Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Jan 1.
Published in final edited form as: J Soc Distress Homeless. 2021 Apr 6;31(2):163–171. doi: 10.1080/10530789.2021.1908487

Application of the Frailty Framework among Vulnerable Populations to Hospitalization Outcomes of Individuals Experiencing Homelessness in Long Beach, California

Dennis G Fisher 1, Grace L Reynolds 1, Noushin Khoiny 1, Loucine Huckabay 1, Debby Rannalli 1
PMCID: PMC9697922  NIHMSID: NIHMS1691861  PMID: 36439946

Abstract

Background:

Individuals experiencing homelessness have a high prevalence of infectious diseases that may result in hospitalization. However, low ability to navigate the healthcare system and lack of health insurance may mean that those who are experiencing homelessness may not receive the healthcare that they need.

Objectives:

This study uses risk factors at baseline to predict hospitalization at follow-up. This paper also presents the associations between reporting homelessness and selected infectious diseases.

Research design:

Longitudinal study of baseline and follow-up conducted August 2000 through July 2014.

Subjects:

4916 Not experiencing homelessness mean age 37.9 years, 29% female, and 2692 experiencing homelessness age 42.1 years, 29% female received services from a research/service center in a low-income, high-crime area of Long Beach, CA.

Measures:

Risk Behavior Assessment, Risk Behavior Follow-up Assessment, laboratory testing for hepatitis A, hepatitis B, hepatitis C, syphilis, chlamydia, and gonorrhea.

Results:

Predictors of hospitalization at follow-up were ever use of crack cocaine, income from Social Security or disability, reporting homelessness, female, and those who identify as Black compared to White race/ethnicity.

Conclusions:

Income from the safety net of Social Security or disability appears to provide the participant with experience that transfers to being able to obtain healthcare. A higher proportion of those experiencing homelessness, compared to those not experiencing homelessness, appear to be hospitalized at follow-up. Women, those who identified as Black, and those who used crack at baseline are more likely to be hospitalized at follow-up whether or not they were experiencing homelessness. We recommend coordination with substance abuse treatment programs for discharge planning for homeless patients. Our findings support use of the Frailty Framework when working with individuals experiencing both homelessness and hospitalization.

Keywords: homelessness, vulnerable populations, public health, substance use disorder, frailty

Introduction

Individuals experiencing homelessness have a high prevalence of infectious diseases compared to a general population which may lead to premature mortality and hospitalization.1 A study of hospital readmissions found that 50.8% of the hospitalizations among those experiencing homelessness resulted in readmission within 30 days of discharge compared to only 18.7% for those note reporting homelessness.2 This is similar to a study of 15–25-year-olds who had been admitted to either an emergency department or an inpatient floor whose address was reported as homeless, none, or a homeless shelter, in that 50% of these patients had been readmitted within 1 year and 43.1% within 30 days.3

Housing instability and homelessness severity have been found to be associated with sexually transmitted infections (STIs) such as gonorrhea. An overall composite STI prevalence is usually reported, but where gonorrhea was reported separately, it appears to be highest (7.8%) among younger adult women reporting that they are homeless.4 Persons who inject drugs (PWIDs) are reported to experience homelessness at proportions over 50% in North America.5 Injection drug use has been associated with hepatitis C (HCV) and hepatitis B (HBV), and HBV is highly transmissible via sexual activity.6 Among individuals experiencing homelessness in North America, prevalence rates of HCV have been reported to be over 55% and rates of HBV have been reported to range from 5% to as high as 30%.1,5 In studies of US veterans, those who reported homelessness had an HBV prevalence of almost 1.0% compared to those who did not report being homeless of 0.4%, and HCV prevalence among veterans experiencing homelessness was 12.1% compared to only 2.7% among veterans not reporting homelessness.6 It is more difficult to find syphilis prevalence rates among populations reporting homelessness in the US. A study of men in Brazil who reported homelessness found that 10.2% were reactive to a Venereal Disease Research Lab (VDRL) test and 5.4% had active syphilis.7

Substance abuse is a major distinguishing characteristic that describes populations experiencing homelessness in comparison to those who do not report being homeless.8 When individuals experiencing homelessness are hospitalized, it is more likely to be for alcohol abuse, drug abuse, and liver disease.9 In a study of veterans experiencing homelessness compared to veterans not reporting homelessness who underwent surgeries, those reporting homelessness were more likely to be readmitted. These readmissions were associated with discharge to someplace other than back into the community, recent alcohol abuse, and an elevated American Society of Anesthesiologists’ classification (a metric used to determine whether someone is healthy enough to tolerate surgery and anesthesia).10

Veterans who had polysubstance use disorders were more likely to identify as Black, report they were currently homeless and to have hepatic disease compared to those who only had a single substance use disorder.11 The risk factors of drug and alcohol abuse were related to health problems among clients at a clinic for those experiencing homelessness in Los Angeles.12 Two-thirds of a sample of adults experiencing homelessness in met the criteria for chronic substance dependence, however only one-fifth had received treatment in the last 60 days.13 Those who had spent less time in places not meant for sleeping were 5 times more likely to have received substance abuse treatment.13 Experiencing homelessness was associated with a lower rate of drug abuse treatment completion and was associated with more treatment episodes being required to complete treatment.14 Experiencing homelessness has also been implicated in substance abusers failing to get into treatment in the first place.15

Theoretical Framework

Our theoretical framework for this study is the Frailty Framework among Vulnerable Populations.16,17 In this framework, frailty domains form a triad of physical, psychological, and social factors, where the presence of all 3 makes an individual frail. Frailty leads to increased health care utilization, hospitalization, health care dependency, disability, and death17. The frailty model has been used previously with samples of individuals experiencing homelessness 18. Individual risk factors for frailty include race/ethnicity, gender, income, education, and incarceration 18,19. In addition, both the experience of homelessness and substance abuse affect frailty, including an increased likelihood of multiple long-term medical conditions 20. Figure 1 depicts the Frailty Framework.

Figure 1.

Figure 1.

Open in a new tab

The Frailty Framework for Vulnerable Populations and Hospitalization Outcome

The purpose of the current study is to apply the Frailty Framework for Vulnerable Populations using information gathered from baseline interviews to predict hospitalizations at follow-up interview and to determine differences between those who reported experiencing homelessness at baseline to those not experiencing homelessness. The current study used variables mirroring those delineated in the Frailty Framework. We elicited information from respondents on drug use of a number of injected and non-injected substances including alcohol, marijuana, crack, powdered cocaine, heroin, speedball (heroin and cocaine used together), illicit methadone, opiates other than heroin, amphetamine and Rohypnol, as well as demographic variables, and history of sexually transmitted and blood-borne infections.

METHODS

Study Setting

The data were collected at a community-based research center affiliated with a local university, located in a low-income area of Long Beach, CA. The center functioned as part of the safety-net in the area through having a food bank, HIV and other STI testing, and other prevention and drug counseling activities available on-site and at no cost.

Study Sample

There were 4916 participants who reported that they were not experiencing homelessness at baseline and 2692 participants who reported that they were experiencing homelessness at baseline. Overall, the sample included 32% of respondents who identified as Black, 34% White, 23% Hispanic, and 10% identified as another race including Asian, Pacific Islander, Native American, and mixed race. The sample was 29% female, almost 50% had a monthly income less than $500, over 50% had a high school graduation, a mean age of 39.3 (SD=11.82) years old, and had spent a mean of 1002 (SD=2049) days incarcerated.

Inclusion criteria included being at least 18 years of age and able to provide informed consent (e.g., not being under the influence of alcohol or drugs at the time of interview). Individuals were recruited via the existing programs at the research center and through word-of-mouth. Participants were compensated with a non-cash gift card in the amount of $10 at both baseline and follow-up.

All participants signed informed consent forms approved by the California State University, Long Beach (CSULB) Institutional Review Board (IRB). All data collection occurred under the protocols approved by the CSULB IRB. All data were protected under Certificates of Confidentiality obtained from the National Institute on Drug Abuse (NIDA).

Measures

Risk Behavior Assessment (RBA)

Baseline data were collected by structured interview conducted with the Risk Behavior Assessment (RBA). The 48-hour test-retest reliability and validity of the drug use have been published.21,22 The question on homelessness was “Do you consider yourself to be homeless?” The reliability of this has been reported.23 The questions on income were: “In the last 30 days, what were your sources of income?” Answers included: “Social Security, disability, Workmen’s Compensation.” The income question was “How much money did you get altogether in the last 30 days?” The reliabilities of the economic questions have been also been published.24

Risk Behavior Follow-Up Assessment (RBFA)

The follow-up data were collected with the Risk Behavior Follow-up Assessment (RBFA).25 The question that was used as the outcome for the logistic regression model was “How much time altogether (days/months) in the last six months have you spent in the hospital as a patient?” This was dichotomized for analysis purposes into ever (1 or more days) versus never (0 days). The reliability of the health service utilization items on the RBFA have been reported.26 They are summarized here for convenience. “How many times in the last six months have you spent time/been treated in an emergency room or in the hospital as a patient?” (n = 259, r = .668, 95% Confidence Limits .593 - .729, p = .0001).

“How much time altogether in the last six months have you spent in the hospital as a patient?” (DAYS n = 66, r = .617, 95% Confidence Limits .437 - .745, p = .0001).

“How much time altogether in the last six months have you spent in the hospital as a patient?” (MONTHS n = 61, r = .892, 95% Confidence Limits .824 - .933, p = .0001).

One study that investigated self-report by respondents who reported experiencing homelessness also reported that they are fairly accurate reporters.27 Procedures for obtaining follow-up interview data were the same as for baseline. All interviews took place at the community-based research center, all participants provided written informed consent, and were paid with a $10 non-cash gift card at the conclusion of the interview. These follow-up interviews were conducted approximately six months after the baseline interview, though the interval between baseline and follow-up varied. No one eligible for a follow-up interview was turned away if they arrived outside of the six-month follow-up. Contact was maintained with participants through a variety of means including weekly encounters with staff at the on-site foodbank, and telephone and mail contacts.

HIV and STI Testing

Whole blood was collected by venipuncture performed by California State licensed phlebotomists following bloodborne pathogens safety procedures, immediately following the baseline or follow-up interviews. The blood was tested for HIV (Bio-Rad Laboratories, Redmond, WA), Hepatitis C (Abbott Laboratories, Abbott Park, IL), two tests for syphilis, the Treponema pallidum-particle agglutination (Fujirebio Inc., Tokyo, Japan), and the Rapid Plasma Reagin (Arlington Scientific Inc., Springville, UT), as well as Hepatitis A (Abbott Laboratories, Abbott Park, IL), and Hepatitis B Core Antigen (Abbott Laboratories, Abbott Park, IL).

Statistical Analysis

For the bivariate analyses, the categorical variables were analyzed as χ2 tests of independence, the ordinal variables were analyzed as Wilcoxon rank sum tests, and the continuous variables were analyzed as student t-tests. The effect sizes reported were Cramer’s V for the χ2 results, and η2 for the t-tests and the Wilcoxon tests. The logistic regression was conducted using procedures from Hosmer et al.28 The methods for the logistic regression models include both a plan for selecting variables and a method for assessing adequacy. Model building seeks to have the most parsimonious model that still reflects the outcome. Having a parsimonious model means that the model will be numerically stable, more easily adopted by others, and will have smaller standard errors.

The first step in the Hosmer et al. purposeful selection method of model building is to start with a careful univariable analysis of each candidate independent variable. Therefore, Table 1 shows the Pearson Chi-square test results for the categorical variables (acceptable according to Hosmer et al.) and also shows the results of either the Wilcoxon two-sample test or the two-sample t-test for the continuous variables (also acceptable according to Hosmer et al.).

TABLE 1.

Bivariate Analysis of Demographics, Disease, and Drug Use [n (%), Mean (SD)] by Homeless Status

Not Homeless Homeless t/χ2/z* df P ES†
Continuous Variables (t-test)
Age (y) 37.9 (12.1) 42.1 (10.7) 15.2 7606 0.0001 0.364
Incarcerated time (days) 692 (1707) 1599 (2474.7) 18.4 7349 0.0001 0.453
Ordinal Varibles (z-test) *
Income (month)‡ 2.4 (1.4) 1.4 (0.7) 32.3 0.0001 0.136
Education§ 4.7 (1.9) 3.8 (1.8) 20.4 0.0001 0.053
Categorical Variables (χ2)
Race
 Black 1342 (26.3) 1196 (43.9)
 White 1801 (35.3) 879 (32.3)
 Hispanic 1360 (26.7) 460 (16.9)
 Other 595 (11.7) 184 (6.7) 290.4 3 0.0001 0.1927
Hepatitis A laboratory test 832 (48.3) 334 (46.6) 0.6 1 0.4277 −0.0161
Hepatitis B laboratory test 460 (17.6) 328 (25.0) 25.5 1 0.0001 0.0868
Hepatitis C laboratory test 430 (40.5) 387 (47.4) 8.8 1 0.0030 0.0684
HIV Elisa 88 (2.1) 39 (1.9) 0.09 1 0.7537 −0.0040
Rapid Plasma Reagin 60 (2.6) 36 (3.4) 1.7 1 0.1941 0.0223
T. pallidum – particle agglut. 167 (7.2) 110 (10.40) 9.7 1 0.0018 0.0537
Chlamydia laboratory test 75 (3.4) 13 (1.5) 8.4 1 0.0037 −0.0524
Gonorrhea laboratory test 20 (1) 5 (0.6) 0.9 1 0.3422 −0.0171
Hepatitis B self-report 335 (6.6) 217 (8.0) 5.3 1 0.0210 0.0262
Gonorrhea self-report 782 (15.4) 611 (22.5) 61.1 1 0.0001 0.0885
Syphilis self-report 284 (5.6) 216 (7.9) 16.5 1 0.0001 0.0461
Genital warts self-report 303 (5.9) 164 (6.0) 0.02 1 0.8752 0.0018
Chlamydia self-report 594 (11.7) 404 (14.9) 16.2 1 0.0001 0.0456
Herpes self-report 169 (3.3) 114 (4.2) 3.9 1 0.0490 0.0223
Trichomonas|| self-report 165 (11.3) 154 (19.6) 28.5 1 0.0001 0.1128
Yeast|| infection self-report 684 (47.4) 465 (59.6) 30.1 1 0.0001 0.1163
Ever used alcohol 4868 (95.5) 2637 (96.9) 9.1 1 0.0026 0.0341
Ever used marijuana 4050 (79.4) 2469 (90.1) 163.7 1 0.0001 0.1447
Ever used crack 1905 (37.4) 1924 (70.7) 790.3 1 0.0001 0.3179
Ever used cocaine 2233 (43.8) 1741 (64.0) 290.9 1 0.0001 0.1929
Ever used heroin 1127 (22.1) 1167 (42.3) 370.9 1 0.0001 0.2178
Ever used speedball 728 (14.3) 715 (26.3) 170.1 1 0.0001 0.1475
Ever used illicit methadone 261 (5.1) 284 (10.4) 77.5 1 0.0001 0.0996
Ever used other opiates 918 (18.0) 838 (30.1) 167.0 1 0.0001 0.1462
Ever used amphetamines 1992 (39.1) 1695 (62.3) 385.7 1 0.0001 0.2221
Ever used Rohypnol 77 (1.8) 72 (3.1) 13.1 1 0.0003 0.0443
Open in a new tab
*

z statistic was reported for Wilcoxon test for ordinal variables.

†

ES=Effect size. For χ2 was Cramer’s V. For t and Wilcoxon was η2.

‡

Income scale was 1=Less than $500, 2=$500 – $999, 3=$1,000–$1,999, 4=$2,000–$3,999, 5=$4,000–$5,999, 6=$6,000 or more per month.

§

Education scale was 00=No formal schooling, 1=Eight grade or less, 2=Less than high school, 3=A GED (high school equivalency), 4= High school graduation, 5=Trade or technical training, 6=Some college, 7=College graduation. ‖Only asked of females.

Speedball = Combination of heroin and cocaine.

Step 2 involves fitting the multivariable model and eliminating variables that do not contribute to the model. Step 3 involves adding back to the model any variables that were eliminated in Step 2 but subsequently found to be important because they provide a needed adjustment of those variables that remain in the model. Step 4 is to continue this process of adding and taking out variables until there is a preliminary main effects model.

Step 5 is to check assumptions and to make sure that continuous variables are linear in the logit resulting in the main effects model. Step 6 is to check for interactions because an interaction implies that the effect of each variable is not constant over the levels of the other variable. This step results in the preliminary final model. Step 7 is to check the model fit using methods such as the Hosmer-Lemeshow goodness-of-fit method 29.

All analyses were performed using SAS statistical software, version 9.4 (TS1M6) (SAS Institute, Cary, NC). The analyses were run on a virtual Windows Server 2016 server.

RESULTS

This study used the Frailty Framework for Vulnerable Populations to determine if the three domains for frailty, psychological, social, and physical could be used to investigate hospitalization among those who reported and those who did not report experiencing homelessness. The baseline and follow-up data were collected August 2000 through July 2014 from 4916 participants who did not report homelessness and whose mean age was 37.9 years, 29% female, and 2692 participants who reported experiencing homelessness and whose mean age was 42.1 years, 29% female. Table 1 shows comparisons between those reporting that they were experiencing homelessness at baseline and those reporting that they were not homeless at baseline. Those who were experiencing homelessness were significantly older, poorer, less well educated, and spent a significantly longer part of their lives being incarcerated. They were also significantly more likely to identify as Black. The large effect sizes were for age, income, incarceration time, and self-identified Black race/ethnicity.

A greater proportion of those experiencing homelessness tested positive for hepatitis B, hepatitis C, and the Treponema pallidum – particle agglutination test, even though they did not for the Rapid Plasma Reagin test for syphilis. Those experiencing homelessness in our sample were not more likely to test positive for hepatitis A, even though that has been reported for other samples reporting homelessness in California.30–32 The non-homeless in our sample were more likely to test positive on the laboratory test for chlamydia, even though the self-report was the opposite in that it was those reporting homelessness who reported more chlamydia. All of the other self-report of the diseases showed those experiencing homelessness being significantly more likely to report that they had been positive for the disease, with the exception of genital warts. The large effect sizes were for Trichomonas and genital warts which were only asked of the female participants.

Table 1 shows that those experiencing homelessness were significantly more likely to report ever use of all illicit drugs listed, with the largest effect sizes for crack, heroin, and amphetamines. There were also large effect sizes for marijuana, cocaine, speedball (combination of heroin and cocaine), and other opiates. The general impression from this part of Table 1 is that it does not matter which illicit drug the reader is looking at, in all cases a significantly larger proportion of those experiencing homelessness reported use of the drug.

Table 2 shows the multivariable logistic regression in which baseline data from the combined sample was used to predict hospitalization at follow-up. The model shows that being experiencing homelessness at baseline is a significant predictor for hospitalization at follow-up. The model also shows that being female and identifying as Black compared to White are predictors of hospitalization. Hispanic and Other race/ethnicity are not significantly different from White as indicated by the probability values and the confidence intervals which include 1. There was an interaction between use of crack and receiving income from Social Security or disability in that if someone used crack but did not receive income from Social Security or disability, then the odds ratio of being hospitalized at follow-up was only 1.52. However, if the person not only used crack, but also received income from Social Security or disability, then the odds ratio went up to 3.03. The model fits very well based on the Hosmer-Lemeshow goodness-of-fit statistics.

TABLE 2.

Logistic Regression Model Predicting Hospitalization at Follow-Up

OR 95% CI P
Homeless 1.60 1.27–2.03 <0.01
Sex
 Male Ref.
 Female 1.35 1.07–1.71 0.01
Race
 White Ref.
 Black 1.31 1.01–1.69 0.04
 Hispanic 0.75 0.53–1.06 0.10
 Other 0.80 0.50–1.71 0.36
Interaction of Crack use and Income from Social Security or Disability* <0.01
 Crack at Income = 0 1.52 1.16–2.00
 Crack at Income = 1 3.03 2.24–4.11
Open in a new tab

Hosmer-Lemeshow goodness-of-fit test χ2 (df=7) = 1.6793, P=0.9754.

*

Where Income = 0 means that no income from Social Security or Disability was obtained. Where Income = 1 means that income was obtained from Social Security or Disability.

OR=odds ratio. Ref.=reference. CI=confidence interval.

DISCUSSION

This manuscript is an application of the Frailty Framework among Vulnerable Populations (FFVP).16 In this application of the theoretical model, our vulnerable population is those reporting that they were experiencing homelessness at baseline. They are being compared to those who did not report experience homelessness at baseline. The FFVP organizes risk factors into individual-level factors, life events, and health-related factors. The individual-level factors include race/ethnicity, gender, income, and education, specifically those who identified as Black, female gender, lower income, and lower education are the individual-level factors that put those experiencing homelessness at risk. Our data is perfectly consistent with this model in that Table 1 shows self-identified Black race and female gender to be associated with experiencing homelessness. Also, those experiencing homelessness in our sample had significantly lower income and lower education than those reporting that they were not experiencing homelessness. Table 2 shows that those who identified as Black and female gender to be risk factors for hospitalization at follow-up.

The FFVP lists life events such as incarceration as a major factor leading to frailty. The notion is that incarceration is associated with poor access to healthcare while incarcerated and lack of access to community resources for healthcare once released, which affect health-related factors that contribute to frailty 33. The portion of our sample which reported experiencing homelessness was incarcerated over 900 days more than those who did not report experiencing homelessness over their lifetime. The greater time incarcerated leads to more difficulty in reintegrating back into society.

The health-related factors are the health conditions that are significant factors in this population. Most of the laboratory tests show those experiencing homelessness in our sample to have significantly higher proportions being positive for the different diseases that were tested. The self-report data show significant associations with experiencing homelessness for almost all of the conditions that were inquired about on the RBA. A recent review of the literature found that rates of sexually transmitted infections among adults experiencing homelessness are high.34 Among unstably housed women, drug use was associated with condomless sex, multiple sex partners, psychiatric co-morbidities and high rates of sexually transmitted infections.35 The one infection that was not significantly different between those experiencing homelessness and non-homeless in our data was hepatitis A. This has been reported by others to be associated with experiencing homelessness even though we did not find it in our Long Beach sample.30–32

The behavioral factors that the FFVP lists include illicit drug use that puts vulnerable populations such as those experiencing homelessness at greater risk for frailty. A study of those experiencing homelessness in the Los Angeles skid row showed drug and alcohol abuse to be related to health problems at a walk-in clinic for those experiencing homelessness.12 The findings on Table 1 shows significantly greater proportions of those experiencing homelessness using all drugs more than those who were not experiencing homelessness which is definitely consistent with the FFVP; the number of illicit drugs listed demonstrates this behavioral risk that those experiencing homelessness engage in that leads to frailty. This is consistent with previous research on hospitalization of those experiencing homelessness being associated with substance use with and without co-occurring mental health disorders. 36 We wanted to present a table that listed all of the drugs and show that they were each associated with experiencing homelessness in our sample. It was not clear from the literature which drugs were included in the various reports on substance use among those experiencing homelessness and whether there was any comparison group in the analysis. Given all of the illicit drugs that were associated with experiencing homelessness in Table 1, it is interesting that it was crack use that was the drug that was included in the multivariable model. The associations between crack use and experience of homelessness are consistent with studies where crack use was both an antecedent of homelessness as well as a consequence of it. 37–39

An adverse event of frailty can be disability according to the FFVP. Our findings show a major effect of obtaining income from Social Security, disability, or workmen’s compensation. There have been associations between disability and reports of homelessness reported in other studies.40 A study of types of discharges among veterans of military service in Afghanistan and Iraq found that disability-discharged veterans had significantly higher rates of experiencing homelessness compared to routine discharges and that having a disability discharge was a risk factor for reporting homelessness.41

One of the antecedents for the FFVP is the Working Framework for Understanding Frailty (WFUF) in which frailty leads to adverse outcomes such as hospitalization.42 The FFVP also makes hospitalization one of the main outcomes of frailty which logically leads to our using hospitalization as the outcome in the logistic regression model in this paper. Dealing with individuals experiencing homelessness has been reported multiple times as a problem for hospitals in that hospital readmission rates are used as national benchmarks of the quality of a hospital and hospitals go to great lengths to reduce their readmission rates. However, having patients who are experiencing homelessness at admission is problematic for hospitals that are striving for excellence in their quality indicators such as readmission rates. In one hospital a matched sample comparing individuals experiencing homelessness to non-homeless, the patients reporting experience of homelessness had nearly 4 times the odds of being readmitted compared to the non-homeless.43 Veterans experiencing homelessness have been reported to be more likely to be readmitted to the hospital in a study of veteran patients.10 It has been deemed to be such a problem that is it termed a “revolving door” in one study that found a strikingly high hospital readmission rate among their patients who reported experiencing homelessness.2

Limitations

There are several limitations to our study. One limitation is that we do not have the admitting diagnosis for the hospital stays among our participants. Previous research has found that substance use and mental disorders are among the most frequent reasons for hospitalization among those experiencing homelessness; however, researchers have also noted that individuals experiencing homelessness are also hospitalized for traumatic injuries resulting from sexual assault and beatings.44,45 Alcohol and drug-related hypothermia in individuals experiencing homelessness can also result in hospitalization or death. 46 Another limitation is that the list of diseases that the RBA asks about and that we had funding to test for were fairly restricted. We also did not confirm the drug use with urinalysis. However, our study took place in a very high-risk location in which the people who came to our center were at high-risk for infectious diseases and low-income. Finally, while the follow-up interviews were usually conducted six months after the baseline interview, the intervals for some participants were much longer than six months. Reasons for this include long-term incarceration and extended periods when individuals were experiencing homelessness, hence they had life events which prevented them from promptly responding to reminders. We did not deny any eligible participant the opportunity to complete their follow-up interview, regardless how late they were.

CONCLUSIONS

This study demonstrates that the Frailty Framework among Vulnerable Populations can be applied to the population experiencing homelessness to better understand their risk factors and adverse outcomes. Risk factors that were significant for those experiencing homelessness in our sample were lower education, lower income, those who identified as Black, and female gender. In addition, findings from this study show the importance of illicit drug use in creating frailty in those experiencing homelessness that leads to hospitalization. When patients reporting homelessness are admitted to a hospital, the medical personnel treat the illness and should also include substance abuse treatment program coordination in the discharge planning for these patients. This may alleviate the “revolving door” problem of readmission of those experiencing homelessness to the hospital. The expansion of International Classification of Diseases (ICD) 10 coding to include Z codes associated with social determinants of health, such as the experience of homelessness, should also facilitate identification of patients experiencing social disparities which can negatively impact their follow-up treatment. 44

Funding Sources:

The project described was supported in part by Award Numbers R01DA030234 from the National Institute on Drug Abuse (NIDA), P20MD003942 from the National Institute of Minority Health and Health Disparities (NIMHD), and ID10-CSULB-008 from the California HIV Research Program (CHRP). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIDA, NIMHD, or CHRP. The NIDA, NIMHD, nor CHRP had no role in the study design, collection or analysis or interpretation of the data, writing of the manuscript, or the decision to submit the paper for publication.

Footnotes

No conflict of interest exists for any author for the past three years.

REFERENCES

  • 1.Fazel S, Geddes JR, Kushel M. The health of homeless people in high-income countries: descriptive epidemiology, health consequences, and clinical and policy recommendations. Lancet. 2014;384(9953):1529–1540. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Doran KM, Ragins KT, Iacomacci AL, Cunningham A, Jubanyik KJ, Jenq GY. The revolving hospital door: hospital readmissions among patients who are homeless. Med Care. 2013;51(9):767–773. [DOI] [PubMed] [Google Scholar]
  • 3.Mackelprang JL, Qiu Q, Rivara FP. Predictors of Emergency Department Visits and Inpatient Admissions Among Homeless and Unstably Housed Adolescents and Young Adults. Med Care. 2015;53(12):1010–1017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Williams SP, Bryant KL. Sexually Transmitted Infection Prevalence among Homeless Adults in the United States: A Systematic Literature Review. Sex Transm Dis. 2018;45(7):494–504. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Degenhardt L, Peacock A, Colledge S, et al. Global prevalence of injecting drug use and sociodemographic characteristics and prevalence of HIV, HBV, and HCV in people who inject drugs: a multistage systematic review. Lancet Glob Health. 2017;5(12):e1192–e1207. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Noska AJ, Belperio PS, Loomis TP, O’Toole TP, Backus LI. Prevalence of Human Immunodeficiency Virus, Hepatitis C Virus, and Hepatitis B Virus Among Homeless and Nonhomeless United States Veterans. Clin Infect Dis. 2017;65(2):252–258. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Barros CVL, Galdino Junior H, Rezza G, et al. Bio-behavioral survey of syphilis in homeless men in Central Brazil: a cross-sectional study. Cad Saude Publica. 2018;34(6):e00033317. [DOI] [PubMed] [Google Scholar]
  • 8.Caton CL, Dominguez B, Schanzer B, et al. Risk factors for long-term homelessness: findings from a longitudinal study of first-time homeless single adults. Am J Public Health. 2005;95(10):1753–1759. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Wadhera RK, Choi E, Shen C, Yeh RW, Joynt Maddox KE. Trends, Causes, and Outcomes of Hospitalizations for Homeless Individuals: A Retrospective Cohort Study. Med Care. 2019;57(1):21–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Titan A, Graham L, Rosen A, et al. Homeless Status, Postdischarge Health Care Utilization, and Readmission After Surgery. Med Care. 2018;56(6):460–469. [DOI] [PubMed] [Google Scholar]
  • 11.Bhalla IP, Stefanovics EA, Rosenheck RA. Clinical Epidemiology of Single Versus Multiple Substance Use Disorders: Polysubstance Use Disorder. Med Care. 2017;55 Suppl 9 Suppl 2:S24–s32. [DOI] [PubMed] [Google Scholar]
  • 12.Vredevoe DL, Brecht ML, Shuler P, Woo M. Risk factors for disease in a homeless population. Public Health Nurs. 1992;9(4):263–269. [DOI] [PubMed] [Google Scholar]
  • 13.Koegel P, Sullivan G, Burnam A, Morton SC, Wenzel S. Utilization of mental health and substance abuse services among homeless adults in Los Angeles. Med Care. 1999;37(3):306–317. [DOI] [PubMed] [Google Scholar]
  • 14.Guerrero EG. Examination of treatment episodes among women and racial and ethnic minorities in addiction treatment. Journal of Social Work Practice in the Addictions. 2013;13(3):227–243. [Google Scholar]
  • 15.Fisher DG, Reynolds GL, D’Anna LH, Hosmer DW, Hardan-Khalil K. Failure to get into substance abuse treatment. J Subst Abuse Treat. 2017;73:55–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Salem BE, Nyamathi A, Phillips LR, Mentes JC, Sarkisian C, Brecht ML. Development of a frailty framework among vulnerable populations. ANS Adv Nurs Sci. 2014;37(1):70–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Salem BE, Nyamathi A, Phillips LR, Mentes J, Sarkisian C, Brecht L. Identifying frailty among vulnerable populations. ANS Advances in Nursing Science. 2014;37(1):70. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Salem BE, Brecht M-L, Ekstrand ML, Faucette M, Nyamathi AM. Correlates of physical, psychological, and social frailty among formerly incarcerated, homeless women. Health Care for Women International. 2019;40(7–9):788–812. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Salem BE, Ma-Pham J, Chen S, Brecht M-L, Antonio AL, Ames M. Impact of a Community-Based Frailty Intervention Among Middle-Aged and Older Prefrail and Frail Homeless Women: A Pilot Randomized Controlled Trial. Community Mental Health Journal. 2017;53(6):688–694. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Rogans-Watson R, Shulman C, Lewer D, Armstrong M, Hudson B. Premature frailty, geriatric conditions and multimorbidity among people experiencing homelessness: a cross-sectional observational study in a London hostel. Housing, Care and Support. 2020;23(3/4):77–91. [Google Scholar]
  • 21.Dowling-Guyer S, Johnson ME, Fisher DG, et al. Reliability of drug users’ self-reported HIV risk behaviors and validity of self-reported recent drug use. Assessment. 1994;1:383–392. [Google Scholar]
  • 22.Needle R, Fisher DG, Weatherby N, et al. Reliability of self-reported HIV risk behaviors of drug users. Psychology of Addictive Behaviors. 1995;9:242–250. [Google Scholar]
  • 23.Klahn JA, Fisher DG, Wood MM, Reynolds GL, Johnson ME. Homelessness among out-of-treatment drug users in Long Beach, CA. Journal of Social Distress and Homeless. 2004;13(3&4):345–368. [Google Scholar]
  • 24.Johnson ME, Fisher DG, Reynolds GL. Reliability of drug users’ self-report of economic variables. Addiction Research. 1999;7:227–238. [Google Scholar]
  • 25.Johnson ME, Fisher DG, Montoya I, et al. Reliability and validity of not-in-treatment drug users’ follow-up self-reports. AIDS and Behavior. 2000;4(4):373–380. [Google Scholar]
  • 26.Fisher DG, Reynolds GL, Johnson ME, Van Otterloo L. Reliability of health utilization questions of illicit drug users. 98th Annual Convention of the Western Psychological Association; 2018; Portland, OR. [Google Scholar]
  • 27.Gelberg L, Siecke N. Accuracy of homeless adults’ self-reports. Med Care. 1997;35(3):287–290. [DOI] [PubMed] [Google Scholar]
  • 28.Hosmer DW, Lemeshow S, Sturdivant RX. Applied logistic regression. Third ed. Hoboken, NJ: John Wiley & Soncs, Inc.; 2013. [Google Scholar]
  • 29.Hosmer DW, Lemeshow S. A goodness-of-fit test for the multiple logistic regression model. Communications in Statistics. 1980;A10:1043–1069. [Google Scholar]
  • 30.Foster M, Ramachandran S, Myatt K, et al. Hepatitis A Virus Outbreaks Associated with Drug Use and Homelessness - California, Kentucky, Michigan, and Utah, 2017. MMWR Morb Mortal Wkly Rep. 2018;67(43):1208–1210. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Hennessey KA, Bangsberg DR, Weinbaum C, Hahn JA. Hepatitis A seroprevalence and risk factors among homeless adults in San Francisco: should homelessness be included in the risk-based strategy for vaccination? Public Health Rep. 2009;124(6):813–817. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Peak CM, Stous SS, Healy JM, et al. Homelessness and Hepatitis A - San Diego County, 2016–2018. Clin Infect Dis. 2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Ahmed RA, Angel C, Martell R, Pyne D, Keenan L. The impact of homelessness and incarceration on women’s health. Journal of Correctional Health Care. 2016;22(1):62–74. [DOI] [PubMed] [Google Scholar]
  • 34.Williams SP, Bryant KL. Sexually Transmitted Infection Prevalence among Homeless Adults in the United States: A Systematic Literature Review. Sexually transmitted diseases. 2018;45(7):494–504. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Meacham MC, Bahorik AL, Shumway M, Marquez C, Riley ED. Condomless Sex and Psychiatric Comorbidity in the Context of Constrained Survival Choices: A Longitudinal Study Among Homeless and Unstably Housed Women. AIDS and behavior. 2019;23(3):802–812. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Lin W-C, Bharel M, Zhang J, O’Connell E, Clark RE. Frequent Emergency Department Visits and Hospitalizations Among Homeless People With Medicaid: Implications for Medicaid Expansion. American Journal of Public Health. 2015;105(S5):S716–S722. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Royse D, Leukefeld C, Logan T, et al. Homelessness and gender in out-of-treatment drug users. The American journal of drug and alcohol abuse. 2000;26(2):283–296. [DOI] [PubMed] [Google Scholar]
  • 38.Orwin RG, Scott CK, Arieira C. Transitions through homelessness and factors that predict them: three-year treatment outcomes. Journal of substance abuse treatment. 2005;28(2):S23–S39. [DOI] [PubMed] [Google Scholar]
  • 39.Johnson RA. African Americans and homelessness: Moving through history. Journal of Black Studies. 2010;40(4):583–605. [Google Scholar]
  • 40.Cimino T, Steinman MA, Mitchell SL, et al. The Course of Functional Impairment in Older Homeless Adults: Disabled on the Street. JAMA Internal Medicine. 2015;175(7):1237–1239. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Fargo JD, Brignone E, Metraux S, et al. Homelessness following disability-related discharges from active duty military service in Afghanistan and Iraq. Disability and health journal. 2017;10(4):592–599. [DOI] [PubMed] [Google Scholar]
  • 42.Bergman H, Béland F, Karunananthan S, Hummel S, Hogan D, Wolfson C. Développement d’un cadre de travail pour comprendre et étudier la fragilité. Gérontologie et Societe. 2004;109:15–29. [Google Scholar]
  • 43.Saab D, Nisenbaum R, Dhalla I, Hwang SW. Hospital Readmissions in a Community-based Sample of Homeless Adults: a Matched-cohort Study. J Gen Intern Med. 2016;31(9):1011–1018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Bensken WP. How do we define homelessness in large health care data? Identifying variation in composition and comorbidities. Health Services and Outcomes Research Methodology. 2020. [Google Scholar]
  • 45.Padgett DK, Struening EL. Victimization and traumatic injuries among the homeless: associations with alcohol, drug, and mental problems. American Journal of Orthopsychiatry. 1992;62(4):525–534. [DOI] [PubMed] [Google Scholar]
  • 46.Turk EE. Hypothermia. Forensic Science, Medicine, and Pathology. 2010;6(2):106–115. [DOI] [PubMed] [Google Scholar]

RESOURCES