Table 2 GRADE analysis for disease activity and secondary outcomes

Disease Activity

SMD = -0.22 [-0.41, -0.02]

10

418

436

Moderate ⨁⨁⨁◯

We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different

Mixture of RCTs and non-RCTs: High quality

60% studies had moderate RoB, 40% had high RoB: no downgrade

Even though we used a random effect model meta-analysis, we consider heterogeneity as an index of inconsistency. I2 = 43%, not considerable (< 75%), no downgrade

Very few studies with disease activity as primary aim. Downgrade 1 level

N = 854 sample size, very large so unlikely to be imprecise. No downgrade

We used an exhaustive searching approach (i.e. scientific databases, grey literature, scientific organizations). Funnel plot is asymmetrical, downgrade 1 level

z score = 2.21, large effect. Upgrade 1 level

N/A

We found no confounding factors that indicate upgrading

Functional Ability (normal)

MD = -0.21 [-0.37, -0.06]

8

482

491

Very Low ⨁◯◯◯

Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect.

Mixture of RCTs, non-RCTs and observational cohort studies: Moderate quality

50% studies had moderate RoB, 50% had high RoB: downgrade 1 level

Even though we used a random effect model meta-analysis, we consider heterogeneity as an index of inconsistency. I2 = 85%, considerable heterogeneity, downgrade 1 level

Studies highly varied in primary aim, with very few with function as primary aim. Downgrade 1 level

n = 973, unlikely to be imprecise, no downgrade

We used an exhaustive searching approach (i.e. scientific databases, grey literature, scientific organizations). No major bias in the funnel plots. No downgrade

z score = 2.66, large effect. Upgrade 1 level

N/A

We found no confounding factors that indicate upgrading

Functional Ability (non-normal)

MD = -0.00 [-0.06, 0.06]

4

209

223

High ⨁⨁⨁⨁

We are very confident that the true effect lies close to that of the estimate of the effect.

Mixture of RCTs and non-RCTs: High quality

100% studies had moderate RoB: no downgrade

Even though we used a random effect model meta-analysis, we consider heterogeneity as an index of inconsistency. I2 = 2%, no heterogeneity

most studies primary aim is function, No downgrade

n = 432, unlikely to be imprecise, no downgrade

We used an exhaustive searching approach (i.e. scientific databases, grey literature, scientific organizations). No major bias in the funnel plots. No downgrade

z score = 0.06, no effect. No upgrade

N/A

We found no confounding factors that indicate upgrading

Pain

SMD = -0.13 [-0.79, 0.53]

10

586

640

Very Low ⨁◯◯◯

We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Mixture of RCTs, non-RCTs and observational cohort studies: Moderate quality

70% studies moderate, 30% high RoB: no downgrade

Even though we used a random effect model meta-analysis, we consider heterogeneity as an index of inconsistency. I2 = 96%, considerable heterogeneity, downgrade 1 level

Studies varied in primary aim, with 6 with pain as primary aim (< 50%). Downgrade 1 level

N = 1226, unlikely to be imprecise, no downgrade

We used an exhaustive searching approach (i.e. scientific databases, grey literature, scientific organizations). Funnel plot is asymmetrical, downgrade 1 level

z = 0.4, no effect. No upgrade

N/A

We found no confounding factors that indicate upgrading

Fatigue

SMD = -0.42 [-0.63, -0.21]

10

534

523

Moderate ⨁⨁⨁◯

We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different

Mixture of RCTs, non-RCTs and observational cohort studies: Moderate quality

60% moderate, 40% high RoB: no downgrade

Even though we used a random effect model meta-analysis, we consider heterogeneity as an index of inconsistency. I2 = 56%, no downgrade

studies had varied primary aims, downgrade 1 level

n = 1057, unlikely to be imprecise, no downgrade

We used an exhaustive searching approach (i.e. scientific databases, grey literature, scientific organizations). No major bias in the funnel plots. No downgrade

z = 3.87, large effect. Upgrade 1 level

N/A

We found no confounding factors that indicate upgrading

Anxiety

SMD = -0.19 [-0.95, 0.57]

3

82

70

Very Low ⨁◯◯◯

We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different

Mixture of RCTs and non-RCTs: High quality

66% moderate, 33% high risk: no downgrade

Even though we used a random effect model meta-analysis, we consider heterogeneity as an index of inconsistency. I2 = 81%, considerable heterogeneity, downgrade 1 level

 < 50% studies had anxiety in their primary aim, downgrade 1 level

n = 152, small sample so likely to be imprecise, downgrade 1 level

We used an exhaustive searching approach (i.e. scientific databases, grey literature, scientific organizations). No major bias in the funnel plots. No downgrade

z = 0.48, little effect. No upgrade

N/A

We found no confounding factors that indicate upgrading

Depression (non-normal)

MD = -0.92 [-2.71, 0.87]

2

62

54

Low ⨁⨁◯◯

We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different

Mixture of RCTs and non-RCTs: High quality

100% studies had moderate RoB: no downgrade

Even though we used a random effect model meta-analysis, we consider heterogeneity as an index of inconsistency. I2 = 49%, no downgrade

 < 50% studies had depression in their primary aim, downgrade 1 level

n = 116, small sample so likely to be imprecise, downgrade 1 level

We used an exhaustive searching approach (i.e. scientific databases, grey literature, scientific organizations). No major bias in the funnel plots. No downgrade

z = 1.01,, little effect. No upgrade

N/A

We found no confounding factors that indicate upgrading

Depression (normal)

SMD = -0.39 [-1.08, 0.30]

4

119

115

Very Low ⨁◯◯◯

We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different

Mixture of RCTs and non-RCTs: High quality

25% studies had moderate Rob, 75% high risk: downgrade 1 level

Even though we used a random effect model meta-analysis, we consider heterogeneity as an index of inconsistency. I2 = 83%, considerable heterogeneity, downgrade 1 level

 < 50% studies had depression in their primary aim, downgrade 1 level

n = 234, small sample so likely to be imprecise, downgrade 1 level

We used an exhaustive searching approach (i.e. scientific databases, grey literature, scientific organizations). No major bias in the funnel plots. No downgrade

z = 1.11, little effect. No upgrade

N/A

We found no confounding factors that indicate upgrading

Quality of Life

SMD = 0.29 [-0.05, 0.62]

9

685

738

Very Low ⨁◯◯◯

We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Mixture of RCTs, non-RCTs and observational cohort studies: Moderate quality

92% studies had moderate RoB, 8% high risk: no downgrade

Even though we used a random effect model meta-analysis, we consider heterogeneity as an index of inconsistency. I2 = 87%, considerable heterogeneity, downgrade 1 level

 < 50% studies had QoL in their primary aim, downgrade 1 level

n = 1423, unlikely to be imprecise, no downgrade

We used an exhaustive searching approach (i.e. scientific databases, grey literature, scientific organizations). Funnel plot is asymmetrical, downgrade 1 level

z = 1.69, little effect. No upgrade

N/A

We found no confounding factors that indicate upgrading

Sedentary Time

MD = -46.80

[-162.30, 68.70]

2

128

130

Very Low ⨁◯◯◯

We are very confident that the true effect lies close to that of the estimate of the effect.

Mixture of RCTs and non-RCTs: High quality

50% studies had moderate RoB, 50% high risk: downgrade 1 level

Even though we used a random effect model meta-analysis, we consider heterogeneity as an index of inconsistency. I2 = 88% considerable heterogeneity, downgrade 1 level

 > 50% studies primary aim was to target SB, no downgrade

n = 258, small sample so likely to be imprecise, downgrade 1 level

We used an exhaustive searching approach (i.e. scientific databases, grey literature, scientific organizations). No major bias in the funnel plots. No downgrade

z = 0.79, no effect. No upgrade

N/A

We found no confounding factors that indicate upgrading

Steps

SMD = 0.30 [0.03, 0.57]

3

130

95

Moderate ⨁⨁⨁◯

We are very confident that the true effect lies close to that of the estimate of the effect.

Mixture of RCTs and non-RCTs: High quality

33% studies had moderate RoB, 67% high risk: downgrade 1 level

Even though we used a random effect model meta-analysis, we consider heterogeneity as an index of inconsistency. I2 = 0%, no heterogeneity, no downgrade

All studies outcomes were some form of PA/SB measure so these are sufficiently similar, no downgrade

n = 225, small sample so likely to be imprecise, downgrade 1 level

We used an exhaustive searching approach (i.e. scientific databases, grey literature, scientific organizations). No major bias in the funnel plots. No downgrade

z = 2.15, large effect. Upgrade 1 level

N/A

We found no confounding factors that indicate upgrading

MVPA

SMD = 1.21 [-0.01, 2.44]

7

307

292

Moderate ⨁⨁⨁◯

We are very confident that the true effect lies close to that of the estimate of the effect.

Mixture of RCTs and non-RCTs: High quality

65% studies moderate risk, 35% high risk: no downgrade

Even though we used a random effect model meta-analysis, we consider heterogeneity as an index of inconsistency. I2 = 98%, considerable heterogeneity, downgrade 1 level

 > 50% studies primary aim was to target MVPA, no downgrade

n = 599, unlikely to be imprecise, no downgrade

We used an exhaustive searching approach (i.e. scientific databases, grey literature, scientific organizations). No major bias in the funnel plots. No downgrade

z = 1.94, little effect. No upgrade

N/A

We found no confounding factors that indicate upgrading

Total PA

SMD = 0.03 [-0.37, 0.43]

4

200

189

Moderate ⨁⨁⨁◯

We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different

Mixture of RCTs and non-RCTs: High quality

50% had high RoB: downgrade 1 level

Even though we used a random effect model meta-analysis, we consider heterogeneity as an index of inconsistency. I2 = 71%, no heterogeneity, no downgrade

 > 50% studies primary aim was to target PA, no downgrade

n = 389, unlikely to be imprecise, no downgrade

We used an exhaustive searching approach (i.e. scientific databases, grey literature, scientific organizations). No major bias in the funnel plots. No downgrade

z = 0.13, no effect. No upgrade

N/A

We found no confounding factors that indicate upgrading

Light/leisure PA

SMD = 0.45 [0.27, 0.64]

4

238

225

High ⨁⨁⨁⨁

We are very confident that the true effect lies close to that of the estimate of the effect.

Mixture of RCTs and non-RCTs: High quality

75% studies had moderate RoB, 25% high risk: no downgrade

Even though we used a random effect model meta-analysis, we consider heterogeneity as an index of inconsistency. I2 = 0%, no heterogeneity

 > 50% studies primary aim was to target some form of PA or SB, no downgrade

n = 463, unlikely to be imprecise, no downgrade

We used an exhaustive searching approach (i.e. scientific databases, grey literature, scientific organizations). No major bias in the funnel plots. No downgrade

z = 4.79, large effect. Upgrade 1 level

N/A

We found no confounding factors that indicate upgrading