This is a small descriptive study including two time periods at four sites only; while it is likely that obstetric practice is similar in other provinces due to the standardised national guidelines and medical regulations, the findings and implications relating to factors influencing practice changes cannot necessarily be generalised to hospitals in other provinces or nationwide. We selected urban centres with a good reputation as these centres are expected to provide optimal performance in the Shanghai region.
A notable finding is the high Caesarean section rate across the four study sites; a pattern that showed little change from 1999–2003. There is an indication that self-decision making for Caesarean section increased over the four years at the study hospitals [30
], and there is limited evidence from elsewhere in China that demand for Caesarean section is associated with women's belief that Caesarean delivery is safe for newborns and is less painful for the woman [31
]. The extent to which maternal requests for caesarean delivery without medical indication are responsible for the increasing rate of caesarean sections is widely debated internationally [32
]; a Cochrane review is planned to inform this debate by comparing the effects of planned vaginal versus planned caesarean delivery [35
We included in the analysis of obstetric procedures only women who delivered vaginally. It is possible that some of the procedures we measured may have influenced whether women delivered vaginally or not; however, the objective of our study was simply to explore routine procedures in vaginal births.
The study demonstrates practice changed in a few obstetric procedures over four years. There appeared to be no pattern in what procedure changed and which hospital it changed in. The direction of the change was, in all but one procedure, towards more evidence-based practice, with uncomfortable procedures such as enemas or rectal examinations being abandoned.
The changes at each hospital were different and the process of change was complex. We initially thought that the feedback to the each hospital after the 1999 survey was responsible for change, but the qualitative data demonstrated that there were multiple factors and the survey appeared not to play a part although it could have done indirectly by influencing the hospital director. Whilst it is to be expected that the hospital director is particularly influential in clinical practice, their influence in these hospitals appears high, and indicates that their endorsement of evidence-based change, and their interpretation of the evidence around each single procedure, is important. The change of hospital directors during the study period at the specialist and county hospitals might have played a particularly important role in the significant practice changes observed; at the specialist hospital, qualitative data suggest that the director's efforts to disseminate international practice standards was influential in changing policy for rectal examinations. It seems that the uptake and enthusiasm for change comes from the directors and other research on 'opinion leaders' has demonstrated this [14
It is interesting that legislation to protect the women has actually led obstetricians to avoid reducing routine episiotomies. The legislation made perineal tear a clinical error and episiotomy rates stayed high; the qualitative data indicated that this was in spite of the evidence that routine episiotomy does not prevent third degree tears [23
]. This change in practice is the equivalent to defensive practice-to avoid litigation rather than in the best interests of the patient.
There appeared to be some emerging evidence from the interviews of a more informed client base questioning practice; on the other hand, obstetric staff, exposed to evidence that does not support routine enemas may actually start offering choice. It is clear from work in South Africa that midwives informed about evidence based standards are frequently keen to abandon an enema, because they make a mess and create work [37
]. It should also be considered that providers are obliged to give women more choice during childbirth, as this is what hospital policy dictates [20
None of the interviews identified the dissemination of evidence-based concepts or information as central to the changes that took place. We cannot therefore easily determine any effect of the regional dissemination activities outlined in table . However, this does not rule out that they are important in sensitising clinicians to the need for change; and the mention in the qualitative data of the use of randomised trials to evaluate practice at the city MCH and district hospitals is an important indication of evidence-informed thinking. Similar findings were reported in the Leeds University maternity audit project, which measured compliance with evidence-based recommendations in 20 UK maternity units. The study found a shift in practice in line with evidence over an eight year period (1988 -1996), with little evidence of any planned dissemination or implementation activities at the maternity units; the authors comment on the difficulty in attributing the change to availability of evidence through the Cochrane Collaboration and national dissemination activities [3