Table 6.

Outcomes quality of evidence according to the Grades of Recommendation, Assessment, Development, and Evaluation (GRADE) approach

Outcomes	Number of studies (number of patients)	Design	Findings and direction	GRADE
Postconcussion symptoms	18 studies (n=2048)	Observational	Important heterogeneity in populations enrolled, definitions of outcome variables and evaluation duration. Only four associations between postconcussion symptoms and a biomarker were statistically significant. Only eight studies used multivariate regression analyses and CIs were often large.	Low
GOS-E and GOS	Nine studies (n=1235)	Observational	Slight discrepancies in definitions, wide differences in populations enrolled, methods quality as well as in evaluation duration and inconsistencies in associations (only three were significant), their direction and strength.	Insufficient
Return to work	Four studies (n=432)	Observational	Slight discrepancies in definitions and reporting but considerable differences in evaluation duration (1 week–1 year). Only one study showed a significant association with increased S-100β protein serum level.	Insufficient

GOS, Glasgow Outcome Scale; GOS-E, Glasgow Outcome Scale-Extended.