Table 3. Quality assessment findings, per AMSTAR 2 tool.

Quality criteria	# of reviews meeting criteria, out of 35 total	%
1. Inclusion criteria covered PICO components	32	92%
2. Review protocol pre-established	22 (of which nine only partially met criteria)	63%
3. Described study design selection	33	94%
4. Used comprehensive literature search strategy	29	83%
5. Study selection performed by multiple reviewers	27	77%
6. Data extraction performed by multiple reviewers	23	66%
7. Described studies in adequate detail (eg, populations, interventions, comparators, outcomes, research designs)	35 (of which 16 only partially met criteria)	100%
8. Assessed study quality/risk of bias in a comprehensive manner (ideally using validated tool)	27 (of which six only partially met criteria)	77%
9. Accounted for study quality/risk of bias in interpretation of results	20	57%
10. Discussed heterogeneity of results and possible reasons for it	25	71%
11. Reported whether any conflicts of interest and any sources of funding for the review	32	91%
# of criteria met*	Out of n=31 reviews:
High quality: met 9–11 criteria	23 (6 met all 11 criteria)	66%
Moderate quality: met 6–8 criteria	12	34%
If meta-analysis, used appropriate methods for the statistical combination of results†	6 out of 6 reviews including meta-analyses	100%

^*Note that we dropped five studies due to poor quality (score of ≤5), reducing the number of included studies from 40 to 35.

^†Appropriate methods for the statistical combination of results: justified combining data via meta-analysis; used weighting technique to combine study results; accounted for heterogeneity between studies; separated findings by randomizedrandomised controlled trials versus non-randomizedrandomised designs.

AMSTAR 2A Measurement Tool to Assess Systematic Reviews, version 2PICOPopulation, Intervention, Comparator group, Outcome