Summary of findings for the main comparison. Shock wave therapy versus placebo for rotator cuff disease with or without calcification.

Shock wave therapy for rotator cuff disease with or without calcification at 3 months
Patient or population: rotator cuff disease with or without calcification Setting: outpatient clinic Intervention: shock wave therapy Comparison: placebo therapy
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with placebo	Risk with shock wave therapy
Pain relief > 50%a Follow‐up: 3 months	375 per 1000	413 per 1000 (232 to 728)	RR 1.10 (0.62 to 1.94)	74 (1 study)	⊕⊕⊝⊝ Lowb,c	Shockwave therapy may provide no improvement in the number of participants with a pain reduction of 50% or more. Absolute change 4% more had relief (19% fewer to 26% more); relative change 10% more had relief (38% fewer to 94% more); NNTB: NAd
Pain Multiple scalese translated to VAS 0–10 (10 was severe pain)f Follow‐up: 3 months	Mean pain in the control group was 3.02 pointsg	Mean pain in the intervention group was 0.78 points better (0.17 better to 1.4 better)	SMD –0.49 (95% CI –0.88 to –0.11)	608 (9 studies)	⊕⊕⊕⊝ Moderateh	Shockwave therapy probably results in little or no clinically important improvement in pain. Mean pain did not appear to differ in participants with and without calcification: test for subgroup differences: Chi² = 0.25, df = 1 (P = 0.62), I² = 0% Absolute change 8% better (2% to 14% better); relative change 14% better (3% better to 25% better);i NNTB: 4 (95% CI 2 to 34)d
Function Multiple scalese translated to Constant 0–100 scale (100 was best function)f Follow‐up: 3 months	Mean function in the control group was 66 pointsg	Mean function in the intervention group was 7.9 points better (1.6 better to 14 better)	SMD 0.62 (95% CI 0.13 to 1.11)	612 (9 studies)	⊕⊕⊕⊝ Moderatej	Shockwave therapy probably results in little or no clinically important improvement in function. Mean function did not appear to differ in participants with and without calcification: test for subgroup differences: Chi² = 1.00, df = 1 (P = 0.32), I² = 0.1% Absolute change: 8% better (1.6% to 14% better); relative change 12% better (3% to 22% better);i NNTB: 3 (95% CI 2 to 18)d
Participant‐reported success Follow‐up: end of studies	255 per 1000	406 per 1000 (222 to 743)	RR 1.59 (0.87 to 2.91)	287 (6 studies)	⊕⊕⊝⊝ Lowb,c	Shockwave therapy may provide no improvement in the number of participants reporting treatment success. Absolute change 15% more had success (3% fewer to 49% more); relative change 59% more (13% fewer to 191% more); NNTB: NAd
Quality of life	—	—	—	—	—	Not measured
Number of participant withdrawals due to adverse events or treatment intolerance	103 per 1000	77 per 1000 (44 to 135)	RR 0.75 (0.43 to 1.31)	581 (7 studies)	⊕⊕⊝⊝ Lowb,c	We are uncertain if shockwave therapy increases withdrawal rates. Absolute change 3% less events (6% less to 3% more); relative change 25% less (57% less to 31% more); NNTH: NAd
Number of participants experiencing any adverse event Follow‐up: 12 months	72 per 1000	260 per 1000 (144 to 469)	RR 3.61 (2.00 to 6.52)	295 (5 studies)	⊕⊕⊝⊝ Lowb,c	We are uncertain if shockwave therapy increases adverse events. Absolute difference: 19% more events (7% more to 40% more); relative change: 261% more (100% more to 552% more); NNTH: NAd
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; NA: not applicable; NNTB: number needed to treat for an additional beneficial outcome; RR: risk ratio; SMD: standardised mean difference; VAS: Visual Analogue Scale.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: 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. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aThe a priori outcome was pain relief 30% or greater, which was not reported in any studies; thus we reported pain relief 50% or greater.
 ^b Downgraded one level due to study limitations (including risk of selection, detection, attrition, and reporting bias).

^cDowngraded one level for imprecision due to wide confidence intervals, or small number of participants or small number of events.

^dNumber needed to treat for an additional beneficial outcome (NNTB), or an additional harmful outcome (NNTH) not applicable (n/a) when result is not statistically significant. NNTB or NNTH for dichotomous outcomes calculated using Cates NNT calculator (www.nntonline.net/visualrx/). NNTB or NNTH for continuous outcomes calculated using Wells Calculator (CMSG editorial office), with an assumed minimal clinical important difference for pain of 1.5 points on 0 to 10 VAS, and for function of 10 points on 0 to 100 Constant score.
 ^ePain scores: VAS 0 to 10, VAS 0 to 100, Constant Score 0 to 15 (also called Constant score); function scores: Constant‐Murley 0 to 100, Shoulder Pain And Disability Index 0 to 100.
 ^fTranslated from SMD and 95% CIs to 0 to 10 VAS for pain and 0 to 100 Constant scale for function by multiplying the SMD by the standard deviation (SD) at baseline in the placebo group from Gerdesmeyer 2003 (values were mean (SD) VAS pain 5.6 (1.6), and mean Constant score (SD) 64.2 (12.8).
 ^gControl group mean (SD) values at 3 months' follow‐up from Gerdesmeyer 2003: values were 3.8 (2.3) on 0 to 10 VAS pain; 74 (15.5) on 0 to 100 Constant function score.
 ^h Downgraded one level due to study limitations (including risk of selection, detection, and attrition bias). Although this outcome had a high I² (80%), the outcome was not downgraded for inconsistency. This high I² was due to one outlier,Hsu 2008 and removing this outlier removes the statistical heterogeneity (I² = 0%) and does not change the direction of the effect

ⁱRelative changes calculated as absolute change (mean difference) divided by mean at baseline in the control group from Gerdesmeyer 2003 (values were 5.6 on 0 to 10 point VAS pain; 64.2 on 0 to 100 Constant score).

^jDowngraded one level due to study limitations (including risk of selection, detection, and attrition bias), and one level due to inconsistency (I² = 91%). Removing the potential extreme outlier reported in Hsu 2008 still left considerable heterogeneity (I² = 72%), additional removal of another, less extreme outlier (Cosentino 2003) resulted in I² = 38%. As we could explain the heterogeneity, we did not downgrade the certainty further.