Table 1. Key findings from intervention co-development meetings.

Component	Outcome/consensus
Online group exercise classes	All agreed that one supervised exercise session per week of 20-60 minutes duration, dependent on fitness level, should be tested for acceptability during the pre-pilot phase. Additional pre-recorded exercise classes planned to provide “on- demand” video material for participants to follow unsupervised at home, up to 2-3 per week. Videos would include a variety of exercises to appeal to all, including: Yoga, Pilates, chair-based and breathing exercise videos. Concerns were expressed by patients that some exercises may not be tolerated by more debilitated patients. The need for robust safety/emergency procedures was also highlighted. People with long-COVID symptoms expressed a loss of confidence about being physically active and anxiety about relapse; also concern about over-exertion. An emphasis on fun over intensity was agreed for first few weeks to promote confidence and adherence. All agreed that light-moderate aerobic exercise (40-70% Heart Rate Reserve (HRR), rating of perceived exertion (RPE) - 11–14) was realistic for most participants after familiarisation. Gentle mobility, coordination, and balance exercises would be needed after long sedentary periods for some participants. Clinical exercise physiologists should test acceptability during the pre-pilot phase. All Long Covid patients agreed that IT literacy might be an obstacle to accessing sessions for older participants.
Concern about post-exertional malaise (PEM)	Members of trial team met with ME groups to discuss the dangers of post-exertional malaise (PEM) in the long-COVID population. The word exercise seemed to have negative connotations, assumed to only mean high intensity/exertion levels which could be detrimental to recovery. Use of term physical activity would be better received. Groups were reassured that (a) intensity would be low-moderate and at the comfort level of the individual and (b) that the risk of PEM would be minimised by closely monitoring participants before, during, and after each session, as well as assessing risk of PEM during the initial 1:1 consultation (c) a pre-determined ‘graded’ approach adopted in previous trials would not be applied to this population. Specific information was provided to participating clinicians during the training sessions highlighting the need to be aware of PEM and how to address it, and a section on PEM was incorporated into the Practitioner Manual provided to all clinicians taking part in REGAIN.
Exercise progression	It was agreed that exercise physiologists should produce 8-12 exercise session templates of different levels of intensity for standardised delivery across the trial. Classes would follow a circuit format with active recovery between exercises. A point was made that quite conceivably a group containing very low and very high ability participants was likely as practitioners would not always be able to arrange groups based on ability level. Therefore, a great degree of versatility was needed with exercise class design and delivery. People with long-COVID symptoms requested ample rest periods and water breaks. Following warm-up and mobility, duration of exercise should be progressed primarily as tolerated, for maximal therapeutic benefit. Researchers requested that exercise physiologists collect ratings of perceived exertion (RPE) to assess intervention fidelity.
Behavioural support materials	Mind-body disconnect was mentioned by people with long-COVID symptoms, so they felt it would be really useful to have an integrated intervention. They felt strongly that long Covid wasn’t simply physical symptoms of fatigue and breathlessness, but also how this affected their thoughts and feelings. Many topics were deemed useful by patients, especially ‘challenging unhelpful thoughts’ and ‘anxiety management’. Pacing of activities was considered crucial. Talking to the trained practitioner before and after group exercise sessions in a supportive environment was thought to be important to ensure continued adherence and group cohesion. This would also promote peer support. Breathing was difficult for some and so being able to stop, take a breath and relax was important. Reinfection was also major source of anxiety/worry. The participant workbook was thought by patients to be a long document, but all content was relevant as this could be read in their own time. The written targets were welcomed (e.g. writing down goals), and making notes during the session was deemed useful. Patients preferred a paper copy to an online PDF. Patients suggested colour-coded sections of the workbook to make it less daunting, and tabs to make easy to find sections for each weekly session. Suggested workbook format included adding the aims of behavioural support, with short clear statements at the beginning of each section of the workbook.
Feasibility of delivery within NHS setting	Exercise physiologists and health psychologists highlighted potential obstacles to online delivery. These included: the need for a second practitioner, or ‘co-pilot’, for safety reasons, IT connectivity , participant retention, risk of busy classes, accurate intensity monitoring, and adequate behavioural intervention training for practitioners less experienced with these techniques. Actions were agreed to mitigate these issues: (1) deliver appropriate training for practitioners led by health psychologist and an experienced clinical research fellow; (2) provide comprehensive reference manuals for participants and practitioners to follow; and (3) ensure sufficient time was spent with each participant during the 1:1 consultation to familiarise them with the online platform and IT requirements.