema
e-mail: shana.dykema@uscmed.sc.edu
and improve care through policy modification that ensures a
focus on a successful, active linkage to outpatient care and
programs specific to the homeless population.
Keywords Homeless persons
Emergency medicine

Health planning
Health services research
Introduction
Homelessness remains a significant social challenge in the
United States. According to data from the Annual Home-
less Assessment Report (AHAR) compiled by the Housing
and Urban Development (HUD), 1,502,196 persons expe-
rienced homelessness in the United States in 2011 [25].
This means that homelessness is more prevalent (0.48 %)
in the United States than any of leading causes of death,
including heart disease, cancer, and stroke [24, 9]. Unlike
these illnesses, there is no national health-based campaign
against homelessness, despite its intricate ties to public
health.
Homelessness is not commonly identified as a health
problem, yet it is both an etiologic factor and outcome of
multiple health issues, directly and indirectly [5, 7, 26].
Homeless persons are more likely to have comorbid con-
ditions, poorer health outcomes, and decreased access to
health care than other population subgroups [2, 6, 8, 19,
20]. Among homeless persons who were sheltered during
the previous year (n = 1,593,794), nearly 43 % reported a
disabling health condition [25]. As many homeless persons
are uninsured or underinsured, any problems accessing care
are further exacerbated by the fact that a relatively small
number of health care systems in the United States are
designed to provide consistent care for these persons. A
healthcare system’s efficacy tends to decrease when
123
686
attempting to provide support to large numbers of unin-
sured persons who also lack stable housing, especially
when those patients have active mental illness issues,
addiction issues, or chronic health conditions [7]. Because
of many of these challenges, homeless persons are much
more likely to present to the emergency department (ED)
for their medical care, which can worsen ED overcrowding
[2, 6, 10, 13, 15, 18, 19, 22].
Not only does homelessness have a major impact on
healthcare systems, but the reverse is also true: healthcare
systems can have negative impacts on homeless persons [7].
The cumulative effects of errors and delays in medical care
can have a negative compounding effect on homeless per-
sons, who may experience barriers to accessing medical care
or have difficulty navigating complex systems. In this light,
medical care systems, both inpatient and outpatient, present
challenges for homeless persons. Active and executed
referrals to outpatient care could greatly reduce emergency
department (ED) use; however, the lack of outpatient care
systems designed to effectively engage homeless persons
contributes to the documented ED overuse and lack of
continuity of medical care for homeless persons [14].
The relocation and mobility of homeless persons is not
well studied, with relatively few published articles in the
last decade. Older research has indicated that the homeless
population is highly mobile, which could lead to problems
with access to care and continuity of care [1, 3, 17]. The
few more recent articles have indicated that the idea of
homeless mobility may be attributable to anecdotal evi-
dence which has created and perpetuated the stereotype of
the ‘‘homeless transient’’ [12, 22].
Empirical research on this subject is needed to inform
health care systems. This project seeks to increase the
understanding in this area and to describe the mobility
characteristics of a homeless population so that providers
may use this knowledge to improve health care delivery to
them. Since it has been shown that the homeless are more
likely to use the ED for care, this research may inform
policies leading to more active linkage of homeless persons
to outpatient primary care, as well as increasing continuity
of care over a homeless patient’s course of treatment.
Methods
This study was approved by the University of South Car-
olina’s Institutional Review Board (IRB). This cross sec-
tional study recruited a convenience sample of homeless
persons from a homeless registry retained from the city’s
largest homeless shelter. Deidentified data from all persons
who stayed in the shelter between November 1, 2009 and
March 31, 2010 were included. Data were extracted from
the Service Point Homeless Management Information
123
J Community Health (2013) 38:685–689
System (HMIS) system and STATA 10 IC was used for
analyses. We defined ‘‘mobility’’ as relocation from the
state of birth, and ‘‘transience’’ as relocation post-
homelessness.
Sociodemographic data included sex, age, race, ethnic-
ity, veteran status, income, income source, non-cash
income, employment status, highest level of education,
domestic violence status, total monthly income, disability
status, years of residence in city/state, and insurance status.
Homeless information included prior housing situation,
length of stay, current housing status, extent of homeless-
ness (frequency and duration), chronic homeless status,
primary and secondary reasons for homelessness, and zip
code of last permanent residence. Chronic homelessness
was defined by Housing and Urban Development (HUD) as
four or more occurrences of homelessness in the last 3 years
or 1 year or more of homelessness among an unaccompa-
nied adult with a disabling condition. Since participants and
study staff may not have understood this definition, this
variable was operationalized during data collection.
For univariate analyses, Chi square tests were used to
analyze differences among categorical variables and t-tests
were used for numeric data. If the cell sizes were small, then
the nonparametric equivalent was used to increase statistical
reliability. Logistic regression was conducted in multivari-
able analyses with -2 log likelihood ratio tests to compare
models ensuring adherence to the rule of parsimony.
Results
Demographics are presented in Table 1. Participants
(n = 674) were primarily (79.53 %) male. 112 persons
(16.62 %) reported military service. A majority (62.76 %)
identified their primary race as Black American. Less than
half (36.80 %) were high school graduates or equivalent,
and nearly one-third (31.90 %) reported a disabling condi-
tion. The median age was 45.37 years. Criminal domestic
survivorship was reported among almost one-third of
women (n = 40, 31.25 %), versus 3.72 % (n = 20) of men.
Homeless markers are reported in Table 2. Over one-
third (n = 260, 38.58 %) reported ‘‘first time homeless’’
and one quarter (n = 171, 25.37 %) reported ‘‘one or two
episodes of homelessness.’’ Chronic homelessness was
reported by 28.93 % of participants. Almost half of par-
ticipants (n = 308, 45.70 %) were born in-state, while the
general adult population reports 40.7 % of persons were
born in-state [16]. The majority of persons sur-
veyed (n = 519, 77.02 %) also lived in South Carolina
when they became homeless. The most commonly reported
prior living situations were outdoors (n = 141, 20.92 %), a
shelter (n = 129, 19.14 %), staying with family (n = 96,
14.39 %) and staying with friends (n = 76, 11.38 %).
J Community Health (2013) 38:685–689 687

Table 1 Basic demographics (n = 674) Table 2 Overview of homelessness (n = 674)

Measure N Percent Measure N Percent

Median age: 45.37 years Zip code of residence at homeless onset

Sex In-state 519 77.00
Male 536 79.79 Out-of-state 149 22.11
Female 132 19.61 Missing/omitted 6 0.89
Missing/omitted 6 0.60 Chronically homelessa
Veteran status No 475 70.47
Non-veteran 542 80.42 Yes 195 28.93
Veteran 112 16.62 Missing/omitted 4 0.59
Missing/omitted 20 2.96 Extent of homelessness
Primary race First time 260 38.58
Black American 426 62.76 1–3 times in past 173 25.67
White American 213 31.60 4 times or more in past 3 years 97 14.39
Other 38 5.64 1 year or more 123 18.25
Hispanic/Latino ethnicity Missing/omitted 21 3.12
No 633 93.92 Prior living situation
Yes 18 2.67 Outdoors 141 20.92
Missing/omitted 23 3.41 Emergency shelter 129 19.14
Survivor of criminal domestic violence Staying with family/friends 172 33.53
No 593 87.98 Renting apartment 47 6.97
Yes 60 8.90 Owned their own home 40 4.94
Missing/omitted 21 3.12 Jail/prison 29 4.30
Report of a disability Otherb 116 17.21
No 449 66.62 Primary reason for homelessness
Yes 215 31.90 Underemployed or low income 199 29.53
Missing/omitted 10 1.48 Loss of job 93 13.80
Born in-state No affordable housing 61 9.13
No 366 54.30 Medical condition 35 5.24
Yes 308 45.70 Otherc 286 42.43
a
‘‘Chronically homeless’’: unaccompanied individual with a dis-
abling condition who has been continuously homeless for a year or
more OR has had at least four episodes of homelessness in the past
Common reasons for homelessness included un/under- 3 years (HUD definition)
employed/low income (n = 199, 29.53 %) and loss of job b
‘‘Other’’ includes: car, care home, doubled up, foster care, hospital,
(n = 93, 13.80 %) along with lack of affordable housing hotel/motel, place not for habitation, psychiatric hospital, refused,
(9.13 %). A chronic medical condition was reported among substandard structure, subsidized housing, substance abuse treatment
5.24 % (n = 35) of participants. Men were more likely to facility, transitional housing, and missing
c
report military service (v2(1) = 18.94, p \ 0.01) and in ‘‘Other’’ includes: criminal activity, domestic violence, health/
safety, loss of child care, loss of public assistance, loss of transpor-
state birth (v2(1) = 6.74, p = 0.03) than women.
tation, mental health issues, foreclosure, substandard housing, sub-
Logistic regressions were modeled for in state birth as stance abuse, released from institution, eviction, and missing
well as further comparisons between in-state born partici-
pants and out-of-state born participants. The odds of being (1.06, 2.18)], male [OR = 1.75, 95 % CI (1.14, 2.66)], and
born in South Carolina for men were 1.68 times that of the Black American [2.87, 95 % CI (2.01, 4.09)].
odds for in-state birth among women [95 % CI (1.13,
2.49)]. The logistic regression used to explore differences
between participants born in state and out of state included Discussion
the covariates: chronic homelessness, gender, and primary
race. The findings are provided in Table 3. While con- This study found that homeless persons were actually less
trolling for the presence of each variable, statistically sig- mobile and less transient than the general state population,
nificant findings indicated that persons born in state were with 45.70 % of the homeless born in-state and 78 %
1.52 times more likely to be chronically homeless [95 % CI reporting their last permanent residence before becoming

123
688
Table 3 Final reduced logistic regression comparing in-state birth to
out-of-state birth
Variable Odds Standard p value 95 % CI
ratio error
Chronically homelessa 1.52 0.28 0.02 (1.06, 2.18)
Male sexb 1.75 0.38 0.01 (1.14, 2.66)
Black American racec 2.87 0.51 0.00 (2.01, 4.09)
a
Reference group: non-chronically homeless
b
Reference group: female sex
c
Reference group: non-Black American race
homeless as in-state. These findings challenge the popular
stereotype of a highly mobile homeless population. These
findings may help dispel the notion among health care
providers that as a result of their mobility and transience,
homeless persons are unlikely to follow up on their medical
care or outside referrals.
Homeless persons who were born in-state were more
likely to be chronically homeless, male, and Black Amer-
ican. The chronically homeless have been found to have
fewer financial resources, poorer physical and mental
health outcomes, and less family support [2]. Their lower
levels of social support and socioeconomic status may
increase retention in the state of birth, as these persons may
lack the necessary resources to relocate. Additionally, the
lack of financial stability may affect their choice of where
to seek medical care, as the chronically homeless are more
likely to utilize the ED [6, 18].
If the perception among clinicians, especially in the ED,
is that homeless persons would not benefit from referral to
a regular outpatient source of primary care, one interven-
tion to combat ED overuse by the homeless is provider
education. Standard practice in many EDs is to advise the
patient to return in case the symptoms that brought them to
the ED persist [4, 11, 21]. However, if patients are pre-
senting for non-emergent care issues and are encouraged to
return because providers do not believe they will follow up
with their care elsewhere, this could create an unending
dependence on ED use and exacerbate overcrowding
issues. Thus, an effective intervention for health care
delivery systems could be an intentional effort to actively
refer these patients to outpatient care providers and retain
those patients there once referred, ensuring continuity of
care. Active referral to non-ED sources of care could also
result in significant cost savings, both for the organization
and the health care system as a whole [23, 18]. In the US,
the Federally Qualified Health Center (FQHC) System is
comprised of publicly funded centers which see patients on
an income based sliding fee scale. In many communities,
these clinics also receive funding to provide health care for
homeless individuals, and are therefore primed to be
123
J Community Health (2013) 38:685–689
involved in such interventions, as well as future research on
health care provision for the homeless.
Considering the individual barriers to care among the
homeless such as cost, transportation, and social stigma, any
intervention aimed at increasing homeless patients’
involvement in medical care while funneling them to more
appropriate sources of said care must be patient-centered and
provide real time referrals. Since nurses are the primary
providers responsible for discharge planning in inpatient and
outpatient settings, an intervention should be designed to
also be clinician focused. Any such intervention to increase
outpatient primary care for the homeless would require a
significant emphasis on and commitment to communication,
integration and sharing of resources and responsibilities.
There are limitations to this study based on the study
design, including the convenience sampling method. The
cross-sectional methodology means that we were unable to
establish causation. Convenience sampling increases the
potential for bias versus random sampling; however, the
sample to population percentage of this project (88 %)
should mitigate bias in the population within the city.
Additionally, given the high percentage of sample to pop-
ulation, generalizability to other homeless populations in
similar cities may be valid, though any extrapolation should
be done with caution. Another limitation was the ability of
the multivariable logistic regression model to fit the data.
While the associations were strong, these data only account
for 5 % of the variability in the data to explain whether or
not a person is born in state. This indicates that there are
other influencing factors not explored in this project which
would more strongly account for the reasons that a homeless
person remains in his/her state of origin.
Future research should further evaluate concepts of
active engagement and direct intervention by shifting
treatment for non-acute and chronic care to outpatient care
providers. Research could include a prospective cohort of
homeless persons measured on multiple markers to include
health, service access, mobility and other key factors that
could improve care.
Acknowledgments Research was funded by the City of Columbia,
SC.
Conflict of interest The authors report no conflict of interest.
References
1. Bachrach, L. L. (1987). Geographic mobility and the homeless
mentally ill. Hospital & Community Psychiatry, 38, 27–28.
2. Baggett, T. P., O’Connell, J. J., Singer, D. E., & Rigotti, N. A.
(2010). The unmet health care needs of homeless adults: A
national study. American Journal of Public Health, 100(7),
1326–1333.
J Community Health (2013) 38:685–689
3. Duchon, L. M., Weitzman, B. C., & Shinn, M. (1999). The
relationship of residential instability to medical care utilization
among poor mothers in New York City. Medical Care, 37(12),
1282–1293.
4. Engel, K. G., Buckley, B. A., Forth, V. E., McCarthy, D. M.,
Ellison, E. P., Schmidt, M. J., et al. (2012). Patient understanding
of emergency department discharge instructions: Where are
knowledge deficits greatest? Academic Emergency Medicine,
19(9), 1035–1044.
5. 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.
6. 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. doi:
10.2105/ajph.92.5.778.
7. 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:10.1001/jama.2009.414.
8. Levy, B. D., & O’Connell, J. J. (2004). Health care for homeless
persons. New England Journal of Medicine, 350(23), 2329–2332.
doi:10.1056/NEJMp038222.
9. Mackun, P., & Wilson, S. (2011). Population distribution and
change: 2000 to 2010. 2010 Census Briefs. US Census Bureau.
10. Martinez, T. E., & Burt, M. R. (2006). Impact of permanent
supportive housing on the use of acute care health services by
homeless adults. Psychiatric Services, 57(7), 992–999. doi:
10.1176/appi.ps.57.7.992.
11. McCarthy, D. M., Engel, K. G., Buckley, B. A., Forth, V. E.,
Schmidt, M. J., Adams, J. G., et al. (2012). Emergency depart-
ment discharge instructions: Lessons learned through developing
new patient education materials. Emergency Medicine Interna-
tional, 2012, 7.
12. McCarthy, J. F., Valenstein, M., & Blow, F. C. (2007). Resi-
dential mobility among patients in the VA health system: asso-
ciations with psychiatric morbidity, geographic accessibility, and
continuity of care. Administration and Policy in Mental Health,
34(5), 448–455. doi:10.1007/s10488-007-0130-2.
13. Oates, G. (2009). A comparison of national emergency depart-
ment use by homeless versus non-homeless people in the United
States. Journal of Health Care for the Poor and Underserved,
20(3), 840–845.
689
14. Parker, D. (2010). Housing as an intervention on hospital use:
Access among chronically homeless persons with disabilities.
Journal of Urban Health, 87(6), 912–919. doi:10.1007/s11524-
010-9504-y.
15. Pearson, D. A., Bruggman, A. R., & Haukoos, J. S. (2007). Out-
of-hospital and emergency department utilization by adult
homeless patients. Annals of Emergency Medicine, 50(6),
646–652. doi:10.1016/j.annemergmed.2007.07.015.
16. Pew Research Center. (2010). Statistical portrait of the foreign-
born population in the United States, 2008. Pew Research Center.
17. Pollio, D. E. (1997). The relationship between transience and
current life situation in the homeless services-using population.
Social Work, 24(6), 541–552.
18. Poulin, S. R., Maguire, M., Metraux, S., & Culhane, D. P. (2010).
Service use and costs for persons experiencing chronic home-
lessness in Philadelphia: A population-based study. Psychiatric
Services, 61(11), 1093–1098. doi:10.1176/appi.ps.61.11.1093.
19. Reid, K. W., Vittinghoff, E., & Kushel, M. B. (2008). Association
between the level of housing instability, economic standing and
health care access: A meta-regression. Journal of Health Care for the
Poor and Underserved, 19(4), 1212–1228. doi:10.1353/hpu.0.0068.
20. Riley, A. J., Harding, G., Underwood, M. R., & Carter, Y. H.
(2003). Homelessness: A problem for primary care? British
Journal of General Practice, 53(491), 473–479.
21. Taylor, D. M., & Cameron, P. A. (2000). Discharge instruction
for emergency department patients: What should we provide?
Journal of Accident and Emergency Medicine, 17, 86–90.
22. Tsai, J., Mares, A. S., & Rosenheck, R. A. (2011). A geographic
analysis of chronically homeless adults before and after enroll-
ment in a multi-site supported housing initiative: Community
characteristics and migration. American Journal of Community
Psychology, 48, 341–351. doi:10.1007/s10464-010-9363-4.
23. Unified Government of Athens-Clarke County. (2005). Athens
area homeless health care cost analysis 2005.Athens, GA:
Human & Economic Development Department. Unified Gov-
ernment of Athens-Clarke County.
24. US Centers for Disease Control. (2011). Deaths: Final data for
2010. National vital statistics reports, 61(4). US Centers for
Disease Control. http://www.cdc.gov/nchs/data/dvs/deaths_2010_
release.pdf.
25. US Department of Housing and Urban Development. (2012). The
2011 annual homeless assessment report to congress. US
Department of Housing and Urban Development. https://www.
onecpd.info/resources/documents/2011AHAR_FinalReport.pdf.
26. Zerger, S. (2005). Research Agenda: Homelessness and health
care, summary of findings. Nashville: National Health Care for
the Homeless Council.
123