Interpretation
Systematic reviews are considered the highest level of evidence regarding the efficacy, effectiveness and safety of prophylactic or therapeutic interventions and play a significant role in guiding health policy, best practices, and evidence-based patient care. However, our quality assessment of the articles reviewed herein found that many did not utilize rigorous methodological approachs when conducting the studies. Thus the introduction of different biases that can confound conclusions and hinder policymaking, ultimately affecting patient-care cannot be excluded.
A moderate level of heterogeneity in the quality assessment of published systematic reviews involving the efficacy, effectiveness and safety of methods for the induction of labour was identified in our assessment with AMSTAR scores ranging from 3 to 9. The number of systematic reviews on this topic has increased over the past decade, with the most (n=11) being published in 2019. However, no significant improvement in methodological quality was found. This observation diverges from other areas in medicine, like critical care and radiology, in which an improvement in methodological quality of systematic reviews has been demonstrated over time56, 57. Thus, results of our AMSTAR assessment of the published systematic reviews revealed a need for greater scientific rigour in the study of the induction of labour.
Systematic reviews funded by hospitals, institutions, philanthropists, government grants or with unreported funding had no difference in methodological quality according to the AMSTAR Checklist (p=0.34). Our findings report that only 36 (75%) of the 48 studies evaluated in our assessment included a conflict of interest statement in both the systematic review and the studies included therein and thus criterion 11 of the AMSTAR Checklist was not satisfied. We suggest this particular weakness in the literature is concerning as there is a need for transparent reporting of conflicts of interest to allow for the judgement of any external influences on study conclusions. The effect of this potential bias is increased in studies that do not have an established a priori protocol where redefined outcome measures and post-hoc analysis could further exacerbate a biased perspective of the evidence. Our findings indicated that 40 of 48 articles (83%) used an a priori study design (AMSTAR criterion #1).
No significant correlation was observed between AMSTAR score and the journal impact factor, suggesting that leading journals may not necessarily evaluate methodological quality more rigorously than others. However, articles from the Cochrane Database of Systematic Reviews on average, score higher on the AMSTAR Checklist than articles from other peer-reviewed journals (p=0.01). This finding supports the generally accepted position that the methods followed by the Cochrane Collaboration sets the highest standard for conducting and publishing systematic reviews. Consistent findings regarding the quality of Cochrane reviews has also been reported for other medical disciplines as well58.
In the present assessment, there was no significant correlation between AMSTAR score and total citations. We sugges that this finding may indicate that authors who utilize systematic reviews to support their findings or hypotheses do not critically evaluate the quality of the studies cited. This is of particular importance if these systematic reviews are being used to generate or refute medical hypotheses.
The highest AMSTAR score of 9/11 was recorded for several Cochrane Collaboration publications (n=7). Three examined the use of oxytocin9,11,14, three others examined mechanical methods of labour induction17,21,47, and one examined the use of mifepristone26. An AMSTAR point for a comprehensive literature search could not be provided for some of these reviews9,11,17,21,26. However, some17,26,47 included grey literature such as evidence from the Cochrane Pregnancy and Childbirth Group’s Trials Register and clinicaltrials.gov which captures data that has not been peer reviewed. All seven high ranking articles also conducted quality assessments of the included studies, which satisfy AMSTAR criteria that we believe to be of greater importance. Important conclusions from these seven studies that scored the highest on the AMSTAR criteria include that a balloon catheter may be less effective than oral misoprostol but have a greater safety profile17. The use of a balloon catheter is likely as effective as the use of intravaginal prostaglandin E2 when inducing labour17. Oxytocin was suggested to be less effective in achieving vaginal birth within 24 hours compared to prostaglandin agents9. High-dose oxytocin (100mU in the first 40 minutes, rising above 600 mU in the first two hours) compared with low-dose oxytocin (below 100mU in the first 40 minutes, rising to below 600 mU in the first two hours) has been shown to increase the risk of uterine hyperstimulation without increasing the rate of vaginal delivery within 24 hours14. However, discontinuing IV oxytocin stimulation after the active phase of labour may reduce the risk of a caesarean delivery11. Another study26concluded that the literature is inadequate to draw conclusions that mifepristone helps induce labour. It is also suggested that the use of membrane sweeping does not provide clinical benefit and that this labour induction method should be balanced against women’s discomfort21. Finally, one study concluded that there was no clear benefit from using acupuncture or acupressure in reducing the rate of caesarean section47.