Table 1: GRADE Quality of Studies – Gastroparesis.

Outcome	Design	Quality	Consistency	Directness	Overall Quality
Reduced vomiting & nausea (1° end point)	RCT (Abell et al. 2003) Case Series	Low Why? Confounders related to diabetes. Possible Type 2 error for subgroup analyses. Subjective self-reported end point. Posthoc change in primary end point analysis. No sample size justification. Concomitant prokinetic/antiemetic therapy. Same as above.	Some uncertainty. Why? Only 1 RCT (with different results for FDA and publication).	Some uncertainty. Why? GES originally hypothesized to correct gastric rhythms, stimulate gastric emptying and therefore eliminate symptoms. Now hypothesized to directly act on neurons to the CNS to control symptoms. Weak correlation between symptoms and gastric emptying. Unclear whether gastric emptying is still considered an end point to investigate.	Low
Improved gastric emptying (2° end point)	Open label portion of Abell et al. 2003 & Case series	Low Why? Same limitations as above. Not a 1° end point. No intent-to-treat analysis.	Wide variation in improvement. Same uncertainties as above.	Same uncertainties as above.	Low/Very Low