2.2.2 All stages of labour
2.2.2.1 Analysis of pain level Five studies assessed incidence of pain at 3 levels including Ⅰ, Ⅱ and Ⅲ, in which pain at level Ⅲ had the greatest impact on participants. For meta-analysis, the incidence of pain at level Ⅲ was converted into a dichotomous variable. We pooled the studies and found that auricular acupressure significantly reduced the incidence of pain at level Ⅲ compared to the controls (RR=0.17, 95%CI [0.13, 0.24], p <0.05), with no heterogeneity (I2=0%) (Figure 8). Sensitivity analysis indicated that the results were robust.
2.2.2.2 Duration of each stage of labour Six studies investigated duration of each stage of labour. Summary results showed a significant reduction in duration of the first stage in the study groups compared to the control groups (MD=-2.57, 95%CI [-3.69, -1.45],p <0.05), with statistically significant heterogeneity (I2=81%, 97%, 99%). Nevertheless, there were no significant differences in the second and third stage of labour (Figure 9). Sensitivity analysis indicated that the results were robust.
2.2.2.3 Postpartum blood loss Two trials reported postpartum blood loss. Pooling the data of the studies showed no significant difference between auricular acupressure intervention and obstetric care routine, with significant heterogeneity (I2=97%) (Figure 10). Sensitivity analysis indicated that the results were robust.
2.2.2.4 Incidence of spontaneous delivery Pooled analysis of four studies that assessed incidence of spontaneous delivery showed no significant difference between the study groups and the control groups with significant heterogeneity (I2=84%) (Figure 11). Sensitivity analysis indicated that Ye et al.[25]appeared to add to significant heterogeneity. Removing this study eliminated this heterogeneity (I2=0%), and changed the effect size of the pooled result (Z = 2.89, P = 0.004), but the difference is still not significant.
2.2.2.4 Incidence of caesarean section Three studies reported incidence of caesarean section. A fixed effect model was applied, for there was no evidence of heterogeneity (I2=0%). Summary results indicated a lower incidence of caesarean section in auricular acupressure groups compared to the control groups (RR=0.36, 95%CI [0.20, 0.65], p =0.0006) (Figure 12). Sensitivity analysis indicated the results were robust.