Rationale, Design, and Characteristics of the VALIANT (COVID‐19 in Older Adults: A Longitudinal Assessment) Cohort

Abstract

Background

Most older adults hospitalized with COVID‐19 survive their acute illness. The impact of COVID‐19 hospitalization on patient‐centered outcomes, including physical function, cognition, and symptoms, is not well understood. To address this knowledge gap, we collected longitudinal data about these issues from a cohort of older survivors of COVID‐19 hospitalization.

Methods

We undertook a prospective study of community‐living persons age ≥ 60 years who were hospitalized with COVID‐19 from June 2020–June 2021. A baseline interview was conducted during or up to 2 weeks after hospitalization. Follow‐up interviews occurred at one, three, and six months post‐discharge. Participants completed comprehensive assessments of physical and cognitive function, symptoms, and psychosocial factors. An abbreviated assessment could be performed with a proxy. Additional information was collected from the electronic health record.

Results

Among 341 participants, the mean age was 71.4 (SD 8.4) years, 51% were women, and 37% were of Black race or Hispanic ethnicity. Median length of hospitalization was 8 (IQR 6–12) days. All but 4% of participants required supplemental oxygen, and 20% required care in an intensive care unit or stepdown unit. At enrollment, nearly half (47%) reported at least one preexisting disability in physical function, 45% demonstrated cognitive impairment, and 67% were pre‐frail or frail. Participants reported a mean of 9 of 14 (SD 3) COVID‐19‐related symptoms. At the six‐month follow‐up interview, more than a third of participants experienced a decline from their pre‐hospitalization function, nearly 20% had cognitive impairment, and burdensome symptoms remained highly prevalent.

Conclusions

We enrolled a diverse cohort of older adults hospitalized with COVID‐19 and followed them after discharge. Functional decline was common, and there were high rates of persistent cognitive impairment and symptoms. Future analyses of these data will advance our understanding of patient‐centered outcomes among older COVID‐19 survivors.

Key points

• Most older adults survive hospitalization with COVID‐19, but little is known about the health outcomes following hospitalization that matter most to older persons.

• We enrolled a racially and ethnically diverse cohort of community‐living adults ≥60 years who were hospitalized with COVID‐19 and completed a detailed assessment that included information about physical function prior to their illness as well as contemporaneous measures of cognitive function and symptoms. We completed follow‐up assessments at one, three, and six months after discharge.

• Participants had high levels of preexisting functional disability and high levels of cognitive impairment and symptoms during the index COVID‐19 admission. At the six‐month follow‐up interview, more than a third of participants reported greater functional disability than before their hospitalization, nearly 20% had cognitive impairment, and persistent symptoms, such as fatigue (52.8%) and dyspnea (37.7%), were common. Future analyses will provide key information about how these issues evolve over time among older survivors of COVID‐19 hospitalization.

Why does this paper matter?

By shedding light on patient‐centered health outcomes and outcome trajectories among older adults who have survived COVID‐19 hospitalization, these unique data will contribute to the evidence base that informs medical decisions for this population and will inform interventions to improve the health of older COVID‐19 survivors.

INTRODUCTION

Older adults are much more likely than younger adults to experience serious illness with COVID‐19. A study in New York City during the first wave of infection found that the odds of hospitalization for persons ≥75 years were almost 40 times higher than for those under age 65. ¹ The most recent data from the Centers for Disease Control and Prevention, reported on November 9, 2022, show that 75% of deaths—more than 800,000—have occurred among Americans age 65 and older. ²

Nonetheless, most older adults survive serious illness (i.e., illness requiring hospitalization) with COVID‐19. ³ There is reason to be concerned, however, that older survivors are at risk for a substantial decline in their health and functional status and may develop impairments that last even after the pandemic wanes. Normal aging confers decreased physiologic reserve and diminished resilience in the face of acute stressors, ⁴ and older adults are much more likely than younger adults to have multiple chronic medical conditions as well as other underlying vulnerabilities, such as frailty. ⁵ Hospitalization has long‐term physical and cognitive consequences for older persons, ⁶ , ⁷ and hospitalization with COVID‐19, specifically, involves long periods of isolation during which patients are unable to leave their hospital rooms and are at increased risk for immobility. ⁸ Delirium is also common. ⁹

Few data exist from those who have been hospitalized with COVID‐19 about the health outcomes that matter most to older patients, including the preservation of cognition, maintaining physical function, and freedom from burdensome symptoms. ¹⁰ This is in part because information about these issues is not collected in routine clinical care. To address this knowledge gap, we designed VALIANT (COVID‐19 in Older Adults: A Longitudinal Assessment), a prospective longitudinal study of adults ≥60 years hospitalized with COVID‐19 at five hospitals in Connecticut. We interviewed patients during or shortly after their hospitalization and then at one, three, and six months after hospital discharge. We collected granular information across multiple health domains and then linked these data to information from the electronic health record. The resulting dataset permits an unusually rich examination of outcomes and outcome trajectories among older adults hospitalized with COVID‐19. Here, we present the characteristics of the VALIANT cohort and provide initial data about how function, cognition, and symptoms evolved over time for study participants.

METHODS

Study overview

VALIANT is a prospective longitudinal study of community‐living persons age ≥60 years who were admitted with COVID‐19 at five hospitals within the Yale‐New Haven Health System (YNHHS) from June 2020–June 2021. During a baseline interview, conducted during hospital admission or within two weeks of discharge, we collected information about participants' physical function before hospitalization as well as contemporaneous data about cognition and symptoms. Data about function, cognition, and symptoms were subsequently collected at follow‐up interviews at one, three, and six months after discharge. We also obtained information about other areas of direct relevance to these patient‐centered outcomes, including medical comorbidities, events during and after hospitalization, and psychosocial and sociodemographic factors (Figure 1). The study was approved by the Institutional Review Board at Yale University. Methods reporting is consistent with the STROBE guidelines. ¹¹

FIGURE 1.

Overview of data collected in VALIANT. Information about three key patient‐centered outcomes (physical function, cognitive function, and symptoms) was supplemented with data from secondary domains of direct relevance to these outcomes.

Patient screening and eligibility

Research coordinators used the electronic health record to identify potential participants. Patients were eligible if they were at least 60 years old, hospitalized at a YNHHS hospital, and had a PCR‐confirmed diagnosis of SARS‐CoV‐2 infection either during or directly prior to hospital admission. We selected an age cutoff of ≥60 years because the Centers for Disease Control and Prevention had identified this group as being at the highest risk for severe disease when the study was being designed. ¹² Patients were excluded if they had previously been hospitalized with COVID‐19 and were subsequently readmitted, regardless of whether their SARS‐CoV‐2 test remained positive. We made an exception for patients who were readmitted within seven days of hospital discharge from their initial (index) COVID‐19 admission, because, in such cases, we considered the two hospitalizations to constitute a single episode of care. Patients were also excluded if they were long‐term residents in a skilled nursing facility, had advanced dementia, did not speak English or Spanish, or had opted out of research. Additionally, because we wished to collect longitudinal data about persons who were not at the end of life, patients were excluded if they had “comfort measures only” orders or a planned discharge to hospice.

For potential participants for whom there was concern about decisional impairment, we conducted the University of California San Diego Brief Assessment for Capacity to Consent. ¹³ If decisional impairment was confirmed, proxies were sought for the informed consent process, with verbal assent from participants. Proxies were also sought for participants who were unable to consent due to the severity of their illness (e.g., those who were intubated).

Participants or proxies were contacted via phone or video conferencing software by research coordinators, who explained the study, answered questions, and completed a verbal informed consent process.

Assessment schedule

All assessments were completed via phone or videoconference for the convenience of participants and the safety of the study team. A schedule of the study interviews appears in Figure S1. Participants underwent a baseline interview, conducted in English or Spanish, during their COVID‐19 hospitalization or within two weeks of hospital discharge. Completion of the baseline interview was required for the participant to be considered enrolled in the study. We collected information about pre‐hospitalization physical function, health status, symptoms, and psychological and lifestyle factors. We also performed cognitive testing, described in detail below. Proxies completed an abbreviated interview that omitted cognitive testing as well as certain measures that could only be reliably reported by patients, such as depressive and anxiety symptoms. Follow‐up interviews containing the same or similar measures occurred at one, three, and six months from the date of hospital discharge or, for readmitted patients, from the date of discharge from the readmission. Follow‐up interviews were completed by patients whenever possible or by proxy. A summary of assessments at each time point appears in Table S1. Details for select assessments appear in Table S2.

Primary measures

Physical function

Physical function was assessed using a 15‐item scale that captures disability in seven basic activities of daily living (bathing, dressing, transferring from a chair, walking inside the house, eating, toileting, and grooming), ¹⁴ five instrumental activities of daily living (shopping, housework, meal preparation, taking medications, and managing finances), ¹⁵ and three mobility activities (walking a quarter mile, climbing a flight of stairs, and lifting or carrying 10 pounds). ¹⁶ During the baseline interview, participants were asked about their disability in these 15 functional activities one month prior to admission. Such information is valid when reported retrospectively and better reflects pre‐hospitalization function than disability measured during acute hospitalization. ¹⁷ Disability in each functional activity was defined as needing help from another person to complete the task. In follow‐up assessments, participants were asked about disability in the same 15 functional activities “at the present time.”

Cognitive function

Cognition was assessed using the Montreal Cognitive Assessment (MoCA) 5‐minute protocol, ¹⁸ an abbreviated version of the MoCA that has been modified for administration by telephone. The MoCA 5‐minute protocol measures attention, orientation, language, verbal learning and memory, and executive function. It displays psychometric properties similar to the full MoCA ¹⁹ and utilizes a cutoff of <22 (range, 0–30) to indicate cognitive impairment. The MoCA 5‐minute protocol was administered during the baseline interview and at each of the follow‐up interviews.

It would have been ideal to have been able to compare each participant's results to an assessment performed prior to hospitalization, in order to distinguish those with established cognitive impairment from those with new deficits in the setting of COVID‐19 and acute illness. Such data, however, are not collected in routine clinical care. To look for evidence of preexisting cognitive impairment, we instead reviewed participants' hospitalization records, primary care notes, and notes from neurologists, geriatricians, and psychiatrists during the 3 years prior to the index admission, using search terms that have been previously validated for identifying cognitive impairment in the electronic health record. ²⁰

Symptoms

We administered a version of the Edmonton Symptom Assessment System (ESAS) ²¹ that we modified by adding symptoms relevant to COVID‐19, including cough, anosmia, and ageusia, to comprise a list of 14 symptoms (Table S2). Given work suggesting that the intensity of symptoms varies considerably among older persons, ²² we also measured symptom severity, with “1” indicating that a symptom was mild, “2” indicating that it was moderate, and “3” indicating that it was severe, yielding a total symptom score ranging from 0 to 42.

Secondary measures

Health status

Using a question from the Short Form‐12, ²³ participants were asked to report their health status one month prior to admission during the baseline interview. Participants were asked about their health status at the present time during follow‐up interviews.

Frailty

Frailty was assessed in accordance with the Fried phenotype, ²⁴ with two questions adapted for remote assessment (self‐reported walking speed in lieu of a gait test and difficulty opening a jar in lieu of a grip strength test ²⁵ ) and the modified questions about low physical activity used in the National Health and Aging Trends Study. ²⁶ Frailty was assessed during the baseline and six‐month interviews.

Sensory impairments

Vision and hearing were measured using questions adapted from the National Health and Aging Trends Study. ²⁷ Hearing impairment was assessed at all time points, whereas vision impairment was assessed only at the baseline interview.

Psychosocial function

Depressive and anxiety symptoms were assessed with the four‐item Patient Health Questionnaire (PHQ‐4). ²⁸ Social support was measured with a five‐item version of the Medical Outcomes Study Social Support Survey. ²⁹

Falls and fractures

Fall history was assessed via a single question regarding a history of falls within the past 12 months. Fracture history was assessed by asking patients whether they had broken a bone since turning 50. Fracture risk was evaluated using selected questions from the Fracture Risk Assessment Tool (FRAX) ³⁰ and the International Osteoporosis Foundation's One Minute Risk Check. ³¹

Post‐discharge recovery and events

At follow‐up interviews, we collected information about health care utilization since the last assessment, including the number and occurrence of hospital admissions and emergency department visits; whether or not the participant had received care in a skilled nursing facility; and use of home‐ and community‐based services, including the location and frequency of physical and occupational therapy. Participants were asked about restricted activity with questions about whether they had cut down on their usual activities or stayed in bed for at least half a day due to illness, injury, or other problems. ¹⁴ We also asked about fractures since the previous interview.

Vaccination status

We added a question to the baseline and follow‐up interviews asking participants if they had received a COVID‐19 vaccine after vaccines became widely available in January 2021 (see the timeline in Figure S2). Participants were considered unvaccinated in interviews conducted before December 15, 2020. Missing data about vaccine status were filled in via chart review.

Medical record data

Patient‐reported data were supplemented with data from the electronic health record. We obtained granular data about the index hospitalization from the Yale Department of Medicine COVID‐19 Data Explorer (DOM‐CovX), ³² which contains health record data from all patients admitted to a YNHHS hospital with a positive SARS‐CoV‐2 test. Data in the DOM‐CovX registry include information about the severity of illness, as determined by the Sequential Organ Failure Assessment (SOFA) score ³³ ; need for advanced respiratory support; and detailed laboratory data. The registry also contains data about specific treatments for COVID‐19, such as corticosteroids, remdesivir, and tocilizumab, that were recommended in the YNHHS COVID‐19 treatment pathway during the study period (Figure S2). The DOM‐CovX registry also provided data about demographics, medical comorbidities, and discharge disposition. Missing data from the registry were filled in via chart abstraction, when possible.

We obtained information about delirium during the index hospitalization using the validated CHART‐DEL and CHART‐DEL‐ICU instruments, which involve looking for diagnoses synonymous with delirium and for words and phrases that suggest a change in the patient's usual mental status. ³⁴ , ³⁵ For every VALIANT participant, abstractors reviewed the medical record for each day of the COVID‐19 hospitalization. They then recorded whether delirium was present at any time and rated their confidence in this assessment (definite, probable, possible, uncertain, or no evidence). To maximize sensitivity, we followed prior work in operationalizing delirium as being present with any confidence level other than “no evidence.”

The Area Deprivation Index, a detailed measure of the socioeconomic conditions in a specific geographical area, was calculated based on each participant's zip code using the Neighborhood Atlas. ³⁶

Statistical analysis

Participant characteristics at the baseline and follow‐up interviews were summarized using means and standard deviations for normally distributed continuous variables, medians and interquartile ranges for ordinal or non‐normally distributed continuous variables, and proportions for categorical variables. Analyses were performed using SAS 9.4 (SAS Institute, Cary, NC).

Data sharing

Data from VALIANT will be made available to the public, for use upon approval by the principal investigators, on January 31, 2023.

RESULTS

Between June 18, 2020 and June 30, 2021, we screened 1747 older adults at five hospitals and determined that 939 were eligible to participate. The most common reasons for exclusion were long‐term residence in a skilled nursing facility (n = 155) and advanced dementia (n = 151). Among those who were eligible, 680 consented, and 341 were enrolled (Figure S3). There were no differences in demographic characteristics between the screened and enrolled participants.

Characteristics of the enrolled participants appear in Table 1. The mean age was 71.4 (standard deviation [SD] 8.4) years, 175 (51.3%) were women, and 125 (36.6%) were of Black race or Hispanic ethnicity. Nearly one‐third of participants had Medicaid as one of their forms of insurance. Most participants lived at home, with 239 (70.5%) reporting that they lived with others and 92 (27.1%) reporting that they lived at home alone. There were 25 participants (7.3%) whose index COVID‐19 illness involved readmission to the hospital within 7 days of discharge.

TABLE 1.

Sociodemographic characteristics of VALIANT participants

Characteristic	VALIANT participants (N = 341) a
Age, years, mean (SD)	71.4 (8.4)
Female sex, n (%)	175 (51.3)
Race and ethnicity, n (%)
White, non‐Hispanic	211 (61.9)
White, Hispanic	8 (2.3)
Black, non‐Hispanic	79 (23.2)
Black, Hispanic	1 (0.3)
Hispanic, race not reported	37 (10.9)
Non‐Hispanic, race not reported	1 (0.3)
Asian, non‐Hispanic	3 (0.9)
Unknown race and ethnicity	1 (0.3)
Primary language, n (%)
English	315 (92.4)
Spanish	26 (7.6)
Highest level of education completed, n (%)
Less than high school	50 (14.9)
High school or equivalent	146 (43.6)
2‐ or 4‐ year degree	100 (29.9)
Graduate or post‐graduate	39 (11.6)
Medicaid recipient, n (%)	107 (31.4)
Living in socioeconomically disadvantaged area, n (%) b	82 (25)
Marital status, n (%)
Married or living with partner	160 (47.2)
Single	64 (18.9)
Divorced or separated	52 (15.3)
Widowed	62 (18.3)
Other	1 (0.3)
Living situation prior to admission, n (%)
At home alone	92 (27.1)
At home with others	239 (70.5)
Short‐term resident at skilled nursing facility	5 (1.5)
Assisted living facility	3 (0.9)
Readmitted within 7 days of index COVID‐19 admission, n (%)	25 (7.3)
Respondent, n (%)
Participant	309 (90.6)
Proxy	32 (9.4)

^a There were missing data for the following variables (number of participants with missing values noted in parentheses): highest level of education completed (6), living in socioeconomically disadvantaged area (13, due to inability to calculate Area Deprivation Index), marital status (2), living situation (2).

^b Area Deprivation Index score > 55.

Clinical and functional characteristics of study participants before their hospital admission with COVID‐19 are provided in Table 2. Participants had a median of three comorbidities (interquartile range [IQR] 1–4), and 157 (46.6%) reported having at least one disability a month prior to hospitalization. For 27 (7.8%) participants, there was documentation of cognitive impairment in the electronic health record prior to admission. Of the 301 participants with complete frailty data, 56 (18.6%) were frail, 147 (48.8%) were pre‐frail, and 98 (32.6%) were not frail. There were 52 participants (15.2%) who had received a COVID‐19 vaccine before being hospitalized.

TABLE 2.

Clinical characteristics of VALIANT participants prior to COVID‐19 hospitalization

Characteristic	VALIANT participants (N = 341) a
General health in month prior to hospitalization, n (%)
Excellent	47 (15.2)
Very good	89 (28.7)
Good	99 (31.9)
Fair	63 (20.3)
Poor	12 (3.9)
Comorbidity count (range, 0–11), median (IQR) b	3 (1–4)
Disability prior to hospitalization
Number of functional activities in which participant experienced disability (range, 0–15), median (IQR) c	0 (0–4)
Disability in any functional activity, n (%)	157 (46.6)
Disability in ≥1 basic activity of daily living, n (%)	48 (14.2)
Disability in ≥1 instrumental activity of daily living, n (%)	123 (36.3)
Disability in ≥1 mobility activity, n (%)	143 (42.3)
Documentation of cognitive impairment in the medical record prior to hospitalization, n (%) d	12 (7.8)
Frailty classification in month prior to hospitalization e
Not frail	98 (32.6)
Pre‐frail	147 (48.8)
Frail	56 (18.6)
Social support in month prior to hospitalization (range, 5–25), median (IQR) f	23 (20–25)
Hearing impairment, n (%) g	120 (35.5)
Vision impairment, n (%) h	50 (14.9)
Fall in 12 months prior to hospitalization, n (%)	101 (29.8)
History of fracture since age 50, n (%)	67 (19.8)
History of tobacco use	161 (52.4)
≥1 COVID‐19 vaccine dose prior to hospitalization, n (%)	52 (15.2)

^a There were missing data for the following variables (number of participants with missing values noted in parentheses): general health (31), disability in functional activities (4), frailty (40, due to missingness in individual components of the 5‐item measure), social support (40), hearing impairment (3), vision impairment (5), history of falls (2), history of fractures (3), history of tobacco use (34).

^b Comorbidities included hypertension, myocardial infarction, heart failure, cerebrovascular disease (stroke, transient ischemic attack, or intracranial hemorrhage), diabetes mellitus, chronic lung disease, chronic kidney disease, end‐stage renal disease, liver disease, immunocompromised status (autoimmune disease, HIV positive, or receipt of solid organ transplant), and malignancy (solid tumors, leukemia or lymphoma, or metastatic disease).

^c Tasks included seven basic activities of daily living (eating, dressing, bathing, toileting, grooming, getting in and out of a chair, walking around indoors); five instrumental activities of daily living (doing housework, going shopping, preparing a meal, taking medications, managing finances); and three mobility activities (walking a quarter of a mile, climbing stairs, and lifting or carrying heavy objects). A higher score indicates greater disability. Four participants had missing data about some items. The denominator for basic activities of daily living and for mobility activities is 338, while the denominator for instrumental activities of daily living is 339.

^d Evidence of preexisting cognitive impairment was ascertained through review of the electronic health record for 3 years prior to the index hospitalization, using previously validated search terms. ²⁰

^e Frailty was assessed using the Fried frailty score, which consists of five items: low physical activity, slow gait speed, muscle weakness, weight loss, and exhaustion. Participants with scores ≥3 were classified as frail, those with scores of 1–2 were classified as pre‐frail, and those with scores of 0 were classified as not frail.

^f Assessed items included having someone to confide in or talk to about problems, having someone to help with daily chores if experiencing illness, having someone to turn to for suggestions about how to deal with a personal problem, having someone to love and make the participant feel wanted, and having someone to get together with for relaxation or fun. Each item was assessed using a 5‐point Likert scale, from 1 (none of the time) to 5 (all of the time). Higher scores indicate greater social support.

^g Using a hearing aid, not hearing well enough to carry on a conversation in a quiet or loud room, or being deaf.

^h Having difficulty reading newspaper print, recognizing a person across the street, or watching television from across the room (with use of corrective lenses, when applicable).

Clinical characteristics of study participants during their hospitalization with COVID‐19 are shown in Table 3. Most were hospitalized on general wards; however, 40 (11.7%) participants required intensive care unit (ICU) admission, and 29 (8.5%) required stepdown unit care. Sixty (17.6%) participants required advanced respiratory support with either high‐flow nasal cannula, non‐invasive positive pressure ventilation, or invasive mechanical ventilation, while 266 (78.0%) required low‐flow supplemental oxygen, and 15 (4.4%) did not require oxygen support. The median hospital length of stay was eight (IQR 6–12) days.

TABLE 3.

Clinical characteristics of VALIANT participants during COVID‐19 hospitalization

Characteristic	VALIANT participants (N = 341) a
Highest level of care, n (%)
Floor	272 (79.8)
Stepdown unit	29 (8.5)
Intensive care unit	40 (11.7)
Highest level of respiratory support, n (%)
None	15 (4.4)
Nasal cannula	266 (78.0)
High flow oxygen	36 (10.6)
Non‐invasive positive pressure ventilation	10 (2.9)
Mechanical ventilation	14 (4.1)
Received vasopressors, n (%)	18 (5.3)
SOFA score, maximum (range, 0–24), median (IQR) b	3 (2–4)
Hospital length of stay, days, median (IQR)	8 (6–12)
ICU length of stay, days, median (IQR)	7 (4–15)
Died in hospital after the baseline interview, n (%)	5 (1.5)
COVID‐specific treatments given, n (%)
Remdesivir	246 (72.1)
Tocilizumab	68 (19.9)
Corticosteroids c	239 (70.1)
Selected laboratory values
High‐sensitivity C‐reactive protein (maximum), mg/L, mean (SD)	109.6 (84.3)
D‐dimer, mg/L FEU (maximum), median (IQR)	1.4 (0.8–3.4)
Interleukin‐6, pg/ml (maximum), median (IQR)	20.6 (8.9–58.6)
Procalcitonin value suggestive of high likelihood of bacterial infection, n (%) d	63 (19.0)
Elevated troponin, n (%) e	118 (44.7)
Body mass index, mean (SD) f	31.0 (7.2)
Number of symptoms endorsed (range, 0–14), mean (SD)	9 (3)
Symptom, n (%)
Fatigue	276 (89.3)
Weakness	265 (85.8)
Cough	257 (83.2)
Sleepiness (daytime)	241 (78.0)
Loss of appetite	230 (74.4)
Shortness of breath	212 (68.6)
Sleeplessness (night‐time)	200 (64.7)
Pain	177 (57.5)
Diarrhea	169 (55.2)
Dizziness	155 (50.2)
Ageusia	150 (48.5)
Trouble thinking	140 (45.3)
Nausea	120 (38.8)
Anosmia	114 (36.9)
Total modified Edmonton Symptom Assessment System score (range, 0–42), mean (SD) g	19 (9)
Cognitive impairment, n (%) h	124 (45.3)
Delirium during hospitalization, n (%) i	62 (18.2)
Depression over past 2 weeks j	93 (30.4)
Anxiety over past 2 weeks k	66 (21.5)
Physical therapy or occupational therapy ordered, n (%)	17 (5.0)
Code status, n (%) l
Full	299 (87.7)
Limited	36 (10.6)
Comfort measures	6 (1.8)
Transferred from an outside hospital, n (%)	6 (1.8)
Discharge destination
Home	176 (51.6)
Home with services	106 (31.1)
Skilled nursing facility	47 (15.0)
Long‐term acute care facility	2 (0.6)
Inpatient hospice	2 (0.6)
Expired in hospital after enrollment	5 (1.5)
Left against medical advice	3 (0.9)

^a There were missing data for the following variables (number of participants with missing values noted in parentheses): cognitive impairment (67), symptoms during COVID‐19 illness (36), depression (35), anxiety (34), high‐sensitivity C‐reactive protein (12), D‐Dimer (1), interleukin‐6 (101), procalcitonin (9), troponin (77), and delirium (26).

^b SOFA, Sequential Organ Failure Assessment. The SOFA score is based on oxygenation, platelet count, Glasgow Coma Scale, bilirubin, mean arterial pressure or receipt of vasoactive agents (dopamine, dobutamine, epinephrine, or norepinephrine), and creatinine. Scores range from 0 to 24, where higher scores indicate worse severity of illness.

^c Among participants, 221 received dexamethasone, 10 received hydrocortisone, 18 received methylprednisolone, and 35 received prednisone; 35 participants received more than 1 corticosteroid.

^d Any procalcitonin value >0.5 ng/mL.

^e Any troponin value above the upper limit of normal, as reported by the manufacturer of the assay used at each study site.

^f Body Mass Index was collected from the index admission or the closest measure within 6 months prior to index admission in the electronic health record.

^g The 14 symptoms above were combined into a modified version of the Edmonton Symptom Assessment System scale that accounts for symptom severity (0 = symptom absent, 1 = mild, 2 = moderate, 3 = severe). A higher score indicates greater symptom burden.

^h Assessed through the Montreal Cognitive Assessment 5‐minute protocol, with a score ≥22 indicating no cognitive impairment and a score <22 indicating cognitive impairment.

ⁱ Ascertained with chart reviews using the CHART‐DEL and CHART‐DEL‐ICU instruments.

^j Assessed through two items on the Patient Health Questionnaire‐4, asking how often over the past 2 weeks the participant felt down, depressed, or hopeless and how often there was little interest or pleasure in doing things. Each item was scored from 0 (not at all) to 3 (nearly every day). A total score ≥3 was interpreted as positive for depression and <3 as negative for depression.

^k Assessed through two items on the Patient Health Questionnaire‐4, asking how often over the past 2 weeks the participant felt nervous, anxious, or on edge, and how often the participant was unable to stop or control worrying. Each item was scored from 0 (not at all) to 3 (nearly every day). A total score ≥3 was interpreted as positive for anxiety and <3 as negative for anxiety.

^l Abstracted through manual chart review. “Limited” indicates that there were restrictions on care, such as a do‐not‐resuscitate or do‐not‐intubate order. When present, comfort measures orders were placed after study enrollment, since patients with such orders at the time of screening were excluded from participation. For participants whose code status changed during hospitalization, the code status with the most restrictions on care was chosen.

The majority of participants (239, or 70.1%) received a corticosteroid during their hospitalization, which in most cases was dexamethasone. Remdesivir was administered to 246 (72.1%) participants and tocilizumab to 68 (19.9%) participants. Considering the maximum value reported for each study participant during hospitalization, the median D‐Dimer was 1.4 (IQR 0.8–3.4) mg/L Fibrinogen Equivalent Units and the median interleukin‐6 (IL‐6) value was 20.6 (IQR 8.9–58.6) pg/ml, although IL‐6 values were missing for 101 patients. During the baseline interview, participants reported a mean of 9 (SD 3) out of 14 possible COVID‐19‐related symptoms and an average symptom burden of 19 (SD 9), out of a maximum of 42, when each symptom was graded according to severity. There were 124 (45.3%) participants who had a score of <22 on the MoCA 5‐minute protocol, suggestive of cognitive impairment. Delirium was present in 62 (18.2%) participants.

Among participants, 252 of 341 (73.9%) completed all three follow‐up interviews, 39 (11.4%) completed two follow‐up interviews, and 20 (5.9%) completed one follow‐up interview. Thirty participants (8.8%) had no follow‐up interviews. Longitudinal data about function, cognition, and selected symptoms appear in Figure 2. Panel A compares the total count of disabilities on the 15‐item physical function scale at each follow‐up interview to the disability count prior to hospitalization. At one month, 42.5% of participants experienced a higher level of disability than before their COVID‐19 hospitalization. At six months, 36.0% of participants had increased functional disability as compared to before hospitalization. Panel B and Panel C describe trends in cognitive impairment and symptom burden among participants with complete data from each follow‐up interview. At one month, 21.4% of participants scored in the impaired range on the MoCA 5‐minute protocol, and 19.7% of participants scored in the impaired range at six months (Panel B). At one month, 53.7% of participants reported fatigue, 44.6% reported dyspnea, and 20.3% reported dizziness. The prevalence of all three symptoms decreased at six months, but remained high (52.8%, 37.7%, and 14.7% of participants, respectively).

FIGURE 2.

Physical function, cognition, and selected symptoms among VALIANT participants. Panel A shows the percentage of participants at each follow‐up interview whose scores on the 15‐item functional disability scale were higher than at enrollment, when they were asked about functional disability prior to hospitalization. These are persons who experienced increased functional impairment following their COVID‐19 hospitalization. Panels B and C show trends in cognition and selected burdensome symptoms among participants with complete data at each follow‐up interview. Panel B shows the percentage of participants at each time point with scores in the impaired range (<22) on the MoCA 5‐minute protocol. Panel C shows the percentage of participants at each time point with three burdensome symptoms (fatigue, dyspnea, and dizziness).

DISCUSSION

We designed VALIANT to provide detailed, high‐quality data about the health outcomes of older adults surviving hospitalization with COVID‐19. We placed particular emphasis on evaluating the outcomes that matter most to older persons, including the preservation of physical function and cognition as well as freedom from burdensome symptoms. ¹⁰ Cohort participants were racially and ethnically diverse, nearly half had a preexisting functional disability, and most were frail or prefrail. Cognitive impairment and burdensome symptoms during hospitalization were common. At the six‐month follow‐up interview, more than a third had worse physical function than before hospitalization, nearly 20% had cognitive impairment, and burdensome symptoms remained highly prevalent.

There is a growing body of literature examining clinical outcomes following COVID‐19 infection. Prior research has used administrative data to examine the post‐acute sequelae of disease ³⁷ and to describe functional impairment among COVID‐19 survivors who receive home health care services. ³⁸ There have been studies assessing persistent symptoms following hospitalization, ³⁹ , ⁴⁰ the development of new disability, ⁴¹ , ⁴² and cognitive impairment, ⁴³ with investigators in the Post‐Hospitalization COVID‐19 Study using cluster analysis to identify recovery phenotypes among a large of cohort of patients discharged from a hospital in the United Kingdom. ⁴⁴ VALIANT adds to this work in several important ways. First, we specifically focused on older patients, who are much more likely to have serious illness with COVID‐19 and are at much greater risk of functional and cognitive impairment following hospitalization. ⁶ , ⁷ The mean age of our cohort, 71.4 years, is substantially higher than in prior studies. Second, existing cohorts of COVID‐19 survivors have nearly all collected data during a single interview. ⁴¹ , ⁴³ , ⁴⁴ One exception involves patients in Wuhan, China, who were hospitalized early in the COVID‐19 pandemic and then evaluated at 6 and 12 months after infection. ⁴⁵ This cohort was considerably younger than ours, however, and participants did not undergo assessment at the time of hospitalization or provide information about function or cognition. VALIANT was designed to overcome these limitations by including a baseline interview, performed during or as close to the acute illness as possible, and by obtaining repeated interviews during the six months following hospital discharge. Since the available evidence suggests that recovery among older persons is highly dynamic, with frequent transitions between states of disability and independence, ⁴⁶ this approach will shed new light on the trajectories of function, cognition, and symptom burden that older adults experience following hospitalization with COVID‐19. We expect that these high‐quality data will guide the development of interventions to improve the health and well‐being of older survivors and will contribute to the evidence base that informs medical decisions for these patients.

Our study has several additional strengths. COVID‐19 has caused serious illness disproportionately in historically marginalized racial and ethnic groups that have often been excluded from medical research. ⁴⁷ The racial and ethnic makeup of the VALIANT cohort approximates the racial and ethnic makeup of the two counties from which most participants were recruited. ⁴⁸ Specifically, more than 20% of participants were Black and more than 13% were Hispanic. In addition, participants were socioeconomically diverse, with over 30% covered by Medicaid, and their health characteristics mirrored the heterogeneity of older adults seen in clinical practice.

There are important limitations to these data. The first is that all assessments were conducted remotely. While the study team initially planned to conduct the final, 6‐month visit in person, this proved infeasible because of restrictions put in place during the pandemic. We instead constructed study interviews with items that could be measured reliably and with high validity via telephone or videoconference. The second is that the data collected in interviews were self‐ or proxy‐reported, which raises the possibility of recall bias. We attempted to minimize this issue by asking participants to look back over the narrowest feasible period. Third, because participants did not routinely undergo assessments of their cognitive function before hospitalization, it is difficult to be certain whether data suggestive of cognitive impairment from the baseline interview, which was performed during acute illness, reflect new or longstanding deficits. We did, however, use a validated method to capture evidence of preexisting cognitive impairment in patients' medical charts. Finally, while most eligible patients agreed to participate, only slightly more than half of the consented patients completed the baseline interview, despite efforts by highly experienced and dedicated research staff. Research personnel reported that many patients simply felt too sick to participate in a research study. This raises the possibility that our cohort is biased towards patients who were less ill during hospitalization, and, indeed, only 20% of VALIANT participants were admitted to an ICU or stepdown unit. However, nearly all required supplemental oxygen, the majority received corticosteroids and remdesivir, and the median hospital length of stay was eight days—characteristics that reflect the presence of severe illness.

In conclusion, VALIANT is a unique longitudinal study of older adults who have survived hospitalization with COVID‐19, designed to provide evidence about the health outcomes that matter most to older persons. Among cohort participants, there was a high prevalence of worsened function, impaired cognition, and persistent symptoms in the months following hospital discharge. Future analyses will examine participant outcomes and outcome trajectories in greater detail. We anticipate that the knowledge gained from this rich resource will ultimately help guide the development of interventions directed at improving the health of older COVID‐19 survivors.

AUTHOR CONTRIBUTIONS

Study concept and design: Cohen, Hajduk, Ferrante. Acquisition of subjects and data: Cohen, Geda, Lee, Hajduk, Ferrante. Analysis and interpretation of data: All authors. Preparation of manuscript: All authors.

FUNDING INFORMATION

This work was supported by the Claude D. Pepper Older Americans Independence Center at Yale University (P30AG21342 and P30AG21342‐18S1). The Yale Department of Medicine COVID‐19 Data Explorer and Repository was supported by the George M. O'Brien Kidney Center at Yale (P30DK079310). Dr. Cohen and Dr. Ferrante were supported by Paul B. Beeson Emerging Leaders in Aging awards (K76AG059987 and K76AG057023) from the National Institute on Aging. Mr. Lee was supported by the Richard K. Gerson fund at Yale School of Medicine.

CONFLICT OF INTEREST

The authors declare that there is no conflict of interest.

SPONSOR'S ROLE

The funding sources were not involved in the design and conduct of the study; the collection, management, analysis, and interpretation of the data; or the preparation, review, and approval of the manuscript.

Supporting information

Figure S1. Schedule of study assessments for VALIANT.

Figure S2. Timeline of events during VALIANT recruitment and enrollment.

Figure S3. Screening and enrollment flow diagram.

Table S1. Interview items completed at each VALIANT assessment.

Table S2. Definition and operationalization of selected variables.

Cohen AB, McAvay GJ, Geda M, et al. Rationale, Design, and Characteristics of the VALIANT (COVID‐19 in Older Adults: A Longitudinal Assessment) Cohort. J Am Geriatr Soc. 2022;1‐13. doi: 10.1111/jgs.18146