key: cord-0827047-r5r5e50v
authors: Clark, Jeffrey R.; Batra, Ayush; Shlobin, Nathan A.; Hoffman, Steven C.; Orban, Zachary S.; Koralnik, Igor J.; Liotta, Eric M.
title: Acute-care hospital reencounters in COVID-19 patients
date: 2021-05-21
journal: GeroScience
DOI: 10.1007/s11357-021-00378-2
sha: 7c06fd53bc37a9166af17326a4c3af0bb57b1104
doc_id: 827047
cord_uid: r5r5e50v

Acute-care hospital reencounters (ACHEs)—encompassing emergency department visits, observation stays, and hospital readmissions—following COVID-19 hospitalization may exacerbate health care system strain and impair recovery from illness. We sought to characterize these reencounters and factors associated with reencounters. We identified the first consecutive 509 patients hospitalized for COVID-19 within an IL hospital network, and examined ACHEs, experienced within 30 days and 4 months of index hospitalization. We identified independent predictors of reencounter using binary logistic regression. Of 509 patients, 466 (91.6%) were discharged alive from index COVID-19 hospitalization. Within 30 days and 4 months, 12.4% and 21.5% of patients, respectively, experienced ACHEs. The median time to first ACHE was 24.2 (IQR 6.5, 55) days. COVID-19 symptom exacerbation was the leading reason for early ACHE (44.8%). Reencounters, both within 30 days and 4 months, were associated with a history of a neurological disorder before COVID-19 (OR 2.78 [95% CI 1.53, 5.03] and OR 2.75 [95% CI 1.67, 4.53], respectively). Older patients and those with diabetes mellitus, chronic obstructive pulmonary disease, or organ transplantation tended towards more frequent ACHEs. Steroid treatment during COVID-19 hospitalization demonstrated reduced odds of 30-day reencounter (OR 0.31 [95% CI 0.091, 0.79]). Forty-nine patients had repeat SARS-CoV-2 nasopharyngeal testing during a reencounter; twelve (24.5%) patients had positive reencounter tests and experienced more frequent reencounters than those testing negative. COVID-19 symptom exacerbation is a leading cause of early ACHE after COVID-19 hospitalization, and steroid use during index hospitalization may reduce early reencounters. Neurologic illness before COVID-19 predicts ACHEs.

strained the capacity and resources of the health care system [2] [3] [4] .

Acute-care hospital reencounters (ACHEs)-encompassing hospital readmissions, visits to the emergency department (ED), and observation unit stays-can compound the strain on health care system resources due to an initial hospitalization and may impair recovery from illness and optimization of patient outcome, especially among elderly patients [5] [6] [7] [8] [9] . ACHEs may also provide insight into the ongoing health challenges of patients following COVID-19 by identifying symptoms impairing health to the point of requiring acute medical attention. Multiple studies have suggested that the sequelae of COVID-19 may persist beyond the initial bout of respiratory illness [10, 11] . Persistent symptoms including fatigue, dyspnea, headache, and cognitive slowing have been reported by "long hauler" patients for weeks or even months after their acute COVID-19 illness, irrespective of whether they were initially hospitalized for COVID-19 [10] [11] [12] [13] . While there have been reports of persistent post-COVID-19 symptoms leading to presentation to EDs or hospital admission [12] , there are limited data regarding the frequency of and reasons for ACHEs in the months following hospitalization for COVID- 19. This study characterizes the rate of short-and longterm ACHEs in a consecutive cohort of patients both within 30 days and within 4 months following COVID-19 hospitalization. Thirty days was selected for shortterm ACHE as that time period represents a standard time metric for tracking short-term reencounters after hospitalization. Four months after COVID-19 hospitalization was selected for long-term ACHE as it represented the longest follow-up available for all patients in our cohort at the time of data collection. We also sought to elucidate the reasons and risk factors for ACHEs. This may assist in efforts to minimize avoidable reencounters as well as provide insight into the ongoing health care challenges experienced by COVID-19 patients following their initial hospitalization.

To identify patients with an initial hospitalization for COVID-19, we used a computerized query of the electronic medical record to determine the first consecutive patients admitted with a confirmed primary diagnosis of COVID-19 to the Northwestern Medicine Healthcare (NMHC) system between 5 March and 6 April 2020, which resulted in a convenience sample of 509 consecutive patients. NMHC consists of one academic medical center and nine other hospitals in the Chicago, IL, area. COVID-19 diagnosis was confirmed by SARS-CoV-2 reverse transcription polymerase chain reaction (RT-PCR) assay of nasopharyngeal swab or bronchoalveolar lavage fluid. All laboratory and radiologic assessments were performed as part of routine clinical care. The study was approved by our institutional review board (STU00212627) with waiver of consent for retrospective analysis.

ACHEs included ED visits, observation stays, or hospital readmissions. We excluded all scheduled reencounters and reencounters pertaining to labor and delivery or accidental trauma. For the analysis of ACHEs, we excluded patients who died during the initial COVID-19 hospitalization. ACHEs following the index hospitalization were identified by query of the NMHC system electronic medical record and by query for reencounters at any of the 20 participating IL hospital systems in the Epic (Epic Systems Corporation, Verona, WI, USA) Care Everywhere Network. We included ED visits and observation stays in addition to hospital readmissions given the literature, suggesting that efforts to reduce hospital readmission in the USA may have shifted a greater portion of acute care to treat-anddischarge visits in the ED and observations stays [5] .

Demographic, medical comorbidity, prehospitalization medication usage, and hospital course data were collected by electronic medical record review, as previously described [14] . Medications received during initial hospitalization, including steroids, were collected by automated electronic query. Given prominent reports of headaches and "brain fog" as persistent post-COVID-19 symptoms and our prior report of frequent neurologic manifestations among hospitalized COVID-19 patients [12, 14] , we included data on the incidence of neurologic manifestations during patients' COVID-19 disease course in our analysis of potential factors contrib-uting to reencounters. Neurologic manifestations were identified by neurologist adjudication, which excluded neurologic manifestations that resulted from medication effects, as previously described [14] . The primary reason for each reencounter was identified by review of physician-documented clinical notes and reported as determined by the attending physician of record for the reencounter. We also identified whether patients with a reencounter had a repeat SARS-CoV-2 RT-PCR assay performed at any point during the reencounter. Patients were categorized as having a positive reencounter RT-PCR if they had a positive result during any reencounter. In the case of multiple reencounter RT-PCR assays, the date of reencounter assay was recorded as the last available positive reencounter RT-PCR or the first available negative reencounter RT-PCR if a patient never had a positive reencounter RT-PCR result. Our RT-PCR assay identifies the presence of SARS-CoV-2 RNA but does not identify whether that RNA is replicative.

Data were summarized as the number of patients/ events (frequency), mean (standard deviation) for normally distributed variables, and median (interquartile range (IQR)) for non-normally distributed variables. Associations were assessed using Fisher's exact test, Spearman's rank correlation test, and Wilcoxon rank-sum test. Binary logistic regression models were developed to identify adjusted factors associated with (1) ACHE within 30 days of discharge from index hospitalization and (2) ACHE within 4 months of discharge from index hospitalization. In each case, we first developed a model using a priori variables of age and sex, which are factors widely appreciated as related to COVID-19 disease severity. Given relatively limited data in the literature regarding additional factors associated with ACHE after COVID-19 hospitalization, we then included additional variables in the model that were univariately associated with ACHE at p ≤ 0.15 and not highly collinear with a variable already included in the model. To avoid overfitting models to the data, we used a backward variable selection algorithm based on Akaike information criterion optimization to develop final parsimoniously adjusted models of ACHE within 30 days and within 4 months. Two-sided p ≤ 0.05 was considered sig-nificant, and all analyses were performed in R version 3.5.0 (R Foundation for Statistical Computing, Vienna, Austria).

Of 509 patients hospitalized with COVID- 19 patients for D-dimer. In each instance, laboratory values were significantly more likely to be available in patients with severe COVID-19 that required mechanical ventilation (in each case, p < 0.02). Since these laboratory measures were not missing at random, they were not considered for inclusion in models of reencounter no association between steroids and reencounters within 4 months. In addition, there was a non-significant tendency towards more frequent encephalopathy during the COVID-19 hospitalization in those who had reencounters within 4 months (35% vs. 26%, p = 0.097).

Of the 100 patients who experienced an ACHE, 49 had at least one SARS-CoV-2 RT-PCR assay performed during a reencounter. Of these, 12 (24.5%) patients had a positive SARS-CoV-2 RT-PCR during a reencounter, including four patients who had a previously documented negative repeat RT-PCR. The time from index hospital admission to the reencounter RT-PCR did not differ significantly between those with positive versus negative assay results (1), infection other than COVID-19 or pneumonia (1), acute renal dysfunction (1), heart failure (1), and stridor and dyspnea due to tracheal stenosis (1).

A total of 67 ACHEs among 58 patients occurred within 30 days of discharge, and an additional 85 ACHEs among 65 patients occurred within 31 days to 4 months of discharge. The median time to first ACHE from index COVID-19 hospitalization discharge was 24.2 (6.5, 55) days. Among reencounters within 30 days, 28 (41.8%) were hospital readmissions, 12 (17.9%) were observations stays, and 27 (40.3%) were ED visits. Among reencounters between 31 days and 4 months, 41 (48.2%) were hospital readmissions, 7 (8.2%) were observations stays, and 37 (43.5%) were ED visits. Two deaths occurred in readmitted patients, both of whom returned within 30 days after index hospitalization with COVID-19 pulmonary symptom exacerbation requiring respiratory support. 

In adjusted models (Table 3) , a history of a neurologic disorder prior to COVID-19 was independently associated with increased odds of a 30-day ACHE (OR 2.78; 95% CI 1.53, 5.03; p < 0.001) while receiving steroids during the index hospitalization was associated with reduced odds of a 30-day ACHE (OR 0.31; 95% CI 0.09, 0.79; p = 0.029). In adjusted models of ACHE within 4 months (Table 4 ), a history of a neurologic disorder prior to COVID-19 was independently associated with increased odds of an ACHE (OR 2.75; 95% CI 1.67, 4.53; p < 0.001) while patients who had a longer time from COVID- 19 symptom onset to index hospitalization demonstrated decreased odds of an ACHE (OR 0.957 per day; 95% CI 0.916, 0.999; p = 0.048). Older patients and those with diabetes mellitus, chronic obstructive pulmonary disease, or organ transplantation had a non-significant tendency towards greater odds of a reencounters within 4 months. Of note, dichotomizing age at 65 years old, rather than treating age as a continuous variable, had no meaningful impact on the effect estimates of either the 30-day or 4-month adjusted models.

This study demonstrates that within 30 days and within 4 months following COVID-19 hospitalization, 12.4% and 21.5% of patients, respectively, experienced ACHEs consisting of ED visits, observation stays, and readmission. Reencounter, within both 30 days and 4 months, was associated with a past medical history of a neurological disorder before COVID-19. Meanwhile, receiving steroid treatment during initial COVID-19 hospitalization reduced the odds of 30-day reencounter, and a longer duration between COVID-19 symptom onset and initial hospitalization was associated with lower likelihood of 4-month reencounter. Hospital readmissions, in general, have been associated with a number of factors such as age, length of stay, and follow-up care [15] [16] [17] [18] . The existing literature on COVID-19 reencounters describes 14-day reencounter rates of 3.6%, most commonly for respiratory complaints and associated with variables including history of hypertension or COPD and shorter initial length of stay [19] ; additionally, 30-day hospital readmission rates of 6.8% were reported in association with medical comorbidities including hypertension, diabetes, chronic pulmonary disease, liver disease, and cancer [20] . Atalla et al. [20] also showed that readmitted patients were less likely to have required ICU care and intubation during initial COVID-19 hospitalization. In contrast, we found no association between need for mechanical ventilation or initial hospital length of stay, as measures of disease severity, and ACHE. In agreement with these other studies, we found that respiratory complaints, including exacerbation of COVID-19 symptoms, were the leading reason for early ACHEs. The differences in findings between our study and the others in the literature might be related to different durations of reencounter followup, inclusion of ED visits and observations stays in addition to hospital readmissions, and identification of ACHEs at multiple hospital systems in our study (using the EPIC Care Everywhere Network). In addition, regional differences, such as COVID-19 case count and reports of hospital resource availability, may have influenced patients' likelihood to represent to the hospital. Similarly, local health care resources and hospital bed availability may have influenced whether health care providers attempted to address ACHEs by ED visits or observation stays rather than hospitalization. In fact, literature prior to the COVID-19 pandemic suggests there may already have been a growing tendency in the USA to reduce hospital readmissions by shifting a greater portion of acute care to treat-and-discharge visits in the ED and observation stays [5] . Our study was conducted in the Chicago, IL, metropolitan area whereas the two COVID-19 reencounter studies cited above were conducted on the East Coast of the United States, all including COVID-19 hospitalization during March and April of 2020. Unlike systems in other regions of the USA, our health care system was not overwhelmed in March and April of 2020, an observation in part reflected in our markedly lower hospital mortality compared to contemporaneous mortality rates at major health care systems based in New York City (8.4% vs. 21.4% hospital mortality) [21] . In fact, it is possible that our study findings are most generalizable to hospital system functioning with residual capacity. It is also noteworthy, and perhaps surprising, that we did not find a significant difference in ACHEs between racial and ethnic groups. While granular socioeconomic data was not available for our cohort, future studies might consider including socioeconomic factors, in addition to race and ethnicity, when investigating disparities in COVID-19 outcomes.

Furthermore, due to growing awareness of and concern for prolonged symptoms and sequelae of COVID-19 [10] [11] [12] , we expanded our investigation to 4 months in order to offer comparison of shorterversus longer-term patterns in ACHEs. In particular, we noted that COVID-19 respiratory complaints were a common reason for reencounters before 30 days, but this was an infrequent reason after 30 days and did not require hospitalization or observation stay in our cohort after 30 days. Our data suggest that the 30-day window is a higher-risk period for reencounter related to COVID-19 symptom recurrence or exacerbation, as well as the time frame during which these concerns are most likely to reach a severity meriting observation or readmission. In addition, the finding that 30-day ACHEs were largely driven by COVID-19 symptom exacerbation, while 4-month ACHEs were not, offers a plausible explanation for the association we identified between steroid treatment during initial COVID-19 hospitalization and a lower likelihood of subsequent 30-day ACHE but a lack of association with 4-month ACHE. Our cohort size was not large enough to consider whether steroid effects might differ across strata of disease severity, but this could be a consideration for future studies as large multicenter cohorts become more readily available.

In addition, we found that nearly a quarter of patients who had a repeat SARS-CoV-2 RT-PCR assay during a reencounter were positive and that positive RT-PCR assays could occur even longer than 3 months after hospitalization for COVID-19. This frequency of prolonged RT-PCR positivity is consistent with a recently reported rate of 18.2% but demonstrates a potential for SARS-CoV-2 assays to remain positive longer than the follow-up dura- tion available in that report [22] . We also found that patients with a positive reencounter RT-PCR had more reencounters within 4 months than those whose reencounter RT-PCR assays were negative. While this may be related to a greater availability for repeated testing in those with more frequent reencounters, this observation would be consistent with an increased vulnerability to symptoms requiring medical attention in those with ongoing expression of viral material. Others have proposed that such a COVID-19 post-viral syndrome might exist and represents both a long-term public health challenge and a potential contributor to poor quality of life recovery for individual COVID-19 patients [23] . While our data are consistent with the existence of postviral syndrome, such a hypothesis requires specific testing in a larger cohort. If a COVID-19 post-viral syndrome exists, the development of specialized outpatient centers for the management of post-acute COVID-19 might be a means to improve patient recovery while decompressing ACHEs by shifting care to the outpatient setting. It may seem surprising that a prior history of neurologic disorders before COVID-19 was associated with increased odds of ACHE since respiratory complaints appeared to be a driving force for early reencounters. However, the effects of COVID-19 on the nervous system are being increasingly reported [14, [24] [25] [26] [27] [28] [29] . We previously demonstrated that encephalopathy, as a neurologic complication of COVID-19 hospitalization, was associated with worse patient functional outcome, and a history of premorbid neurologic disorder was associated with both increased odds of encephalopathy during COVID-19 hospitalization and worse functional outcome [14] . Therefore, premorbid neurologic disease may represent a unique vulnerability in the recovery from COVID-19, which should be investigated further. While we only observed a tendency towards increased 4-month reencounter in those who experienced encephalopathy during their COVID-19 hospitalization, it is possible this was due to inadequate cohort size to detect a statistically significant relationship. Furthermore, the size of our cohort was insufficient to examine an association between ACHE and each subtype of premorbid neurologic disease; as the size of COVID-19 cohorts increases, future studies might examine the associations between specific premorbid neurologic diagnoses and various outcome endpoints following acute COVID-19.

Our study has limitations that should be acknowledged. First, it is possible that some patients experienced ACHEs that were not identified by our study. However, we mitigated this limitation by searching for ACHEs at the 20 IL hospital systems participating in the EPIC Care Everywhere Network. Furthermore, missed ACHEs outside the EPIC network would represent a non-systematic source of bias that would not be expected to alter the distribution of reasons for ACHEs or factors associated with ACHE occurrence, but could lead to an overall underestimation of ACHE frequency by our study. Secondly, ACHEs may underrepresent patients who experienced severe COVID-19 illness due to a propensity for these patients to enter inpatient rehabilitation or nursing facilities following discharge where some medical concerns may have been addressed without return to the hospital. It is also possible that severely ill patients might die from a complication prior to reaching a hospital and would not therefore register as an ACHE. As such, while reencounters illustrate common ongoing, reemergent, or new concerns of COVID-19 patients, they may not be the most informative metric to gauge the medical needs of severely ill COVID-19 patients following discharge. Future research might examine the medical needs experienced by severely affected COVID-19 patients who require inpatient rehabilitation or nursing facility stays. Additionally, fear of reexposure to the virus or increased access to telehealth resources may have prevented some reencounters for issues that typically would have resulted in a reencounter prior to the COVID-19 pandemic. In fact, the novel features of COVID-19 and the considerable impact of the pandemic on health care operations make meaningful comparison to historical reencounter data challenging. Due to the retrospective nature of our study, it is possible that not all risk factors for reencounter were examined. For example, followup care after discharge has been identified as influential for rates of readmission [17, 18] , a variable which was not investigated in this study. However, given the novelty of COVID-19 and the unclear natural history after acute illness, it may be more challenging to implement effective post-hospital follow-up recommendations than for more well-known diseases. It is also possible that our study was underpowered to detect associations between uncommon comorbidities, such as organ transplantation, and ACHE.

In summary, COVID-19 symptom exacerbations along with other respiratory complaints are leading causes of early ACHE after index COVID-19 hospitalization, and steroid use during index hospitalization may reduce the odds of early ACHEs. A history of neurologic illness before COVID-19 is associated with increased likelihood of ACHE, in addition to worse functional outcome from COVID-19. A potential unique role for neurologic factors in acute illness and recovery from COVID-19 requires further investigation.

WHO Coronavirus Disease (COVID-19) Dashboard

COVIDView: a weekly surveillance summary of U.S. COVID-19 activity

Fair allocation of scarce medical resources in the time of Covid-19

Covid-19 -implications for the health care system

Hospital revisits within 30 days after discharge for medical conditions targeted by the Hospital Readmissions Reduction Program in the United States: national retrospective analysis

Association of 30-day all-cause readmission with long-term outcomes in hospitalized older Medicare beneficiaries with heart failure

Early hospital readmission is a predictor of one-year mortality in community-dwelling older Medicare beneficiaries

Fewer emergency readmissions and better quality of life for older adults at risk of hospital readmission: a randomized controlled trial to determine the effectiveness of a 24-week exercise and telephone follow-up program

Post-hospital syndrome-an acquired, transient condition of generalized risk

Persistent symptoms in patients after acute COVID-19

Persistent symptoms 3 months after a SARS-CoV-2 infection: the post-COVID-19 syndrome?

As Their Numbers Grow, COVID-19 "Long Haulers" Stump Experts

Persistent neurologic symptoms and cognitive dysfunction in non-hospitalized Covid-19 "long haulers

Epub ahead of print

Frequent neurologic manifestations and encephalopathy-associated morbidity in Covid-19 patients

Readmission risk factors after hospital discharge among the elderly

Hospital length of stay and readmission: an early investigation

Why do patients keep coming back? Results of a readmitted patient survey

Relationship between early physician follow-up and 30-day readmission among Medicare beneficiaries hospitalized for heart failure

Characterization of patients who return to hospital following discharge from hospitalization for COVID-19

Readmissions among patients with COVID-19

Reader Response: A Prospective Study of Neurologic Disorders in Hospitalized Patients With COVID-19 in New York City

A tale of three cities: the state of racial justice in Chicago

Assessment of SARS-CoV-2 RNA Test Results Among Patients Who Recovered From COVID-19 With Prior Negative Results

url: Doctors examining possible link between chronic fatigue syndrome and COVID-19 (msn.com). Accessed

Neurologic manifestations of hospitalized patients with coronavirus disease 2019 in Wuhan

Neurologic manifestations in hospitalized patients with COVID-19: The ALBACOVID registry. Neurology

Neurological associations of COVID-19

Neurological and neuropsychiatric complications of COVID-19 in 153 patients: a UK-wide surveillance study. Lancet Psychiatry

COVID-19: a global threat to the nervous system

Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations

Acknowledgements Our author group participates in the Global Consortium Study of Neurologic dysfunction in COVID-19 (GCS-NeuroCOVID).Consent for publication All authors have been included, and all authors have provided their approval for submission of the manuscript for consideration.

Author contribution Jeffrey R. Clark collected the patient data; assisted in the data management, data analysis, and study interpretation; and drafted the manuscript.Ayush Batra assisted in the study design, collected the patient data, assisted in the study interpretation, revised the manuscript for critical intellectual content, and assisted in the study supervision.Nathan A. Shlobin and Steven C. Hoffman collected the patient data, assisted in the study interpretation, and revised the manuscript for critical intellectual content.Zachary S. Orban collected the patient data, assisted in the data management and data analysis, and revised the manuscript for critical intellectual content.Igor J. Koralnik assisted in the study design, collected the patient data, assisted in the study interpretation, revised the manuscript for critical intellectual content, and assisted in the study supervision.Eric M. Liotta conceived the study, led the study design, collected the patient data, performed the primary data analysis and interpretation, drafted the manuscript, supervised the study conduct, and took responsibility for the manuscript as a whole.Funding Dr. Liotta is supported by National Institutes of Health grant L30 NS098427.Data availability De-identified data and statistical code can be made available to appropriately qualified individuals for purposes of replication and at reasonable request.

Ethics approval The study was approved by our institutional review board (STU00212627) with waiver of consent for retrospective analysis.

The authors declare no competing interests.