PEH had significantly higher rates of delirium (10.6% vs. 8.1%; P< 0.01), as shown in Table 1. However, PEH had fewer respiratory complications, including pneumonia (48.5% vs. 81.9%; P<0.001) and respiratory failure (28.7% vs. 61.9%; P<0.001), and lower in-hospital mortality compared to the baseline group (3.3% vs. 9.5%; P<0.001). With respect to mean length of stay (LOS), there was no significant difference between the two groups (8.3 days vs. 8.6 days; P > 0.1).

As shown in Table 1, PEH had significantly higher rates of hepatitis, chronic kidney disease, liver disease and traumatic brain injury. The incidence of TBI was about four times higher for PEH compared with the baseline group (1.3% vs. 0.3%; P<0.001). PEH were more likely to have Medicaid insurance and less likely to have commercial insurance. PEH had significantly higher rates of alcohol, cannabis, and opioid use disorders. Twenty-nine percent of PEH had at least one substance use disorder (SUD). By contrast, only four percent of the baseline group had one or more SUDs. The prevalence of SUDs was six to nine times higher among PEH. To summarize, seven of the delirium-associated risk factors were more prevalent among PEH compared to baseline.

Only two characteristics of the homeless population were associated with a reduced risk of delirium. PEH were younger compared to the baseline group (50.2 vs. 52.3 years, P<0.01). Second, PEH had slightly higher emergency admissions and slightly lower admissions from a skilled nursing facility, although these differences were not significant. Treatment with Remdesivir was associated with a lower risk of delirium. However, PEH were less likely to be treated with Remdesivir (8.4% vs. 26.7%; P <0.01).

The most influential patient categories, in terms of their explanatory power, were chronic comorbidities, followed by source of payment, and substance use disorders (SUDs). As shown in Table 2, the most significant predictors of delirium were dementia (AOR = 3.09; CI: 2.58, 3.70, opioid use disorder (AOR = 2.77; CI: 2.39, 3.21), and traumatic brain injury (AOR = 2.66; CI: 2.06, 3.42). Transfer from a skilled nursing facility (AOR = 2.49; CI: 2.08, 2.97) was also a risk factor for delirium. As shown in Table 2, protective effects for delirium included treatment with Remdesivir (AOR = 0.63; CI: 0.6, 0.66), female gender (AOR = 0.91; CI: 0.88, 0.94), Hispanic ethnicity (AOR = 0.86; CI: 0.82, 0.89), commercial insurance (AOR = 0.83; CI: 0.79, 0.86), and emergency department admission (AOR = 0.65; CI: 0.62, 0.68). Based on the predicted values from the logistic model, PEH also had a higher risk of delirium compared to baseline (11.7% vs. 8.1%; P<0.001).

Compared to patients with only one risk factor, patients with two risk factors had about twice the risk of delirium (16%). The risk of delirium was 21% for persons with three or more risk factors. Regardless of housing status, delirium was associated with significant morbidity and mortality. Mean length of stay (LOS) was about ten days longer for delirious patients compared with non-delirious patients (18.4 days vs. 7.7 days; P<0.001). For PEH, the average LOS was five days longer for delirious patients compared with non-delirious patients (12.9 days vs. 7.7 days; P<0.001). Based on the regression results, for ICU patients, having delirium increased their LOS by 9.5 days (CI: 9.3, 9.7). For acute-care patients, having delirium increased their LOS by 5.3 days (CI: 5.1, 5.5). In addition, eight percent of delirious patients required mechanical ventilation. More than one-fourth (28.5%) of delirious patients died in the hospital compared to 7.8% of non-delirious patients. Delirium was also found to greatly increase the risk of mortality (AOR = 3.8; CI = 3.6, 4.0). The additional cost per case of delirium was estimated to be $18,177 (CI: $16,872, $19,702).

Delirium is a well-known complication of respiratory disease among older adults. The causes of delirium are multifactorial, and these can be categorized as predisposing, precipitating, and environmental [1]. Predisposing factors include age, chronic medical conditions, and preexisting cognitive impairment. Precipitating factors include the acute illness itself, substance use and withdrawal, and the use of certain medications, such as benzodiazepines. Environmental effects, such as isolation and lack of cognitive stimulation, can also play a role [3]. This is the first study to find that PEH who were hospitalized with COVID-19 had higher rates of delirium compared to the baseline group.

PEH had lower rates of respiratory complications and in-hospital mortality, and these findings are consistent with a prior study from the Center for Disease Control (CDC) that found similar results [20]. We sought to better understand these findings. Obesity was found to be a major risk factor for pneumonia among COVID-19 patients [21]. With respect to the present study, obesity rates were more than twice as high in the baseline group (38.2%) compared to PEH (15.7%; P<0.001), where obesity was defined as a Body Mass Index (BMI) of 30 or greater. Because PEH had significantly lower rates of obesity, this was likely a contributing factor to their lower rates of respiratory complications, such as pneumonia and respiratory failure.

Among patients with delirium, PEH were significantly younger compared to baseline, as measured by median age (55 vs. 62 years; P<0.001.) Ostensibly, younger patients should have a reduced risk of delirium. However, the age distribution of the PEH in this study may reflect both accelerated aging and premature mortality. Previous studies have found homelessness to be associated with a shorter lifespan. For younger PEH, drug overdose was found to be the leading cause of death [13]. By contrast, older PEH, i.e., those in their forties and fifties, die from similar causes as the housed population; however, they die 10–15 years earlier [14]. Another study of 615 PEH found the average age at death to be 56 for males and 52 for females [22].

A recent systematic review found that older PEH are susceptible to accelerated aging and geriatric conditions, such as frailty, cognitive impairment, and functional decline [5]. Another study found that PEH had accelerated physical limitations, balance impairments, frailty, and lack of mobility, regardless of their age [23]. With respect to the baseline group, dementia (AOR = 3.09; CI: 2.58, 3.7) was found to be the most significant risk factor for delirium, as shown in Table 2. However, this was not the case for PEH. Among this group, the most significant risk factors were traumatic brain injury (TBI) (AOR = 6.55; CI: 1.85, 23.24) and AUD (AOR = 1.74; CI: 1.04, 2.90). AUD, as consistent with prior literature, was also associated with an increased risk of liver disease and hepatitis. In addition, PEH were more likely to have a diagnosis of chronic pain than baseline (0.7% vs. 0.1%; P < 0.001). Chronic pain has also been associated with an increased risk of delirium [1].

A recent systematic review and meta-analysis of 66 studies emphasized that delirium is a frequent complication of COVID-19 in older adults. However, it did not address whether PEH are at increased risk for delirium [24]. Since most of these studies focused on patients aged 65 and older, most PEH were likely excluded from the analyses. In the current study, approximately 80 percent (n = 774) of the PEH in the database (n = 974) were younger than 65 years.

Due to various factors, the onset of the pandemic coincided with an increase in SUDs and smoking, especially among vulnerable populations, such as PEH [8, 25]. Nyamathi et al., (2022), conducted in-depth interviews with 21 PEH to identify the reasons for the increase in SUDs [8]. The pandemic and subsequent restrictions increased feelings of isolation, anxiety, and depression. Some PEH used alcohol and smoking as a coping strategy, which were easy to obtain and were perceived as a "safe alternative" to street drugs, such as opioids. There were also cutbacks and reduced staffing for social services, such as mental health services, drug treatment programs, and needle exchange programs [26]. As only about ten percent of this population have cell phones, most PEH were unable to access telehealth services. Among the PEH in our study who had AUD (n = 162), about half (51%) were smokers. In addition, the percentage of smokers among PEH increased from 32% at the onset of the pandemic to 46% by June, 2021. It is likely that the true prevalence of smokers in this population was even higher [25].

Regardless of a patient’s housing status, delirium was associated with significant morbidity and mortality. Mean length of stay (LOS) was about ten days longer for delirious patients compared with non-delirious patients (18.4 days vs. 7.7 days; P<0.001). In addition, eight percent of delirious patients required mechanical ventilation. The in-hospital mortality rate for persons with delirium was 28.5% compared to 7.8% for non-delirious patients (P < 0.001). A separate logistic model was used to estimate the impact of delirium on in-hospital mortality. The mortality risk was 3.8 times higher (CI = 3.6, 4.0) for delirious patients compared to non-delirious patients.

Remdesivir was found to have a protective effect against delirium. This is consistent with prior research which found that the risk of in-hospital mortality was reduced by 17 percent for patients treated with Remdesivir [18]. Remdesivir was also found to have a protective effect against mortality (AOR = 0.68; CI: 0.65, 0.71). Patients with commercial insurance (P < 0.001) were more likely to be treated with Remdesivir. Those with Medicaid insurance, charity care, and substance use disorders (P < 0.001) were less likely to receive the treatment. We calculated mortality risk based on the risk factors in Table 2, excluding Remdesivir. Patients who received the drug had a lower a priori risk of mortality compared to those who did not receive it (8.6% vs. 9.7%; P < 0.001). Therefore, the observed differences in mortality may be partly due to differences in the risk profile of each group.

For PEH (n = 878), the risk of mortality was more than eight times higher for delirious patients as compared to those without delirium (AOR = 8.4; CI: 3.7–18.7). For PEH (n = 29) who died from COVID-19, delirium was present in more than half (52%) of cases. Hence delirium was found to greatly increase the risk of mortality for this vulnerable population. For delirious patients (n = 14,410), there was no significant difference in mortality between those who did or did not receive Remdesivir (28.9% vs. 28.4%; P > 0.05).

During the pandemic, many hospitals implemented infection control measures, such as isolation precautions, restrictions on family visitation, and limited physical contact with hospital staff. These measures often led to social isolation, lack of mobility, and even the use of physical and chemical restraints [3]. These environmental conditions may have exacerbated the risk of delirium and prolonged its severity and duration. A recent study of more than 33,000 hospitalized patients found that visitor restrictions, and the subsequent lack of human contact, imposed during the COVID-19 pandemic were associated with a 35 percent increase in the risk of delirium [27].

The incremental cost per case of delirium was estimated to be $18,177 (CI: $16,872, $19,702). This is somewhat consistent with prior literature [12]. During the period under study, the cost of the full course of treatment with the antiviral Remdesivir, received over five days, was about $3,100. Hence the savings that could be achieved by preventing a single case of delirium could pay for five patients to be treated with Remdesivir.

The most recent systematic review of the economic costs of delirium at the national level was based on multiple studies conducted before the pandemic. This review estimated that in 2019, the total cost of delirium at US hospitals could be up to $82.4 billion [12]. The corresponding incremental cost of delirium per hospital stay was estimated to be between $1,529 and $19,050. Ample evidence indicates that delirium increases costs across all settings, such as hospitals, long-term care facilities, and home care. Delirium is also a cause of dementia, and the cost of long-term care for persons with dementia is even higher [28]. The true costs to society are likely to be significantly more than these estimates because they do not include opportunity costs. Because patients with delirium had a longer LOS and treatment intensity, this also increased the demand for critical hospital resources, such as ventilators, ICU beds, and nursing staff. By greatly increasing hospital costs and resource requirements, delirium also put additional financial strain on hospitals [29]. Additional opportunity costs include years of lost productivity by the young elderly who are unable to work and must be institutionalized [28].

Among the strengths of this study, we examined homelessness as a potential risk factor for delirium and mortality for patients hospitalized with COVID-19. We also demonstrated that several risk factors for delirium are more prevalent among this vulnerable population. We performed multivariable analysis that included several potential confounders. The study also estimated the incremental cost of delirium based on changes in patients’ mean LOS. This study underscores that delirium leads to increased mortality and morbidity, a significantly longer LOS, and higher hospital costs. In addition, these findings suggest that the antiviral Remdesivir was effective in reducing the risk of delirium.

This study is subject to some limitations. Data on the patients’ housing status is likely to be missing from medical records. Hence the prevalence of PEH who have COVID-19 is likely to be underestimated, especially in the early stages of the pandemic when testing was limited. Because PEH may have limited access to both primary care and specialty care, their true prevalence of chronic conditions may be underdiagnosed. Hence these study findings may not be generalizable to the broader population of PEH who are hospitalized with COVID-19. We did not have information on the length of time a person was homeless. Moreover, delirium is often under-documented in administrative databases [30].

Long-term effects of delirium are not uncommon among patients who survive and are discharged from the hospital.¹ According to previous studies, twenty percent of patients still show some symptoms of delirium six months after hospital discharge. For patients hospitalized with COVID-19, delirium has been associated with long-term cognitive decline and increased risk of dementia [4, 28].