Table 2. Risk of bias assessment of included randomized controlled trials.

Bias domain	Source of bias	Lipton et al. 2009 (PRISM study) [16]	Saper et al. 2011 (ONSTIM study) [17]	Silberstein et al. 2012 [19]	Serra and Marchioretto, 2012 [18]	Slotty et al. 2014 [31]
Selection bias	Random sequence generation	Unclear risk	Low risk	Low risk	Unclear risk	Low risk
	Allocation concealment	Unclear risk	Low risk	Low risk	Unclear risk	Low risk
Performance bias	Blinding of participants	Unclear risk	Unclear risk (high risk for medication management group)	Unclear risk	High risk	Low risk (high risk for suprathreshold stimulation)
	Blinding of study personnel	Unclear risk	Unclear risk	Unclear risk	High risk	Low risk
Detection bias	Blinding of outcome assessment: patient reported outcomes	Unclear risk	Unclear risk (high risk for medication management group)	Unclear risk	High risk	Low risk (high risk for suprathreshold stimulation)
	Blinding of outcome assessment: investigator assessed outcomes	Low risk	Low risk	Low risk	High risk	Low risk
Attrition bias	Incomplete outcome data	Unclear risk	High risk	Low risk	Low risk	Low risk
Reporting bias	Selective reporting	Unclear risk	High risk	Unclear	Low risk	Low risk
Other bias	Any other important concerns about bias not covered in the other domains above	Based on conference abstract with very limited information; manufacturer-sponsored study	Manufacturer-sponsored study	Manufacturer-sponsored study	High risk (weakness related to crossover design—see below)	High risk (patients already had good treatment response; lack of washout—see below)
	Measurement of effectiveness of blinding and/or patients’ expectation of treatment effectiveness	Not done	Not done	Not done	Not done	Not done
Crossover design	Analysis of paired data	Not applicable	Not applicable	Not applicable	Not done	Yes
	Assessment of carryover effects and/or justification of washout period	Not applicable	Not applicable	Not applicable	Not done	Not done