Table 5.

GRADE strength of evidence for different neuromonitoring approaches.

Table 5a: GRADE strength of evidence for SSEP
Quality criteria	Rating (circle one for each criterion)	Footnotes (explain reasons for up- or downgrading)	Quality of the evidence (circle one per outcome)
Risk of bias	No serious (-1) very serious (-2)	All studies observational with majority of them being low risk of bias	⊕⊕⊕⊕ High ⊕⊕⊕□ Moderate ⊕⊕□□ Low ⊕□□□ Very low
Inconsistency	No serious (-1) very serious (-2)	Inconsistency in results/estimates/effects very likely due to differences in included population/pathology type (deformity vs tumor vs degenerative vs mixed population)
Indirectness	No serious (-1) very serious (-2)	All studies employed a postoperative neurologic examination as a reference standard using which the utility/diagnostic efficacy of neuromonitoring was assessed
Imprecision	No serious (-1) very serious (-2)	Due to low number of events (true positive + false negatives), most studies have very large confidence intervals, particularly for sensitivity
Publication bias	Unlikely likely (-1) very likely (-2)	Test for asymmetry not statistically significant
Large effect	Large (+1) Very large (+2)	No
Dose-response gradient	No Yes (+1)	NA
Plausible confounding would change the effect	No Yes (+1)	NA