Is Deqi an Indicator of Clinical Efficacy of Acupuncture? A Systematic Review

Table 6

Summary of the findings table for the evidence of comparative effects of AWD versus AOD for Bell Palsy.

AWD compared with AOD for Bell Palsy
Study ID	Outcomes	Illustrative comparative risks (95% CI)		Relative effect (95% CI)	No. of participants (studies)	Quality of the evidence (grade)
		AOD (assumed risk)	AWD (corresponding risk)
Mei et al. 2010 [62]	Effective rate (followup: 3 months); Assessment of changes in facial nerve functions based on House-Brackmann scale^1	Study population		RR 1.40  (1.00 to 1.97)	50 (1 study)	⊕⊝⊝⊝; very low^2,3,4,5
		636 per 1000	890 per 1000  (636 to 1000)
Xu et al 2013 [63]	Complete recovery rate (followup: 6 months); House-Brackmann score graded by 3 assessors	Study population		RR 1.27  (1.14 to 1.42)	338 (1 study)	⊕⊕⊕⊕; high^3,5,6,7
		708 per 1000	899 per 1000  (807 to 1000)

1Patient-important outcome.
²Randomization methods and allocation concealment not mentioned. Stratified and randomized assignment and binding of the patient were mentioned. For acupuncture trials, blinding of the practitioner is impossible. None lost to followup. No selective outcome reporting.
³This item was omitted here because we assessed one single study.
⁴Subjective assessment based on any observed improvement on HB scale for facial nerve function. RR has a wide CI; it almost equals effect size and covers 1.0.
⁵A single study is very likely to be biased. However, it was omitted here to avoid all evidence being “very low” in quality and therefore indistinguishable.
⁶Computer-generated random number sequence, randomized assignment, allocation concealment (sealed opaque envelope, and a designated personnel kept it) and blinding of the patient, recruiter, and assessor were described. For acupuncture trials, blinding of the practitioner is impossible; 22/338 dropouts, ITT analysis done. No selective outcome reporting.
⁷Subjective outcome, but rigorously controlled. Specifically, three skilled experts rated scores according to the House-Brackmann scale. For RR, narrow CI equals 1/10 effect size.