This was a prospective, single centre, randomised pilot trial that was undertaken between March 2021 and August 2021.The reporting of this pilot trial followed the Consolidated Standards of Reporting Trials for pilot and feasibility studies (CONSORT) and Template for Intervention Description and Replication (TIDieR) guidelines.The trial was approved by the Sydney Local Health District Human Research Ethics Committee (X20-0545 & 2020/ETH03228) and was registered prospectively withthe Australian and New Zealand Clinical Trials Registry (Trial ID: ACTRN12621000031864).
This trial was a sub-study of an ongoinglongitudinal observational cohort investigation of people with confirmed SARS-CoV-2 infection led by St Vincent’s Hospital, Sydney Australia (ADAPT). The ADAPT study included people with COVID-19 disease, regardless of severity, who participated in a broad range of assessments to comprehensively characterize the immune-pathobiological effects of COVID-19 disease.From the ADAPT cohort, participants who reported ongoing symptoms, defined as persistent dyspnoea, fatigue or chest pain, at any time-point after 110 days post diagnosis were screened for inclusion in this trial.
Participants were eligible for enrolment if they had consented to the ADAPT study, were aged 16 years and over and had ongoing symptoms of Long-COVID,defined as persistent dyspnoea, fatigue or chest pain at any time beyond 110 days post SARS-Cov-2 diagnosisPatients were excluded if they had pre-existing lung disease amenable to pulmonary rehabilitation eg COPD, had completed supervised exercise training in the 3- month period prior, had a household member already recruited to this sub-study, or were unable to participate due to other medical conditions or unwilling to participate intelerehabilitation. The reason for declining participation was recorded.
Participants were contacted via telephone by a member of the investigating team to be screened for eligibility. Consenting eligible participants underwent a standardised in-person baseline assessment during the enrolment visit.
After the enrolment visit, participants were randomised (1:1) using a computer generated randomisation (www.sealedenvelope.com) with concealed allocation to one of two groups: 1) telerehabilitation that included supervised exercise training over videoconferencing; and 2) a control group who performed no supervised exercise training over the study period.
Telerehabilitation Intervention Group
The group telerehabilitation exercise intervention was based on previous telerehabilitation programs that have beenreported to be effective in patients with COPD.Participants allocated to the intervention groupwere given instructions via email to set up a home-based exercise area and advised of equipment requirements; a walking track free of clutter, weighted objects to use as hand weights and a stable chair. An instructional copy of the Borg dyspnoea scale was also provided along with links to join the videoconferencing platform. All telerehabilitation sessions were conducted by physiotherapists at St Vincent’s Private Hospital Sydney, in groups of up to six participants, twice weekly for ten weeks with participants needing to complete at least 16 sessions to meet an exercise adherence definition and to allow for sessions missed due to illness or other commitments. The telehealth exercise sessions consisted of walking training, upper and lower limb strengthening and combination aerobic movements with modifications made so that the participants were exercising at a rating of three to five on the modified Borg scale (0-10) for dyspnoea and perceived exertion.
A detailed example of an exercise session is shown in Supplement 1. Participants were not given any specific exercise prescription to complete on other days.
Participants allocated to the control group received usual medical care and participation in other observational components of the larger longitudinal ADAPT study. The control group did not participate in supervised exercise training and were not given any advice regarding exercise training. The control group were offered access to the same tele-rehablilitation sessions or individual physiotherapist advice after completion of their control period.
All participants attended one in-person visit at enrolment and one at the completion of the study at St Vincent’s Private Hospital Sydney. On both visits, all outcome measures were collected by physiotherapists who were blinded to the group allocation with no involvement in any components of the telerehabilitation program. Participants in the trial and the physiotherapists who were blinded to the group allocation were instructed not to discuss the intervention received when undertaking the outcome assessment procedures.
The primary feasibility outcomes for this pilot trial were i)study recruitment rate, ii)adherence rate to the telerehabilitation program, iii) completion rate, iv) adverse events during the telerehabilitation program and v) technological issues limiting participant participation. The recruitment rate wasdefined as the number of patients that met the inclusion criteria who then were enrolled in the study divided by the number of patients contacted to be invited to the study. The adherence to the telerehabilitation program was measured by the percentage of enrolled patients who completed at least 16 of the 20 exercise sessions. Completion was measured as the percentage of participants who were followed-up successfully after completing the follow-up assessments during the follow-up visit. Safety was measured by the number of adverse events reported during the telerehabilitation period. Technological issues limiting connectivity to and participation in the telerehabilitation sessions were also recorded.
Exploratory Clinical Outcomes:
Exploratory clinical data-collection measurements included i) six-minute walk test, ii) health-related quality of life, iii) perceived level of fatigue, iv) five repetition sit to stand test, v) gait speed and vi) handgrip grip strength.
Six-minute Walk Test
Physiological functional capacity was estimated using the six-minute walk test (6MWT). Two 6MWTs were completed, at least 30 minutes apart,and followed standardised procedures. The longest distance recorded from either of the two 6MWTs was used for analysis. The 6MWT result was compared to 6MWT estimates for healthy Australian individuals. Peripheral oxygen saturation (Sp02) and heart rate (HR) were continuously monitored during the test with a pulse oximeter (Masimo-Rad-5v, Masimo Corporation, Irvine, Ca, USA). Perceived level of dyspnoea was determined before and after completion of each 6MWT using the modified 0-10 Borg scale of perceived dyspnoea. 
Health related quality of life
Health-related quality of life (HRQoL) was assessed by the St George’s Respiratory Questionnaire (SGRQ). The SGRQ rates the HRQoL of people with chronic respiratory disease on a 0-100 scale with a increasing scores indicating decreasing HRQoL. 
Level of fatigue was assessed using the 0-52 point FACIT-Fatigue scale, with a score of less than 34 points indicating severe levels of fatigue. [23, 24]
Five repetition sit to stand test
The functional capacity of the quadriceps was measured using the five repetition sit to stand (5STS) test. The participants were asked to stand up and sit down from a 48cm chair, without using their arms, five times as fast as possible. The fastest of two tests was recorded for subsequent analysis. Participants who could not stand up without using their arms did not complete the 5STS test. 
Gait speed was assessed using the four-metre gait speed test (4MGS), with the testcompleted three times using a standardised procedure. The shortest time recorded was used for subsequent analysis.
Handgrip strength was measured using a handgrip dynamometer (Jamar Plus Dynamometer, Cedaburg Wisconsin USA), with each hand tested alternatively three times. The average result for the dominant hand was used for subsequent analysis.
As a pilot feasibility study, the objective was to test trial procedures, recruitment potential and feasibility and safety. The future trial arising from this pilot will aim for a medium effect size with 80% power; as suchBell et al.  recommend 10 participants per arm in the pilot study.
Demographic and symptom variables were summarised using means and standard deviations for numerical variables and count with percentage for categorical variables. For each exploratory clinical outcome measure, mean scores and 95% confidence intervals (CI) were reported for each group together with the 95% CI for the mean difference between groups and difference in mean change. The latter intervals are provided for use with future trial sample size estimation and are not intended for inferential purposes. All statistical analyses were performed using R version 4.1.0.