Abstract
Background
Stroke, whether or not accompanied by concurrent COVID‐19 infection, has been associated with varying acute care outcomes, with patients who are COVID‐19 positive typically faring worse. This study aims to explore the functional outcomes of patients with stroke —with and without simultaneous COVID‐19 infection—who survived their acute care stay and progressed to acute inpatient rehabilitation (IRF).
Objectives
To identify differences in sociodemographic factors, medical complexity, and rehabilitation outcomes from an IRF stay between patients with stroke with concurrent COVID‐19 and those without.
Design
A retrospective, observational cohort study using electronic medical records (EMR) data.
Setting
EMR data from 138 IRFs across 34 states of the United States and Puerto Rico involving 40,282 individuals following stroke discharged between April 1, 2020 and May 31, 2021 of whom 1483 (3.7%) were COVID‐19 positive.
Participants
NA.
Interventions
NA.
Main Outcome Measures
Variables collected included sociodemographic and medical complexity along with outcome variable categories included functional complexity, process outcomes, and functional abilities.
Results
Significant differences were found between the two groups using standardized effect sizes >0.2. COVID‐19 positive patients with stroke had more comorbidities (94.1% vs. 51.8%, standardized effect = 1.1), lower admission mobility (26 vs. 30, standardized effect = 0.27), lower discharge mobility scores (56 vs. 65, standardized effect = 0.27), and a longer IRF stay (17 days vs. 14 days, standardized effect = 0.30). They also were less likely to return to the community (65.5% vs. 78.3%) but had a higher acute care transfer rate (19.1% vs. 10.6%). Logistic regression showed that Hispanic COVID‐positive individuals and those with higher mobility scores were more likely to be discharged to the community.
Conclusions
There are meaningful differences in rehabilitation outcomes between COVID–19 positive and negative individuals with stroke that clinicians can use to better understand, anticipate and mitigate outcome challenges facing the COVID‐19 positive population with stroke.
INTRODUCTION
There is a body of evidence demonstrating the association between COVID‐19 and stroke that highlights the difference in clinical presentation as well as outcomes between individuals with stroke who were COVID‐19 positive versus COVID‐19 negative. The incidence of COVID‐19‐related neurological manifestations including stroke is reported to range from 1%–6% among patients hospitalized with COVID‐19. 1 , 2 , 3 A higher incidence of stroke was found in the setting of more severe COVID‐19 infections. 1 , 2 , 3 , 4 , 5 , 6 It is worth noting that these figures may have been skewed in the initial stages of the pandemic secondary to possible reluctance in seeking hospital care due to fear of COVID exposure as well as inconsistent testing indication especially if asymptomatic of respiratory illness. 3 , 6 A COVID‐ 19 positive status does not change the typical distribution of stroke etiology, namely ischemic (87.4%) and hemorrhagic (11.6%), but the ischemic subtypes are notably different. 7 Of the subtypes of ischemic stroke etiology including large artery atherosclerosis, cardio embolism, small vessel occlusion, other, and stroke of undetermined etiology, the latter was more common in COVID‐19 positive patients (51.8%) compared to COVID‐19 negative (22.3%). 8 , 9 This is likely a reflection of the physiological changes associated with COVID‐19 including coagulopathy, inflammation, platelet activation, and alterations to the vascular endothelium. The pathogenesis and optimal primary and secondary prevention strategies for stroke with COVID‐19 remain unclear. 8 Studies have also described atypical large vessel involvement affecting younger patients who lack traditional risk factors or typical manifestations of COVID‐19 at stroke onset as having a disproportionate impact on male and non‐White individuals. 10 , 11 , 12 , 13 For those strokes treated with endovascular thrombectomy, comorbid COVID‐19 infection was associated with younger age, male gender, diabetes, Black race, Hispanic ethnicity, a longer hospital length of stay (LOS), and higher rate of in‐hospital death. 11 These etiological differences may relate to the multiple pathogenic mechanisms of COVID‐19‐related stroke. 14
Multiple studies from acute care data have found that patients who had concomitant COVID and stroke diagnoses had increased morbidity and mortality compared to patients who had COVID alone or had a stroke unrelated to COVID. 3 , 11 , 15 , 16 , 17 , 18 , 19 In these studies, COVID‐19 positive patients with stroke compared to COVID‐19 negative patients with stroke were older (65.3 years), male (62.4%), and more likely to have hypertension, diabetes mellitus, dyslipidemia, coronary artery disease, and severe COVID infection with an acute care hospital mortality rate as high as 31.5%. 7 The overall COVID‐19 acute care in‐hospital mortality rate range during a similar but shorter time frame as these study data were reported at 15.3%–24.5%. 20
The COVID‐19 pandemic posed challenges to modern medical care. Often the postacute care disposition of patients was influenced by a number of factors including availability of home services, acute inpatient rehabilitation facilities (IRFs), long‐term acute care hospitals, and skilled nursing facilities (SNFs). 21 Throughout this evolving landscape of COVID‐19 care, patients continued to benefit from inpatient rehabilitation. Looking at COVID status only, patients who were admitted to an IRF with a diagnosis of COVID‐19, with or without an additional rehabilitation diagnosis, were found to have similar or better functional improvements compared to the non‐COVID admissions. 22 , 23
There are few data to date on the IRF outcomes of individuals with stroke with COVID‐19 compared to those without COVID‐19. The current study's aim is to define the sociodemographic and medical complexity along with outcome variables including process outcomes, such as LOS, fall occurrence, discharge disposition, acute care transfers, mortality, patient satisfaction, and functional outcomes, for example, therapy intensity, mobility, self‐care gain, and efficiency between COVID‐19 positive and COVID‐19 negative patients with stroke.
METHODS
Study design
This was a country‐level retrospective, observational cohort study using a standard electronic medical record (EMR) as the source dataset for IRF data. The study was designated as non‐human subjects research by an institutional review board because study team members did not have access to any identifiable data. Informed consent was not applicable based on the nature of the study.
This study sample consisted of a smaller subset of a larger database already used for prior publication. 24 That larger data set comprised all individuals discharged from 138 IRFs from April 1, 2020, to May 31, 2021. Patients with age <18 years (n = 3) and cases with incomplete functional data (n = 3484) were excluded. Functional abilities outcomes by rule are missing for short or incomplete stays (n = 623) associated with discharge to the acute care setting, discharge against medical advice, an LOS <3 days, and inpatient mortality. These patients were excluded when analyzing functioning abilities outcomes only and were retained for data analysis of other variable categories. Active or resolving COVID‐19 diagnoses were identified by applicable International Classification of Diseases, Tenth Revision codes. 24
Four categories of variables were collected, namely, sociodemographic, medical and functional complexity, process outcomes, and functional abilities outcomes. The subset of data focusing on individuals with stroke was captured by analyzing those who were contained in the Rehabilitation Impairment Category (RIC) assigned to stroke. 24 The study population so identified included 40,282 patients in the stroke subset of this larger group who met the aforementioned inclusion and exclusion criteria.
Data collection
Sociodemographic variables
Sociodemographic variables including age, gender, and race or ethnicity were obtained at admission as coded by data entry into the EMR at registration.
Medical and functional complexity variables
Medical and functional complexity variables at admission including Case Mix Index, 25 etiologic diagnosis specified by RIC using only the Stroke RIC, and associated Case Mix Group (CMG) 26 , 27 were determined and tabulated according to Centers for Medicare and Medicaid Services regulatory requirements. The CMG is determined using a combination of data including age and functional abilities for mobility on admission. Levels on medical and functional complexity for stroke are defined according to the assigned CMG ranging from 0101 to 0106 with increasing age and/or fewer functional abilities for mobility as the number increases. Comorbid conditions documented by physicians during the rehabilitation encounter were tabulated and categorized according to the convention of a report of the National Center for Health Statistics, National Vital Statistics System. 27 These comorbidities are tiered into categories reflecting severity and thus higher resource use with Tier B being the highest followed by decreasing severity to Tiers C, then D and Tier A, which indicated no tier‐changing comorbid conditions. The time to onset, defined as the duration (days) from the acute care hospital admission to the IRF admission date, is also reported. A higher time to onset may indicate patient medical complexity with more days needed in acute care to stabilize for an IRF transition. 28
Process outcome variables
Process outcome variables included metrics routinely collected outside of the EMR as part of clinical care quality. These included LOS, defined as the duration between the IRF admission and discharge date; inpatient mortality; discharge disposition options including acute care transfer, discharge to community (home), discharge to SNF and other being hospice; and fall rate, captured as the number of falls per patient days. The net promoter score was collected by patient questionnaires after discharge to describe patient experience for IRF care. 29 As entered into the EMR by clinicians, therapy intensity was measured as hours of therapy completed on average per day over a 7‐day period with the standard goal being 3 hours per day over 5 days or modified as 15 hours over 7 days.
Functional abilities
Functional abilities outcomes were captured at admission and discharge using Section GG of the Inpatient Rehabilitation Facility‐Patient Assessment Instrument 28 as required by the Centers for Medicare and Medicaid Services. Section GG scores were recorded for each item within the mobility (bed mobility, sit to stand, transfer from toilet, chair, and car, walk, manage stairs) and self‐care (eating, oral hygiene, toilet hygiene, shower/bathing, upper body dressing, lower body dressing, and putting on/taking off footwear) domains. Each mobility and self‐care item has a set definition with an associated scoring system where 6 = independent; 5 = setup; 4 = supervision/touching; 3 = partial assistance; 2 = substantial assistance; 1 = dependent for the task. A total mobility score can range from 15 to 90 with a total self‐care score ranging from 7 to 47. The higher the total score the more functional abilities the patient has for completing these activities of daily living tasks independently. 30 The difference in functional abilities for each domain was calculated as the difference between respective discharge and admission GG scores. At admission, clinicians assessed and documented depression as assessed by Patient Health Questionnaire (PHQ2 and PHQ9) 31 and cognitive status by Brief Interview for Mental Status Scale (BIMS). 32 , 33
Data analysis
Demographic and clinical characteristics were assessed according to the presence or absence of COVID‐19. Continuous variables were presented as mean and SD or median and interquartile range (IQR), and categorical variables were presented as frequencies and percentages. Differences between COVID‐19 status were investigated with unadjusted analyses. For these, between‐group differences were analyzed using t‐tests for continuous variables and chi‐square tests for categorical variables. Because even small differences can be statistically significant with large sample sizes, the decision was made to present measures of effect in addition to traditional p values. For unadjusted analyses, Cohen's d and phi were used for continuous and categorical variables, respectively. Standardized effect sizes >0.2 were used to indicate a meaningful difference. The final goal of this project was to investigate factors associated with a positive outcome (ie, discharge to community) for the subset of COVID‐19 positive individuals. Multiple logistic regression was used to regress discharge status (community vs. anything else) on age, gender, race, CMG, and comorbidity tier. These factors were selected a priori for their assumed relationship with discharge status. Due to similar outcomes, CMG 105 and 106 were combined and used as the reference category. Similarly, comorbidity Tier B (ie, the highest level of comorbidities) was used as the reference category. Odds ratios (OR) and their associated confidence intervals were reported. For this analysis, ORs >1.5 or <0.67 indicated a meaningful association. Because this study relied more on effect sizes than significance tests, no adjustments were made for multiple comparisons. Missing data were minimal. Mobility scores and the BIMS had the largest amounts of missing data (12% and 14%, respectively), and all other variables had <1% missing data. Missing data were handled using listwise deletion. Data were analyzed using SAS v 9.4 (SAS Institute Inc., Cary, NC).
RESULTS
Presented in Table 1 are the sociodemographic, medical, and functional complexity variables for the sample, categorized by COVID‐19 status. The sample consisted of 40,282 individuals with stroke, of whom 38,799 (96.3%) tested negative for COVID‐19 and 1483 (3.7%) tested positive. The median age of the sample was 70 years (IQR, 61–79), with 47.6% of the sample being female and 72.6% being White. There was a notable difference between the two groups in terms of comorbidity tier (standardized effect = 1.1, p < .001). Specifically, COVID‐19 negative individuals were more likely to be in comorbidity Tier A (48.3% vs. 5.9%) indicating fewer to no comorbidities, whereas COVID‐19 positive individuals were more likely to be in comorbidity Tier D (86.6% vs. 46.2%) indicating at least one comorbid condition. Although there was no difference in average age, COVID‐19 positive individuals were more likely to fall under CMG 106 than COVID‐19 negative individuals (48.9% vs. 37.9%; standardized effect = 0.3, p < .001) indicating a lower admission mobility level. Lastly, COVID‐19 positive individuals took longer to be transferred to the IRF setting as measured by the time to onset in days (median [IQR]: 9 [5–18] vs. 6 [4–11]; standardized effect = 0.4, p < .001). The two groups showed no meaningful variations in the 11 most common comorbidities reported including hypertension, hyperlipidemia, left hemiplegia, right hemiplegia, dysphagia, dysarthria, atherosclerotic coronary artery disease, aphasia, gastroesophageal reflux disease, diabetes mellitus type II, and other cognitive symptoms.
TABLE 1.
Overall differences between the COVID‐19 negative and positive groups in sociodemographic characteristics and medical and functional complexity.
| Overall sample | COVID‐19 negative | COVID‐19 positive | p value | Effect size | |||
|---|---|---|---|---|---|---|---|
| n = 40,282 | n = 38,799 | n = 1483 | |||||
| Age (y), median [IQR] | 70 [61–79] | 70 [61–79] | 71 [61–78] | .566 | 0.015 | ||
| Gender | .030 | 0.056 | |||||
| Male | 21,108 (52.4%) | 20,290 (52.3%) | 818 (55.2%) | ||||
| Female | 19,174 (47.6%) | 18,509 (47.4%) | 665 (44.8%) | ||||
| Race or ethnicity | .121 | 0.07 | |||||
| White | 29,240 (72.6%) | 28,209 (72.7%) | 1031 (69.5%) | ||||
| Black | 7227 (17.9%) | 6933 (17.9%) | 294 (19.8%) | ||||
| Hispanic | 2522 (6.3%) | 2412 (6.2%) | 110 (7.4%) | ||||
| Islander | 541 (1.3%) | 523 (1.4%) | 18 (1.2%) | ||||
| Indian | 101 (0.3%) | 97 (0.3%) | 4 (0.3%) | ||||
| Other | 651 (1.6%) | 625 (1.6%) | 26 (1.8%) | ||||
| Comorbidity tier | <.001 | 1.079 | |||||
| A | 18,816 (46.7%) | 18,729 (48.3%) | 87 (5.9%) | ||||
| B | 1581 (3.9%) | 1507 (3.9%) | 74 (5.0%) | ||||
| C | 677 (1.7%) | 639 (1.7%) | 38 (2.5%) | ||||
| D | 19,208 (47.7%) | 17,924 (46.2%) | 1284 (86.6%) | ||||
| Case mix group | <.001 | 0.3 | |||||
| Motor score | Age | CMG | |||||
| ≥72.5 | 101 | 2202 (5.5%) | 2165 (5.6%) | 37 (2.5%) | |||
| ≥63.5 and <72.5 | 102 | 4685 (11.6%) | 4561 (11.8%) | 124 (8.4%) | |||
| ≥50.5 and <63.5 | 103 | 8869 (22.0%) | 8589 (22.1%) | 280 (18.9%) | |||
| ≥41.5 and <50.5 | 104 | 6606 (16.4%) | 6375 (16.4%) | 231 (15.6%) | |||
| <41.5 | ≥84.5 | 105 | 2484 (6.2%) | 2398 (6.2%) | 86 (5.8%) | ||
| <41.5 | <84.5 | 106 | 15,436 (38.3%) | 14,711 (37.9%) | 725 (48.9%) | ||
| Time to onset (days), median [IQR] | 6 [4–11] | 6 [4–11] | 9 [5–18] | <.001 | 0.415 | ||
| Comorbidities | |||||||
| Hypertension | 23,593 (58.6%) | 22,741 (58.6%) | 852 (57.5%) | .380 | 0.004 | ||
| Hyperlipidemia | 22,999 (57.1%) | 22,145 (57.1%) | 854 (57.6%) | .686 | 0.002 | ||
| Hemiplegia (L) | 12,622 (31.3%) | 12,150 (31.3%) | 472 (31.8%) | .676 | 0.002 | ||
| Hemiplegia (R) | 12,025 (29.9%) | 11,630 (30.0%) | 395 (26.6%) | .006 | 0.014 | ||
| Dysphagia | 11,370 (28.2%) | 10,944 (28.2%) | 426 (28.7%) | .666 | 0.002 | ||
| Dysarthria | 10,815 (26.9%) | 10,522 (27.1%) | 293 (19.8%) | <.001 | 0.031 | ||
| Atherosclerotic heart disease | 10,131 (25.2%) | 9786 (25.2%) | 345 (23.3%) | .089 | 0.009 | ||
| Aphasia | 9908 (24.6%) | 9551 (24.6%) | 357 (24.1%) | .632 | 0.002 | ||
| Gastroesophageal reflux | 8714 (21.6%) | 8403 (21.7%) | 311 (21.0%) | .53 | 0.003 | ||
| Diabetes (type 2) | 8136 (20.2%) | 7748 (20.0%) | 388 (26.2%) | <.001 | 0.029 | ||
| Other cognitive symptoms | 7656 (19.0%) | 7447 (19.2%) | 209 (14.1%) | <.001 | 0.025 | ||
Abbreviations: CMG, Case Mix Group; IQR, interquartile range.
Table 2 presents the process and satisfaction outcomes for the sample, categorized by COVID‐19 status. There was a meaningful difference in the IRF LOS between COVID‐19 positive and COVID‐19 negative individuals. The standardized effect was 0.3 (p < .001), indicating that COVID‐19 positive individuals stayed longer than COVID‐19 negative individuals. The median [IQR] LOS for COVID‐19 positive individuals was 17 [12–23] days, whereas for COVID‐19 negative individuals it was 14 [10–20] days. Moreover, there was a meaningful difference in the discharge destination between the two groups (standardized effect = 0.3, p < .001). COVID‐19 negative individuals were more likely to be discharged to a community setting, with a rate of 78.3%. In contrast, only 65.5% of COVID‐19 positive individuals were discharged to a community setting. This percentile difference is due to the higher discharge to acute care and skilled nursing facilities for the COVID‐19 positive individuals. The rate of discharge to acute care for COVID‐19 positive individuals was 19.1%, whereas for COVID‐19 negative individuals, it was 10.6%. The rate of discharge to an SNF for COVID‐19 positive individuals was 14.4% versus 10.1% for COVID‐19 negative individuals. Both groups saw similar rates of falls, with a rate of 12.4%. Finally, there were no meaningful differences in the patient satisfaction measures of net promoter score or their recommendation of the hospital between the two groups.
TABLE 2.
Overall differences between the COVID‐19 negative and positive groups in process outcomes.
| Overall sample | COVID‐19 negative | COVID‐19 positive | p value | Effect Size | |
|---|---|---|---|---|---|
| n = 40,282 | n = 38,799 | n = 1483 | |||
| Length of stay (days), median [IQR] | 14 [10–20] | 14 [10–20] | 17 [12–23] | <.001 | 0.335 |
| Fell | .007 | 0.069 | |||
| No | 35,284 (87.6%) | 34,021 (87.7%) | 1263 (85.3%) | ||
| Yes | 4988 (12.4%) | 4771 (12.3%) | 217 (14.7%) | ||
| Discharge alive | .991 | 0 | |||
| No | 81 (0.2%) | 78 (0.2%) | 3 (0.2%) | ||
| Yes | 40,201 (99.8%) | 38,721 (99.8%) | 1480 (99.8%) | ||
| Discharge | <.001 | 0.297 | |||
| Community | 31,341 (77.8%) | 30,370 (78.3%) | 971 (65.5%) | ||
| Acute | 4397 (10.9%) | 4114 (10.6%) | 283 (19.1%) | ||
| SNF | 4144 (10.3%) | 3930 (10.1%) | 214 (14.4%) | ||
| Other | 400 (1.0%) | 385 (1.0%) | 15 (1.0%) | ||
| Patient satisfaction survey | |||||
| NPS breakdown | .208 | 0.02 | |||
| Blank | 1055 (9.2%) | 1014 (9.1%) | 41 (12.4%) | ||
| Detractor | 1388 (12.1%) | 1349 (12.1%) | 39 (11.8%) | ||
| Passive | 1080 (9.4%) | 1047 (9.4%) | 33 (10.0%) | ||
| Promoter | 7931 (69.2%) | 7714 (69.4%) | 217 (65.8%) | ||
| Recommend hospital, median [IQR] | 10 [9–10] | 10 [9–10] | 10 [9–10] | .654 | 0.027 |
Abbreviations: IQR, interquartile range; NPS, net promoter score; SNF, skilled nursing facility.
The functional abilities, cognitive status, and depression measures of the sample are presented in Table 3, categorized by their COVID‐19 status. COVID‐19 positive individuals had lower mobility scores at both admission and discharge compared to COVID‐19 negative individuals. The median admission and discharge mobility scores for COVID‐19 positive individuals were 26 and 56, respectively, and for COVID‐19 negative individuals, they were 30 and 65. The standardized effects for both admission and discharge were meaningful at 0.27 (p < .001). Although the COVID‐19 positive individuals had a smaller change in mobility scores than the COVID‐19 negative individuals, the difference was not large enough to be considered meaningful, with a standardized effect of 0.15 (p < .001). At admission, the COVID‐19 positive individuals had meaningfully worse self‐care scores than the COVID‐19 negative individuals with a standardized effect of 0.22 (p < .001). However, there was no meaningful difference at discharge, and the change from admission to discharge was also not considered meaningful, with a standardized effect of 0.05 (p = .067). COVID‐19 positive individuals underwent more therapy minutes than COVID‐19 negative individuals, with a median of 2292.5 minutes compared to 2022 minutes, respectively. The standardized effect for therapy minutes was 0.24 (p < .001). There was no meaningful difference on admission in depression as measured by PHQ2 or cognitive status as measured by BIMS between the two groups.
TABLE 3.
Overall differences between the COVID‐19 negative and positive groups in functional abilities, cognitive status, and depression.
| Overall sample | COVID‐19 negative | COVID‐19 positive | p value | Effect size | |
|---|---|---|---|---|---|
| n = 40,282 | n = 38,799 | n = 1483 | |||
| Total mobility, median [IQR] | |||||
| Admission | 29 [21–42] | 30 [21–42] | 26 [19–36] | <.001 | 0.269 |
| Discharge | 65 [42–82] | 65 [43–82] | 56 [34–75] | <.001 | 0.266 |
| Change | 29 [16–40] | 29 [16–40] | 26 [12–39] | <.001 | 0.146 |
| Total self‐care, median [IQR] | |||||
| Admission | 20 [15–25] | 20 [15–26] | 18 [14–24] | <.001 | 0.224 |
| Discharge | 34 [25–42] | 34 [25–42] | 32 [22–41] | <.001 | 0.173 |
| Change | 12 [8–16] | 12 [8–16] | 12 [6–17] | .067 | 0.052 |
| Therapy minutes | 2030 [1440–2839.5] | 2022 [1440–2825] | 2292.5 [1625–3202.5] | <.001 | 0.241 |
| BIMS 500 | 14 [11–15] | 14 [11–15] | 13 [10–15] | .030 | 0.059 |
| PHQ2 | >.999 | 0 | |||
| 0 | 33 339 (83.0%) | 32 112 (83.0%) | 1227 (82.9%) | ||
| 1+ | 6854 (17.0%) | 6601 (17.0%) | 253 (17.1%) | ||
Abbreviations: BIMS, Brief Interview for Mental Status Scale; IQR, interquartile range; PHQ, Patient Health Questionnaire.
Table 4 displays the results of the multiple logistic regression assessing discharge status (discharge to community vs. all other discharge) in the subset of COVID‐19 positive individuals. The overall model was significant (χ2 (13) = 5033, p < .001). Several factors were associated with ORs >1.5, suggesting that they were meaningfully associated with discharge to community. Hispanic individuals had a higher likelihood of being discharged to community compared to White individuals (OR, 1.78 [95% confidence interval (CI), 1.59–1.99]). individuals with higher GG mobility scores in CMG categories 101, 102, and 103 were significantly more likely to be discharged to community compared to the less mobile reference group of CMG 105 and 106 (See Table 4 for ORs). Individuals with no or fewer comorbidities including comorbidity Tier A OR 1.93 (95% CI, 1.72–2.18) and Tier D OR 1.55 (95% CI, 1.38–1.75) were more likely to be discharged to the community compared to those with the most comorbidities in Tier B. These results suggest that mobility defined in CMG classification and comorbidity tiers are significantly related to the likelihood of discharge to community settings, with specific subgroups such as CMG 101–103 and comorbidity Tier A and Tier D individuals being much more likely to experience a positive discharge.
TABLE 4.
Multiple logistic regression predicting discharge to home/community among COVID‐19 positive patients.
| Predictor | Odds ratio (95% CI) | p value |
|---|---|---|
| Age | 0.91 (0.89–0.99) | <.001 |
| Gender | <.001 | |
| Male | 0.85 (0.81–0.90) | |
| Female | Reference | |
| Race or ethnicity | <.001 | |
| Black | 1.18 (1.1–1.26) | |
| Hispanic | 1.78 (1.59–1.99) | |
| Other | 1.24 (1.07–1.44) | |
| White | Reference | |
| CMG | <.001 | |
| 101 | 16.52 (13.01–20.98) | |
| 102 | 8.18 (7.25–9.22) | |
| 103 | 5.35 (4.96–5.78) | |
| 104 | 3.14 (2.92–3.38) | |
| 105 and 106 | Reference | |
| Comorbidity tier | <.001 | |
| A | 1.93 (1.72–2.18) | |
| D | 1.55 (1.38–1.75) | |
| C | 1.33 (1.09–1.63) | |
| B | Reference |
Note: Full model: χ2(13) = 5033, p < .001. Odds ratio for age is interpreted as an increase in 10 years.
Abbreviations: CI, confidence interval; CMG, Case Mix Group.
DISCUSSION
This analysis defines the sociodemographic, medical and outcome variables between COVID‐19 positive and COVID‐19 negative individuals with stroke as measured in an IRF setting. Data were not available to define the stroke etiology, ischemic versus hemorrhagic. There were no meaningful differences in age, gender, or race between the COVID‐19 negative and positive individuals. The median age in the present study of COVID‐19 positive individuals with stroke at 71 years old (IQR, 61–78) was older than the 65.3 years cited in data for acute care COVID‐19 positive patients with stroke, but there was the same male predominance in both. 7 COVID‐19 positive individuals with stroke were nearly twice as likely to have one or more comorbid conditions compared to the COVID‐19 negative individuals with stroke. The types of comorbidities among the groups did not vary but the number of comorbidities was higher in the COVID–19 positive individuals with stroke. This finding is consistent with studies that have shown an increasing number of comorbidities as predictors of the severity and progression of COVID‐19 illness. 34 Known medical conditions associated with increased stroke risk such as hypertension, diabetes, and atherosclerotic heart disease were also found in both the COVID‐19 positive and negative individuals with stroke. 7 , 35
Creech et al. 22 found any COVID‐19 positive patients admitted to the intensive care unit had a longer acute care LOS. A longer interval between acute care and transition to a postacute venue may both reflect patient medical complexity and have a negative impact on mobility. This may also be influenced by nonpatient‐related factors such as local management of COVID with lack of typical access to IRF for these individuals particularly in the early days of the pandemic. Creech et al. also found a direct correlation between longer LOS in acute care, regardless of ICU stay, and a higher incidence of discharge to SNF for all COVID‐19 positive individuals. The present study shows COVID‐19 positive individuals with stroke took a longer time to transition to IRF and had a higher discharge to SNF rate than the COVID‐19 negative individuals with stroke. COVID‐19 positive individuals with stroke also had a higher incidence of acute care transfers from IRF, which would be expected given the combination of stroke, active or resolving COVID‐19 positive status, and higher number of comorbidities.
The COVID–19 positive individuals with stroke were less mobile on admission. COVID‐19 positive individuals with stroke had a meaningfully lower total admission and discharge mobility score indicating a greater need for physical assistance for mobility tasks such as ability to perform transfers and to ambulate. This along with the smaller improvement in mobility score would be expected to raise the burden of care for COVID‐19 positive individuals with stroke and negatively influence discharge to community, consistent with the present data. Admission self‐care scores were lower for COVID‐19 positive individuals with stroke but discharge score and change totals were not meaningfully different. This indicates an equal need for assistance at discharge for activities of daily living such as bathing, dressing, and toileting for both COVID‐19 positive and negative groups with stroke. These functional activity outcomes deficits in mobility and self‐care occurred despite a meaningful effect size difference in total therapy minutes during the IRF stay with more therapy minutes provided to the COVID‐19 positive individual with stroke.
LIMITATIONS
Limitations of this study are like any of those found with the use of data from an EMR for research purposes. Given the size of the data set, significance by p value would be easily found thus using an effect size with significance at 0.2 was selected to minimize this bias. There exists a selection bias that favors the more severely involved COVID‐19 positive individuals with stroke who survived and required intensive rehabilitation services available only in the IRF setting. Potentially not all patients with stroke were tested for COVID; therefore, some categorized in the stroke COVID‐negative group may have indeed been undiagnosed, asymptomatic, or minimally symptomatic COVID positive. Severity of illness is a major factor in determining eligibility and insurance approval for IRF; thus, results from this study should not be generalized to other postacute care settings where the acuity may not match. The results of this data set reflect health care knowledge and management strategies for stroke and COVID during the period of the data collection.
CONCLUSION
This country‐level retrospective cohort study findings reinforced the benefits of IRF care for all individuals post stroke especially in terms of improving self‐care, better mobility, and overall satisfaction, along with increasing the likelihood of patients returning to their homes as supported by clinical guidelines. 36 They also showed a meaningful difference in rehabilitation outcomes between COVID –19 positive and negative individuals with stroke. COVID –19 positive individuals with stroke had a longer LOS, lower mobility but similar self‐care scores on discharge, and were less likely to discharge to home. Clinicians can use this work to better understand the challenges facing the rehabilitation planning for the COVID‐19 positive population with stroke. These data reflect process and functional outcomes for stroke recovery that occur during the phase of inpatient rehabilitation at the IRF level.
DISCLOSURES
The authors have no disclosures.
ACKNOWLEDGMENTS
Poster presentation/abstract at the Association of Academic Physiatrist Annual Meeting February 2024 in Orlando, FL.
Ziems A, McLouth CJ, Elwert N, Charbonneau E, Stillo J, McDowell S. Impact of COVID‐19 positive status on outcome for individuals with stroke treated in acute inpatient rehabilitation. PM&R. 2026;18(1):24‐32. doi: 10.1002/pmrj.13435
REFERENCES
- 1. Mao L, Jin H, Wang M, et al. Neurologic manifestations of hospitalized patients with coronavirus disease 2019 in Wuhan, China. JAMA Neurol. 2020;77(6):683‐690. doi: 10.1001/jamaneurol.2020.1127 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Merkler AE, Parikh NS, Mir S, et al. Risk of ischemic stroke in patients with coronavirus disease 2019 (covid‐19) vs patients with influenza. JAMA Neurol. 2020;77:1‐7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Yaghi S, Ishida K, Torres J, et al. Sars‐cov‐2 and stroke in a New York healthcare system. Stroke. 2020;51:2002‐2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Tsivgoulis G, Katsanos AH, Ornello R, Sacco S. Ischemic stroke epidemiology during the covid‐19 pandemic: navigating uncharted waters with changing tides. Stroke. 2020;51:1924‐1926. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Lodigiani C, Iapichino G, Carenzo L, et al. Venous and arterial thromboembolic complications in covid‐19 patients admitted to an academic hospital in Milan, Italy. Thromb Res. 2020;191:9‐14. doi: 10.1016/j.thromres.2020.04.024 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Klok FA, Kruip M, van der Meer NJM, et al. Incidence of thrombotic complications in critically ill ICU patients with covid‐19. Thromb Res. 2020;191:145‐147. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Nannoni S, de Groot R, Bell S, Markus HS. Stroke in covid‐19: a systematic review and meta‐analysis. Int J Stroke. 2021;16:137‐149. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Stein LK, Mayman NA, Dhamoon MS, Fifi JT. The emerging association between covid‐19 and acute stroke. Trends Neurosci. 2021;44:527‐537. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Adams HP, Bendixen BH, Kappelle LJ, et al. Classification of subtype of acute ischemic stroke – definitions for use in a multicenter clinical‐trial. Stroke. 1993;24:35‐41. [DOI] [PubMed] [Google Scholar]
- 10. Oxley TJ, Mocco J, Majidi S, et al. Large‐vessel stroke as a presenting feature of covid‐19 in the young. N Engl J Med. 2020; 382(20):e60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. de Havenon A, Yaghi S, Mistry EA. Endovascular thrombectomy in acute ischemic stroke patients with covid‐19: prevalence, demographics, and outcomes. J Neurointerv Surg. 2020;12(11):1045‐1048. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Majidi S, Fifi JT, Ladner TR, et al. Emergent large vessel occlusion stroke during new York City's covid‐19 outbreak clinical characteristics and paraclinical findings. Stroke. 2020;51:2656‐2663. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Escalard S, Maïer B, Redjem H, et al. Treatment of acute ischemic stroke due to large vessel occlusion with covid‐19 experience from Paris. Stroke. 2020;51:2540‐2543. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Spence JD, de Freitas GR, Pettigrew LC, et al. Mechanisms of stroke in covid‐19. Cerebrovasc Dis. 2020;49:451‐458. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Katz JM, Libman RB, Wang JJ, et al. Cerebrovascular complications of COVID‐19. Stroke. 2020;51:E227‐E231. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Benussi A, Pilotto A, Premi E, et al. Clinical characteristics and outcomes of inpatients with neurologic disease and COVID‐19 in Brescia, Lombardy, Italy. Neurology. 2020;95(7):E910‐E920. doi: 10.1212/WNL.0000000000009848 [DOI] [PubMed] [Google Scholar]
- 17. Srivastava PK, Zhang SQ, Xian Y, et al. Treatment and outcomes of patients with ischemic stroke during COVID‐19 an analysis from get with the guidelines‐stroke. Stroke. 2021;52:3225‐3232. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Burns SP, Fleming TK, Webb SS, et al. Stroke recovery during the COVID‐19 pandemic: a position paper on recommendations for rehabilitation. Arch Phys Med Rehabil. 2022;103:1874‐1882. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Martí‐Fàbregas J, Guisado‐Alonso D, Delgado‐Mederos R, et al. Impact of COVID‐19 infection on the outcome of patients with ischemic stroke. Stroke. 2021;52:3908‐3917. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Nguyen NT, Chinn J, Nahmias J, et al. Outcomes and mortality among adults hospitalized with COVID‐19 at US medical centers. JAMA Netw Open. 2021;4(3):e210417. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Yonter SJ, Alter K, Bartels MN, et al. What now for rehabilitation specialists? Coronavirus disease 2019 questions and answers. Arch Phys Med Rehabil. 2020;101:2233‐2242. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Creech Z, Truong GT, Vivekanandan R, Destache C, Tierney M, Velagapudi M. Discharge disposition of COVID‐19 patients admitted to the intensive care unit. Crit Care Med. 2022;50:94. [Google Scholar]
- 23. Kelly AA, Lewis CA, Escalon MX. Inpatient rehabilitation issues related to COVID‐19. Phys Med Rehabil Clin N Am. 2023;34:513‐522. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Charbonneau EJ, Grover P, Johns JS, McDowell SM, Stillo J. Baseline characteristics and outcomes for people with and without COVID‐19 diagnoses receiving inpatient rehabilitation care across the US in 2020‐2021. Arch Rehabil Res Clin Transl. 2023;5(3):100281. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Medicare hospital prospective payment system how DRG rates are calculated and updated. 2001.
- 26. Services USCfMM . IRF grouper – case mix group (CMG). 2021. https://www.cms.gov/medicare/payment/prospective‐payment‐systems/inpatient‐rehabilitation/grouper‐case‐mix‐group
- 27. Conditions contributing to deaths involving COVID‐19, by age group, United States. Week ending 2/1/2020 to 12/5/2020. National Vital Statistics System. 2020.
- 28. Services USCfMM. Inpatient rehabilitation facility patient assessment instrument (IRF‐PAI) and IRF‐PAI manual. https://www.cms.gov/medicare/quality/inpatient‐rehabilitation‐facility/irf‐pai‐and‐irf‐qrp‐manual
- 29. Proper use and attribution for the net promoter system. https://www.bain.com/consulting‐services/customer‐strategy‐and‐marketing/net‐promoter‐score‐system/
- 30. U.S. Centers for Medicare & Medicaid Services . Impact Act of 2014 data standardization & cross setting measures. https://www.cms.gov/medicare/quality/initiatives/pac‐quality‐initiatives/impact‐act‐2014‐data‐standardization‐cross‐setting‐measures/impact‐act‐standardized‐patient‐assessment‐data‐elements
- 31. Kroenke K, Spitzer RL, Williams JBW, Löwe B. The Patient Health Questionnaire somatic, anxiety, and depressive symptom scales: a systematic review. Gen Hosp Psychiatry. 2010;32:345‐359. [DOI] [PubMed] [Google Scholar]
- 32. Mansbach WE, Mace RA, Clark KM. Differentiating levels of cognitive functioning: a comparison of the Brief Interview for Mental Status (BIMS) and the Brief Cognitive Assessment Tool (BCAT) in a nursing home sample. Aging Ment Health. 2014;18:921‐928. [DOI] [PubMed] [Google Scholar]
- 33. Saliba D, Buchanan J, Edelen MO, et al. MDS 3.0: brief interview for mental status. J Am Med Dir Assoc. 2012;13:611‐617. [DOI] [PubMed] [Google Scholar]
- 34. Chatterjee S, Nalla LV, Sharma M, et al. Association of COVID‐19 with comorbidities: an update. ACS Pharmacol Transl Sci. 2023;6(3):334‐354. doi: 10.1021/acsptsci.2c00181 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Boehme AK, Esenwa C, Elkind MSV. Stroke risk factors, genetics, and prevention. Circ Res. 2017;120:472‐495. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Powers WJ, Rabinstein AA, Ackerson T, Adeoye OM, Bambakidis NC, et al. Guidelines for the early management of patients with acute ischemic stroke: 2019 update to the 2018 guidelines for the early management of acute ischemic stroke: a guideline for healthcare professionals from the American Heart Association/American Stroke Association. Stroke. 2019;50(12):e440‐e441. [DOI] [PubMed] [Google Scholar]