3.1. General characteristics
We gathered 1531 studies from databases and other hand search sources. Eight of 1531 studies (Eight full-text papers under six trials) had eligibility criteria and were included in qualitative and quantitative analysis [33-40]. Three of them had the same ethic code (910732) (approval code of research ethics committee) and were different reports of one trial [33-35]. Also two other studies were similar and were different reports of one trial [36-37]. We included all different outcomes (which considered in our systematic review) of the different reports of each trial to the meta-analysis (if the trial had multiple reports). The flowchart of searching and inclusion process of studies is shown in Figure 1(According to PRISMA statement). Fore of included studies had used the date palm fruit [33, 36, 39-40], and one study had used date fruit honey [38]. Descriptions and characteristics of the reviewed studies are shown in Table 1. No side effects were reported for date palm in any of the studies. The minimum and maximum intervention durations of founded studies were two days and four weeks, respectively. Date palm and its extracts were orally utilized in all of the studies. Sample sizes of these studies were 89 to 210 participants. 653 participants were included in the analysis, 325 participants in intervention group, and 328 in control group. Ages of participants were ranged from 20 to 40. All of these studies had control groups consisting of placebo and routine cares in one study [38], and routine cares in other studies [33, 36, 39-40]. However, placebo control group was not included in meta-analysis and qualitative analysis.
3.2. Risk of bias within studies (Figs. 2, 3)
The details of risk of bias within included studies and authors judgment are listed in Table 2.
Random sequence generation
Four studies of 5 studies had used random number table, random generator or computer programmed random sequencing; and thus, they were rated as low risk of bias. Another study used no reliable randomization method and was evaluated as high risk of bias.
Allocation concealment
One study of included studies had used sealed-envelopes method to the allocation concealment and was evaluated as low risk of bias. Three trials of five included trials did not determine the method of allocation concealment, and were evaluated as unclear risk of bias. One trial had not concealment and was evaluated as high risk of bias.
Blinding of participants and personnel
Due to the consumption of date fruit as intervention and routine care as control, all of the included studies had not performed blinding and were rated as high risk of bias.
Blinding of outcome assessment
All of these studies were assessed as high risk of bias; because they had no evidence of blinding of outcome assessors.
Incomplete outcome data
Out of 5 trials, 2 trials have mentioned the dropped out and analyzed the intention to treat, and 2 trials had no dropped out or lost to follow up; therefore they were rated as low risk of bias. Also, one study had attrition for missing participants; however, the statistical analysis was not followed by the intention to treat.
Selective reporting
One study had registered protocol; but the reported outcomes did not match with registered outcomes and were given the high risk of bias. Other included studies had reported their expected outcomes and were assessed as low risk of bias.
Other bias
Three trials had registered protocol, specified funding source, appropriate ethical criteria, inclusion and exclusion criteria, specified sample size calculating method, and declaration of conflict of interest; therefore, they were rated as low risk of bias. Other trials did not have some of mentioned cases and were rated as high risk of bias.
3.3. Outcomes
We considered common outcome among the included studies in quantitative integration, included active phase of labor duration, first stage of labor duration, second stage of labor duration, and third stage of labor duration as primary outcomes; and bishop score and frequency of caesarian section as secondary outcomes. Also, we performed the sensitivity analysis, because one trial of included trials was quasi-randomized study. Sensitivity analysis did not change the results of primary meta-analysis. The summary of sensitivity analysis is shown in Table 3. The forest plots of sensitivity analyses are available in Appendix 2-6 [See Additional file 2]. Moreover, we performed subgroup analysis for two subgroups (Intervention during labor, and intervention during pregnancy). The results of subgroup analysis showed significant changes in the third stage of labor meta-analysis results. The forest plots of subgroup analyses are available in Appendix 7-10. [See Additional file 3]
Active phase of labor
Three trials with 380 participants were included (190 in intervention group and 190 in control group). There was moderate heterogeneity among the studies (I2= 89%, P = 0.0002). The quantitative synthesis showed that date consumption significantly reduce the duration of active phase of labor compared with control group (MD= -109.3, 95%CI (-196.32, -22.29), P=0.01). (Fig. 4)
First stage of labor
Two studies reported the duration of first stage of labor. Totally, 211 participants were included (103 in intervention group and 108 in control group). Moderate heterogeneity accompanied (I2= 71%, P = 0.06). There was no significant difference between two groups (MD= -76.16, 95%CI (-198.51, 46.18), P=0.22). (Fig. 5)
Second stage of labor
We collected data from 4 trials with 481 participants (238 in intervention group and 243 in control group). The heterogeneity was high (I2= 93%, P <0.00001). There was no significant difference between two groups (MD= -6.41, 95%CI [-22.67, 9.86], P=0.44). (Fig. 6)
Third stage of labor
Three trials with 421 participants (208 in intervention group and 213 in control group) described the duration of the third stage of labor. The intervention cannot decrease the duration of the third stage of labor (MD = 0.39, 95%CI (-2.92, 3.71), P = 0.82). The heterogeneity was high (I2= 89%, P=0.0001). (Fig. 7) The subgroup analysis showed that the intervention during pregnancy significantly reduces the duration of the third stage of labor compared with control group. (Appendix 9) [See Additional file 3]
Bishop score
Two trials reported data on the bishop score with 320 participants (160 in intervention group and 160 in control group). There was no heterogeneity (I2= 0%, P=0.60). The intervention significantly improved the bishop score, compared to control group (MD = 2.45, 95%CI (1.87, 3.04), P<0.00001). (Fig. 8)
Frequency of caesarian section
Three trials had showed the effect of intervention on frequency of caesarian section. 474 participants were included in this outcome meta-analysis (237 in intervention group and 237 in control group). Low heterogeneity was accompanied (I2= 0%, P=0.38). There was no significant difference between two groups. (Risk Ratio= 0.80, 95%CI (0.56, 1.15), P=0.23). (Fig. 9)
Adverse effects
No side effects have been reported in any of the included studies.