Unexpected reduction in the incidence of birth trauma and birth asphyxia related to instrumental deliveries during the study period: was this the Hawthorne effect?

Authors


* Dr W. C. Leung, Department of Obstetrics and Gynaecology, Queen Mary Hospital, 102, Pokfulam Road, Hong Kong, China.

Abstract

Objective The study was originally designed to identify the risk factors that could predict those difficult instrumental deliveries resulting in birth trauma and birth asphyxia.

Design A prospective study on all singleton deliveries in cephalic presentation with an attempt of instrumental delivery over a 12-month period (13 March 2000 to 12 March 2001).

Setting A local teaching hospital.

Sample Six hundred and seventy deliveries.

Methods A codesheet was designed to record the demographic data, characteristics of first and second stages of labour and neonatal outcome. In particular, the doctor had to enter the pelvic examination findings before the attempt of instrumental delivery.

Main outcome measures Birth trauma and birth asphyxia.

Results There was a significant reduction in the incidence of birth trauma and birth asphyxia related to instrumental deliveries during the study period (0.6%) when compared with that (2.8%) in the pre-study period (1998 and 1999) (RR 0.27, 95% CI 0.11–0.70). There was more trial of instrumental deliveries in the operating theatre although this was not statistically significant (RR 1.19, 95% CI 0.88–1.60). The instrumental delivery rate decreased during the study period (RR 0.88, 95% CI 0.82–0.94). The caesarean section rate for no progress of labour, the incidence of direct second stage caesarean section and the incidence of failed instrumental delivery did not increase during the study period.

Conclusions Apart from the merits of regular audit exercise and increasing experience of the staff, the Hawthorne effect might be the major contributing factor in the reduction of birth trauma and birth asphyxia related to instrumental deliveries during the study period.

Introduction

The true success of a trial of instrumental delivery is the outcome, by whatever route, of the healthiest child and mother. The true failure is not when vaginal delivery is not achieved but when avoidable injury is inflicted1. Our hospital is a local teaching hospital with about 4000 deliveries per year. There is a report system that every single case of birth trauma (except cephalohaematoma and clavicular fractures) and severe birth asphyxia would be reviewed by a Birth Trauma Panel. The Panel consists of two consultants, one associate professor and the chief nursing officer of the labour ward. Birth trauma is ascertained by the paediatrician who has attended the delivery or examined the baby afterwards. Severe birth asphyxia refers to those cases with an Apgar score on the first minute of ≤3 requiring admission to neonatal intensive care unit. There were altogether 30 cases of birth trauma and 13 cases of birth asphyxia related to instrumental deliveries in 1998 and 1999 in our hospital (Table 1). The 13 cases of birth asphyxia referred to those with the Birth Trauma Panel's conclusion that the instrumental delivery was contributing to the asphyxia. The 30 cases of birth trauma included: Erb's palsy (10 cases); fractured skull (one case); fractured skull with Erb's palsy (one case); fractured skull with facial nerve palsy (one case); fractured skull with subaponeurotic haemorrhage (three cases); fractured skull with subaponeurotic and subdural haemorrhage (one case); fractured skull with subaponeurotic and subarachnoid haemorrhage (one case); subaponeurotic haemorrhage (six cases); subaponeurotic and subarachnoid haemorrhage with facial nerve palsy (one case); subarachnoid haemorrhage (one case); subarachnoid haemorrhage with Erb's palsy (one case); facial nerve palsy (one case); facial laceration (one case); and laceration of eyebrow (one case).

Table 1.  The hospital obstetric statistics in the pre-study period (1998 and 1999) and the study period (13 March 2000 to 12 March 2001).
 Pre-study period (24 months)Study period (12 months)RR [95% CI]
  1. * The denominator is the total number of attempts of instrumental deliveries. Values in parentheses are given as (%).

  2. ** The denominator is the total number of deliveries. Values in parentheses are given as (%).

  3. *** The denominator is the total number of assisted deliveries in the second stage of labour, including direct second stage caesarean sections. Values in parentheses are given as (%).

Birth trauma and birth asphyxia related to instrumental deliveries*43/1510 (2.8)4/670 (0.6)0.27 [0.11–0.70]
Instrumental delivery rate**1494/7671 (19.5)663/3996 (16.6)0.88 [0.82–0.94]
Overall caesarean section rate**1348/7671 (17.6)795/3996 (19.9)1.10 [1.04–1.17]
Caesarean section rate for no progress of labour**268/7671 (3.5)132/3996 (3.3)0.96 [0.83–1.11]
Failed instrumental delivery*16/1510 (1.1)7/670 (1.0)0.99 [0.53–1.84]
Trial of instrumental delivery in operating theatre*51/1510 (3.4)29/670 (4.3)1.19 [0.88–1.60]
Direct second stage caesarean section***11/1521 (0.7)3/673 (0.4)0.70 [0.26–1.91]

Our study was originally designed to identify the risk factors that could predict those difficult instrumental deliveries resulting in birth trauma and birth asphyxia. The ultimate goal is to reduce the incidence of birth trauma and birth asphyxia related to instrumental deliveries. The rationale is that if these difficult instrumental deliveries can be predicted, one can then perform the instrumental delivery in the operating theatre with full preparation for caesarean section, or even to go ahead for a second stage caesarean section directly.

Methods

A prospective study was performed on all singleton deliveries in cephalic presentation with an attempt of instrumental deliveries in a local teaching hospital over a 12-month period from 13 March 2000 to 12 March 2001. There are about 4000 deliveries each year. About 20% of them (800) are instrumental deliveries. The ratio of vacuum extraction to low forceps deliveries is 3:1. Majority of the instrumental deliveries are performed by the residents who are undergoing the specialist training. The rest of them are performed by specialists covering the labour ward.

A codesheet (this form will be available on request from the authors) was designed to record the demographic data, characteristics of first and second stages of labour and neonatal outcome. In particular, the doctor had to enter the pelvic examination findings before the attempt of instrumental delivery. This is very important because retrospective entry of pelvic examination findings is often biased by the outcome of the instrumental delivery. In order not to cause any delay in the instrumental delivery, particularly in case of fetal distress, the doctor did not have to write down the pelvic examination findings himself or herself. For every instrumental delivery, a midwife in the labour ward would come up and ask the one who performed the delivery about the pelvic examination findings (and record them on the codesheet) before the actual attempt of the instrumental delivery.

The research protocol was approved by the Hospital Ethics Committee.

Results

During the 12-month study period, there were only two cases of birth trauma, subaponeurotic haemorrhage (one case) and subarachnoid haemorrhage (one case), and two cases of birth asphyxia with the Birth Trauma Panel's conclusion that the instrumental delivery was contributing to the asphyxia. It was a significant reduction when compared with that in the pre-study period (1998 and 1999) (RR 0.27, 95% CI 0.11–0.70) (Table 1). As there were only four cases of birth trauma and birth asphyxia in total during the study period, it would not be worthwhile or appropriate to analyse the risk factors for these four difficult instrumental deliveries. On the other hand, it would be interesting to investigate why there was a significant reduction in birth trauma and birth asphyxia related to instrumental deliveries during the study period.

There was no change in the obstetric practice, staff structure and coverage system of the labour ward as well as the report system to the Birth Trauma Panel during the study period when compared with that in the pre-study period. Table 1 compared the instrumental delivery rate, overall caesarean section rate and that for no progress of labour, incidence of failed instrumental delivery, trial of instrumental delivery in operating theatre and direct second stage caesarean section during the study period with those in the pre-study period. Relative risk with 95% confidence interval was used to compare the various incidences using the Instat software package. There was a significant decrease in instrumental delivery rate during the study period (RR 0.88, 95% CI 0.82–0.94). The overall caesarean section rate increased from 17.6% in the pre-study period to 19.9% during the study period (RR 1.10, 95% CI 1.04–1.17) but the caesarean section rate for no progress of labour (including cephalopelvic disproportion, persistent OT/OP position and uterine dysfunction) remained the same. There was no significant difference in the incidence of failed instrumental delivery and direct second stage caesarean section between the study period and those in the pre-study period. The incidence of trial of instrumental delivery in the operating theatre increased during the study period when compared with that in the pre-study period, although this did not reach statistical significance (RR 1.19, 95% CI 0.88–1.60).

Discussion

The true success of a trial of instrumental delivery is the outcome, by whatever route, of the healthiest child and mother. The true failure is not when vaginal delivery is not achieved but when avoidable injury is inflicted1. Vacuum delivery is associated with greater neonatal morbidity and mortality than was previously recognised, the adage that the vacuum is designed to come off before infant damage occurs appears unsubstantiated2. Out of the 43 cases of birth trauma and birth asphyxia related to instrumental deliveries in the pre-study period, 38 of them (88%) were by vacuum extraction. The majority of the instrumental deliveries resulting in birth trauma and severe birth asphyxia occurred in the labour ward (36/43 = 84%) when the difficulty was not anticipated. If these difficult instrumental deliveries could be anticipated, there was the possibility that earlier resort to caesarean section or trial of instrumental delivery in the operating theatre with full preparation to proceed to caesarean section might reduce the incidence of birth trauma and birth asphyxia. The neonatal and maternal morbidity from carefully conducted trial of instrumental delivery in the operating theatre had been shown to be comparable to direct caesarean section for prolonged second stage1,3. On the other hand, birth trauma and severe birth asphyxia could still occur even when difficulty was anticipated and trial of instrumental delivery was already performed in the operating theatre (7/43 = 16%). One possible explanation was that if the obstetrician was not willing to abandon the trial of instrumental delivery appropriately, the advantage of being able to predict these difficult deliveries and to conduct the trial in the operating theatre would be lost.

Contrary to our belief that these cases of birth trauma and birth asphyxia are avoidable, Towner et al.4 showed that the common risk factor for neonatal intracranial haemorrhage is abnormal labour rather than the mode of delivery. The significant reduction in the incidence of birth trauma and birth asphyxia related to instrumental deliveries during our study period did prove that they are avoidable. Although there was only 1 out of the 43 cases of birth trauma and birth asphyxia related to instrumental deliveries in the pre-study period which had a long term sequel of cerebral palsy, the short term sufferings to the neonates and their parents in the other 42 cases could not be ignored.

The interesting question is what contributes to the significant reduction in the incidence of birth trauma and birth asphyxia related to instrumental deliveries during our study period. There was no change in the obstetric practice, staff structure and coverage system of the labour ward as well as the report system to the Birth Trauma Panel during the study period when compared with that in the pre-study period. The 43 cases of difficult instrumental deliveries in the pre-study period were not limited to those performed by a few individual staff. Almost every specialist trainee and specialist who worked in the labour ward had their own share, with their number of cases involved roughly proportional to the total number of instrumental deliveries they had performed. We do not think we can ignore the increasing experience of the staff over the time as well as the merits of the Birth Trauma Panel, which conducts regular audit exercises on all the cases of birth trauma and severe birth asphyxia. But all these are ongoing processes and we could not explain why the reduction in the incidence of birth trauma and birth asphyxia related to instrumental deliveries only occurred when our study began.

We hypothesise the Hawthorne effect5 to be a major contributing factor in the reduction of birth trauma and birth asphyxia related to instrumental deliveries during the study period. The Hawthorne effect refers to the tendency to improve performance because of the awareness of being studied. Furthermore, the design of our study is that the one to perform the instrumental delivery has to enter the pelvic examination findings before the actual attempt. Some of these pelvic examination findings, in particular, the degree of flexion of fetal head, the presence of asynclitism and the descent of fetal head with maternal pushing were poorly documented when we reviewed the records of the 43 cases of difficult instrumental deliveries in the pre-study period. Nevertheless, these pelvic examination findings before the actual attempt of instrumental delivery might help to identify those potentially difficult instrumental deliveries. This was reflected by the increase in the incidence of trial of instrumental delivery in the operating theatre during the study period although this did not reach statistical significance (RR 1.19, 95% CI 0.88–1.60). Interestingly, when we presented our study protocol in the hospital research meeting before the study began, one of the colleagues did raise the question that we might fail to identify the risk factors for these difficult instrumental deliveries (the original objective of our study) because of the Hawthorne effect. We did not think that matters because the ultimate goal is to reduce the incidence of birth trauma and birth asphyxia related to instrumental deliveries, whatever the means to achieve this goal.

Did we perform more caesarean sections either before or at the second stage of labour in order to avoid a difficult instrumental delivery? Were we more often to abandon an instrumental delivery and turn to a caesarean section in order to avoid the potential birth trauma and birth asphyxia related to a difficult instrumental delivery? The hospital obstetric statistics told us that this was not the case (Table 1). Although the overall caesarean section rate increased from 17.6% in the pre-study period to 19.9% during the study period but the caesarean section rate for no progress of labour remained the same. There was also no significant difference in the incidence of direct second stage caesarean section and failed instrumental delivery between the study period and those in the pre-study period. In fact the incidence of instrumental deliveries even decreased in the study period (RR 0.88, 95% CI 0.82–0.94). We postulate that the more careful and detailed assessment in the second stage of labour (part of the Hawthorne effect) had converted some of the unnecessary instrumental deliveries into normal vaginal deliveries.

In conclusion, we hypothesise the Hawthorne effect to be the major contributing factor in the reduction of the incidence of birth trauma and birth asphyxia related to instrumental deliveries during the study period. It is interesting to observe how long would this Hawthorne effect last after the study period. A well-designed randomised controlled trial would be necessary to study the real impact of this Hawthorne effect.

Acknowledgements

The authors would like to thank Ms Amy Chow of the Information Technology Department, Hospital Authority, HKSAR, for providing the hospital obstetric statistics and the midwives in our labour ward who helped in the study.

Ancillary