Ricerc@Sapienza

The first point highlighted by Tang et al. is about our search strategy. Particularly, the authors demonstrate concerns on searching the literature only through PubMed. We totally agree with this concern, but we have to point out that this is not the case of our systematic review. We performed our systematic research in “Pubmed”, “PsycINFO” and “MEDLINE” databases. Cochrane Collaboration Guidelines for searching literature [1] were followed. Consistently, several search strategies were used in addition to the standard literature databases. We identified further studies from the references of the screened full-texts and we collected data from non-published studies (searching in the grey literature, non published literature and contacting authors in the field to obtain further data). This extensive search strategy was adopted by our group in several previous publications [3-5]. Indeed, we did not search relevant literature in Embase database. Nevertheless, evidences are available that indicate that low risks of biases are associated by searching literature in Medline but not Embase [2].

Furthermore, our keywords were chosen accurately and they were comprehensive of synonymous and “*” as to include all existing suffixes: (((pregnan*[Title/Abstract] OR “expect* mother*” [Title/Abstract] OR “mother*”[Title/Abstract])) AND ((insomnia*[Title/Abstract] OR sleep*[Title/Abstract])) AND ((treat* [Title/Abstract] OR therap*[Title/Abstract] OR intervention [Title/ Abstract]))). Indeed, these terms were not searched as “Keywords”, but as indicated as any word in the title or the abstract. For this reason the provided keywords search is actually much more extensive than the relatively broad categorization by MeSH.
Finally, we include articles in the following languages: English, Italian, German, Spanish, and French. Indeed, we did not search relevant literature in Chinese databases as unfortunately we did not have the possibility to read in Chinese. Nevertheless, our systematic review was not focused on Chinese medicine treatments.
A further comment by Tang et al. was on the inappropriateness of our interventions’ classification. As stated by the authors “Complementary medicine is not standard therapy, which, instead, is utilized along with standard medical therapy. Alternative medicine is considered as the replacement of standard medical therapy”. In our systematic review, only one study was found that evaluated the efficacy of Acupuncture on insomnia [6] and the authors used acupuncture in an alternative way (instead of conventional medical treatment) and not in a complementary way (in combination with conventional medical treatment) [7]. For this reason, this study was categorized in the alternative interventions section.
Tang and colleagues also commented on the inappropriateness of our choice to evaluate with Risk of Bias Cochrane Collaboration’s tool [8] only for Randomized Controlled Trial and they stated that we have used the Newcastle-Ottawa Scale (NOS) for non-randomized clinical trial studies’ evaluation. Actually, we did not use the Newcastle-Ottawa Scale (NOS) for non-randomized studies, but we decided to provide a qualitative critical evaluation of the level of evidence of non-randomized controlled trials. That is, the risk of bias of the interpretation of the results of non-randomized trials included in our systematic review was qualitatively assessed and considered.
The authors comment on the necessity to evaluate in our systematic review the safety of the interventions. This was actually done whenever information about safety was reported (e.g. for pharmacotherapy and acupuncture).
Finally, Tang and colleagues commented on the similarity of two of the included studies in our systematic review [9,10]. These studies were conducted by the same research group. In one of this study [11], hospitalized pregnant women were evaluated, whether in the other study, non-hospit