a b s t r a c t

Study objective: During the COVID-19 pandemic, prescribing supplemental oxygen was a common reason for hos- pitalization of patients. We evaluated outcomes of COVID-19 patients discharged from the Emergency Depart- ment (ED) with home oxygen as part of a program to decrease hospital admissions.

Methods: We retrospectively observed COVID-19 patients with an ED visit resulting in direct discharge or obser- vation from April 2020 to January 2022 at 14 hospitals in a single healthcare system. The cohort included those discharged with new oxygen supplementation, a Pulse oximeter, and return instructions. Our primary outcome was subsequent hospitalization or death outside the hospital within 30 days of ED or observation discharge.

Results: Among 28,960 patients visiting the ED for COVID-19, providers admitted 11,508 (39.7%) to the hospital, placed 907 (3.1%) in Observation status, and discharged 16,545 (57.1%) to home. A total of 614 COVID-19 patients (535 discharge to home and 97 observation unit) went home on new oxygen therapy. We observed the primary outcome in 151 (24.6%, CI 21.3-28.1%) patients. There were 148 (24.1%) patients subsequently hospitalized and 3 (0.5%) patients who died outside the hospital. The subsequent hospitalized mortality rate was 29.7% with 44 of the 148 patients admitted to the hospital dying. Mortality all cause at 30 days in the entire cohort was 7.7%.

Conclusions: Most patients discharged to home with new oxygen for COVID-19 safely avoid later hospitalization and few patients die within 30 days. This suggests the feasibility of the approach and offers support for ongoing research and implementation efforts.

(C) 2023

1. Background

The COVID-19 pandemic placed a burden on the United States healthcare system with peaks of up to 160,000 simultaneously hospital- ized patients and over 1,100,000 deaths thus far [1]. The majority of COVID-19 illness is mild or moderate and requires limited medical ther- apy for recovery [2].

Oxygen supplementation is one of the most common therapies dur- ing hospitalization for COVID-19 [3,4]. Because many patients require only supplemental oxygen until recovery, and since oxygenation moni- toring is readily available outside the hospital, home oxygen therapy is an alternative to hospitalization that could reduce the overall stress on hospital capacity and allow recovery at home, both attractive options. Home oxygen therapy reduces hospitalization in children with RSV [5],

* Corresponding author at: 3550 Terrace Street, Alan Magee Scaife Hall, Suite 600, Pittsburgh, PA 15213, USA.

E-mail address: [email protected] (A. Schoenling).

but there are few data and no trials about the feasibility and safety of home oxygen for acutely hypoxic adults with COVID-19.

In response to the rapid demand for healthcare services, our healthcare system developed an ED-initiated home oxygen program for COVID-19 patients. We anticipated that most patients receiving home oxygen therapy would recover with few progressing to hospital- ization or death. This study aims to document the primary outcome of subsequent hospitalization or death outside a hospital for patients en- rolled in our discharge with home oxygen program to inform current and future efforts on patient care during events that stress Healthcare resources.

1. Methods
   1. Study design

We retrospectively examined patients discharged from 14 hospitals in a single Pennsylvania healthcare system from April 2020 to January 2022 in accordance with STROBE reporting guidelines for observational

https://doi.org/10.1016/j.ajem.2023.03.012

0735-6757/(C) 2023

studies [6]. The University’s Human Research Protection Office Institu- tional Review Board approved a waiver of informed consent to review these medical records. One emergency physician and one research as- sistant manually abstracted data by electronic health record chart review. Chart reviewers were not blinded to the purpose of the study. We performed cross-review on 6% of charts and found a Cohen’s Kappa of 0.9 and 95% agreement between abstractors. All data was stored in a secure REDCap database [7].

• 1. Study population

We included patients aged 18 years and older discharged with a new order of home oxygen therapy following ED or observation visit for COVID-19 respiratory illness - identifying these patients by a systemwide record of Durable Medical Equipment (DME) orders and associated ICD coding. We adjudicated Covid-19 respiratory illness by reviewing EHR documentation and Diagnostic tests. We recorded in- system PCR SARS-CoV-2 testing, patient reported out-of-system testing and clinician diagnosis by history. We excluded any people on chronic oxygen therapy, persons lacking a diagnosed COVID-19 infection (diag- nosis of infection either per history and exam, patient-reported outside testing, or in-system SARS-CoV-2 testing), or if home oxygen was unre- lated to COVID-19.

• 1. Outcome measures

The primary patient outcome was death outside the hospital or hos- pitalization within 30 days determined by EHR records of return ED visits or hospitalizations, outpatient visits, ensuing surgeries or proce- dures, external laboratory or Imaging results, pharmacy filled prescrip- tions, and review of online obituaries. We considered patients to be alive without hospitalization from COVID-19 if there was documented resolution of respiratory symptoms, discontinuation of oxygen therapy or completion of the teleHealth monitoring program even if data beyond 30 days was absent. We reported on patient deaths after transitioning to hospice care within 30 days of ED discharge but excluded this from our primary outcome. We determined that home oxygen use in hospice transition preserved inpatient healthcare resources while allowing pa- tients to pursue medical therapies that most aligned with their values. Enrolling in hospice provided evidence that the patient had adequate time and re-evaluation outside the hospital such that they could have returned for hospital admission if patients and/or family desired. We defined lost-to-follow-up as those having an index visit without subse- quent EHR data points or reported death by online obituary during final EHR review. Complete follow up occurred by experiencing the primary outcome or having evidence of living without a primary outcome at 30 days. We analyzed the data without imputation.

• 1. Home oxygen program overview

Clinicians independently determined individual patient appropri- ateness for discharge with home oxygen absent any imposed or protocolized treatment guidelines (as none existed early in the pan- demic). The healthcare system sent scheduled reminders to all physi- cians to consider home oxygen as an alternative. The DME supplier provided services Monday to Friday from 8 AM to 6 PM and Saturday to Sunday from 8 AM to 4:30 PM. EDs either provided an oxygen cylin- der on discharge with planned DME delivery for patients treated outside usual hours of operation or placed patients in Observation Units until as- sured home oxygen delivery. All EDs discharged patients with a per- sonal home pulse oximeter and return guidelines provided by the treating physician. Primary care or other treating physicians and ad- vanced practice providers assumed follow up for patients after dis- charge. The system provided an option for scheduled tele-health follow up or daily at home tele-health monitoring programs based on availability and physician discretion.

• 1. Data collection

We collected the total number of ED patients evaluated for COVID- 19 by an EHR query using ICD-10 code U07.1 for COVID-19 illness. The query noted the disposition of ED patient encounters as admission, placed in observation, or ED discharge. We used the total number of pa- tients admitted plus the addition of our home oxygen discharge cohort as a denominator. We then used the number of home oxygen discharges without subsequent hospitalization from our cohort as a numerator di- vided by the denominator to evaluate the impact home oxygen dis- charges had on COVID-19 hospital admissions.

We collected data from our inpatient and outpatient electronic re- cords, identical across our system. We supplemented internal data with external shared records found in the Pennsylvania CommonWell network Health Information Exchange, EPIC Care Everywhere and on- line obituaries [8,9]. We collected patient demographics and comorbid- ities from the EHR. We defined immunosuppression as those with human immunodeficiency virus, active hematologic or solid organ can- cer, organ transplant on antirejection medication, chronic steroid use, immune-mediated inflammatory disorders, and patients receiving che- motherapy or immunomodulating drugs. Cancer history included any history of cancer or active cancer. Reviewers abstracted vital signs from the EHR for the initial ED visit. When only one set of recorded vital signs existed, these were used as both admission and discharge values.

We collected imaging results from the initial ED episode. We classi- fied results “abnormal” if radiologist interpretation included evidence of opacities, infiltrates, ground glass, airspace disease or reported an in- ability to rule out, suspicion, consistent, or indication of pneumonia. We documented newly prescribed COVID-19 therapies (steroids, anti- biotics, and anticoagulation) from the discharge clinical summary docu- ment. The system guidelines recommended against outpatient antiviral and monoclonal antibodies for patients receiving oxygen therapy and remdesivir was not available as an outpatient infusion; we did not col- lect data on these outpatient therapies for this reason. We counted pre- existing patient prescriptions for COVID-19 therapies as an ED discharge therapy if the ED physician noted the prior prescription as a reason not to provide a prescription in their medical decision making.

• 1. Statistical analysis

We report primary data as count, frequencies, ranges, and rates. We report our variables as median and interquartile ranges (IQR) since there were no normally distributed variables. STATA SE v16.1 was used to perform statistical analysis.

1. Results

Among 28,960 patients evaluated for COVID-19 at the 14 EDs during the study period, providers admitted 11,508 (39.7%) to the hospital, placed 907 (3.1%) in observation units, and discharged 16,545 (57.1%) directly home. Among those discharged home, 632 had new home oxy- gen therapy: 535 from the ED after a median of 6.3 h (IQR 4.6, 8.8), and 97 from an observation unit after a median of 21.8 h (IQR 18.8, 26.8). Eighteen (2.8%) patients were lost-to-follow-up. We analyzed out- comes for 614 (97.2%) patients with complete follow up.

The demographics and comorbidities for the 614 patients are listed in Table 1. ED index visit characteristics are listed in Table 2 and radio- graphic results are detailed in Table 3. EDs confirmed COVID-19 illness by PCR testing for 302 (49.2%) patients, and 296 (48.2%) patients had a patient-reported positive COVID-19 test prior to ED visit. Physicians diagnosed COVID-19 clinically in 16 (2.6%) patients by history and exam without testing. One patient had a segmental pulmonary embo- lism detected among 140 chest computed tomography scan with con- trast.

Table 1

Cohort Demographics with Breakdown by Primary Outcome.

Total (N = 614) Primary Outcome (N = 151) No Primary Outcome (N = 463)

	Median	IQR		Median	IQR		Median	IQR
Age	61.4	50.3, 71.4		64.9	55.0, 74.5		60.2	49.5,70.0
Gender	N	%		N	%		N	%
Male	331	53.9%		87	57.6%		244	52.7%
Female	283	46.1%		64	42.4%		219	47.3%
Race	N	%		N	%		N	%
White	558	90.9%		132	87.4%		426	92.0%
Black/African	33	5.4%		12	8.0%		21	4.5%
Asian	4	0.7%		2	1.3%		2	0.4%
American Indian	1	0.2%		1	0.7%		0	0.0%
Native Hawaiian	0	0.0%		0	0.0%		0	0.0%
NA	18	2.9%		4	2.7%		14	3.0%
Comorbidities	N	%		N	%		N	%
Hypertension	304	49.5%		83	55.0%		221	47.7%
Diabetes	120	19.5%		26	17.2%		94	20.3%
Heart Failure	26	4.2%		8	5.3%		18	3.9%
CAD	42	6.8%		15	9.9%		27	5.8%
CVA/TIA	30	4.9%		4	2.7%		26	5.6%
Asthma	87	14.2%		20	13.3%		67	14.5%
COPD	48	7.8%		10	6.6%		38	8.2%
CKD	17	2.8%		5	3.3%		12	2.6%
ESRD	1	0.2%		0	0.0%		1	0.2%
Cancer Hx	62	10.1%		19	12.6%		43	9.3%
Immunosuppressed	31	5.1%		15	9.9%		16	3.5%

N = Number of Patients, % = Percentage of Patients, IQR = Interquartile Range, NA = not available, CAD = Coronary Artery Disease, CVA = Cerebrovascular Accident, TIA = Transient Ischemic Event, COPD = Chronic Obstructive Pulmonary Disease, CKD = Chronic Kidney Disease, ESRD = end stage renal disease.

Table 2

ED Index Visit Characteristics with Breakdown by Primary Outcome.

Total (N = 614) Primary Outcome (N = 151) No Primary Outcome (N = 463)

Vital Signs	Median	IQR		Median	IQR		Median	IQR
Triage RR	20	18, 22		20	18, 22		20	18, 22
Disch RR	19	18, 22		20	18, 22		19	18, 21
Disch O2%	95	93, 96		95	93, 96		95	93, 96
Disch LPM	2	0, 2		2	2, 2		2	0, 2
SpO2/FiO2	331	317, 424		328	303, 341		331	321, 433
Lowest O2%	89	88, 91		89	88, 91		89	88, 91
BMI	32	27.4, 37.9		32.7	26.9, 38.2		32.0	27.7, 37.8
Symptoms (d)	8	6, 11		7	5, 10		8	6, 11
Visit Type	N	%		N	%		N	%
Emergency	519	84.5%		126	83.4%		393	84.9%
Observation	95	15.5%		25	16.6%		70	15.1%
Discharge Medications	N	%		N	%		N	%
Steroids	318	51.8%		79	52.3%		239	51.6%
Antibiotics	45	7.3%		12	8.0%		33	7.1%
Anticoagulation	4	0.7%		0	0.0%		4	0.9%

N = Number of Patients, % = Percentage of Patients, IQR = Interquartile Range, Disch = Discharge, RR = Respiratory Rate, O2% = Peripheral Oxygen Saturation, LPM = Liters Per Minute, BMI = Body Mass Index, (d) = Days.

The primary outcome occurred in 151 (24.6%, CI 21.3-28.1%) pa- tients in the cohort: 148 (24.1%) were hospitalized and 3 (0.5%) died outside the hospital. Separately, two patients (0.3%) died in hospice fol- lowing ED discharge, one after three days and the other after nine days. There were 461 patients (75.1%) who recovered without requiring

subsequent hospitalization or experiencing death within 30 days of ED home oxygen discharge. All patient outcomes following ED index visit discharge are listed in Table 4. The 148 patients who returned and were admitted spent a median of 3 days (IQR 2, 5) outside the hospital. Their median Hospital Length of Stay was 8 days (IQR 5, 15). The

Table 3

ED Index Visit Radiography with Breakdown by Primary Outcome.

Total (N = 614) Primary Outcome (N = 151) No Primary Outcome (N = 463)

Radiology	N	%	N Abn (%)		N	%	N Abn (%)		N	%	N Abn (%)
None	24	3.9%	-		4	2.6%	-		20	4.3%	-
CXR	439	71.5%	339 (77.2%)		117	77.5%	95 (81.2%)		322	69.5%	244 (75.8%)
CT Chest	151	24.6%	136 (90.1%)		30	19.9%	27 (90%)		121	26.1%	109 (90.1%)

N = Number of Patients, % = Percentage of Patients, N Abn (%) = Number of Abnormal Studies and Percentage, CXR = Chest Radiograph, CT Chest = Computed Tomography of the Chest.

Table 4

Primary Outcomes and Composite Breakdown. Outcomes

	N = 614	%
Primary Outcome	151	24.6%
Hospitalization	148	24.1%
Death	3	0.5%
Recovery	461	75.1%
Hospice	2	0.3%

Legend: N = Number of Patients, % = Percentage of Patient.

Table 5

Admitted Outcome Characteristics.

Characteristic	N = 148	%
HHFNC/NIPPV	85	56.7%
ICU	60	40.5%
IMV	33	22.3%
ECMO	1	0.7%
Mortality	44	29.7%
	Median	IQR
Hosp LOS	8	5, 15
ED to Admit (d)	3	2, 5

N = Number of Patients, % = Percentage of Patients, HHFNC/ NIPPV = Heated High Flow Nasal Canula or non-invasive positive pressure ventilation, ICU = Intensive Care Unit, IMV = Invasive Mechanical Ventilation, ECMO = Extra-Corporeal Membrane Oxygenation, Hosp LOS = Hospital Length of Stay, (d) = Days.

mortality rate of those admitted to the hospital was 29.7% with 44 pa- tients dying. The cohort’s overall mortality rate was 7.7% with 44 (7.2%) patients dying after hospitalization and 3 (0.5%) patients dying outside the hospital. Outcomes and therapies for patients with subse- quent hospitalization are listed in Table 5. The index visit characteristics for ED discharge compared to observation discharge are provided in Table 6.

1. Discussion

Our experience shows it is possible to create and implement an ED- based new home oxygen therapy pathway for COVID 19 patients, and the crude safety signals were not distressing. Most (75.1%) patients pro- vided new home oxygen from the ED recovered outside the hospital with supportive care. For comparison, Terp et al. expanded on the Banerjee et al. study to report on 360 ED patients discharged from a sin- gle center with pulse oximetry, home oxygen and telemedicine follow up. Their study found a rate of 15.8% 30-day unscheduled hospitaliza- tion due to COVID-19 with one patient dying without evidence of

Table 6

ED Index Visit Characteristics by Visit Type.

Emergency	N = 535	%	Observation	N = 97	%
Primary Outcome	126	23.6%	Primary Outcome	25	25.8%
Vital Signs	Median	IQR	Vital Signs	Median	IQR
Triage RR	20	18, 22	Triage RR	20	18, 22
Disch RR	19	18, 22	Disch RR	18	18, 20
Disch O2%	95	93, 96	Disch O2%	94	93, 96
Disch LPM	2	0, 2	Disch LPM	2	2, 3
Lowest O2%	89	88, 91	Lowest O2%	89	87, 91
BMI	32	27.3, 37.8	BMI	31.6	28.4, 38.2
Symptoms (d)	8	6, 10	Symptoms (d)	8	5, 12

N = Number of Patients, % = Percentage of Patients, IQR = Interquartile Range, Disch = Discharge, RR = Respiratory Rate, O2% = Peripheral Oxygen Saturation, LPM = Liters Per Minute, BMI = Body Mass Index, (d) = Days.

being hospitalized at the time. These rates are lower than our reported rates of 24.1% 30-day hospitalization and 0.5% known to die outside a hospital. The findings in this paper do fall within the confidence inter- vals of the worst-case sensitivity analysis performed by Terp et al. Our study describes a cohort that is older, with a median age of 61.4 years compared to 51 years old, with more hypertension and less diabetes. We also reported on patients from 14 different EDs including non- urban centers where Access to healthcare may vary from a single urban center [10,11].

The differences in our findings, compared to the Terp et al. study, may be related to their study having a set clinical pathway with criteria for enrollment and dedicated telephone follow up after discharge. Tele- medicine was available at certain centers within our study but was not provided to every patient. Support for telemedicine was observed in a Guadeloupean study finding subsequent Hospital admission rates of 4.8% and 20% respectively when ED patients discharged with home ox- ygen were followed by a dedicated COVID-19 telehealth team as com- pared to usual care [12]. Similarly, a study in the Netherlands found a 30-day hospital admission rate of 15.9% from a small single center co- hort of ED patients discharged with telemonitoring on less than 3 l of supplemental oxygen [13]. A separate study reported a hospital admis- sion rate of 12.7% within 7 days of ED discharge for severe COVID-19 with virtual follow-up. Almost a quarter of these patients were provided home oxygen for exertional hypoxia defined as SpO2 90-91% during a 1-min walk [14]. Nonetheless, revisiting an ED or later admission alone is not a strong safety concern unless linked to other missed care opportunities.

Our study is not able assess the potential benefits or risks conferred to patients if they had been hospitalized to receive inpatient care versus discharge with home oxygen. Our cohort’s overall mortality was 7.7%, lower than some other reported inpatient mortalities ranging from 10.8%-13.6 [15,16]. It is higher than higher than the 1.4% overall 30- day mortality rate found by Terp et al. but our cohort and implementa- tion of home oxygen pathway are decidedly different [10]. Our lower outpatient mortality rate, compared to reported in-hospital mortality, could be the result of ED physicians selecting a relatively healthier cohort, but could also be associated with avoiding inPatient harms [17,18]. Discharged patients that were subsequently admitted experi- enced a relatively high eventual mortality rate of 29.7%, clearly in part evidence of a sicker subpopulation. It is unknown if inpatient monitor- ing and earlier escalation of care would have impacted these patients’ outcome.

Many patients had oxygen prescribed absent recorded hypoxemia. Additionally, some of these patients were not receiving oxygen therapy during their ED or observation visit. We cannot know why clinicians de- cided home oxygen therapy was indicated for these patients. However, clinician gestalt likely drove home oxygen prescribing in such cases, as the primary outcome was met in 19.3% of those patients with normoxia without in-ED supplemental oxygen recorded at discharge and 25.5% with recorded in-ED supplemental oxygen.

We carefully assessed all patients in our cohort dying within 30 days after ED care. There were only three patients who died outside the hos- pital without returning for care. One patient was evaluated by an inter- nal medicine team in the ED who recommended discharge since the patient was not requiring oxygen during ED evaluation or at discharge. Another patient awaited a bed for 38 h before electing discharge with two liters home oxygen, subsequently dying within 24 h of discharge. The final patient died following a 50-h observation visit and discharge with two liters oxygen to continue recovery at home. Our chart review of these cases did not find any clinical care concerns that could have been acted on to avoid death. Two patients entered hospice following home oxygen discharge dying three and nine days after ED discharge. Discharge with oxygen was potentially in the best interest of these two patients as it likely provided additional time with family in a famil- iar environment prior to death, in accordance with the care and values these patients sought.

Our findings provide additional perspective on a unique cohort that has not been extensively studied since home oxygen therapy has not been a common discharge therapy from the ED. We estimate that 4% of expected hospitalizations were avoided using the program assuming patients discharged with new home oxygen would have otherwise been admitted. This may underestimate the achievable impact as new prac- tice patterns take time for wide adoption among providers. Even small reductions in admissions have an important impact on ED and Hospital overcrowding. Home oxygen therapy is an important tool for future COVID-19 surges and other respiratory pandemics. We believe our find- ings support the need for further investigation into the impact of ED home oxygen prescribing on patient outcomes and healthcare utilization.

• 1. Limitations

Our study is limited by the observational retrospective nature based on electronic medical record review. Our findings may not be generaliz- able as we reported on an overwhelmingly white (90.9%) cohort, aligned with racial demographics of the included regions. [19] Data on the vaccination status of patients was not available, which could have an impact on the likelihood of our primary outcome. Lastly, the stan- dards of care during the pandemic varied based on regional prevalence and resources as well as secular trends.

1. Conclusion

Home oxygen therapy is feasible for COVID-19 patients after emer- gency department discharge and later admission is low while death at 30-days is very low. The safety and feasibility of this practice will likely increase as advances in home monitoring and telehealth continue to progress. The addition of home oxygen to accepted outpatient COVID- 19 therapies is another tool to relieve healthcare system overcrowding as it appears COVID-19 may persist longer than initially suspected.

Grants

There were no grants or funding used to produce perform this re- search project or produce the manuscript.

Meetings

Andrew Schoenling presented a poster on a proportion of the cohort prior to final data collection. The poster was presented at the Society of Academic Emergency Medicine New Orleans, May 2022.

Author contributions

Andrew Schoenling conceived of the study, obtained IRB approval, collected the data, performed statistics, and wrote the manuscript. Adam Frisch, Alex Weissman, Clifton Callaway, and Donald Yealy guided the study design, assisted in acquisition of data, provided interpretation of the data, revised the manuscript, and approved the final version.

CRediT authorship contribution statement

Andrew Schoenling: Writing - review & editing, Writing - original draft, Methodology, Funding acquisition, Formal analysis, Data curation, Conceptualization. Adam Frisch: Writing - review & editing, Resources, Data curation. Clifton W. Callaway: Writing - review & editing, Super- vision, Methodology, Formal analysis. Donald M. Yealy: Writing - review & editing, Supervision, Formal analysis. Alexandra Weissman: Writing - review & editing, Supervision, Project administration, Methodology, Formal analysis, Conceptualization.

Conflicts of interest

None of the authors have conflicts of interest to report in relation to this research project.

Acknowledgements

Thanks to Dr. David Huang for mentorship to Dr. Andrew Schoenling during the research track of his Critical Care Fellowship. Thanks to Angela Batko for providing data pulls from the DME database. Thanks to Dr. Oscar Marroquin for system wide COVID-19 data access.

References

1. United States - COVID-19 overview - Johns Hopkins. (accessed 19 Jan 2023). https:// coronavirus.jhu.edu/region/united-states.
2. Wu Z, McGoogan JM. Characteristics of and important lessons from the coronavirus disease 2019 (COVID-19) outbreak in China: summary of a report of 72314 cases from the Chinese Center for Disease Control and Prevention. JAMA - J Am Med Assoc. 2020;323:1239-42. https://doi.org/10.1001/jama.2020.2648.
3. Peltan ID, Caldwell E, Admon AJ, et al. Characteristics and outcomes of US patients hospitalized with COVID-19 download. Am J Crit Care. 2022;31:146-57. https:// doi.org/10.4037/ajcc2022549.
4. Fillmore NR, La J, Zheng C, et al. The COVID-19 hospitalization metric in the pre- and post-vaccination eras as a measure of pandemic severity: a retrospective, nation- wide cohort study. Infect Control Hosp Epidemiol. 2022:1-6. https://doi.org/10. 1017/ice.2022.13.
5. Bajaj L, Turner CG, Bothner J. A randomized trial of home oxygen therapy from the emergency department for acute bronchiolitis. Pediatrics. 2006;117:633-40. https://doi.org/10.1542/peds.2005-1322.
6. von Elm E, Altman DG, Egger M, et al. The strengthening the reporting of observa- tional studies in epidemiology (STROBE) statement: guidelines for reporting obser- vational studies. Int J Surg. 2014;12:1495-9. https://doi.org/10.1016/j.ijsu.2014.07. 013.
7. Harris PA, Taylor R, Minor BL, et al. The REDCap consortium: building an interna- tional community of software platform partners. J Biomed Inform. 2020:1-24. https://doi.org/10.1016/j.jbi.2019.103208.The.
8. CommonWell Health Alliance | Cerner. (accessed 2 Aug 2022). https://www.cerner. com/commonwell-health-alliance.
9. Software | Epic. (accessed 2 Aug 2022). https://www.epic.com/software#Overview.
10. Terp S, Reichert Z, Burner E, et al. Characteristics and outcomes of 360 consecutive COVID-19 patients discharged from the emergency department with supplemental oxygen. Ann Emerg Med. 2022;81:14-9. https://doi.org/10.1016/j.annemergmed. 2022.08.449.
11. Banerjee J, Canamar CP, Voyageur C, et al. Mortality and readmission rates among patients with COVID-19 after discharge from acute care setting with supplemental oxygen. JAMA Netw Open. 2021;4:1-6. https://doi.org/10.1001/jamanetworkopen. 2021.3990.
12. Viel S, Markowicz S, Ait-Medjber L, et al. Dedicated team to ambulatory care for pa- tients with COVID-19 requiring oxygen: low rate of Hospital readmission. Int J Infect Dis. 2022;123:92-6. https://doi.org/10.1016/j.ijid.2022.07.057.
13. van den Berg R, Meccanici C, de Graaf N, et al. Starting home telemonitoring and ox- ygen therapy directly after emergency department assessment appears to be safe in COVID-19 patients. J Clin Med. 2022:11. https://doi.org/10.3390/jcm11237236.
14. Steel PAD, Siegal J, Zhang Y, et al. Telehealth follow up in emergency department pa- tients discharged with COVID-like illness and exertional hypoxia. Am J Emerg Med. 2021;49:426-30. https://doi.org/10.1016/j.ajem.2021.02.052.
15. Roth GA, Emmons-Bell S, Alger HM, et al. Trends in patient characteristics and COVID-19 in-hospital mortality in the United States during the COVID-19 pandemic. JAMA Netw Open. 2021;4:3-11. https://doi.org/10.1001/jamanetworkopen.2021. 8828.
16. Nguyen NT, Chinn J, Nahmias J, et al. Outcomes and mortality among adults hospital- ized with COVID-19 at US medical centers. JAMA Netw Open. 2021;4:20-3. https:// doi.org/10.1001/jamanetworkopen.2021.0417.
17. James JT. A new, evidence-based estimate of patient harms associated with hospital care. J Patient Saf. 2013;9:122-8. https://doi.org/10.1097/PTS.0b013e3182948a69.
18. Adler L, Yi D, Li M, et al. Impact of inpatient harms on Hospital finances and patient clinical outcomes. J Patient Saf. 2018;14:67-73. https://doi.org/10.1097/PTS. 0000000000000171.
19. U.S. Census Bureau QuickFacts: Venango County, Pennsylvania; Lawrence County, Pennsylvania; Mercer County, Pennsylvania; Blair County, Pennsylvania; Allegheny County, Pennsylvania; Pennsylvania. (accessed 2 Aug 2022). https://www.census. gov/quickfacts/fact/table/venangocountypennsylvania, lawrencecountypennsylvania,mercercountypennsylvania,blaircountypennsylvania, alleghenycountypennsylvania,PA/PST045221.