Over the last few years, the episiotomy rate significantly decreased at national and departmental level. In 2014, for non-operative vaginal deliveries, the national rate and the rates for all of the geographic departments except one were below 30%. These results suggest that the recommendations have been seriously taken into account and that proactive changes in practices to restrict the use of episiotomy have been implemented nationwide.
One of the strengths of our study was to include nearly all deliveries, thanks to national discharge abstract data, as almost all deliveries occur in hospitals in France: the difference in the total number of deliveries when compared with the national civil registry, which records all births in France, was only 0.3% [19]. The National Perinatal Survey in 2010 showed a vaginal delivery rate of 79.0% and a non-operative vaginal delivery rate of 66.9% [4]. Our results were 79.4% for all vaginal deliveries and 67.7% for non-operative vaginal deliveries.
In 2012, at the individual level, a validation study was performed to evaluate the metrological quality of hospital discharge abstracts for perinatal indicators. The validity study concerned the same data but only from three university hospitals which agreed to provide a comparison between hospital discharge abstracts and medical records. For vaginal deliveries, the positive predictive value (PPV) was 99.5% [98.5–100] and the sensitivity (Se) was 100%. For episiotomy, irrespective of the vaginal mode of delivery, the PPV was 88.9% [79.7–98.1] and the Se was 90.9% [82.4–99.4]. For perineal tears in vaginal deliveries, the PPV was 94.3% [89.9–98.7] and the Se was 88.6% [82.8–94.4] [18]. In France, no data from validation studies are available for severe perineal tears in cases of non-operative vaginal deliveries. However, in 2010, the rate of severe perineal tears for vaginal deliveries was 0.6% in our study, while the rate was estimated at 0.8% ([0.6–0.9]) in the National Perinatal Survey [4].
Moreover, our population-based study allowed us to examine some groups like breech vaginal deliveries or multiple deliveries with large sub-populations. These national data also allowed us to take into account the effects of the health facility or the variability in hospital medical practices as we included a specific level for hospitals in our multilevel model (level 2 of the model).
The main limitation of this study was related to differences in coding practices. In the validation study in 2012, we examined divergent cases for the discussion. During interviews, we were able to discern that some physicians could seldom report the code for episiotomy in cases of delivery with instrumental assistance, considering that episiotomy is a classical part of this delivery procedure. For this reason, we decided to focus on non-operative vaginal deliveries. All tables and figures were restricted to these cases. The results for all vaginal deliveries are only given for information (Additional file 3), as we know that they may be underestimated. However, when an episiotomy is coded, this episiotomy is generally performed.
Our results are consistent with previous studies regarding factors associated with the use of episiotomy. We retrieved the usual risk factors: primiparous women, multiple pregnancies, breech vaginal deliveries, epidural analgesia, non-reassuring fetal heart rate, newborn weight > 4000 g [20,21,22,23]. In France, episiotomy is not systematically used in breech vaginal deliveries. Indeed, the episiotomy rate decreased from 57% in 1994 to 28.4% in 2009–2010 [24]. Although the restrictive practice of episiotomy has been established by evidence-based medicine, the indications to perform episiotomy are still a matter of debate.
In France, even though, to our knowledge, no national rates have been published for all hospitals, some studies have been performed, and their results seem to agree with ours. A first study in 2007, which concerned vaginal deliveries in university hospitals, estimated a national episiotomy rate of 32.4%. It is not surprising that our estimation for vaginal deliveries in 2007 (26.7%) was lower as we considered all hospital types [25]. In 2009, another study in one hospital estimated the episiotomy rate in vaginal deliveries at 7.6%. This figure is slightly below our estimation (11.3%, all over the corresponding department) but the 2009 study included only one hospital and only single pregnancies and cephalic presentations [26]. Another study based on Burgundy Perinatal Network data showed similar results to ours in the four departments included in this region [27]. A study conducted in the south of France reported a decrease in the rate of episiotomy (from 35.8% in 2003–2005 to 16.7% in 2012–2014) [20]. All these studies highlight a high disparity in episiotomy rates not only between departments but also between hospitals. In our study, a decrease in episiotomy rates was shown for the vast majority of French departments from 2007 to 2014. Over the same period, severe perineal tears (third and fourth degrees) significantly increased in women who had non-operative vaginal deliveries. These results were consistent with those of the Euro-peristat project, which described an increase in the rate of severe perineal tears for all vaginal deliveries between 2004 and 2010 in all European countries, except Germany and Norway [28]. This issue is still the subject of a controversial debate [7, 9, 20, 29,30,31,32]. Randomized trials showed no increase in severe perineal tears related to the restrictive use of episiotomy [3]. On the contrary our results showed that the restrictive practice of episiotomy was associated with a greater risk of severe perineal tears. However, the rate in France is lower than the mean rate found in EURO-PERISTAT, which suggests that the rate is under-estimated in France. Given that the rate found in our study is in keeping with that found in the national perinatal survey, it is likely that this under-estimation does not stem from coding problems in hospital data. Another explanation could be non-diagnosis [33], which nonetheless seems to be diminishing over time. In fact, we observed an increase in severe perineal tears in France which may be related to improvements in vigilance and the training of professionals in the diagnosis and suturing of these severe perineal tears which are responsible for urinary and fecal incontinence [29].
We can notice that, for non-operative vaginal deliveries, the rate decreased markedly in all departments (decrease of 25 to 75%), even when initial rates were about 30%. In 2014, 14 of the 97 departments presented an episiotomy rate below 10% for non-operative vaginal deliveries. The rate of severe perineal tears was unavailable for two of these departments, between 0.15 and 0.58% for ten departments and above 1% for two departments.
It seems difficult to define what could be the right episiotomy rate in France. The WHO recommended a target of 10% for episiotomy. This recommendation cannot be generalized as it was based on a case-controlled study that included only non-induced labor for a single pregnancy at over 37 weeks of amenorrhea. Moreover, it took into account the high infection rate in developing countries, which is not the case in France. The restrictive practice of episiotomy must provide an episiotomy rate that is optimal for children’s and mothers’ health and ensure low rates of severe perineal tears, which are very harmful for women. Our results suggest that a rate below 15% for non-operative vaginal deliveries was obtained in 57% of French departments with a rate of severe perineal tears not more than 1%. As a consequence, one could hypothesize than a rate of 15% could be reached by most departments in a reasonable time. Further research is of course needed to confirm this hypothesis.
This target rate can be considered achievable for all French departments, though a national program is necessary. A passive approach after the publication of guidelines is not enough and the implementation of evidence-based practices remains a real challenge. Previous publications have shown that the impact of guidelines is greater if they are worked on with the teams concerned, particularly in obstetrics [34]. At the national level, a community of practices could promote the dissemination of experience, and thus decrease the episiotomy rate without increasing severe perineal tears.
Our study suggests that the action plan should now be looking at the individual level. Some authors have described how a private and confidential feedback from physicians about their own practices can induce a decrease in the use of episiotomy [35, 36]. Ambassadors with communication and training skills may effectively facilitate changes in their teams [37]. In the same way, audits, risk management approaches, continuous care quality improvement programs and the understanding of professionals’ behavior with regard to perineum protection should lead to the standardization of good practices for selective episiotomy.
The continuous training of physicians and midwives is an important lever to improve quality of care. Perinatal networks, which aim to inform, train and motivate practitioners, gain their support in implementing a restrictive practice policy for episiotomy. Perinatal networks have a role to play in standardizing the restrictive use of episiotomy in the areas they cover.