Studies were selected for inclusion based on the population, intervention, comparison, outcome measures and study design.

Population: Women in labour with a singleton, non-anomalous, live birth. Three of the included studies described rotation of OP positions, while four described rotation of both OP and OT positions.

Intervention: Manual rotation after 7 cm dilatation.

Comparison: Four of the included studies compared MR with expectant management and three compared successful MR with unsuccessful MR.

Outcome measures: Vacuum, forceps or caesarean delivery.

Study design: Primary studies.

Exclusion criteria were MR or DR used on other cephalic presentations and unpublished literature.

The systematic search was conducted in the Medline, Ovid Medline, SveMed+, Cinahl, Cochrane Library, Maternity and infant care and Up To Date databases for the period 1946 to January 2015 with guidance from a specialist librarian employed at the Medical Library of Oslo University hospital, Ullevål. The following keywords were employed: manual rotation and digital rotation, labour presentation, obstetric labour complications, occiput posterior and version foetal. The search and findings were reported according to the method described by Aveyard [24] . An example of the search in one of the main databases, Ovid Medline is presented in Table 1.

A total of 330 papers were retrieved from the database search, of which seven were included in the thematic analysis (Figure 1).

Seven studies were included in the literature review, none of which was a randomized controlled trial (RCT). There were two Cohort studies [14] [27] one Case-control study [13] and one Pilot study [28] . The rest were a mix of retrospective and prospective studies [29] [30] [31] .

The article by Graham et al. [28] is a pilot study. Although it is of medium methodological quality, the results are not be of great importance for this paper due to the small population, that makes the study underpowered to test the efficiency of prophylactic DR. Nevertheless, it is a valuable study because it shows that a large double-blinded multi-centre RCT is feasible. Although the study by Sen et al. [31] was not excluded despite its low methodological quality, papers by Le Ray et al. [13] , Le Ray et al. [29] and Shafferet al. [14] were given the highest methodological were prioritised in the thematic analysis. The studies by Shaffer et al. [27] and Reichman et al. [30] were of medium quality.

Six themes emerged: How can MR reduce VD, FD and CD?; Mode of delivery after MR versus expectant management; Mode of delivery after successful or unsuccessful MR; Technique and success rate; Midwives’ involvement in the MR; and Predictors of successful or unsuccessful MR.

All of the included studies suggest that MR is a procedure to consider for reducing VD, FD and CD rates.

Having studied the conclusions of each study, there is a consensus that MR is a procedure worth considering as it can contribute to reducing VD, FV and CD.

The authors based their reasons for arguing that MR can reduce VD, FD and CD on two different findings. Reichman et al. [30] , Shaffer et al. [14] , Le Ray et al. [29] and Graham et al. [28] were influenced by the findings related to mode of delivery after MR versus expectant management.

Shaffer et al. [27] , Le Ray et al. [13] and Sen et al. [31] referred to the findings related to mode of delivery after successful versus unsuccessful MR. The results will be presented separately as: mode of delivery after MR versus expectant management (Table 4) and mode of delivery after successful versus unsuccessful MR (Table 5).

Three of the included studies described rotation of OP positions [28] [30] [31] , Four described rotation of both OP and OT positions [13] [14] [27] [29] . Only Le Ray et al. [29] referred to the success rate when comparing rotation from the two positions, although they did not observe any difference in the success of rotations from OT and OP positions.

Reichman et al. [30] revealed that none of the women who underwent MR had a CD (Table 4). Shaffer et al. [14] reported a CD reduction rate of 32.8 % after MR. Although Le Ray et al. [29] only found a CD reduction rate of 1.7% after MR, there was a 15.4% reduction in the operative delivery rate. Both Reichman et al. [30] and Le Ray et al. [29] found an increase in instrumental delivery rates after MR of 27.4% and 13.8%. Shaffer et al. [14] reported almost identical instrumental delivery rates after MR and expectant management. Reichman et al. [30] and Le Ray et al. [29] found an increase in spontaneous vaginal delivery of 50.1% and 15.5% after MR. The study by Graham et al. [28] lacks the power to provide answers on rates.

The studies on MR versus expectant management differed in many ways, but an important similarity between Reichman et al. [30] , Le Ray et al. [29] and Graham et al. [28] was that the diagnosis of the OP or OT position was made in the second stage of labour.

The expectant management group in the study by Shaffer et al. [14] reported POP positions diagnosed at delivery, which can explain the very high CD rate that might be due to failure to take account of OP position that would have

Caesarean Delivery = CD, Instrumental Delivery = ID, Operative Delivery = OP, Forceps Delivery = FD, Spontaneous Vaginal Delivery = SVD, Vaginal Delivery = VD, Undefined Vaginal Delivery =UVD, NI = No Information.

spontaneously rotated had they been diagnosed in the second stage of labour. OP positions are associated with a large proportion of emergency CD [18] . Of all emergency CDs, 18% are due to dystocia related to OP presentation [21] .

Shaffer et al. [14] found almost equal instrumental rates after MR and expectant management. The high rate in the expectant management group was no surprise, as all the deliveries involved POP positions, a large proportion of which require a VD, FD or CD [3] . However, the high instrumental delivery rate in the MR group was not anticipated [14] and could be related to the relatively low success rate of 74%. Furthermore, there is no way of knowing whether the instrumental delivery took place after a successful or unsuccessful rotation. For this reason, there might have been many unsuccessful rotations that led to the need for an instrumental delivery. A high rate in successful MR could be related to the fact that MR corrects the foetal head in the birth canal, making VD or FD easier to implement.

The main results show an increase of spontaneous vaginal delivery after MR. While the CD and instrumental rates vary, there seems to be an overall decrease in the operative delivery rate after MR (Table 4).

Le Ray et al. [13] and Shaffer et al. [27] found that women with successful MR had a significantly lower CD rate (p < 0.001) (Table 5). In the study by Sen et al. [31] none of the successful MRs needed a CD.

The instrumental rate reported by Le Ray et al. [13] was higher in the successful MR group, but the difference was not significant.

Le Ray et al. [27] also demonstrated that women with a successful MR had a 50.9% increase in the spontaneous vaginal delivery rate. Shaffer et al. [27] show- ed that women who had a successful MR had nearly a three-fourths increase in undefined vaginal delivery.

Sen et al. [31] did not analyse MR as an individual procedure, but 17 cases of MR and 13 Posture changes. Although the small population led to a lack of power,

Caesarean Delivery = CD, Instrumental Delivery = ID, Undefined Vaginal Delivery = UVD, NI = No Information.

the conclusion is interesting. Only eight of the 17 cases had a successful MR, but none of the women with a successful MR needed a CD.

The instrumental delivery rate in Le Ray et al. [13] was higher in the successful MR group, but the difference was not significant, for which no explanation was given. However, it could be due to the fact that MR corrects the foetal head in the birth canal, making VD or FD easier to implement.

It is important to investigate whether the MR was successful or unsuccessful in order to understand the impact of rotation on mode of delivery. The results from the included studies on successful versus unsuccessful MR show a significant reduction in CD and an increase of spontaneous vaginal delivery after successful MR (Table 5).

The success rate varied from 47% to 93.3% (Table 6). Two main techniques, MR and DR, were used in the studies. Some of the studies used MR or DR as the main term for the rotation procedure, even though the other technique was also performed.

Having studied the results it appears that the choice of technique does not have a major influence on the success rate.

In all of the included studies midwives supervised the labour and attended the physician performing MR. Midwives who experienced in MR also performed the procedure in all studies, with the exception of Sen et al. [31] and Graham et al. [28] .

MR has the best chance of success when performed at full dilatation [13] after engagement of the head [13] [31] and when done prophylactically before failure to progress [13] (Table 7).

All of the included studies suggest MR as a procedure to consider for reducing VD, FD and CD rates.

The main aim of midwives and obstetricians in maternity care is to protect the health of mother and child. The two professions are based on two fundamentally different models of care [32] . Midwives are responsible for normal labours, due to their expertise in protecting, supporting and enhancing the normal physiology of labour [32] . Physicians have a background in medicine and are responsible for the pathologic aspects of labour, with often involve diagnosing and treating complications [32] . In reality, despite the fact that these two basic differences are non-debatable, the two professions share information, expert opinions and research, working together to holistically safeguard both the physiological and the pathological labour.

Labours with OP and OT positions often start normally but pathology can develop during the process. These malpositions are diversities of foetal rotation and while not pathological, involve a risk of pathology. Obstetricians and midwives work together to protect the health of mother and child, but how to manage an individual labour and decide when to intervene can be a subject of debate, especially between midwives and obstetricians.

It is relevant to argue that most OP positions rotate spontaneously and for that reason expectant management is the best approach. Nevertheless, OP and OT positions are associated with a greater risk of prolonged second stage labour and maternal morbidity [1] . As many as 5.5 to 8.3% of OP positions are persistent [5] [16] of which 68% require an OD [3] . Many OP positions rotate during the first stage of labour [15] .

Le Ray et al. [13] and Sen et al. [31] state that a predictor of successful MR is performing the procedure at full dilatation. This indicates that MR should be performed in stage two of labour, at which point many of the OP presentations that spontaneously rotate have already done so. Le Ray et al. [13] found that MR has a better chance of success when performed prophylactically before failure to progress. In expectant management, the indication for MR is likely to be failure to progress, which reduces the chance of successful MR.

Because midwives are responsible for monitoring normal labours and most labours with OP and OT positions start normally, it is their duty to diagnose the OP or OT position. Many midwives believe that expectant management is the best way to manage labour because they understand it as a normal physiological process and therefore try to avoid unnecessary interventions. However, for expectant management to support natural physiological labour it is necessary to monitor the rotation and presentation of the foetus when assessing progress. The results from the included studies show a reduction in the OD rate and an increase of SVD after MR versus expectant management (Table 4). For this reason MR can contribute to protecting, supporting and enhancing the normal physiology of labour by facilitating SVD.

Midwives’ and physicians’ different models of care [32] can also have consequences for the mode of delivery, in that a midwife performs MR to support and enhance the normal physiology of labour. In other words, the midwife performs MR and then waits to see if the labour proceeds normally, as this is the basic philosophy of the midwifery profession. On the other hand, obstetricians might find it easier to convert the MR procedure into an ID due to their underling expectation to treat and the fact that performing obstetric interventions is their area of expertise. This can explain the high level of ID after successful rotations in Le Ray et al. [13] .

The included studies demonstrate the great importance of successful MR for reducing VD, FD and CD (Table 6). For this reason, we investigated the predictors of successful rotation, only to find that none of the studies highlighted the actual MR technique as a predictor of success. As MR is a handcraft with wide interpretations, it was interesting to explore the different techniques used and the success rate in each study (Table 5). However, the choice of technique did not seem to influence the success rate and instead experience appears to be the decisive factor. In most studies experienced personnel performed MR or supervised less experienced colleagues. This is a traditional and valuable way of working as it provides inexperienced personnel with hands on experience, an opportunity to ask about the procedure and receive advice [33] . However, gaining experience requires midwives and physicians to perform MR. Two large surveys were conducted in Australia and New Zealand to increase understanding of midwives’ and obstetricians’ attitudes towards MR [34] [35] . It was found that while only a minority of obstetricians and midwives perform MR, most consider the procedure acceptable. Midwives and Obstetricians would be willing to perform MR if it was known to decrease OD from 68% to a median of 50%. From my own experience as a midwife, I believe that this survey is relevant and also reflects Norwegian Midwives’ and Obstetricians’ attitudes towards MR. Experience is of great importance for achieving the best possible results in most handcrafts [7] . Le Ray et al. [13] states that the low failure rate in the study is probably related to the efficacy of this technique when used routinely.

OP and OT positions are associated with a prolonged first and second stage of labour [1] [5] . Admission to the maternity ward can be a long and painful experience for the woman and a challenge for the midwife. In addition to emotional support, the midwife provides pain relief and information, as well as monitoring the progress of both mother and foetus. Recent research confirms that continuous care in labour contributes to increased SVD, decreases the need for analgesia, leads to less dissatisfaction and can reduce the need for CD and instrumental delivery [23] . There are many ways of supporting women in labour. The best support is dependent on what becomes a challenge in each individual labour and the woman’s motivation and resources, both physical and psychological. The study on continuous support for women during childbirth [23] does not define the concept. Other than suggesting that it can be emotional support, comforting measures, information and advocacy, it was also stated that the best support was given by someone other than hospital staff. This indicates that the midwife’s special competence regarding the labour process may not be the only support needed to facilitate spontaneous vaginal delivery. However, because of the risk factors involved, when faced with OP or OT positions, the midwife’s competence is important for providing adequate comfort measures, information and advocacy to prevent maternal complications or to support the woman when the obstetrician has to intervene. Few other professions in the obstetric field work as close to the individual women in labour, which gives midwives a huge advantage in terms of their responsibility to monitor the birth process and therefore also the rotation of the foetus during labour. A diagnosis of OP position made early in labour is important because the MR rotation may be easier to perform before the labour fails to progress [13] . The midwife attending the labour can perform the MRor in the event of lacking the necessary skill, can ensure that the situation is evaluated at the right time by a midwife or an obstetrician with competence in performing MR.

Midwives supervised labours and attended doctors performing MR in all of the included studies. They also performed MR in all studies with the exception of Sen et al. [31] and Graham et al. [28] .

This literature review has several limitations. A meta-analysis was not possible due to the diversity of the aims, design and population in the included studies [24] . None of the studies were RCTs and of the seven included studies, only three had a high methodological quality. When conducting a review, the researcher is in danger of making choices regarding the selection of studies and data extraction that could lead to the risk of bias [36] .