Quality improvement report•

Reducing maternal infection after assisted vaginal birth in a diverse and deprived population

Abstract

Postpartum maternal sepsis is a leading cause of maternal mortality and morbidity. A single dose of prophylactic antibiotics following assisted vaginal births has been shown to significantly reduce postpartum maternal infection in a landmark multicentre randomised controlled trial, which led to its national recommendation. This project aimed to improve the local implementation of prophylactic antibiotics following assisted vaginal births to reduce postnatal maternal infections.

Using a prospectively collated birth register, data were collected retrospectively on prophylactic antibiotics administration and postnatal maternal infection rates after assisted vaginal births at the Sandwell and West Birmingham Hospitals National Health Service Trust in North-West Birmingham of the UK. The data were collected from routinely used electronic health records over three audit cycles (n=287) between 2020 and 2023.

A mixed-method approach was used to improve the use of prophylactic antibiotics: (1) evidence-based journal clubs targeting doctors in training, (2) presentations of results after all three audit cycles at our and (3) expedited a formal change of local guidelines to support prophylactic antibiotics use.

Prophylactic antibiotic administration increased from 13.2% (December 2021) to 90.7% (July 2023), associated with a reduction in maternal infection rates (18.2% when prophylaxis was given vs 22.2% when no prophylaxis was given). However, we observed a gradual increase in the overall postnatal maternal infection rates during the project period.

Our repeat audit identified prophylactic antibiotics were regularly omitted after deliveries in labour ward rooms (59.3%), compared with 100% of those achieved in theatre. After further interventions, prophylactic antibiotics administration rates were comparable between these clinical areas (>90%) in 2023.

Together, we have demonstrated a simple set of interventions that induced sustainable changes in practice. Further evaluation of other modifiable risk factors and infection rates following all deliveries is warranted in view of the gradual increase in the overall postnatal maternal infection rates.

What is already known on this topic

  • Prophylactic antibiotics reduce maternal infection rates after assisted vaginal birth.

What this study adds

  • Simple local intervention can improve implementation rates, which was associated with improved maternal infective outcomes.

How this study might affect research, practice or policy

  • This study acts as an exemplar of clinical quality improvement in maternity, which is an often overlooked area of healthcare.

Problem

The Sandwell and West Birmingham National Health Service Trust (SWBH) is a network of hospitals serving one of the most diverse and deprived patient populations in the UK; both factors have been associated with higher maternal peripartum morbidities and mortalities.1 SWBH has a catchment of over 500 000 people in North-West Birmingham and all of the towns within Sandwell.

The maternity unit at SWBH delivers approximately 5000–5500 births per annum. Assisted vaginal births are the vaginal births of babies performed with the help of forceps or vacuum devices. The rates of assisted vaginal births at SWBH were consistent between the three audit periods (11.4%–12.4%).

The use of prophylactic antibiotics following assisted vaginal births has been associated with a significant reduction in postnatal maternal infections in a landmark randomised controlled trial.2 3 This has led to an update of the national guidance in 2020 to recommend a single prophylactic dose of intravenous antibiotics following assisted vaginal births after the babies were born.3 However, at the start of this project, the use of prophylactic antibiotics following instrumental deliveries has not been incorporated into SWBH’s guideline on assisted vaginal birth.

This quality improvement project aimed to reduce postnatal maternal infection rates by improving the use of prophylactic antibiotics after assisted vaginal deliveries, as recommended by national guidance and high-quality evidence.

Background

Postpartum maternal sepsis continues to be a leading cause of maternal mortality and morbidity, responsible for around 11% of maternal deaths both within the UK1 and globally.2 For each maternal death, it is estimated a further 70 women survive severe infection with significant long-term morbidity.2

The use of prophylactic antibiotics to reduce maternal infection is well established after caesarean sections—with a significant risk reduction by 60%–70% (baseline risks of 20%–25%).4 More recently, a landmark multicentre randomised controlled trial (the ANODE trial) demonstrated that a single dose of prophylactic intravenous antibiotics (amoxicillin and clavulanic acid) after assisted vaginal births reduced overall maternal infective complications by over 40% (from the baseline risk of 16%), particularly perineal wound infections (with a relative risk reduction of approximately 50%).

The new evidence was incorporated in the national guidance update in 20203 and a widely used evidence-based multiprofessional training package for obstetric emergencies (PROMPT; PRactical Obstetric Multi-Professional Training). During the local mandatory PROMPT training, the study team identified this specific discrepancy between local and national guidelines and practice, which prompted the start of this quality improvement project to improve the use of prophylactic antibiotics after assisted vaginal births to reduce maternal infections.

Measurements

We collected baseline data to evaluate local clinical practice, with the following questions asked:

  • How many patients received prophylactic antibiotics after assisted vaginal births (primary measurement)?

  • What were the factors associated with omission of prophylactic antibiotics (secondary measurements)?

  • What was the incidence of maternal infections after assisted vaginal births at SWBH during the study period (exploratory measurements)?

To maximise the reliability of our measurements, routinely collected data from BadgerNet Maternity, an established and widely used electronic maternity healthcare record system, and the SWBH online prescribing systems were used to collect the majority of data in this study. Clinically significant postnatal maternal infections were evaluated by retrospective evaluation of electronic clinical notes.

Operational definitions:

  • Prophylactic antibiotics administration was defined as a single dose of intravenous antibiotics given and recorded on the electronic medical records to those who had no other indications for peripartum antibiotics use (eg, intrapartum sepsis or obstetric anal sphincter injuries).

  • Clinically significant postnatal maternal infections were defined as those who have recorded use of antibiotics (both in clinical notes text and prescribing records at SWBH) for new infective symptoms within 6 weeks (42 days) of their delivery dates. This was the only measurement that required professional interpretation of clinical notes.

The results of baseline measurements (1 October 2020–31 December 2020):

In total, 156 assisted vaginal deliveries were performed. After excluding those who had peripartum antibiotics for other reasons (eg, intrapartum sepsis and obstetric anal sphincter injuries), 114 patients were included in the analysis. Only 15/114 (13.2%) received prophylactic antibiotics. Lower rates of prophylactic antibiotics use were observed in those who had ventouse deliveries (4/59, 6.8%), delivered in labour ward rooms (instead of operating theatres; 6/92 6.5%) and by more junior doctors (below specialty training year 5 or equivalent; 2/36, 5.6%). The rate of clinically significant postnatal maternal infections was 12/114 (10.5%). Consistent with previous reports, forceps deliveries were associated with higher risks of infections compared with ventouse deliveries (9/55, 16.4% vs 3/59, 5.1%).

Design

A mixed-method approach was used with the aim to improve the use of prophylactic antibiotics. The methods were planned by two obstetrics consultants and a senior trainee/resident after two virtual meetings; the second meeting also included the data collection team.

We use educational events to encourage our multidisciplinary team to discuss and understand the benefits of prophylactic antibiotics after assisted vaginal deliveries. The aim was to encourage senior management to prioritise the change of local guidelines. We believed a formal change of local guidance was the most effective way to induce sustainable changes in practice.

Data collection was performed by junior trainees/residents rotating through the unit. When possible, categorised and routinely collected clinical data was used. In addition, one-to-one induction of the group of junior trainees by the same senior trainee leading this study ensured the consistence of our data collection.

Strategy

Our objective

We set our specific objective to significantly improve the use of prophylactic antibiotics with the goal of 100% of eligible patients receiving a single dose of prophylactic antibiotics after vaginal-assisted deliveries within 12 months of our interventions.

PDSA cycle 1

First, an evidence-based journal club on the ANODE trial and the related national guideline update was held by a senior trainee/resident with experience in critical appraisal and teaching. Second, the baseline data were presented at the regular quality improvement half-day (QIHD) meeting for all members of the multidisciplinary team (QIHD) and the labour ward forum (primarily for obstetrics staff with high attendance of midwifery coordinators). At the time of the initial presentation, we included an online quiz on assisted vaginal delivery and our local data on maternal infection rates and antibiotics use with a nominal prize to maximise attendance. Third, the local guideline for assisted vaginal deliveries was updated to incorporate the recommendation to give prophylactic antibiotics after assisted vaginal delivery. These interventions increased the use of prophylactic antibiotics from 13.2% to 68.4% (52/76; 26 December 2021–25 February 2022).

PDSA cycle 2

Although our initial interventions achieved significant improvements in the use of prophylactic antibiotics after assisted vaginal deliveries, it fell short of our 100% prescription rates. Further analysis identified implementational gaps for assisted deliveries that were conducted in labour ward rooms (25/59; 59.3%), compared with those conducted in theatre (17/17; 100%). The initial interventions were successful to engage staff from all grades. We were particularly successful in improving the practice of junior trainees—20 of 27 (74.1%) of patients who had been assisted by junior trainees (defined as those with ≤4 years of experience in obstetrics) were appropriately given prophylactic antibiotics, vs births assisted by senior trainees (defined as trainees with ≥5 years of experience in obstetrics; 24/38; 65.8%) and consultants (2/3, 66.7%).

Our findings were presented at the regular QIHD. We also highlighted them to senior midwifery staff (labour ward coordinators) to encourage prophylactic antibiotics prescribing by obstetricians who conducted deliveries in labour ward rooms to prescribe prophylactic antibiotics. These interventions increased the use of prophylactic antibiotics from 68.4% to 90.7% (88/97; 1 May 2023–31 July 2023).

Results

In total, data from 287 patients eligible for prophylactic antibiotics after assisted vaginal deliveries were included in this audit. The majority were South Asians (ie, Indian, Pakistani or Bangladeshi; 149/287, 51.9%; figure 1), nulliparous (192/287 66.9%), started labour spontaneously (166/287, 57.8%) after their estimated delivery date. The median age was 28 years (IQR: 25–33), body mass index was 24 kg/m2 (IQR: 21–27 kg/m2). The majority of patients had forceps deliveries (165/287, 57.5%) and the median number of pulls was 2, with the instrument application time of 3 min (IQR: 2–4 min).

Figure 1
Request permissionrequest permission icon
Figure 1

The ethnicity of the patients included during the study periods (n=287). Patients were proportionately represented.

We demonstrated a sustainable increase in the use of prophylactic antibiotics after assisted vaginal deliveries in all birth locations (labour ward rooms and obstetrics theatres; figure 2A). The overall implementation rate was 90.7% in 2023, compared with 13.2% at baseline in 2021.

Figure 2
Request permissionrequest permission icon
Figure 2

(A) The changes of prophylactic antibiotics use over time by percentage of eligible patients, according to the location of birth (birthing rooms vs operating theatres). Total=the proportion of all patients given prophylactic antibiotics; Room=the proportion of patients delivered in birthing rooms given prophylactic antibiotics; Theatre=the proportion of patients delivered in operating theatre given prophylactic antibiotics. (B) The changes of prophylactic antibiotics use over time by percentage of eligible patients, according to the seniority of the clinicians. Total=the proportion of all patients given prophylactic antibiotics; Senior trainees=the proportion of patients delivered by senior trainees (with ≥5 years of experience in obstetrics) given prophylactic antibiotics; Junior trainees=the proportion of patients delivered in operating theatre given prophylactic antibiotics (with ≤4 years of experience in obstetrics).

We have identified an implementation gap in the second audit cycle, which showed prophylactic antibiotics were often omitted when assisted vaginal births were achieved in delivery rooms (59.3%), compared with assisted vaginal births achieved in operating theatres (100%). We successfully addressed this implementation gap by targeted education at QIHD in 2022; comparable rates of prophylaxis use were achieved in different birth locations (90.2% and 94.3% when deliveries happened in the room and in the theatre) in our last audit cycle.

The repeated educational effort has also improved the adherence to guidelines over time (figure 2B). The improvement was particularly seen in junior trainees (from 5.6% at baseline to 89.5% in 2023), a group that tends to engage with departmental educational activities more regularly.

The use of prophylactic antibiotics was associated with a reduction in maternal infections in our latest audit cycle in 2023 (figure 3; 18.2% when prophylaxis was given vs 22.2% when no prophylaxis was given). This is likely to be reflective of the now well-implemented routine prophylaxis guidance, compared with when clinicians use their clinical judgement and give prophylaxis to those at the highest risk of infections. We also identified the gradual increase in the total clinically significant maternal infections over time (from 10.5% in 2021 to 18.6% in 2023). We did not identify a difference in postpartum maternal infection by ethnicity groups (p=0.815). We have reflected on these findings in the next sections.

Figure 3
Request permissionrequest permission icon
Figure 3

The changes of maternal infections over time, according to whether prophylactic antibiotics were given. Total=the proportion of all patients who experienced postnatal maternal infections; No prophylaxis=the proportion of patients who were not given prophylactic antibiotics and experienced postnatal maternal infections; Prophylaxis given=the proportion of patients who were given prophylactic antibiotics and experienced postnatal maternal infections.

Lessons and limitations

We presented our experience on a targeted intervention to align national and local guidance on reducing maternal infection rates post partum–a frequent and often overlooked area of obstetrics care. We demonstrated a sustainable improvement over time, and in our last audit, a reduction in maternal infection rates was demonstrated in the prophylaxis group, compared with those who were not given prophylaxis. This is consistent with a large randomised controlled trial (the ANODE trial), which demonstrated a reduction in infection rates when antibiotics prophylaxis was given routinely.

We evaluated neonatal infection rates during our baseline audit—there were very few neonatal infections in the eligible group for inclusions (ie, no other indications for intrapartum antibiotics). Moreover, prophylactic antibiotics were given after the births in most cases (the babies were already born before antibiotics were given). As evaluating minor neonatal infective symptoms is challenging, we pragmatically decided not to include this data point in our subsequent audit cycles.

Although we had robust prospective documentation of the intrapartum events via our electronic medical records, clinically significant maternal infections were defined as antibiotics prescriptions or documentation in our secondary care clinical notes. Despite our records also being used by our community midwives, we had no additional access to the general practitioner and community pharmacy dispensing records. Therefore, if patients sought medical attention for maternal infection via their general practitioners, this would be missed. We also acknowledge that not all maternal infections require antibiotics treatment, and there could be overtreatments when patients present to secondary care as an emergency.

Although we have demonstrated a sustainable improvement, we were also conscious of the fact that doctors in training rotate between different obstetrics units. Our observed improvement could be part of the wider implementation of this recommendation nationally. However, we also ensure continuing professional development of midwifery staff, especially labour ward coordinators, at the heart of this project. This has supported the address of our implementation gap in deliveries that happened in labour ward rooms in our second Plan, Do, Study, Act (PDSA) cycle. Further multidisciplinary continuing professional development events with evaluations, in addition to regular infection rate updates, will further support ongoing quality improvement efforts in obstetrics more broadly.

We identified a potentially worrying trend towards increased risks of maternal infections over time. However, our baseline data were collected during the COVID-19 pandemic, when we actively encouraged postpartum patients to seek support via their primary care team to avoid viral exposure in the hospital. After the pandemic recovery, our patients also increasingly reported difficulties in getting primary care appointments, which led to subsequent increased use of obstetrics emergency services in hospitals. It is recognised locally that patients who did not speak English as their first language found that particularly challenging to obtain primary care appointments (ie, the majority of our maternity patients). Although these changes may have contributed to the apparent increase in maternal infections (relative risk increased by 76.3%, from 10.5% to 18.6%), further evaluation of postpartum infection rates after all deliveries and additional strategies to prevent postpartum maternal infection are warranted.

Conclusion

The project team had identified a discrepancy between local and national guidelines prior to this study. Through lobbying and educational events, we expedited the incorporation of prophylactic antibiotics after assisted vaginal deliveries into local practice and guidance. We further identified an implementation gap in a specific clinical area (labour ward) and implemented a targeted and simple educational note (the prescription of prophylactic antibiotics after assisted deliveries in the room). These have contributed to the sustainable improvement of evidence-based practice and are associated with a reduction in postpartum maternal infection.

We also observed a gradual increase in overall maternal infection rates in patients who had assisted vaginal deliveries. The results were fed back to the Trust’s obstetrics team via the consultant lead of this project (EE); further assessment will need to be performed to clarify if these are consistently high and to identify modifiable risk factors. For example, the use of statistical control charts of routinely recorded data (eg, the prescription of antibiotics post partum) may be helpful to identify trends of maternal infection rates in the department over time. In parallel, further work is required to ensure there is adequate awareness about postpartum infection for both staff and the public, taking into account the diverse population we serve.

Article metrics
Altmetricopen-url
Dimensionsopen-url