Table 4: Evidence qualities of included studies.

No. of studiesQuality assessmentNo. of patientsEffect Quality Importance
DesignRisk of biasInconsistencyIndirectnessImprecisionOther considerationsChinese herbal medicineControlRelative (95% CI)Absolute

Cough symptom score (better indicated by lower values)
5Randomised trialsSerious1,2No serious inconsistencyNo serious indirectnessNo serious imprecisionNone216162Not pooled ◯ 
Moderate
Critical

Cough relief time (cough relief time) (better indicated by lower values)
7Randomised trialsSerious3No serious inconsistencyNo serious indirectnessNo serious imprecisionNone585382Not pooled ◯ 
Moderate
Critical

Cough disappearance time (better indicated by lower values)
2Randomised trialsNo serious risk of biasNo serious inconsistencyNo serious indirectnessNo serious imprecisionNone322128Not pooled
High
Critical

TCM syndrome clinical effective rate
9Randomised trialsSerious3,4No serious inconsistencyNo serious indirectnessNo serious imprecisionNone687/764
(89.9%)
357/485
(73.6%)
Not pooledNot pooled ◯ 
Moderate
Important

Quality-of-life (QoL) score (better indicated by lower values)
2Randomised trialsNo serious risk of biasNo serious inconsistencyNo serious indirectnessNo serious imprecisionNone14489Not pooled
High
Important

Fujimori et al. stopped the study when the significant differences between two groups were detected, which may lead to an overestimation of intervention in the trial group.  
2Irifune et al. conducted an open-label trial, which would introduce an influence on subjective patient-report results.  
3An MC and Wang YF conducted an open-lable study, which would introduce a bias to subjective patient-report results.  
4Huang MH conducted a quasirandom method, which would introduce selection bias.