Morbidity and Mortality Among Adults Experiencing Homelessness Hospitalized With COVID-19

Abstract

People experiencing homelessness (PEH) are at higher risk for chronic health conditions, but clinical characteristics and outcomes for PEH hospitalized with coronavirus disease 2019 (COVID-19) are not known. We analyzed population-based surveillance data of COVID-19–associated hospitalizations during 1 March to 31 May 2020. Two percent of the people hospitalized with COVID-19 for whom a housing status was recorded were homeless. Of 199 cases in the analytic sample, most were of racial/ethnic minority groups and had underlying health conditions. Clinical outcomes such as ICU admission, respiratory support including mechanical ventilation, and deaths were documented. Hispanic and non-Hispanic black persons accounted for most mechanical ventilation and deaths. Severe illness was common among persons experiencing homelessness who were hospitalized with COVID-19.

homelessness, COVID-19, SARS-CoV-2, clinical outcomes

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the virus that causes coronavirus disease 2019 (COVID-19), can spread easily among people in close contact with one another, causing infections ranging from asymptomatic to fatal [1]. Older adults and those with underlying health conditions are at increased risk for severe illness from COVID-19 [2]. People experiencing homelessness often have health conditions common to older adults (eg, cognitive impairment, frailty, depression, and difficulties with daily living activities), but develop these conditions at a younger age than the general population [3]. Additionally, chronic conditions like hypertension, diabetes, respiratory diseases, and cardiovascular disease are more prevalent among people experiencing homelessness than in the general population, and are often untreated [4, 5]. Thus, people experiencing homelessness, particularly those living in close quarters, could be at increased risk for severe COVID-19.

Point-prevalence COVID-19 testing among people experiencing homelessness have shown a high proportion of asymptomatic infections at the time of testing, raising questions about the severity of COVID-19 in this population [6, 7]. However, drawing conclusions is difficult; there have been few comparable surveys conducted in the general population. One study at an urban medical center showed that people experiencing homelessness were more likely to be hospitalized than the general population, but other outcomes were similar [8]. Further data are needed to elucidate the clinical course of COVID-19 among people experiencing homelessness.

We used data from the Centers for Disease Control and Prevention (CDC) COVID-19–Associated Hospitalization Surveillance Network (COVID-NET) to evaluate the clinical characteristics and outcomes of COVID-19–associated hospitalizations among persons experiencing homelessness.

METHODS

COVID-NET is a population-based surveillance system that collects data on laboratory-confirmed COVID-19–associated hospitalizations [9, 10]. The current network includes 99 counties in 10 Emerging Infections Program sites (California, Colorado, Connecticut, Georgia, Maryland, Minnesota, New Mexico, New York, Oregon, and Tennessee) and 4 additional states through the Influenza Hospitalization Surveillance Project (Iowa, Michigan, Ohio, and Utah) [10]. COVID-NET represents approximately 10% of the US population. In participating sites, trained surveillance officers identified laboratory-confirmed COVID-19–associated hospitalizations by reviewing hospital, laboratory, and admission databases, and infection control logs for hospitalized patients. They completed standardized case report forms based on medical record reviews for a random sample of identified cases. Cases in the current analysis had a documented positive SARS-CoV-2 result by reverse transcription-polymerase chain reaction during hospitalization or up to 14 days before admission.

We included data from patients aged ≥ 18 years hospitalized with COVID-19 during 1 March to 31 May 2020 and identified in the medical record as being homeless or residing at a shelter at the time of hospitalization in 13 US states (Iowa did not report cases with documented homelessness). We assessed patient demographic and clinical characteristics (eg, age, race/ethnicity, underlying health conditions, symptoms at admission), and outcomes: intensive care unit (ICU) admission, length of hospitalization, ventilator use, vasopressor support, systemic corticosteroid use, and death during hospitalization. We assessed outcome distributions by key characteristics. Descriptive statistics including unweighted frequencies and weighted percentages were generated for population-based estimates; weights were applied to reflect the probability of being sampled for complete chart abstractions. Confidence intervals (95% CI) were calculated using Taylor series linearization method. Analyses were performed using SAS 9.4.

This activity was reviewed by CDC and was conducted consistent with applicable federal law and CDC policy (see eg, 45 CFR part 46.102(l)(2), 21 CFR part 56; 42 USC §241(d); 5 USC §552a; 44 USC §3501, et seq). Participating sites obtained approvals for the COVID-NET surveillance protocol from their Institutional Review Boards, as applicable.

RESULTS

Among 28 917 cases reported in COVID-NET during 1 March to 31 May 2020, 70% (20 189) were missing data on homelessness status. Of 8728 cases with information about housing at the time of hospital admission, 201 (2%; 95% CI, 1.5%–2.4%) were identified as experiencing homelessness through complete chart abstractions. We excluded 2 cases aged < 18 years, leaving n = 199 for analysis. The median age of patients experiencing homelessness hospitalized with COVID-19 was 53 years (interquartile range [IQR] = 48–58). Most were male (n = 165, 84%), non-Hispanic black, Hispanic, or non-Hispanic other race/ethnicity (n = 126, 62%), and had health insurance (n = 163, 85%; Table 1). At least 1 underlying health condition was recorded for 83% of patients (n = 155); the most common health conditions were hypertension (n = 77, 44%), obesity (body mass index [BMI] ≥ 30 kg/m²; n = 59, 24%), chronic lung disease (n = 43, 22%), cardiovascular disease (n = 36, 25%), and diabetes (n = 32, 16%). Current tobacco smoking (n = 105, 46%) and alcohol abuse (n = 70, 34%) were common, and almost 1 in 12 was diagnosed with a mental health disorder (n = 38, 8%) based on International Classification of Diseases 10th revision (ICD-10) discharge diagnosis codes.

Table 1.

Characteristics of People Experiencing Homelessness Hospitalized With COVID-19, 1 March to 31 May 2020 (n = 199)

Characteristics	Unweighted, n	Weighted, %
Age, y, median, IQR	53 (48–58)
18–49	101	40.6
50–64	69	38.5
65+	29	20.9
Race/ethnicity
Non-Hispanic white	68	36.9
Non-Hispanic black	69	33.4
Hispanic or Latino	38	23.8
Non-Hispanic othera	19	4.3
Unknown	5	1.6
Sex
Male	165	83.5
Female	34	16.5
Geographic regionb
West	83	49.5
Midwest	53	10.6
South	44	31.8
Northeast	19	8.1
Insurance
Yes	163	84.5
No	34	15.2
Unknown	2	0.4
Smoking
Current	105	45.9
Former	33	21.2
Never	61	32.9
Alcohol abuse
Current	70	33.7
Former	17	10.4
Never	112	55.9
Substance abuse or dependencec
Yes	17	3.1
No	182	96.9
Mental disorderd
Yes	38	8.4
No	161	91.6
Body mass indexe
Underweight	10	5.1
Normal	50	27.6
Overweight	57	33.2
Obese	59	23.6
Unknown	23	10.5
Pregnantf	1	3.3
Underlying health conditions
Any condition	155	83.0
No condition	43	15.4
Hypertension	77	43.9
Chronic lung disease	43	22.3
Asthma	26	16.7
Emphysema/COPD	19	9.5
Chronic metabolic disease	37	18.6
Diabetes mellitus	32	15.9
Cardiovascular diseaseg	36	24.9
Coronary artery disease	12	6.5
Congestive heart failure	12	10.3
Cardiomyopathy	5	4.6
Myocardial infarction	4	2.8
Pulmonary embolism	4	4.1
Peripheral vascular disease	4	4.6
Cerebral vascular accident	3	3.1
Peripheral artery disease	1	1.9
Gastrointestinal/liver disease	34	18.7
Hepatitis C (HCV)	21	12.8
Cirrhosis/end-stage liver disease	10	7.8
Alcoholic hepatitis	5	0.9
Hepatitis B (HBV)	3	0.5
Chronic liver disease	1	0.2
Neurologic disorder	28	23.8
Epilepsy/seizure/seizure disorder	14	13.2
Traumatic brain injury	9	7.8
Dementia/Alzheimer disease	3	2.1
Cognitive dysfunction	2	0.4
Immunosuppressive condition	6	2.0
Renal disease	9	5.3
Blood disorders/hemoglobinopathy	6	4.5
Rheumatologic/autoimmune disease	2	1.3
Other
Wheelchair dependent	4	5.6

Weights were applied to reflect the probability of being sampled for complete chart abstraction; weighted percentages and unweighted case counts are presented.

Abbreviations: BMI, body mass index; COPD, chronic obstructive pulmonary disease; COVID-19, coronavirus disease 2019; HBV, hepatitis B virus; HCV, hepatitis C virus; ICD-10, International Classification of Diseases 10th revision; IQR, interquartile range.

^aOther race groups were Asian or Pacific Islander, American Indian or Alaska Native, Multiracial.

^bGeographic regions are west (California, Colorado, New Mexico, Oregon, Utah), Midwest (Michigan, Minnesota, Ohio), northeast (Connecticut, New York), south (Georgia, Maryland, Tennessee).

^cSubstance abuse or dependence based on ICD-10 discharge diagnosis codes for abuse/dependence of opioids, cocaine, or other stimulants.

^dMental disorder based on ICD-10 discharge diagnosis codes for schizophrenia, schizoaffective disorder, bipolar disorder, major depressive disorder, mood disorder, other anxiety disorder, posttraumatic stress disorder, personality disorder, attention deficit hyperactivity disorder.

^eBMI based on calculated weight and height (kg/m²), and if BMI is missing, by ICD-10 discharge diagnosis codes: BMI underweight (less than 18.5), normal (18.5–24.9), overweight (25.0–29.9), obese (30.0 and above).

^fPregnant, restricted to women aged 18–49 years (n = 16).

^gCardiovascular disease excludes hypertension.

Most patients had signs and symptoms at admission (n = 168, 92%), including cough (n = 107, 54%), fever/chills (n = 102, 53%), and shortness of breath (n = 89, 51%) (findings not shown in tables). Other symptoms such as chest pain, headache, nausea/vomiting, and myalgias were less common (20%–23%). Half of patients (n = 102, 54%) were hospitalized for >4 days (median = 4, IQR = 2–12), and 1 in 6 persons (n = 37, 17%) were admitted to the ICU (Table 2). Invasive mechanical ventilation (n = 18, 11%), vasopressor support (n = 11, 6%), systemic steroids (n = 20, 16%), and renal replacement therapy or dialysis (n = 4, 4%) were documented. Invasive mechanical ventilation was most common among patients aged ≥ 65 years (n = 3, 20%), and patients with no underlying health conditions (n = 8, 21%). Six patients (1%) died—all were symptomatic; most were male, had underlying health conditions, and were aged ≥ 50 years. Hispanic persons and non-Hispanic black persons accounted for most documented mechanical ventilation and deaths.

Table 2.

Prevalence of Clinical Outcomes by Key Characteristics Among People Experiencing Homelessness Hospitalized With COVID-19, 1 March to 31 May 2020

Characteristics	Hospital Length of Stay Above Median, >4 da	ICU Admission	Highest Level of Respiratory Support				Vasopressors	Renal Replacement Therapy or Dialysis	Systemic Steroids	Death During Hospitalization
			Invasive Mechanical Ventilation	BiPAP/CPAP	HFNC	Any Respiratory Support, HFNC, BiPAP/CPAP, Ventilator
Overall, (n = 199)	102 (54.0)	37 (16.9)	18 (11.3)	3 (2.6)	6 (1.1)	27 (15.0)	11 (6.4)	4 (3.6)	20 (15.5)	6 (1.1)
Race and ethnicity
White NH (n = 68)	33 (49.5)	13 (16.0)	4 (11.6)	1 (0.5)	2 (1.0)	7 (13.0)	5 (7.1)	1 (0.5)	10 (17.0)	2 (1.0)
Black NH (n = 69)	38 (58.3)	12 (15.1)	7 (11.6)	2 (7.2)	0 (0.0)	9 (18.7)	4 (4.7)	3 (10.2)	6 (8.5)	2 (1.1)
Hispanic or Latino (n = 38)	18 (59.1)	9 (22.5)	6 (12.6)	0 (0.0)	2 (1.5)	8 (14.1)	2 (9.5)	0 (0.0)	3 (26.2)	2 (1.5)
Other NH (n = 19)	11 (46.7)	2 (8.5)	1 (4.2)	0 (0.0)	1 (4.2)	2 (8.5)	0 (0.0)	0 (0.0)	1 (4.2)	0 (0.0)
Age, y
18–49 (n = 101)	50 (54.9)	18 (22.1)	8 (8.1)	2 (5.9)	4 (1.8)	14 (15.8)	3 (1.3)	2 (5.9)	8 (22.1)	1 (0.4)
50–64 (n = 69)	34 (43.1)	14 (13.2)	7 (9.9)	0 (0.0)	1 (0.5)	8 (10.4)	6 (9.4)	2 (3.1)	9 (11.1)	3 (1.4)
≥65 (n = 29)	18 (71.4)	5 (13.4)	3 (20.1)	1 (0.9)	1 (0.9)	5 (21.8)	2 (10.8)	0 (0.0)	3 (11.0)	2 (1.7)
Sex
Male (n = 165)	84 (52.4)	31 (14.6)	16 (11.0)	2 (2.9)	5 (1.1)	23 (15.0)	9 (5.2)	4 (4.3)	14 (14.2)	5 (1.1)
Female (n = 34)	18 (61.9)	6 (28.1)	2 (12.6)	1 (1.1)	1 (1.1)	4 (14.8)	2 (12.6)	0 (0.0)	6 (22.6)	1 (1.1)
Overweight or obese BMIb
Yes (n = 116)	58 (50.5)	23 (18.4)	10 (11.0)	0 (0.0)	4 (1.3)	14 (12.3)	7 (10.1)	2 (2.1)	12 (18.5)	4 (1.3)
No (n = 60)	36 (64.1)	11 (12.4)	7 (14.9)	3 (7.9)	1 (0.6)	11 (23.3)	4 (2.2)	2 (7.4)	6 (8.7)	1 (0.6)
Any symptoms
Yes (n = 168)	85 (53.6)	29 (16.8)	15 (11.7)	2 (2.6)	5 (1.0)	22 (15.3)	11 (7.0)	4 (3.9)	20 (16.9)	6 (1.2)
No (n = 31)	17 (59.2)	8 (18.1)	3 (6.8)	1 (2.3)	1 (2.3)	5 (11.3)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Any underlying health conditions
Yes (n = 155)	77 (54.3)	24 (17.5)	10 (9.7)	2 (2.9)	4 (0.9)	16 (13.4)	8 (7.1)	4 (4.3)	17 (18.1)	5 (1.1)
No (n = 43)	24 (46.5)	13 (15.4)	8 (21.3)	1 (1.2)	2 (2.4)	11 (24.9)	3 (3.5)	0 (0.0)	3 (3.5)	1 (1.2)

Data are unweighted case counts (weighted %).

Weights were applied to reflect the probability of being sampled for complete chart abstraction. For mechanical ventilation, BiPAP/CPAP, and HFNC, patients were assigned based on the highest level of respiratory support required during hospitalization (ie, invasive mechanical ventilation, then BiPAP or CPAP, then HFNC). Subgroup totals may not add to 199 due to missing data.

Abbreviations: BiPAP, bilevel positive airway pressure; BMI, body mass index; COVID-19, coronavirus disease 2019; CPAP, continuous positive airway pressure; HFNC, high flow nasal cannula; ICU, intensive care unit; NH, non-Hispanic.

^aHospital length of stay is calculated from earliest admission date to latest discharge date in days.

^bBMI based on calculated weight and height (kg/m²), and if BMI is missing, by ICD-10 discharge diagnosis codes: BMI underweight (less than 18.5), normal (18.5–24.9), overweight (25.0–29.9), obese (30.0 and above).

Discussion

We used weighted population estimates to evaluate characteristics and outcomes of hospitalized patients with COVID-19 experiencing homelessness. Most patients were male, from racial/ethnic minority groups, and had underlying health conditions. Most patients were symptomatic and were often hospitalized for >4 days.

In a previously published COVID-NET analysis, hospitalized COVID-19 cases in the general population had a higher prevalence of underlying health conditions (92%), but were generally older; among 2491 hospitalizations, median age was 62 years (IQR = 50–75) [11]. In our analysis of patients experiencing homelessness, the median age was 53 years, and there were greater proportions of people from racial/ethnic minorities and current smokers. Despite the younger age, it is notable that 83% of our sample had underlying health conditions. Furthermore, outcomes in the general population, namely ICU admission and mechanical ventilation, did not differ by race/ethnicity, and white patients had higher prevalence of in-hospital death than other racial/ethnic groups. In our analysis, Hispanic and non-Hispanic black persons accounted for most documented mechanical ventilation and deaths. Future studies could explore the incremental contribution of homelessness to hospitalizations and clinical outcomes related to COVID-19. While sample size was too low to include children and youth experiencing homelessness in the current analysis, additional information is needed to better understand the collective influence of homelessness on COVID-19 from a public health and clinical perspective.

This analysis is subject to limitations. Data on housing status were missing from most patients at the time of analysis, as hospital records often do not have housing information. The number of hospitalized COVID-19 patients experiencing homelessness was likely underestimated because of undiagnosed cases and limited testing early in the pandemic in the United States. Among people experiencing homelessness, the true prevalence of chronic conditions might be underdiagnosed due to limited access to health care. People experiencing homelessness might be admitted for inpatient hospitalization for other reasons in addition to medical reasons. Lastly, results do not account for alternative factors that could affect the outcomes such as testing coverage, quality of care, and barriers to health care (eg, transportation, stigma).

Most hospitalized COVID-19 patients experiencing homelessness had underlying health conditions. Clinical outcomes such as ICU admissions, respiratory support like mechanical ventilation, and deaths were documented. Residing in congregate settings like homeless shelters could increase risk of COVID-19 if there are difficulties with implementing infection prevention and control measures; the high proportion of underlying conditions might also increase the risk of severe disease [7, 12]. These findings fill a critical gap in understanding morbidity and mortality among people hospitalized with COVID-19 experiencing homelessness and highlight the importance of targeting effective prevention strategies including COVID-19 vaccination efforts among marginalized groups [13, 14].

Notes

Acknowledgments. Susan Brooks, MPH, Brooke Heidenga, MPH, Kareena Hundal, MPH, and Pam Daily Kirley, MPH, California Emerging Infections Program, Oakland, CA; Rachel Herlihy, MD, and Nisha Alden, MPH, Colorado Department of Public Health and Environment, Denver, CO; Kimberly Yousey-Hindes, MPH, Amber Maslar, MPA, Adam Misiorski, MPH, Carol Lyons, MPH, Danyel Olson, MPH, Paula Clogher, MPH, Hazal Kayalioglu, BS, Gaggan Barr, MD, and Christina Parisi, MPH, Connecticut Emerging Infections Program, Yale School of Public Health, New Haven, CT; Emily Fawcett, MPH, Jeremiah Williams, MPH, Katelyn Lengacher, MPH, and Kyle Openo, DrPH, Emerging Infections Program, Georgia Department of Health, Veterans Affairs Medical Center, Foundation for Atlanta Veterans Education and Research, Atlanta, GA; Monica Farley, MD, Emerging Infections Program, Georgia Department of Public Health, Atlanta, GA, Veterans Affairs Medical Center, Atlanta, GA, and Department of Medicine, Emory University School of Medicine, Atlanta, GA; Tyra Goss, RN, Iowa Department of Public Health, Des Moines, IA; Maya Monroe, MPH, Maryland Department of Health, Baltimore, MD; Jim Collins, MPH, Justin Henderson, MPH, Shannon Johnson, MPH, Libby Reeg, MPH, Alex Kohrman, MPH, Sam Hawkins, BS, and Val Tellez Nunez, BS, Michigan Department of Health and Human Services, Lansing, MI; Austin Bell, MS, Kayla Bilski, BA, Emma Contestabile, BS, Claire Henrichsen, BS, Katherine Schleiss, MPH, Samantha Siebman, MPH, Emily Holodick, MPH, Lisa Nguyen, BS, Kristen Ehresmann, MPH, Ruth Lynfield, MD, Erica Bye, MPH, and Richard Danila, PhD, Minnesota Department of Health, St Paul, MN; Kathy M. Angeles, MPH, Emily B. Hancock, MS, Yadira Salazar-Sanchez, MPH, Meaghan Novi, MPH, Sarah A. Khanlian, MPH, Nancy Eisenberg, MPH, Melissa Christian, BA, and Mayvilynne Poblete, MA, New Mexico Emerging Infections Program, Santa Fe, NM; Nancy Spina, MPH, Grant Barney, MPH, and Adam Rowe, BA, New York State Department of Health, Albany, NY; Sophrena Bushey, MHS, Virginia Cafferky, BS, Christina Felsen, MPH, Maria Gaitan, RaeAnne Kurtz, BS, Christine Long, MPH, Kevin Popham, MPH, and Marissa Tracy, MPH, University of Rochester School of Medicine and Dentistry, Rochester, NY; Jess Shiltz, MPH, Ohio Department of Health, Columbus, OH; Ama Owusu-Dommey, MPH, and Nicole West, MPH, Public Health Division, Oregon Health Authority, Portland, OR; Karen Leib, RN, Katie Dyer, Terri McMinn, Danielle Ndi, MPH, Anise Elie, MPH, Kathy Billings, MPH, Manideepthi Pemmaraju, MPH, and John Ujwok, MPH, Vanderbilt University Medical Center, Nashville, TN; Andrea Price, LPN, Ashley Swain, CHES, Laine McCullough, MPH, Jake Ortega, MPH, Ryan Chatelain, MPH, Tyler Riedesel, MPH, Ian Buchta, MPH, Caitlin Shaw, BS, Andrew Haraghey, BS, and Amanda Carter, BS, Salt Lake County Health Department, Salt Lake City, UT; Keegan McCaffrey, BS, Utah Department of Health, Salt Lake City, UT; Michael Whitaker, MPH, Alissa O’Halloran, MPH, and Jennifer Milucky, MSPH, COVID-NET Team, CDC, Atlanta, GA; Kadam Patel, MPH, and Onika Anglin, MPH, COVID-NET Team, CDC, Atlanta, GA, General Dynamics Information Technology, Falls Church, VA.

COVID-NET Surveillance Team Authors: Shua J. Chai, MD, MPH, California Emerging Infections Program, Oakland, CA, and Center for Preparedness and Response, CDC; Breanna Kawasaki, MPH, Colorado Department of Public Health and Environment, Denver, CO; James Meek, MPH, Connecticut Emerging Infections Program, Yale School of Public Health, New Haven, CT; Evan J. Anderson, MD, Departments of Pediatrics and Medicine, Emory University School of Medicine, and Emerging Infections Program, Atlanta Veterans Affairs Medical Center, Atlanta, GA; Andrew Weigel, MSW, Iowa Department of Public Health, Des Moines, IA; Patricia Ryan, MSc, Maryland Department of Health, Baltimore, MD; Sue Kim, MPH, Michigan Department of Health and Human Services, Lansing, MI; Kathryn Como-Sabetti, MPH, Minnesota Department of Health, St. Paul, MN; Salina Torres, PhD, MPH, New Mexico Department of Health, Santa Fe, NM; Alison Muse, MPH, New York State Department of Health, Albany, NY; Nancy M. Bennett, MD, MS, University of Rochester School of Medicine and Dentistry, Rochester, NY; Laurie M. Billing, MPH, Ohio Department of Health, Columbus, OH; Melissa Sutton, MD, MPH, Public Health Division, Oregon Health Authority, Portland, OR; H. Keipp Talbot, MD, Vanderbilt University Medical Center, Nashville, TN; and Ilene Risk, MPA, Salt Lake County Health Department, Salt Lake City, UT.

Disclaimer. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention (CDC). Funding support had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data, preparation, review, or approval of the manuscript; and decision to submit the manuscript for submission.

Financial support. J. M., E. A., A. W., P. R., L. B., M. S., and H. K. T. receive grant support from the CDC. A. W., S. K., and L. B. receives grant support from the Council of State and Territorial Epidemiologists.

Potentialconflicts of interest. E. A. reports consultantships for Abbvie, Pfizer, Sanofi-Pasteur, and Kentucky BioProcessing, Inc.; and grants/grants pending for clinical trials from MedImmune, Regeneron, PaxVax, Pfizer, GSK, Merck, Novavax, Sanofi-Pasteur, Micron, and Janssen. All other authors report no potential conflicts.

All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.