Study design

This prospective cohort study is a subgroup analysis study of Irish data collected as part of a pan European project (EU-COGER), the protocol for which has been published elsewhere [27]. See Supplementary Material 1 for EU-COGER consortium list. The conduct and reporting of this study were in accordance with the STROBE checklist for cohort studies [28]. See Supplementary Material 2 for a copy of this checklist.

Ethics approval

Ethical approval was obtained from the Health Service Executive Mid-Western Region Research Ethics Committee for University Hospital Limerick (UHL) on the 27th of April 2021 and amended approval was obtained on the 5th of August 2021 (025/2021) to include participation in the EU-COGER study. Ethical approval was obtained from Beaumont Hospital Ethics (Medical Research) committee on the 9th of April 2021 (21/31—COVID) and included approval to participate in the EU-COGER study also.

Participants

Recruitment to this study took place at two acute hospital sites in Ireland: Beaumont Hospital and University Hospital Limerick (UHL). Beaumont hospital (BH), located in Dublin city, provides emergency and acute care services to a largely urban population of approximately 290,000 people. UHL is an acute hospital in the Mid-West of Ireland serving an urban and rural population. It provides acute and emergency services to a region with a population of approximately 473,269 people.

Participants consecutively admitted to the acute hospital and who met the following inclusion criteria were recruited between April, 2021, and November, 2021:

Recovering from acute COVID-19 infection requiring treatment in an acute hospital, confirmed with polymerase chain reaction for viral RNA, serology for antibodies against SARS-CoV-2, or an approved alternative test.

Accepted for geriatric rehabilitation following acute COVID-19 infection in the hospital setting.

Due to the study design, study sample size calculation was inappropriate and the decision to recruit over the course of 6 months was in keeping with the recruitment period for the EU-COGER study. Patients from UHL were recruited by a clinical research nurse (GC) and patients from BH were recruited by a geriatrician (FD) and an occupational therapist (KMcL).

Patients were excluded if they had a severe cognitive impairment that would impact their decision-making ability to participate in a study and/or those who did not provide written informed consent. The presence of a severe cognitive impairment was determined by a comprehensive chart review by the clinical research nurse and deemed positive if the patient met any one of the following criteria:

The interventions provided were determined by each patients’ unique health and rehabilitation needs which included, but was not limited to Occupational Therapy, Physiotherapy, Clinical Nutrition and Dietetics, and Specialist Nursing.

Procedures and outcome measures

A comprehensive list of study variables and time points can be found in the EU-COGER study protocol [27]. Outcomes reported for the purpose of this secondary analysis are as follows: demographics and baseline clinical characteristics including age, sex, pre-morbid residence, date of admission, medical history and co-morbidities, date of admission, date of discharge, acute hospital length of stay, number of days engaged in rehabilitation, discharge destination, and COVID-19 status including testing method.

Study timepoints for clinical outcomes were as follows:

T0: pre-morbid (retrospective evaluation, where applicable)

T1: on admission to rehabilitation

T2: on discharge from rehabilitation

T3: 6 weeks after discharge from rehabilitation (obtained via telephone interview)

T4: 6 months after discharge from rehabilitation (obtained via telephone interview)

For this study, pre-morbid ratings refer to performance/presentation before acute viral illness with COVID-19. All UHL data were gathered by the study’s first author (AMcC) and a clinical research nurse (GC). All BH data were gathered by the EU-COGER national coordinator for Ireland who was also a geriatrician (FD) and an occupational therapist (KM).

Functional ability was measured through use of the Barthel index [29] given its ability to measure change in function in the older adult population [30]. The Barthel is a 10-item ordinal instrument reporting patient’s performance of activities of daily living, including bathing, toileting, and dressing. The Barthel was administered at all timepoints (T0–T4).

Patient reported symptoms were measured through use of the self-report COVID-19 Yorkshire Rehabilitation Screen (C-19 YRS). The original version of the screen was used [31]. The C-19 YRS is a patient-reported outcome measure and was developed by healthcare professionals in the UK’s National Health Service who were practicing in the rehabilitation of patients with COVID-19. The scale asks patients to rate their perceived symptom severity or functional difficulty “now” and “pre-COVID” in the domains of breathlessness, mobility, fatigue, personal care, usual activities, pain/discomfort, anxiety, depression, and global perceived health status. New onset symptoms pertaining to laryngeal complications, voice, swallowing, nutrition, continence, cognition, and communication are also recorded. The perceptions of family and care givers regarding symptomatology and presentation are also sought. Ordinal items of the C-19 YRS are rated on a Likert scale of 0–10. Zero indicates no difficulty/absence of symptoms and 10 indicates severe difficulty/severe symptoms, except for global perceived health where higher ratings indicate better health and lower ratings indicate poorer health. Categorical items are rated with a yes/no response. No overall score can be generated in this earliest version of the screen. The C-19 YRS was administered at T2 and T4. Five participants had the assessment of the C-19 YRS at T3 and T4.

QoL was measured through use of the EQ-5D-5L [32]. The tool comprises of five domains: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression. The respondent rates each domain on five levels: no problems, slight problems, moderate problems, severe problems, and extreme problems. The tool also requires respondents to rate their overall health on a visual analogue scale. The tool generates a profile score based on the respondents rating of the five domains. From the profile score, an index score can be generated based on country-specific tariffs. The derived index scores range from −0.59 to 1. The maximum index score of 1 indicates full health and lower scores indicate poorer health. The EQ-5D-5L was administered at all timepoints (T0–T4).

Statistical analysis

Descriptive statistics were used to profile the baseline characteristics of the cohort. Categorical measures, for example gender, were analysed in terms of frequencies and percentages; continuous measures, for example Barthel index, were analysed in terms of means and standard deviations.

Function and quality of Life

Separate repeated measures ANOVAs were conducted to explore differences in Barthel index scores and EQ-5D-5L index scores across the following time points: T0, T1, and T4. Data from these three timepoints were used given the focus of our research question on long-term outcomes. Where necessary, violations of sphericity resulted in use of a greenhouse Geisser correction. All analyses were conducted using JASP v.0.16.2. Post-hoc analysis was also carried out, if analysis of variance demonstrated a significant result, to determine the significance of paired outcomes between timepoints. Data were graphically represented using raincloud plots that combine a cloud of data points with a box plot and a one-sided violin plot.

Patient-reported symptoms

Paired samples t-test or their non-parametric equivalent—Wilcoxon matched pairs signed rank tests—was conducted to examine C19-YRS ratings “now” and “pre-COVID” on the following domains: breathlessness at rest, dressing, and stairs, problems with walking, personal care, and usual activities, fatigue, pain, anxiety, depression, and overall health.