Table 1.

Details of interventions promoting physical activity engagement or reducing sedentary behaviour in RA; study aim, design and content, participant characteristics, methodology and results.

Author and study aim	Design and content	Participant characteristics (years, M ± SD) 1. Age, 2. RA duration, 3. Gender (% female)	Assessment of PA (OB, SR)	Outcome and results (PA)
Minor et al., 1989 RCT - to evaluate the effects of 12-week exercise programme on exercise tolerance, disease-outcomes and self-reported health status Minor et al., 1993 – predictors of exercise adherence at 3, 6, and 18 months.	IG1 vs. IG2 vs. CG n = 40 (group n = not reported) IG1= aerobic exercise (walking) IG2 = aerobic exercise (aquatics) CG = non-aerobic ROM exercise Length of intervention: 12 weeks	1. 54 ± 14 2. 11 ± 8 3. 85%	SR: PA diary. Specific questions/data handling not described. Not validated in RA. Assessments: baseline, 12 weeks, 6 and 12 months.	Aerobic exercise (min/week) - no significant differences between IG and CG at 12 weeks, 6 or 12 months. Hrs/day ambulating – IG reported significantly more time in ambulation at 12 weeks compared to CG.
Brus et al., 1998 RCT - to determine the effects of patient education on compliance (with treatment regimens) and health in patients with RA	IG (n = 25) vs. CG (n = 30) IG = education programme on medication/PA compliance CG = RA brochure Length of intervention: 1 year	1. 60 ± 15 (IG) 59 ± 9 (CG) 2. Not reported 3. 80%	SR: Questionnaire - report performance of prescribed physical exercises and endurance activities (walking, swimming, cycling). Not validated in RA Assessments; baseline, 3, 6, and 12 months.	Physical exercise (min/week) – significantly higher in the IG at 3 months compared to the CG. Endurance activities (min/week)- no significant difference between IG and CG at any time point.
Feldthusen et al., 2003 RCT - examining effects of a person-centred physical therapy intervention focussed on health-enhancing PA and balancing life activities on fatigue and fatigue-related variables.	IG (n = 36) vs. CG (n = 34) IG = person centred physiotherapy to tailor PA and balancing life activities. CG = usual care – including visits with rheumatologist and rehabilitation as prescribed. Length of intervention: 12 weeks	1. 54 ± 9 (IG) 53 ± 11 (CG) 2. 14 ± 11 (IG) 12 ± 8 (CG) 3. 89%	SR: Leisure Time Physical Activity Index – reported previous 7-day PA. Specific questions/data handling not described. Not validated in RA. Assessments: baseline, 12 weeks, and 6 months	MVPA (hours/week) – significantly greater improvement in MVPA in the IG group compared to the CG, at 12 weeks and 6 months.
Van den Berg et al., 2006 RCT - to compare the effectiveness of 2 Internet-based PA interventions for patients with RA Hurkmans et al., 2010 - 2 year follow-up; n = 110 (56/54)	IG (n = 82) vs. CG (n = 78) IG = internet-based PA programme with individual guidance, bicycle ergometer and group contacts (via website) CG = internet-based programme providing only general information on exercise and PA (via website). Length of intervention: 12 months	1. 50 ± 13 (IG) 50 ± 14 (CG) 2. 8 ± 9 (IG) 6 ± 11 (CG) 3. 76%	**SR: Questionnaire – report days/week doing; 1) a moderate PA for >=30 minutes in succession, and 2) a vigorous PA for >=20 minutes in succession. Not validated in RA. OB: Accelerometer, Actilog 3. Assessed PA in 5-minute epochs over 5 days. 3 days of data used in analysis. Not validated in RA. Assessments; baseline, 3, 6, 9, 12 and 24 months	Proportion of participants meeting moderate and vigorous PA recommendations (SR) - significantly greater in the IG for moderate PA at 6 and 9 months (and vigorous PA at 6, 9 and 12 months), compared to the CG. No significant differences between groups at 2 years. Average number of accelerations/5-min period/day and total number of 5-min peak activity periods/day (OB) - no significant differences between IG and CG at any time point.
Mayoux-benhamou et al., 2008 RCT - to determine the effect of education on the exercise habits of patients with RA at 6 and 12 months	IG (n = 104) vs. CG (n = 104) IG = Multi-disciplinary education programme, including training in home-based exercise and guidelines for leisure time PA. CG = usual care + booklet with home-based exercises and leisure PA recommendations. Length of intervention: 12 months	1. 55 ± 12 (IG) 54 ± 14 (CG) 2. 12 ± 10 (IG) 14 ± 10 (CG) 3. 89%	SR: Baecke questionnaire. Specific questions/data handling not described. Not validated in RA. Assessments; baseline, 6 and 12 months.	Compliance with leisure time PA (increased score by >=20% from baseline) – significantly more participants in the IG group complied with leisure PA at 6 months, compared to CG. No significant difference at 12 months. Overall level of leisure PA - significantly increased at 6 months only in the IG. Decrease observed in both groups at 12 months.
Brodin et al, 2008 RCT - to investigate the effect of a 1-year coaching program for healthy PA on perceived health status, body function, and activity limitation in patients with early RA. Sjoquist et al., 2011 – 2 year follow-up; n = 228	IG (n = 94) vs. CG (n = 134) IG = coaching programme aimed at implementing healthy PA CG = access to physiotherapy, including education, treatment with physical modalities and organised exercise 2 × week. Length of intervention: 1 year	1. 54 ± 14 (IG) 56 ± 14 (CG) 2. 21 ± 5 (IG) 22 ± 4 (CG) 3. 74%	SR: Questionnaire – 3 × questions, regarding frequency of low, moderate and high-intensity PA. Response options; a) never / occasionally, b) 1–3 times/week, c) 4 –5 times/week, and d) 6–7 times/week. Not validated in RA. Assessments: baseline and 1 year	Number of participants classified as undertaking ‘healthy PA’ (>=30 mins of moderate or vigorous PA >= 4 days/week) - no significant differences between IG and CG at 1 year. No significant different at 2 years (Sjoquist et al., 2011).
Baxter et al., 2015 RCT (feasibility study) - to determine whether an RA walking programme successfully facilitated regular PA, without detriment to pain levels	IG (n = 11) vs. CG (n = 12) IG = Walking intervention – instructions on a walking route, to be completed 3–4 times/week CG = Nutrition education session Length of intervention: 6 weeks	1. 67 ± 10 (IG) 59 ± 13 (CG) 2. 9 ± 2 (IG) 6 ± 5 (CG) 3. 80%	OB: Pedometer. Worn for the duration of their involvement in the study. Assessments: baseline and 6-weeks.	Steps/day – both groups increased their step count at 6 weeks, but there were no significant difference between IG and CG.
Knittle et al., 2015 RCT - to evaluate the effects of targeting both the motivation and action phases of behaviour change in a 5-week intervention to increase PA among patients with RA not meeting current PA recommendations. Knittle et al., 2016 – theoretical process evaluation.	IG (n = 38) vs. CG (n = 40) IG = group-based patient education session + motivational interview (from physiotherapist) + 2 × self-regulation coaching sessions from a rheumatology nurse. CG = group-based patient education session. Length of intervention: 5 weeks (with follow-up phone calls focussed on self-regulation for the IG, at weeks 6, 12 and 18).	1. 61 ± 12 (IG) 65 ± 12 (CG) 2. Not reported 3. 67%	**SR: 1) Short Questionnaire to Assess Health-Enhancing PA. Report days/week and min/day undertaking walking, cycling and sporting activities; 2) report days/week engaged in >= 30 mins of >=moderate PA in the past month. Not validated in RA. Assessments: baseline, 6 and 32 weeks.	Leisure time PA (min/week) – significant group × time interaction over the 32 weeks of the study. However, no significant difference in change scores from baseline between IG and CG at either 6, or 32 weeks. Days/week >=30 mins of moderate PA (PA recommendations) – significant group × time interaction over 32 weeks of the study. A higher % of participants in the IG met the 5 × 30 min/week moderate PA recommendations, compared to the CG, at 6 and 32 weeks.
Nordgren et al., 2015 Not RCT – to document adherence to and changes in health-enhancing physical activity (HEPA) and functioning, and to explore aspects of adherence and response during the first year of an outsourced 2-year HEPA programme in people with RA. Nordgren et al., 2018 – 2 year follow-up; n = 117.	All participants received the intervention, n = 220 IG = encouragement for daily PA (pedometer, website), twice weekly circuit training and bi- weekly support group meetings to support PA behaviour change. Length of intervention: 2 years	1. 59 ± 9 2. 12 ± 10 3. 81%	SR: International Physical Activity Questionnaire (IPAQ) Short-Form. Categorised as adherers vs non- adherers based on 70% participation in HEPA (5 × 30 mins moderate PA/week). Low validity for moderate PA in RA. SR: Exercise Stage Assessment Instrument (ESAI), to determine adherence to HEPA (including muscle strength training >= 2 × week) for > 6 months. Assessments: baseline, 1 year, 2 years.	Current (adherence to) HEPA - significantly increased from baseline, to 1 year (55% to 82% adherence). Significantly decreased from year 1to year 2 (82% to 75%) (Nordgeren et al., 2018). Maintained (>6 months) of HEPA – significantly increased from baseline, to 1 year (0 to 37%). Significantly decreased from year 1, to year 2 (41% to 27%) Nordgeren et al., 2018).
Garner et al., 2018 RCT (feasibility) - to test the feasibility and effect of a brief individualised counselling intervention on PA levels and fitness, and dietary intake, compared with standard care in RA patients <1 year from diagnosis.	IG (n = 14) vs. CG (n = 14) IG = 2 × individualised nutrition and exercise counselling sessions. CG = usual care Length of intervention: 3 months (only 2 sessions).	1. 49 ± 14 (IG) 45 ± 10 (CG) 2. 21 days (IG); 23 days (CG) 3. 82%	OB: Pedometer. Worn for 7 days. Assessments: baseline and 6-months.	Steps/week - no significant differences between IG and CG.
Gilbert et al., 2018 RCT - to test the efficacy of the IMPAACT intervention for persons with RA (and OA) in improving arthritis-specific and generic self-reported pain and physical function outcomes, observed measures of function, and objectively measured and self-reported PA levels.	IG (n = 93) vs. CG (n = 92) IG = brief physician recommendation to increase PA to meet national guidelines + motivational interviewing for PA at baseline, 3, 6 and 12-months (year 1, + 2 × sessions in year 2). CG = brief physician recommendation to increase PA to meet national guidelines. Length of intervention: 2 years	1. 55 ± 14 (IG) 55 ± 14 (CG) 2. 13 ± 10 (IG) 13 ± 10 (CG) 3. 84%	OB: GT1M Actigraph accelerometer (7 day wear, 60 second epochs, valid wear requirement; >=4 days with >= 10 hours wear/day. MVPA, >=2020 accelerometer counts/min (Troiano et al., 2008). Cut-points not validated for use in RA SR: Yale Physical Activity Scale. Report time in PA during a typical week from the past month, and overall estimates of PA over the entire past month. Some evidence for validity in RA. Assessments: baseline, 3, 6, 12, 18 and 24 months.	Moderate to vigorous PA (OB) - no significant differences between IG and CG. 1) total time index; 2) energy expenditure index; 3) activity dimensions summary Index (SR) - no significant differences between IG and CG.
Katz et al., 2018 RCT - To test the effect of a pedometer- based intervention on increasing PA and decreasing fatigue among individuals with RA.	IG1 (n = 34) vs. IG2 (n = 34) vs. CG (28) IG1 = education brochure with suggestions of ways to increase PA.+ pedometer + step monitoring diary + step targets (10% above baseline levels, every 2 weeks). IG2 = education + pedometer + step monitoring diary CG = education only Length of intervention: 20 weeks	1. 55 ± 13 2. 15 ± 12 3. 88%	**OB: Pedometer (Jawbone Up). Worn over 7-days. Assessments; baseline, 10-weeks (questionnaires only), and 21-weeks.	Steps/day - both IGs significantly increased steps/day from baseline, to 21 weeks. The CG significantly decreased steps, from baseline to 21 weeks. Changes within the IGs significantly differed from those in the CG.
Sedentary behaviour
Thomsen et al., 2017 RCT - to investigate the efficacy of an individually tailored, theory-based motivational counselling intervention on reducing daily sitting time in RA. Thomsen et al., 2016 – randomised feasibility study of the same intervention; n = 20	IG (n = 75) vs. CG (n = 75) IG = 3 × motivational interviewing – counselling sessions + individual text message reminders, aimed at reducing daily sitting time. CG = encouraged to maintain usual lifestyle Length of intervention: 16 weeks	1. 60 ± 11 (IG) 60 ± 13 (CG) 2. 12 (range, 7 – 20) 3. 69%	**OB: ActivPAL 3 TM activity monitor. Worn over 7 consecutive days. Validated for use in RA SR: Physical Activity Scale 2.1 (PAS 2.1). Reported; 1) number of hours/minutes in an average 24 hour day spent sitting at work and during leisure time; 2) the longest continuous time with uninterrupted sitting during work/leisure time. Not validated for use in RA. Assessments; baseline and 16 weeks,	Time spent sitting (hours/day, OB - decreased in the IG (by 1.61 hours/day ) and increased in CG (by 0.59 hours/day), at 16 weeks. Significantly greater difference in change in the IG compared to CG. Time spent standing/stepping (hours/day, OB) –daily sitting was replaced by increased standing and stepping. Between-group differences in change of 1.52 and 0.55 hours/day, respectively. Daily sitting time (at work) – significant differences in favour of the IG.

Note: IG = intervention group, CG = control group, PA = physical activity, RA = Rheumatoid arthritis; MVPA = moderate-to-vigorous physical activity, SR – self-report, OB = objective.

^**indicates PA as study primary outcome.

Values are presented as intervention/control where information for both study arms is provided separately. Where studies included RA + OA participants (Minor et al., 1989, 1993; Knittle et al., 2015), only information for RA subsamples are reported. For participant characteristics (values are presented for the overall sample, or the IG or CG separately where overall aggregates for the sample are not available). Participant characteristics no reported for follow-up studies or theoretical process evaluation as not significant different from baseline assessments.