Risk Factors for Admission Within a Hospital-Based COVID-19 Home Monitoring Program.

Background: Despite increasing vaccination rates, coronavirus disease 2019 (COVID-19) continues to overwhelm heath systems worldwide. Few studies follow outpatients diagnosed with COVID-19 to understand risks for subsequent admissions. We sought to identify hospital admission risk factors in individuals with COVID-19 to guide outpatient follow-up and prioritization for novel therapeutics.
Methods: We prospectively designed data collection templates and remotely monitored patients after a COVID-19 diagnosis, then retrospectively analyzed data to identify risk factors for 30-day admission for those initially managed outpatient and for 30-day re-admissions for those monitored after an initial COVID-19 admission. We included all patients followed by our COVID-19 follow-up monitoring program from April 2020 to February 2021.
Results: Among 4070 individuals followed by the program, older age (adjusted odds ratio [aOR], 1.05; 95% CI, 1.03-1.06), multiple comorbidities (1-2: aOR, 5.88; 95% CI, 2.07-16.72; ≥3: aOR, 20.40; 95% CI, 7.23-57.54), presence of fever (aOR, 2.70; 95% CI, 1.65-4.42), respiratory symptoms (aOR, 2.46; 95% CI, 1.53-3.94), and gastrointestinal symptoms (aOR, 2.19; 95% CI, 1.53-3.94) at initial contact were associated with increased risk of COVID-19-related 30-day admission among those initially managed outpatient. Loss of taste/smell was associated with decreased admission risk (aOR, 0.46; 95% CI, 0.25-0.85). For postdischarge patients, older age was also associated with increased re-admission risk (aOR, 1.04; 95% CI, 1.01-1.06). Conclusions: This study reveals that in addition to older age and specific comorbidities, the number of high-risk conditions, fever, respiratory symptoms, and gastrointestinal symptoms at diagnosis all increased odds of COVID-19-related admission. These data could enhance patient prioritization for early treatment interventions and ongoing surveillance.

With >530 million cases of coronavirus disease 2019 (COVID-19) worldwide and >84 million cases and >1 million deaths in the United States [1], COVID-19 remains a critical public health issue and ongoing pandemic. Despite rapid vaccine development, most of the world remains unvaccinated, and variant strains continue to emerge [1-3].

Although the majority of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)–positive patients experience mild to moderate disease manageable at home, hospitalization rates among unvaccinated individuals are still up to 23 times the rates of those who have been fully vaccinated and received a booster dose [4]. Outpatient risk assessment for novel therapeutics and monitoring postdiagnosis could be essential to proactively prevent rapid decompensation and hospitalizations. Targeted telehealth programs have emerged to monitor patients remotely, optimize at-home care, and guide outreach frequency to facilitate transitions to higher levels of care [5-8].

For programs to prioritize limited outpatient treatments and proactively identify severe disease progression, recognizing demographic and clinical factors that predict re-admission is essential. Although many studies have described risk factors for severe COVID-19 in hospitalized patients, few studies have addressed hospitalization risk factors among patients diagnosed and followed in ambulatory settings or included symptom data in their analyses [8-10].

COVID-19 has exacerbated preexisting racial/ethnic and socioeconomic health and health care disparities in the United States and disproportionally impacted people of color, notably individuals who identify as Black or Hispanic [11]. Inequities in social determinants of health such as higher housing density, lower education levels, lower income and savings, and decreased health care access, as well as higher likelihoods of working essential jobs and experiencing ongoing discrimination, likely contribute to the increased disease and mortality risk observed among these groups [12].

We created a robust, institutional COVID-19 home monitoring program to improve health care access and closely monitor all patients diagnosed with COVID-19 in our diverse patient population. This study aimed to identify clinical and sociodemographic predictors of COVID-19-related hospital admissions among confirmed SARS-CoV-2-positive patients followed by this program.

METHODS

Setting

Boston Medical Center (BMC) is an urban academic medical center and the largest safety-net hospital in New England [13]. BMC primarily serves neighborhoods with high social vulnerability indexes (SVIs), with many of BMC’s patients identifying as Hispanic (>20%), Black (>50%), and nonprimary English-speaking (>32%), or experiencing housing instability [14]. In response to the first COVID-19 surge in March 2020, our institution established dedicated teams to notify patients of COVID-19 test results and conduct prospective clinical monitoring of patients diagnosed with COVID-19 (hereafter referred to as the COVID-19 Follow-up Program). Led by infectious diseases and general internal medicine physicians and an operations manager, the program includes physicians, advanced practice providers (APPs), nurses, and medical assistants. The program has adapted to the rapidly changing national and local guidance, staffing, and clinical changes as the pandemic has evolved. Nurses or medical assistants call patients to report initial results, and nurses, APPs, and/or physicians conduct follow-up telehealth visits.

The program notifies all nonadmitted patients tested at BMC of positive COVID-19 results; negative test results are viewed via patient electronic health record (EHR) access. Using an EHR report, the program also contacts patients with COVID-19 upon hospital discharge. The team then utilizes EHR templates to record symptoms, clinical trajectory, and determine risk stratification (low, moderate, or high risk for severe disease progression) (Appendix 1). The highest-risk individuals are advised to present to the emergency department (ED). Telehealth follow-up frequency for those at moderate to high risk is every 1–2 days, and for low-risk patients, every 3 days, until near resolution of symptoms. All EHR templates also include current Centers for Disease Control and Prevention (CDC) and local department of health recommendations to advise patients on household infection control practices and isolation and close contact quarantine guidance and to provide contact information for questions or worsening symptoms. The COVID-19 Follow-up Program continues monitoring all BMC primary care patients and those without established primary care; patients with outside primary care are referred to their providers after initial contact for further follow-up per local health center preferences.

Study Population

We included all patients who had ≥1 COVID-19 Follow-up Program contact from April 2020 to February 2021 within 4 weeks of a polymerase chain reaction (PCR)–confirmed SARS-CoV-2 infection (Appendix 2). We classified patients as “outpatient” if their initial positive SARS-CoV-2 PCR test occurred in an outpatient or ED setting without hospital admission before COVID-19 Follow-up Program contact and as “postdischarge” if their initial contact with the Follow-up Program succeeded discharge from a hospital admission.

Data Collection

We retrospectively extracted the following data from the BMC Clinical Data Warehouse, which houses the health system’s EHR: demographics, underlying medical conditions (using International Classification of Diseases, 10th edition [ICD-10], codes from active problem lists), hospital admissions, symptoms, risk assessment, clinical trajectory, comorbidities, imaging findings, and number of calls from the COVID-19 Follow-up Program’s EHR templates. We recorded medical conditions associated with increased risk of severe COVID-19 as potential confounders for initial or repeat hospital admission: diabetes, chronic lung disease, cirrhosis, hypertension, chronic kidney disease, sickle cell anemia, coronary artery disease, heart failure, obesity, smoking (current or former), rheumatologic disease, and immunocompromising conditions [15]. We defined severe immunosuppression as documented active chemotherapy, high-dose steroids, or individuals with HIV with a CD4 cell count <200 cells/mm3 within 6 months prior. We grouped the following clinically similar COVID-19 symptoms: fatigue/generally feeling sick, headaches/body aches, nasal congestion/sore throat, and loss of taste/smell. We calculated SVI using patient home zip codes and categorized scores as >0.5 (socially vulnerable) or ≤0.5 [16,17]. The primary outcome was COVID-19-associated hospital admission (primary diagnosis of COVID-19 or a potential complication) (Appendix 3) within 30 days of first Follow-up Program contact.

The Boston University Medical Campus Institutional Review Board approved this study as exempt human subject research for EHR data review only.

Analysis

We used descriptive statistics to characterize patients in each of the outpatient and postdischarge subsets given baseline differences in disease severity and course between groups. We used logistic regression to determine associations between covariates (demographics, medical history, COVID-19 symptoms, and trajectory) and subsequent hospital admissions. Variables with unadjusted P values <.2 were considered for the multivariable model. Age and race/ethnicity were forced in the multivariable model regardless of unadjusted P value given literature reports of the significance of these factors [8-10]. We used pairwise chi-square comparisons to test correlations between categorical independent variables due to significant associations suspected or reported in the literature. When 2 variables were highly correlated (Pearson’s correlation coefficient >0.4), we included only the variable with greater clinical impact in the multivariable model or used 2 separate models if each variable was clinically important. In the postdischarge group, we had only 33 outcome observations; therefore, we limited this multivariable model to 3 covariates: age, sex, and race/ethnicity. We analyzed risk stratification in a separate analysis combining both groups given missing data in 33%.

We performed a sensitivity analysis including individuals with >1 COVID-19 Follow-up Program contact to isolate individuals with continued home monitoring using the same analysis plan. With closer follow-up, this longitudinal group may have been less likely to present to an outside hospital ED, resulting in more reliable acute health care utilization data and ability to track symptom evolution. Due to small numbers in this subgroup, we examined a composite outcome that included any acute care utilization (ED visit, observation stay, or hospital admission) within 30 days of first Follow-up Program contact. Finally, to ensure results were not confounded by analyzing children and adults together, we repeated the analysis excluding children.

All analyses were conducted using SAS, version 9.4 (SAS Institute Inc, Cary, NC, USA).

RESULTS

Altogether, 4070 patients met inclusion criteria (age range 0–99+ years old, including 356 children aged 0–17 years): 3185 in the outpatient group and 885 patients in the postdischarge group, for a total of 11 627 COVID-19 Follow-up Program contacts. In the outpatient group, 53.7% were female with a mean age of 41.6 years; 43.6% identified as non-Hispanic Black or African American, 33.3% as Hispanic, and 9.9% as non-Hispanic White (Table 1). Over one-third (37.7%) listed a language other than English as their primary language. More than half (55.3%) had Medicaid insurance, 11.1% Medicare, and 26.1% had private insurance. The mean SVI was 0.682; 80.0% lived in zip codes considered socially vulnerable (SVI >0.5). Sixty-six percent had at least 1 comorbidity, and 67.9% reported ≥1 symptom at initial contact (most commonly respiratory symptoms, 40.1%). In total, 245 patients (7.7%) had a COVID-19-related acute care utilization, and 83 (2.6%) had a COVID-19-related hospital admission within 30 days of first COVID-19 Follow-up Program contact.

Table 1.

Demographic Characteristics, Comorbidities, Clinical Disease Data, and Follow-up Outreach Among All Patients, Stratified by Initial Testing Site

Variable	Outpatient[a] (n = 3185), No. (%)	Postdischarge[a] (n = 885), No. (%)	P Value
Age, mean ± SD, y	41.6 ± 19.9	53.2 ± 19.0	<.001
Female	1710 (53.7)	446 (50.4)	.08
Race/ethnicity
Non-Hispanic Black or African American	1387 (43.6)	365 (41.2)	.22
Non-Hispanic White	316 (9.9)	119 (13.5)	.003
Hispanic	1061 (33.3)	306 (34.6)	.48
Another[b]	133 (4.2)	49 (5.5)	.08
Unknown	288 (9.0)	46 (5.2)	.002
Primary language
English	1986 (62.4)	479 (54.1)	<.001
Spanish	672 (21.1)	232 (26.2)	.001
Other[c]	527 (16.6)	174 (19.7)	.03
Insurance
Medicaid	1762 (55.3)	461 (52.1)	.09
Medicare	353 (11.1)	238 (26.9)	<.001
Private	832 (26.1)	144 (16.3)	<.001
Otherd	238 (7.5)	42 (4.8)	.005
SVI, mean ± SD	0.682 ± 0.254	0.697 ± 0.226	.11
SVI >0.5	2548 (80.0)	757 (85.5)	<.001
Comorbidities
Diabetes	487 (15.3)	263 (29.7)	<.001
Coronary artery disease	94 (3.0)	68 (7.7)	<.001
Chronic lung disease	506 (15.9)	131 (14.8)	.43
Cirrhosis	32 (1.0)	15 (1.7)	.09
Hypertension	918 (28.8)	405 (45.8)	<.001
Chronic kidney disease	78 (2.5)	86 (9.7)	<.001
Obesity	1317 (41.4)	482 (54.5)	<.001
Heart failure	70 (2.2)	68 (7.7)	<.001
Current or former smoker	629 (19.8)	307 (34.7)	<.001
HIV	44 (1.4)	21 (2.4)	.04
Sickle cell	4 (0.1)	6 (0.7)	.003
Rheumatologic disease	44 (1.4)	17 (1.9)	.24
Total comorbidities
0	1095 (34.4)	125 (14.1)	<.001
1–2	1497 (47.0)	418 (47.2)	.90
3–4	497 (15.6)	22 (29.6)	<.001
5+	96 (3.0)	80 (9.0)	<.001
Abnormal CXR/chest CT	43 (1.4)	93 (10.5)	<.001
Severely immunosuppressed[e]	34 (1.1)	40 (4.5)	<.001
Symptoms
Fatigue/generally feeling sick	908 (28.5)	166 (18.8)	<.001
Headache/body aches	1206 (37.9)	102 (11.5)	<.001
Runny nose/sore throat	979 (30.7)	56 (6.3)	<.001
Loss of taste or smell	813 (25.5)	52 (5.9)	<.001
Any fever[f]	642 (20.2)	71 (8.0)	<.001
Any GI symptom[g]	385 (12.1)	52 (5.9)	<.001
Any respiratory symptom[g]	1277 (40.1)	214 (24.2)	<.001
Scaled number of symptoms
0	1021 (32.1)	574 (64.9)	<.001
1	586 (18.4)	110 (12.4)	<.001
2	502 (15.8)	92 (10.4)	<.001
3	438 (13.8)	48 (5.4)	<.001
4+	638 (20.0)	61 (6.9)	<.001
Clinical trajectory
Symptoms resolved	1106 (34.7)	227 (25.6)	<.001
Improving	742 (23.3)	193 (21.8)	.35
Staying the same	718 (22.5)	52 (5.9)	<.001
Worsening	55 (1.7)	3 (0.3)	.002
Unknown	564 (17.7)	410 (46.3)	<.001
Total follow-up calls within 30 d
<3	1898 (59.6)	544 (61.5)	.31
3–5	967 (30.4)	213 (24.1)	<.001
6+	320 (10.1)	128 (14.5)	<.001
Initial call risk stratification
Already re-admitted	7 (0.2)	8 (0.9)	.003
Sent to ED	14 (0.4)	0 (0.0)	–
High (next call tomorrow)	185 (5.8)	64 (7.2)	.12
Moderate (next call in 2 d)	447 (14.0)	107 (12.1)	.14
Low (next call in 3 d)	699 (22.0)	89 (10.6)	<.001
Symptoms resolved, no follow-up needed	983 (30.8)	212 (24.0)	<.001
Community health center patient, no follow-up needed	571 (17.9)	304 (34.4)	<.001
Unable to reach after 3 attempts	279 (8.8)	101 (11.4)	.02

Abbreviations: COVID-19, coronavirus disease 2019; CT, computed tomography scan; CXR, chest x-ray; ED, emergency department; EHR, electronic health record; GI, gastrointestinal; PCR, polymerase chain reaction; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; SVI, social vulnerability index.

Patients stratified by initial COVID-19 testing site—outpatient (including patients tested in and discharged from the ED and those with SARS-CoV-2 PCR completed elsewhere but captured in the EHR) or postdischarge (those tested during an inpatient stay or admitted before first COVID-19 follow-up program contact).

Includes Asian, American Indian/Native American, Native Hawaiian/Pacific Islander, and “other.”

Includes Haitian Creole, Cape Verdean/Port Creole, Portuguese, Vietnamese, Amharic/Ethiopian, and other languages with <3 patients each.

Includes no insurance on file, worker’s compensation, Veterans Affairs insurance, and grant-funded medical coverage.

Includes those receiving active chemotherapy, high-dose steroids, or with HIV with CD4 cell count <200 cells/mm3.

Subjectively reported or recorded temperature >100.4°F.

See Appendix 1 for details of the questions asked to elicit whether any gastrointestinal or respiratory symptoms were present; those with mild, moderate, or severe symptoms were grouped together as having any symptoms in the category “present.”

Compared with the outpatient group, the postdischarge group had a higher mean age (53.2 years), mean SVI (0.697), proportion that was Medicare-insured (26.9%), and proportion with any comorbidities (85.9%). Most postdischarge patients (64.9%) reported no symptoms at initial contact. Sixty-six postdischarge patients (7.5%) had any COVID-19-related acute care utilization, and 33 (3.7%) a COVID-19-related re-admission within 30 days.

In unadjusted analyses within the outpatient group, older age, Medicare insurance, reporting fatigue/generally feeling sick, headaches/body aches, fever, gastrointestinal symptoms, respiratory symptoms, or presence of ≥4 symptoms, and having ≥1 comorbidity or severe immunosuppression were associated with increased odds of admission (Table 2). Adjusting for sex, race/ethnicity, SVI, loss of taste/smell, and severe immunosuppression, older age (aOR, 1.05; 95% CI, 1.03–1.06), presence of gastrointestinal symptoms (aOR, 2.19; 95% CI, 1.53–3.94), and presence of respiratory symptoms (aOR, 2.46; 95% CI, 1.53–3.94) were significantly associated with hospital admission. Language, insurance, number of symptoms, and total comorbidities were excluded from the main multivariable model due to strong correlations with race/ethnicity, age, individual symptoms, and age, respectively.

Table 2.

Hospital Admissions Within 30 Days of Initial Contact by COVID-19 Follow-up Program in All Patients Initially Tested in Outpatient Setting

Variable	No Admission Within 30 Days[a] (n = 3102), No. (%)	Admission Within 30 Days[a] (n = 83), No. (%)	Unadjusted OR (95% CI)	With Respiratory Symptoms, Adjusted OR (95% CI)[b]	With Fever, Adjusted OR (95% CI)[c]	With Total Comorbidities, Adjusted OR (95% CI)[d]
Age, mean ± SD, y	41.1 ± 19.7	58.1 ± 19.2	1.05 (1.03–1.06)	1.05 (1.03–1.06)	1.05 (1.04–1.06)	–
Female	1672 (53.9)	38 (45.8)	0.72 (0.47–1.12)	0.67 (0.43–1.05)	0.69 (0.44–1.10)	0.71 (0.45–1.12)
Race/ethnicity
Non-Hispanic Black or African American	1355 (43.7)	32 (38.6)	0.91 (0.42–1.99)	1.07 (0.46–2.49)	0.97 (0.30–2.23)	1.02 (0.44–2.37)
Non-Hispanic White	308 (9.9)	8 (9.6)	Ref	Ref	Ref	Ref
Hispanic	1027 (33.1)	34 (41.0)	1.28 (0.58–2.78)	1.62 (0.71–3.72)	1.42 (0.91–2.26)	1.67 (0.73–3.83)
Another[e]	130 (4.2)	3 (3.6)	0.89 (0.23–3.40)	0.83 (0.21–3.27)	0.72 (0.18–2.84)	1.18 (0.30–4.71)
Unknown	282 (9.1)	6 (7.2)	0.82 (0.28–2.39)	1.05 (0.34–3.19)	0.91 (0.30–2.76)	1.31 (0.43–4.01)
Language
English	1941 (62.6)	45 (54.2)	Ref	–	–	–
Spanish	507 (16.3)	18 (21.7)	1.19 (0.68–2.07)	–	–	–
Other[f]	507 (16.3)	20 (24.1)	1.70 (1.00–2.91)	–	–	–
Insurance
Medicaid	1730 (55.8)	32 (38.6)	0.75 (0.43–1.32)	–	–	–
Medicare	325 (10.5)	28 (33.7)	3.50 (1.94–6.30)	–	–	–
Private	812 (26.2)	20 (24.1)	Ref	–	–	–
Other[g]	235 (7.6)	3 (3.6)	0.52 (0.15–1.76)	–	–	–
SVI, mean ± SD	0.681 ± 0.253	0.693 ± 0.261	1.21 (0.50–2.90)	1.23 (0.50–3.31)	1.31 (0.51–3.35)	1.01 (0.39–2.58)
SVI category
≤0.5	620 (20.0)	17 (20.5)	Ref	–	–	–
SVI >0.5	2482 (80.0)	66 (79.5)	0.97 (0.57–1.66)	–	–	–
Total comorbidities
0	1091 (35.2)	4 (4.8)	Ref	–	–	Ref
1–2	1464 (47.2)	33 (39.8)	6.15 (2.17–17.41)	–	–	5.88 (2.07–16.72)
3+	547 (17.7)	29 (34.9)	22.94 (8.21–64.05)	–	–	20.40 (7.23–57.54)
Severely immunosuppressed[h]	31 (1.0)	3 (3.6)	3.72 (1.11–12.40)	2.07 (0.58–7.37)	2.13 (0.61–7.49)	1.58 (0.44–5.64)
Symptoms
Fatigue/generally feeling sick	866 (27.9)	42 (50.6)	2.65 (1.71–4.10)	–	–	–
Headache/body aches	1164 (37.5)	42 (50.6)	1.71 (1.10–2.64)	–	–	–
Nose/throat symptoms	961 (31.0)	18 (21.7)	0.62 (0.36–1.05)	–	–	–
Loss of taste or smell	799 (25.8)	14 (16.9)	0.59 (0.33–1.05)	0.56 (0.30–1.04)	0.55 (0.30–1.03)	0.46 (0.25–0.85)
Any fever[i]	610 (19.7)	32 (38.6)	2.56 (1.63–4.02)	–	2.70 (1.65–4.42)	–
Any GI symptom[j]	366 (11.8)	19 (22.9)	2.22 (1.32–3.75)	2.19 (1.53–3.94)	2.05 (1.16–3.63)	1.79 (1.02–3.15)
Any respiratory symptom[j]	1225 (39.5)	52 (62.7)	2.57 (1.64–4.03)	2.46 (1.53–3.94)	–	2.18 (1.35–3.51)
Scaled number of symptoms
0	1002 (32.3)	19 (22.9)	Ref	–	–	–
1	576 (18.6)	10 (12.1)	0.92 (0.42–1.98)	–	–	–
2	493 (15.9)	9 (10.8)	0.96 (0.43–2.14)	–	–	–
3	424 (13.7)	14 (16.9)	1.74 (0.87–3.51)	–	–	–
4+	607 (19.6)	31 (37.4)	2.69 (1.51–4.81)	–	–	–
Clinical trajectory
Symptoms resolved	1079 (34.8)	27 (32.5)	Ref	–	–	–
Improving	718 (23.2)	24 (28.9)	1.34 (0.77–2.33)	–	–	–
Staying the same	697 (22.5)	21 (25.3)	1.20 (0.68–2.15)	–	–	–
Worsening	54 (1.7)	1 (1.2)	0.74 (0.10–5.55)	–	–	–
Unknown	554 (17.9)	10 (12.1)	0.72 (0.35–1.50)	–	–	–

Abbreviations: COVID-19, coronavirus disease 2019; GI, gastroinstestinal; OR, odds ratio; SVI, social vulnerability index.

Within 30 days of first COVID-19 follow-up team contact.

Multivariate model includes age, female, race/ethnicity, SVI, loss of taste or smell, any respiratory symptom, any gastro symptom, and severely immunosuppressed variables.

Multivariate model includes age, female, race/ethnicity, SVI, loss of taste or smell, any fever, any GI symptom, and severely immunosuppressed variables.

Multivariate model includes female, race/ethnicity, SVI, loss of taste or smell, any respiratory symptom, any GI symptom, total comorbidities, and severely immunosuppressed variables.

Includes Asian, American Indian/Native American, Native Hawaiian/Pacific Islander, and “other.”

Includes Haitian Creole, Cape Verdean/Port Creole, Portuguese, Vietnamese, Amharic/Ethiopian, and other languages with <3 patients each.

Includes no insurance on file, worker’s compensation, Veterans Affairs insurance, and grant-funded medical coverage.

Includes those receiving active chemotherapy, high-dose steroids, or with HIV with CD4 cell count <200 cells/mm3.

Subjectively reported or recorded temperature >100.4°F.

See Appendix 1 for details of the questions asked to elicit whether any gastrointestinal or respiratory symptoms were present; those with mild, moderate, or severe symptoms were grouped together as having any symptoms in the category present.

Fever was significantly associated with COVID-19-related admissions (aOR, 2.70; 95% CI, 1.65–4.42) when included instead of respiratory symptoms. Using total comorbidities instead of age also yielded similar results, and those with 1–2 (aOR, 5.88; 95% CI, 2.07–16.72) and ≥3 (aOR, 20.40; 95% CI, 7.23–57.54) comorbidities had significantly higher odds of being admitted. Loss of taste/smell was significantly associated with decreased odds of admission (aOR, 0.46; 95% CI, 0.25–0.85). Severe immunosuppression, race/ethnicity, sex, SVI, and clinical trajectory were not significantly associated with admission in any model.

Age, Medicare insurance, and severe immunosuppression were all significantly associated with re-admission in the postdischarge group unadjusted analyses (Table 3); however, we retained age and excluded insurance for the adjusted analyses due to the high correlation we found between age and Medicare status. Older age predicted higher likelihood of re-admission in the adjusted model (aOR, 1.04 for each additional year of age; 95% CI, 1.01–1.06). A second adjusted model including severe immunosuppression also demonstrated significantly decreased re-admission odds with non-Hispanic Black race/ethnicity (aOR, 0.33; 95% CI, 0.11–0.95). Severe immunosuppression and having ≥3 comorbidities each had a large effect size in the adjusted model but did not meet statistical significance.

Table 3.

Hospital Admissions Within 30 Days of Initial Contact by COVID-19 Follow-up Program in All Postdischarge Patients

Variable	Not Admitted Within 30 Days[a] (n = 852), No. (%)	Admitted Within 30 Days[a] (n = 33), No. (%)	Unadjusted OR (95% CI)	With Sex, Adjusted OR (95% CI)b	With Severely Immunosuppressed, Adjusted OR (95% CI)[c]	With Total Comorbidities, Adjusted OR (95% CI)[d]
Age, mean ± SD, y	52.8 ± 19.0	62.9 ± 18.0	1.03 (1.01–1.05)	1.04 (1.01–1.06)	1.04 (1.01–1.06)	–
Female	434 (50.9)	12 (36.4)	0.55 (0.27–1.14)	0.54 (0.26–1.14)	–	–
Race/ethnicity
Non-Hispanic Black or African American	3507 (41.9)	8 (24.2)	0.36 (0.13–1.01)	0.37 (0.13–1.05)	0.33 (0.11–0.95)	0.34 (0.12–0.96)
Non-Hispanic White	112 (13.2)	7 (21.2)	Ref	Ref	Ref	Ref
Hispanic	292 (34.3)	14 (42.4)	0.77 (0.30–1.95)	0.95 (0.36–3.85)	0.87 (0.34–2.26)	0.84 (0.33–2.17)
Another[e]	45 (5.3)	4 (12.1)	1.42 (0.40–5.10)	1.47 (0.40–5.44)	1.20 (0.33–4.40)	1.81 (0.48–6.90)
Unknown	46 (5.4)	0 (0.0)	N/A	N/A	N/A	N/A
language
English	461 (54.1)	18 (54.6)	Ref	–	–	–
Spanish	222 (26.1)	10 (30.3)	1.15 (0.52–2.54)	–	–	–
Other[f]	169 (19.8)	5 (15.2)	0.76 (0.28–2.07)	–	–	–
Insurance
Medicaid	446 (52.4)	15 (45.5)	4.81 (0.63–36.69)	–	–	–
Medicare	222 (26.1)	16 (48.5)	10.3 (1.35–78.47)	–	–	–
Private	143 (16.8)	1 (3.0)	Ref	–	–	–
Other[g]	41 (4.8)	1 (3.0)	3.49 (0.21–56.92)	–	–	–
SVI, mean ± SD	0.696 ± 0.226	0.718 ± 0.218	1.55 (0.31–7.81)	–	–	–
SVI >0.5	728 (85.5)	39 (87.9)	1.24 (0.43–3.57)	–	–	–
Total comorbidities
0	123 (14.4)	2 (6.1)	Ref	–	–	Ref
1–2	405 (47.5)	13 (39.4)	1.97 (0.44–8.86)	–	–	2.07 (0.45–9.59)
3+	324 (38.1)	18 (54.5)	3.42 (0.78–14.93)	–	–	4.25 (0.93–19.50)
Abnormal CXR/chest CT	91 (10.7)	2 (6.1)	0.54 (0.13–2.29)	–	–	–
Severely immunosuppressed[h]	36 (4.2)	4 (12.1)	3.13 (1.04–9.37)	–	3.06 (0.99–9.43)	2.89 (0.93–19.50)
Symptoms
Fatigue/generally feeling sick	159 (18.7)	7 (21.2)	1.17 (0.50–2.75)	–	–	–
Headache/body aches	98 (11.5)	4 (12.1)	1.06 (0.37–3.08)	–	–	–
Nose/throat symptoms	56 (6.6)	0 (0.0)	N/A	–	–	–
Loss of taste or smell	50 (5.9)	2 (6.1)	1.04 (0.24–4.45)	–	–	–
Any fever[i]	68 (8.0)	3 (9.1)	1.15 (0.34–3.88)	–	–	–
Any gastro symptom[j]	50 (5.9)	2 (6.1)	1.04 (0.24–4.45)	–	–	–
Any respiratory symptom[j]	208 (24.4)	6 (18.2)	0.69 (0.28–1.69)	–	–	–
Scaled number of symptoms
0	551 (64.7)	23 (69.7)	Ref	–	–	–
1	106 (12.4)	4 (12.1)	0.90 (0.31–2.67)	–	–	–
2	90 (10.6)	2 (6.1)	0.53 (0.12–2.30)	–	–	–
3	47 (5.5)	1 (3.0)	0.51 (0.07–3.86)	–	–	–
4+	58 (6.8)	3 (9.1)	1.24 (0.36–4.25)	–	–	–
Clinical trajectory
Symptoms resolved	217 (25.5)	10 (30.3)	Ref	–	–	–
Improving	187 (22.0)	6 (18.2)	0.70 (0.25–1.95)	–	–	–
Staying the same	48 (5.6)	4 (12.1)	1.81 (0.54–6.01)	–	–	–
Worsening	3 (0.4)	0 (0.0)	N/A	–	–	–
Unknown	397 (46.6)	13 (39.4)	0.71 (0.31–1.65)	–	–	–

Abbreviations: COVID-19, coronavirus disease 2019; CT, computed tomography scan; CXR, chest x-ray; OR, odds ratio; SVI, social vulnerability index.

Within 30 days of first COVID-19 follow-up team contact.

Only controlled for 3 variables (age, female, and severely immunosuppressed) due to low number of outcome variables.

Only controlled for 3 variables (race/ethnicity, total comorbidities, and severely immunosuppressed) due to low number of outcome variables.

Includes Asian, American Indian/Native American, Native Hawaiian/Pacific Islander, and “other.”

Includes Haitian Creole, Cape Verdean/Port Creole, Portuguese, Vietnamese, Amharic/Ethiopian, and other languages with <3 patients each.

Includes no insurance on file, worker’s compensation, Veterans Affairs insurance, and grant-funded medical coverage.

Includes those receiving active chemotherapy, high-dose steroids, or with HIV with CD4 cell count <200 cells/mm3.

Subjectively reported or recorded temperature >100.4°F.

See Appendix 1 for details of the questions asked to elicit whether any gastrointestinal or respiratory symptoms were present; those with mild, moderate, or severe symptoms were grouped together as having any symptoms in the category “present.”

Those stratified as moderate or high risk had twice the odds of 30-day admission or re-admission compared with those who reported resolution of symptoms (OR, 1.93; 95% CI, 0.97–3.83; aOR, 2.02; 95% CI, 1.19–3.44, respectively) (Table 4).

Table 4.

Hospital Admissions Within 30 Days of Initial Contact by Initial Call Risk Stratification

Variable	Not Admitted Within 30 Days[a] (n = 3954), No. (%)	Admitted Within 30 Days[a] (n = 116), No. (%)	Unadjusted OR (95% CI)[b]
Initial call risk stratification
Sent to emergency department	14	0 (0.0)	N/A
High (next call tomorrow)	237 (6.0)	12 (10.3)	1.93 (0.97–3.83)
Moderate (next call in 2 d)	526 (13.3)	28 (24.1)	2.02 (1.19–3.44)
Low (next call in 3 d)	769 (19.4)	19 (16.4)	0.94 (0.52–1.69)
Symptoms resolved, no follow-up needed	1103 (27.9)	29 (25.0)	Ref
Unknown	1305 (33.0)	28 (24.1)	–

Abbreviations: COVID-19, coronavirus disease 2019; OR, odds ratio.

Within 30 days of first COVID-19 Follow-up Program contact.

Only patients with known initial risk stratification were included in regression model.

The longitudinal group sensitivity analysis included 1683 individuals (see Appendix 4 for characteristics of those excluded). Compared with the full cohort, the longitudinally followed group had higher prevalence of each individual symptom and comorbidity and greater risk of any acute care utilization within 30 days (9.6% vs 7.6%). Older age, Medicare insurance, report of fatigue/generally feeling sick, headaches/body aches, fever, gastrointestinal or respiratory symptoms, >3 symptoms, any comorbidities, and severe immunosuppression were all associated in unadjusted analyses with increased odds of any acute care utilization in the outpatient group (Appendix 5). We retained age and excluded insurance again for the adjusted analyses due to the high correlation we found between age and Medicare status. The adjusted model demonstrated increased age (aOR, 1.03; 95% CI, 1.01–1.05), presence of fever (aOR, 1.87; 95% CI, 1.22–2.86) or gastrointestinal symptoms (aOR, 2.00; 95% CI, 1.29–3.09), ≥5 comorbidities (aOR, 5.84; 95% CI, 2.45–13.91), and severe immunosuppression (aOR, 6.24; 95% CI, 2.44–15.98) to each be significantly associated with increased odds of COVID-19-related acute care utilization. Among the postdischarge longitudinal group, older age and the presence of gastrointestinal symptoms were associated with increased adjusted odds of acute care utilization (aOR, 1.03; 95% CI, 1.00–1.05; and aOR, 2.88; 95% CI, 1.13–7.36, respectively) (Appendix 6). Non-Hispanic Black race was associated with decreased risk of acute care utilization (aOR, 0.20; 95% CI, 0.07–0.63).

Pediatric characteristics are shown in Appendix 7. Analyses excluding children revealed no additional significant risk factors for hospital admission (data not shown).

DISCUSSION

In this large study of outpatients monitored for ongoing COVID-19 symptoms after an initial positive SARS-CoV-2 PCR test, we found that older age, the presence of fever, respiratory symptoms, or gastrointestinal symptoms at initial contact, and high-risk comorbidities (particularly having ≥3) were associated with increased risk of initial COVID-19-related 30-day hospital admission. Loss of taste/smell, conversely, was associated with decreased admission risk. Among patients followed postdischarge from a COVID-19-related admission, older age was also associated with re-admission risk, and non-Hispanic Black race/ethnicity was associated with decreased re-admission risk. Clinical trajectory was not associated with admission or re-admission risk in any analyses but may have been limited by few individuals reporting worsening disease at initial contact (<4%).

Severely immunosuppressed individuals had more admissions, re-admissions, or any acute care utilizations in the 30 days following initial follow-up team contact in all groups, but only had significantly increased odds for any acute care utilization in the longitudinally followed group, perhaps due to the small number of severely immunosuppressed individuals overall.

These findings are consistent with European studies and 1 other US study, which also found that age >45 years, male sex, obesity, cancer, diabetes, chronic renal, liver, respiratory and/or heart disease, immunosuppression, fever, and shortness of breath were associated with increased risk of hospital admission and that loss of taste/smell was associated with decreased hospitalization risk [8-10]. Studies differed on the role of race as a risk factor. Our study adds to this evidence base and further contributes our findings that primary language, SVI, and Medicaid insurance were not significantly associated with risk of COVID-19-related acute care utilization. Although many social determinants of health can affect risk of infection from COVID-19 and other conditions [12,18], our results do not reflect increased admissions or re-admissions due to these factors in our study population. We also found, consistent with another US study [10], that those who identify as non-Hispanic Black were less likely to be re-admitted in the postdischarge group. Interestingly, other studies in the United States have found much higher rates of hospitalization among Black and Latino patients compared with our program data [19,20]. Although causality cannot be determined with the present data, our program’s lack of higher hospitalization rates among these patients could indicate that this type of equity-centric home monitoring program improves health care access among these groups and helps combat disproportionately high hospitalization rates. This lack of higher hospitalization rate among Black and Latino patients also may be impacted BMC’s many outreach programs, wide interpreter use, and health care equity improvement efforts overall [21]. Alternatively, this may be because our study only examines individuals already diagnosed with COVID-19, and therefore does not capture differences in initial health care utilization or COVID-19 infection risk.

This study has several limitations. The retrospective design relied on accuracy and completeness of EHR template data. We did not prospectively follow patients to capture non-BMC admissions or out-of-hospital deaths. Despite a large sample size (n = 4070), the study was limited by few acute care utilizations, particularly in the postdischarge and longitudinal groups. We therefore could not control for many variables, including individual comorbidities rather than total comorbidities combined, in any 1 model and had limited power. However, we still were able to observe significant associations with several variables across groups. Another possible limitation is selection bias; we excluded those unable to be reached by the COVID-19 Follow-up Program as their data were incomplete. Obtaining greater data on those not contacted by the COVID-19 Follow-Up Program would be beneficial to the evaluation of our program’s impact, but we are limited by the data available to abstract. Additionally, as successfully contacted patients may have differed from those who did not seek care or respond to calls, this study likely missed milder cases and did not capture sociodemographic differences of those not reached. Other unmeasured variables, such as homelessness, which can affect admission rates, could have confounded our results. This was mitigated in our population by standard use of nonhospital recuperation and isolation units set up for unstably housed individuals with COVID-19 [22].

Finally, this study occurred during circulation of primarily the SARS-CoV-2 Alpha strain and before widespread SARS-CoV-2 vaccination. In Massachusetts, vaccine rollout for those aged 65+ and those with 2 or more certain comorbidities began in February 2021, when our study period ended [23]. Risk predictors may differ for other variants and in vaccinated hosts. However, the majority of severe COVID-19 infections and hospitalizations continue to occur in unvaccinated individuals [24].

Our data highlight acute care utilization risk factors for COVID-19 outpatient treatment and home monitoring programs to prioritize patients who will benefit from therapeutics and/or close follow-up. These findings add to prior literature to confirm the importance of age but also to underline the impact the number of high-risk comorbidities and particular symptoms may have in predicting hospital admission risk. These data are critical to incorporate into priority tiering of patients for limited-availability antivirals and monoclonal antibodies at this time [5-7]. Monitoring programs may also benefit patients, reduce disparities in health care utilization, and reduce unneeded health care utilization through ongoing proactive assessments. Although vaccination rates are increasing worldwide, the end of the COVID-19 pandemic is still not in sight. Particularly given disparities in vaccine access and hesitancy [25,26] and limited hospital capacity nationally and worldwide, we must continue to support programs and research to understand how to best equitably prevent COVID-19 health care utilizations.

Appendix 1A.

Clinical Assessment Guide for Symptom Stratification Used by BMC COVID-19 Follow-up Program

Symptom Stratification Clinical Assessment
Symptom Assessment	Mild	Moderate	Severe
How is your breathing?	New cough and no SOB In patient with chronic cough, cough worse and no SOB	Cough with mild SOB Aware of breathing but comfortable Able to complete sentence without taking a breath midsentence Able to climb a flight of stairs without losing breath. If at baseline has dyspnea with climbing stairs, worse from baseline	SOB with 1 flight of stairs Any chest pain Unable to speak in full sentences
What is your oxygen saturation	O2 Sat ≥94% and has not fallen by >4% in last 4–8 h	O2 Sat 91%–94% and has not fallen by >4% in last 4–8 h	O2 Sat ≤90% OR O2 sat has fallen by >4% in last 4–8 h
What is your temperature?	Temperature <100.4°F OR subjective no fever	Temperature 100.4°F–102.5°F but responding to fever medicine OR subjective fever	Temperature >102.5°F or >100.4°F and not responsive to fever medicine OR subjective fever unresponsive to fever medicine and/or confusion
How is your intake of liquids?Are you having vomiting or diarrhea?	Mild vomiting/diarrhea Able to drink liquids Urinating every 4–6 h	Moderate vomiting or diarrhea Decreased fluid intake (<50% usual) Urinating at least 3x daily, has tears	Severe vomiting or diarrhea Unable to keep fluids down Decreased urine output to <3x daily Syncope or near syncope
Are you (or your family member) more confused than usual?	Mentation normal/at baseline	Mentation is normal/at baseline	Mentation not at baseline: confused, waxing and waning consciousness, not able to concentrate, hallucinating
Have you had a change in your mobility or a fall?	Function is normal Able to perform ADLs without change in level of assistance	Function is mildly reduced but patient is able to manage daily function safely Needs some increased assistance in performing ADLs from baseline	Sustained a fall Function severely reduced Needs significantly increased assistance in performing ADLs from baseline

Abbreviations: ADLs, activities of daily living; SOB, shortness of breath.

Appendix 1B.

Clinical Assessment Guide for Symptom Stratification Used by BMC COVID-19 Follow-up Program

Symptoms & Clinical Trajectory[a]	With Risk Factors	No Risk Factors
Any severe symptom(s)	Send to ED	Send to ED
Moderate respiratory + worsening	Send to ED	Send to ED vs high risk (next day)
Moderate respiratory + stable	Send to ED vs high risk (next day)	Send to ED vs high risk (next day) vs moderate risk (2nd day)
Moderate respiratory + improving	Moderate risk (2nd day)	Moderate risk (2nd day)
Mild respiratory + worsening	Send to ED vs high risk (next day)	High risk (next day) vs moderate risk (2nd day) vs send to ED
Mild respiratory + stable/improving	Moderate risk (2nd day) vs graduate[b]	Low risk (3rd day) vs graduate[b]
Moderate fever/GI symptoms/mobility only	Moderate risk (2nd day)	Moderate risk (2nd day)
Mild fever/GI symptoms/mobility only	Low risk (3rd day) vs graduate[b]	Low risk (3rd day) vs graduate[b]
No symptoms	Low risk (3rd day) vs graduate[b]	Low risk (3rd day) vs graduate[b]

Abbreviations: BMC, Boston Medical Center; COVID-19, coronavirus disease 2019; ED, emergency department; GI, gastrointestinal; ICU, intensive care unit.

These are guides as to appropriate follow-up; clinical judgment takes precedence when determining action.

Criteria for graduation (we stop following patients in the team; patient is cleared from isolation): currently inpatient (admitted); cannot reach the patient by any means for 3 days; meet symptom-based clearance for isolation (More than 10 days have passed since symptom onset, 1 day without fever (without use of antipyretics), AND 1 day with improvement in other symptoms; OR for patients admitted to the ICU or receiving a biologic, they must isolate for 20 days from symptom onset).

Appendix 3.

Primary Diagnoses and Associated ICD-10 Codes Used for Determining COVID-19-associated Hospital Admissions

Primary Diagnosis	ICD-10 Code
Other specified sepsis	A41.89
Sepsis, unspecified organism	A41.9
Hb-SS disease with acute chest syndrome	D57.01
Neutropenia, unspecified	D70.9
Insomnia, unspecified	G47.00
Other specified cardiac arrhythmias	I49.8
Pneumonia due to COVID-19	J12.82
Moderate persistent asthma with (acute) exacerbation	J45.41
Decreased fetal movements, third trimester, not applicable or unspecified	O36.8130
Other viral diseases complicating childbirth	O98.52
Cough	R05
Dyspnea, unspecified	R06.00
Shortness of breath	R06.02
Precordial pain	R07.2
Other chest pain	R07.89
Epigastric pain	R10.13
Right lower quadrant pain	R10.31
Unspecified abdominal pain	R10.9
Nausea with vomiting, unspecified	R11.2
Delirium	R41.0
Altered mental status, unspecified	R41.82
Dizziness and giddiness	R42
Fever, unspecified	R50.9
Localized enlarged lymph nodes	R59.0
COVID-19	U07.1
Contact with and (suspected) exposure to COVID-19	Z20.822
Contact with and (suspected) exposure to other viral communicable diseases	Z20.828

Abbreviations: COVID-19, coronavirus disease 2019; Hb-SS, sickle cell anemia ; ICD-10, International Classification of Diseases, 10th Edition.

Appendix 4.

Baseline Characteristics of Those Included in Longitudinal Group Sensitivity Analysis Compared With Those who Were Excluded (Community Health Center Patients, Those Asymptomatic by First Contact and Not Needing Follow-up, and Those Unable to Be Reached)

Variable	Longitudinal Group (n = 1683), No. (%)	Community Health Center Patients (n = 875), No. (%)	Asymptomatic by First Contact(n = 1058), No. (%)	Unable to Be Reached(n = 454), No. (%)
Age, mean ± SD, y	45.9 ± 21.2	44.7 ± 20.1	40.5 ± 18.9	44.3 ± 19.3
Female	923 (54.0)	465 (52.8)	550 (52.0)	218 (48.2)
Race/ethnicity
Non-Hispanic Black or African American	835 (47.0)	278 (28.3)	430 (37.9)	209 (42.4)
Non-Hispanic White	136 (8.3)	117 (13.9)	118 (11.0)	64 (14.1)
Hispanic	519 (32.5)	357 (42.6)	363 (36.0)	128 (30.0)
Another[a]	59 (3.9)	65 (7.8)	47 (4.7)	11 (2.8)
Unknown	134 (8.3)	58 (7.3)	100 (10.5)	42 (10.6)
Language
English	1045 (59.1)	487 (54.2)	646 (58.3)	287 (59.9)
Spanish	333 (21.7)	248 (29.5)	241 (25.0)	82 (19.4)
Other[b]	305 (19.3)	140 (16.3)	171 (16.7)	85 (20.6)
Insurance
Medicaid	889 (56.3)	498 (59.1)	586 (57.0)	250 (57.7)
Medicare	309 (15.4)	120 (11.4)	90 (6.4)	72 (12.8)
Private	398 (22.7)	184 (21.2)	291 (27.0)	103 (22.4)
Other[c]	87 (5.6)	73 (8.3)	91 (9.6)	29 (7.2)
SVI, mean ± SD	0.697 ± 0.247	0.686 ± 0.240	0.674 ± 0.251	0.667 ± 0.260
SVI >0.5	1381 (81.3)	725 (81.6)	845 (79.6)	354 (76.7)

Abbreviations: SVI, social vulnerability index; VA, Veterans Affairs.

Includes Asian, American Indian/Native American, Native Hawaiian/Pacific Islander, and “other.”

Includes Haitian Creole, Cape Verdean/Port Creole, Portuguese, Vietnamese, Amharic/Ethiopian, and other languages with <3 patients each.

Includes no insurance on file, worker’s compensation, VA insurance, and grant-funded medical coverage.

Appendix 5.

Acute Care Utilization (ED Visit, Observation Stay, or Hospital Admission) Within 30 Days of Initial Contact by COVID-19 Follow-up Program in All Patients Initially Tested in Outpatient Setting

Outpatient Group Only
Variable	No Acute Care Utilization Within 30 Days[a] (n = 1322), No. (%)	Any Acute Care Utilization Within 30 Days[a] (n = 130), No. (%)	Unadjusted OR (95% CI)	Adjusted OR (95% CI)
Age, mean ± SD, y	43.2 ± 20.9	50.45 ± 20.43	1.02 (1.01–1.03)	1.03 (1.01–1.05)
Female	754 (57.0)	68 (52.3)	0.83 (0.57–1.19)	0.72 (0.49–1.08)
Race/ethnicity
Non-Hispanic Black or African American	647 (48.9)	56 (43.1)	0.92 (0.46–1.87)	1.18 (0.54–2.57)
Non-Hispanic White	104 (7.9)	10 (7.7)	Ref	Ref
Hispanic	413 (31.2)	49 (37.7)	1.28 (0.63–2.62)	1.53 (0.70–3.35)
Another[b]	47 (3.6)	5 (3.9)	1.11 (0.36–3.44)	1.68 (0.51–5.54)
Unknown	111 (8.4)	10 (7.7)	0.94 (0.38–2.36)	1.48 (0.55–3.94)
language
English	842 (63.7)	73 (56.2)	Ref	–
Spanish	256 (19.4)	32 (24.6)	1.47 (0.95–2.29)	–
Otherc	224 (16.9)	25 (19.2)	1.27 (0.79–2.04)	–
Insurance
Medicaid	727 (55.0)	65 (50.0)	0.97 (0.62–1.52)	–
Medicare	187 (14.2)	30 (23.1)	1.75 (1.02–2.98)	–
Private	336 (25.4)	31 (23.9)	Ref	–
Other[d]	72 (5.5)	4 (3.1)	0.58 (0.20–1.70)	–
SVI, mean ± SD	0.691 ± 0.252	0.693 ± 0.243	1.05 (0.51–2.15)	1.02 (0.46–2.25)
SVI >0.5	1071 (81.0)	107 (82.3)	1.11 (0.69–1.78)	–
Total comorbidities
0	351 (26.6)	17 (3.1)	Ref	Ref
1–2	652 (49.3)	59 (45.4)	1.91 (1.10–3.33)	1.76 (0.96–3.21)
3–4	265 (20.0)	34 (26.2)	2.75 (1.50–5.04)	2.00 (0.99–4.05)
5+	54 (4.1)	20 (15.4)	8.26 (4.05–16.87)	5.84 (2.45–13.91)
Severely immunosuppressed[e]	20 (1.5)	10 (7.7)	7.88 (3.38–18.35)	6.24 (2.44–15.98)
Symptoms
Fatigue/generally feeling sick	520 (39.3)	78 (66.2)	2.44 (1.69–3.53)	1.45 (0.94–2.24)
Headache/body aches	636 (48.1)	86 (66.2)	2.24 (1.53–3.27)	1.39 (0.88–2.19)
Nose/throat symptoms	508 (38.4)	49 (37.7)	1.00 (0.69–1.45)	–
Loss of taste or smell	436 (33.0)	48 (36.9)	1.23 (0.85–1.79)	–
Any fever[f]	361 (27.3)	64 (49.2)	2.78 (1.93–4.01)	1.87 (1.22–2.86)
Any gastro symptom[g]	217 (16.4)	45 (34.6)	2.95 (1.99–4.37)	2.00 (1.29–3.09)
Any respiratory symptom[g]	695 (52.6)	90 (69.3)	2.12 (1.44–3.13)	1.27 (0.82–1.97)
Scaled number of symptoms
0	249 (18.8)	11 (8.5)	Ref	–
1	238 (18.0)	12 (9.2)	1.14 (0.49–2.64)	–
2	221 (16.7)	19 (14.6)	2.03 (0.94–4.35)	–
3	223 (16.9)	21 (16.2)	2.23 (1.05–4.73)	–
4+	391 (29.6)	67 (51.5)	4.27 (2.21–8.25)	–
Clinical trajectory
Symptoms resolved	397 (30.0)	43 (33.1)	Ref	–
Improving	334 (25.3)	39 (30.0)	1.06 (0.67–1.68)	–
Staying the same	370 (28.0)	36 (27.7)	0.90 (0.56–1.43)	–
Worsening	40 (3.0)	5 (3.9)	1.22 (0.46–3.29)	–
Unknown	181 (13.7)	7 (5.4)	0.33 (0.15–0.75)	–

Abbreviations: COVID-19, coronavirus disease 2019; ED, emergency department; OR, odds ratio; SVI, social vulnerability index; VA, Veterans Affairs.

Within 30 days of first COVID-19 follow-up team contact.

Includes Asian, American Indian/Native American, Native Hawaiian/Pacific Islander, and “other.”

cIncludes Haitian Creole, Cape Verdean/Port Creole, Portuguese, Vietnamese, Amharic/Ethiopian, and other languages with <3 patients each.

Includes no insurance on file, worker’s compensation, VA insurance, and grant-funded medical coverage.

Includes those receiving active chemotherapy, high-dose steroids, and/or having HIV with CD4 <200.

Subjectively reported or recorded temperature >100.4°F.

See Appendix 1A for details of the questions asked to elicit whether any gastrointestinal or respiratory symptoms were present; those with mild, moderate, or severe symptoms were grouped together as having any symptoms in the category “present.”

Appendix 6.

Acute Care Utilization (ED Visit, Observation Stay, or Hospital Admission) Within 30 Days of Initial Contact by COVID-19 Follow-up Program in All Postdischarge Patients

Postdischarge Group Only
Variable	No Acute Care Utilization Within 30 Days[a] (n = 298), No. (%)	Any Acute Care Utilization Within 30 Days[a] (n = 31), No. (%)	Unadjusted OR (95% CI)	Adjusted OR (95% CI)[b]
Age, mean ± SD, y	55.1 ± 19.0	61.61 ± 14.99	1.02 (1.00–1.04)	1.03 (1.00–1.05)
Female	141 (47.3)	13 (41.9)	0.79 (0.37–1.67)	–
Race/ethnicity
Non-Hispanic Black or African American	169 (56.7)	11 (35.5)	0.23 (0.08–0.69)	0.20 (0.07–0.63)
Non-Hispanic White	25 (8.4)	6 (19.4)	Ref	Ref
Hispanic	78 (26.2)	13 (41.9)	0.59 (0.20–1.73)	0.56 (0.18–1.72)
Another[c]	8 (2.7)	1 (3.2)	0.42 (0.04–4.03)	0.50 (0.05–5.07)
Unknown	18 (6.0)	0 (0.0)	N/A	N/A
Language
English	173 (58.1)	18 (58.1)	Ref	–
Spanish	66 (22.2)	9 (29.0)	1.45 (0.62–3.40)	–
Other[d]	59 (19.8)	4 (12.9)	0.58 (0.19–1.78)	–
Insurance
Medicaid	139 (46.6)	11 (35.5)	0.63 (0.22–1.79)	–
Medicare	96 (32.2)	13 (41.9)	1.07 (0.38–3.01)	–
Private	51 (17.1)	6 (19.4)	Ref	–
Other[e]	12 (4.0)	1 (3.2)	0.63 (0.07–5.70)	–
SVI, mean ± SD	0.722 ± 0.223	0.725 ± 0.185	1.04 (0.20–5.58)	–
SVI >0.5	260 (87.2)	28 (90.3)	1.42 (0.41–4.91)	–
Total comorbidities
0	24 (8.1)	3 (9.7)	Ref	–
1–2	119 (39.9)	10 (32.3)	0.64 (0.16–2.53)	–
3–4	117 (39.3)	15 (48.4)	1.02 (0.27–3.82)	–
5+	38 (12.8)	3 (9.7)	0.61 (0.11–3.30)	–
Severely immunosuppressed[f]	20 (6.7)	4 (12.9)	2.04 (0.65–6.42)	–
Symptoms
Fatigue/generally feeling sick	127 (42.6)	15 (48.4)	1.31 (0.62–2.75)	–
Headache/body aches	69 (23.2)	7 (22.6)	0.97 (0.40–2.34)	–
Nose/throat symptoms	45 (15.1)	5 (16.1)	1.15 (0.42–3.18)	–
Loss of taste or smell	44 (14.8)	6 (19.4)	1.44 (0.56–3.73)	–
Any fever[g]	53 (17.8)	6 (19.4)	1.13 (0.44–2.91)	–
Any gastro symptom[h]	40 (13.4)	8 (25.8)	2.60 (1.08–6.29)	2.88 (1.13–7.36)
Any respiratory symptom[h]	159 (53.4)	16 (51.6)	0.95 (0.45–2.00)	–
Scaled number of symptoms
0	79 (26.5)	7 (22.6)	Ref	–
1	63 (21.1)	9 (29.0)	1.71 (0.60–4.85)	–
2	72 (24.2)	3 (9.7)	0.47 (0.12–1.89)	–
3	33 (11.1)	4 (12.9)	1.47 (0.40–5.38)	–
4+	51 (17.1)	8 (25.8)	1.96 (0.67–5.75)	–
Clinical trajectory
Symptoms resolved	79 (26.5)	8 (25.8)	Ref	–
Improving	140 (47.0)	13 (41.9)	1.05 (0.40–2.75)	–
Staying the same	42 (14.1)	7 (22.6)	1.90 (0.62–5.80)	–
Worsening	3 (1.0)	0 (0.0)	N/A	–
Unknown	34 (11.4)	3 (9.7)	0.93 (0.23–3.83)	–

Abbreviations: COVID-19, coronavirus disease 2019; ED, emergency department; OR, odds ratio; SVI, social vulnerability index; VA, Veterans Affairs.

Within 30 days of first COVID-19 follow-up team contact.

Only controlled for 3 variables due to low number of outcome variables.

Includes Asian, American Indian/Native American, Native Hawaiian/Pacific Islander, and “other.”

Includes Haitian Creole, Cape Verdean/Port Creole, Portuguese, Vietnamese, Amharic/Ethiopian, and other languages with <3 patients each.

Includes no insurance on file, worker’s compensation, VA insurance, and grant-funded medical coverage.

Includes those receiving active chemotherapy, high-dose steroids, and/or having HIV with CD4 <200.

Subjectively reported or recorded temperature >100.4°F.

See Appendix 1A for details of the questions asked to elicit whether any gastrointestinal or respiratory symptoms were present; those with mild, moderate, or severe symptoms were grouped together as having any symptoms in the category “present.”

Appendix 7.

Demographic Characteristics, Comorbidities, Clinical Disease Data, and Follow-up Outreach Among All Patients Aged 0–17 Years

Variable	Patients Aged 0–17 (n = 356), No. (%)
Age, mean ± SD, y	8.3 ± 5.8
Female	182 (51.1)
Race/ethnicity
Non-Hispanic Black or African American	145 (40.7)
Non-Hispanic White	23 (6.5)
Hispanic	117 (32.9)
Another[a]	17 (4.8)
Unknown	54 (15.2)
Primary language
English	235 (66.0)
Spanish	53 (14.9)
Other[b]	68 (19.1)
Insurance
Medicaid	292 (82.0)
Private	52 (14.6)
Other[c]	12 (3.4)
SVI, mean ± SD	0.695 ± 0.256
SVI >0.5	285 (80.1)
Comorbidities[d]
Chronic lung disease	66 (18.5)
Obesity	65 (18.3)
Total comorbidities
0	244 (68.5)
1–2	110 (30.9)
3–4	2 (0.6)
Symptoms
Fatigue/generally feeling sick	63 (17.7)
Headache/body aches	70 (19.7)
Nose/throat symptoms	105 (29.5)
Loss of taste or smell	44 (12.4)
Any fever[e]	39 (11.0)
Any gastro symptom[f]	24 (6.7)
Any respiratory symptomg	92 (25.8)
Scaled number of symptoms
0	165 (46.4)
1	71 (19.9)
2	58 (16.3)
3	31 (8.7)
4+	31 (8.7)
Clinical trajectory
Symptoms resolved	1106 (34.7)
Improving	742 (23.3)
Staying the same	718 (22.5)
Worsening	55 (1.7)
Unknown	564 (17.7)
Total follow-up calls within 30 d
<3	225 (63.2)
3–5	115 (32.3)
6+	16 (4.5)
Initial call risk stratification
High (next call tomorrow)	9 (2.5)
Moderate (next call in 2 d)	47 (13.2)
Low (next call in 3 d)	108 (30.3)
Symptoms resolved, no follow-up needed	104 (29.2)
CHC patient, no follow-up needed	65 (18.3)
Unable to reach after 3 attempts	23 (6.5)

Includes Asian, American Indian/Native American, Native Hawaiian/Pacific Islander, and “other.”

Includes Haitian Creole, Cape Verdean/Port Creole, Portuguese, Vietnamese, Amharic/Ethiopia, and other languages with <3 patients each.

Includes no insurance on file, worker’s compensation, Veteran’s Administration insurance, and grant-funded medical coverage.

We only listed those comorbidities with 5 or more individuals; <5 individuals had severe immunosuppression.

Subjectively reported or recorded temperature >100.4°F.

See Appendix 1A for details of the questions asked to elicit whether any gastrointestinal or respiratory symptoms were present; those with mild, moderate, or severe symptoms were grouped together as having any symptoms in the category “present.”