Table 5. Quality of evidence assessment based on Grading of Recommendations, Assessment, Development and Evaluation (GRADE) criteria.

Movement variability based on the task preformed	Number of LBP patients (studies)	Risk of bias (D&B scale score)	Inconsistency	Indirectness	Imprecision	Number of participants		Publication bias	Overall Quality of Evidence
						LBP group	Control group
Repeated Flexion and Extension tasks	91 (3) Graham 2014 Mokhtarinia et al. 2016 Bauer et al. 2017	Not serious 67% 75% 83%	Not serious	Seriousb ↓	Seriousc ↓	10 22 59	10 22 27	Suspectedd	Very low
Lifting tasks	184 (5) Dideriksen et al. 2014 Bauer et al. 2015 Moreno Catalá et al. 2018 Pranata et al. 2018 Fujii et al. 2022	Not serious 75% 83% 67% 67% 67%	Seriousa ↓	Seriousb ↓	Seriousc ↓	17 63 30 43 31	17 31 29 29 20	Suspectedd	Very low
Gait	89 (5) Vogt et al. 2001 Lamoth et al. 2006a Lamoth et al. 2006b Seay et al. 2011 Ebrahimi et al. 2017	Not serious 67% 67% 67% 67% 67%	Seriousa ↓	Seriousb ↓	Seriousc ↓	34 19 12 14 10	22 14 12 14 10	Suspectedd	Very low
Sit to stand to sit (STS) task	16 (1) Ippersiel et al. 2018	Not serious 83%	Seriousa ↓	NA*	Seriousc ↓	16	21	Suspectedd	Very low

GRADE: High quality: very confident that the true effect lies close to that of the estimate of the effect and further research is very unlikely to change. Moderate quality: moderately confident in the effect estimate and further research may change the estimate. Low quality: limited confident in the effect estimate and further research is very likely to have an important impact to change the estimate. Very low quality: Any estimate of effect is very uncertain.

^a ↓High level of heterogeneity; we downgrade one level due to variability in results or inconsistency between effects.

^b ↓Evidence was not directly comparable to the question of interest; we downgrade one level due to the variety of outcome measurement tools including the used motion analysis systems and the analysis metrics.

^c ↓Few participants and few events: we downgrade one level due to the wide confidence interval (CI) around the estimate of the effect (as estimated by forest plot).

^dLimited number of observational studies with small sample size as well as asymmetrical distribution of observations on the funnel plot

NA* not applicable: Only one study.