The Growth Assessment Protocol: a major cause of declining stillbirth rates in the UK

Iliodromiti and colleagues1 compared the reduction in stillbirth rate in Scotland with that in England and Wales from 2000–2015. They claimed to find that, after 2010, the stillbirth rate declined faster in Scotland even though fewer maternity units had implemented the Growth Assessment Protocol (GAP), and concluded, with apparent certainty, that GAP could not have been responsible for the observed reduction in stillbirth rate in either nation. They speculate that there may instead be an underlying downward trend in stillbirth rates due to other factors, such as the introduction of smoking cessation policies in 2006/2007 or the 2008 National Institute of Health and Care Excellence (NICE) antenatal guidelines, but list no actual evidence in support. In contrast, the rationale underpinning GAP, as well as the benefits of its implementation, are based on good evidence2. The program was developed with a focus on improving antenatal detection of fetal growth restriction, after a series of studies which showed that (1) most stillbirths classified as ‘unexplained’ were in fact small-for-gestational age (SGA)3, (2) according to standardized clinical outcome reviews of root causes, most SGA stillbirths were avoidable4, and (3) antenatal detection of the SGA baby can halve the risk of stillbirth5. The GAP program includes multidisciplinary training and implementation of evidence-based protocols and tools for risk assessment, antenatal surveillance and audit. Early risk assessment identifies pregnancies that require serial scanning in the third trimester. Customized growth-chart software, recommended by guidelines of the Royal College of Obstetricians and Gynaecologists6, is provided for plotting fundal height and estimated fetal weight, and establishing the birth-weight centile. The charts adjust for constitutional variation and hence help to reduce false positives7 while improving the clinician's ability to identify the fetus that is pathologically small and apply the appropriate management2. Rollout of GAP started in 2008/9, with uptake being entirely voluntary. Over the next 2 years, the majority of maternity units in three English health regions had implemented the program, and in 2013, we were able to report that, compared with other, non-participating regions, the GAP regions had a significant downward trend in stillbirth rate, and that this was also responsible for a national drop in stillbirth rate in England8. While the study was observational, fulfilment of all nine Bradford Hill causality criteria allowed us to conclude that the reduction in stillbirths was due to the GAP program8. Subsequent independent evaluation9 of NHS England's Saving Babies' Lives Care Bundle10 showed that, of its four elements (smoking cessation, fetal growth surveillance, decreased fetal movement, intrapartum monitoring), implementation of the fetal growth element was the only likely explanation for the observed reduction in stillbirth rate over the 2014–2017 evaluation period, and was associated with a significant increase in antenatal detection of SGA fetuses. Most (15/17; 88%) of the hospitals in which SGA detection rates were audited were in the GAP program. It is curious that Iliodromiti et al.1 analyze data only up to 2015, even though stillbirth rates up to 2018 were published and available from the National Records of Scotland (NRS)11 and the Office of National Statistics12 since well before they submitted their study for publication. Data from the missing 3 years show that the crude stillbirth rate in Scotland rose in 2016 and 2017, and even exceeded that in England, before dropping again in 2018 (Figure 1). Applying a slopes (Clogg Z) test to the 2010–2015 data demonstrated a significant difference in trend between Scotland and England (P = 0.03); however, if data from the omitted 3 years (2016–2018) are added, the difference disappears (P = 0.97). It is therefore not accurate to state that the decline in stillbirth rate in Scotland was faster. The reduction in stillbirth rate in England is more sustained and shows a consistent, year-on-year decline from 2011, after implementation of GAP in the first English regions8. However, according to health-service policy, such programs are not mandated, and the extent of implementation varied between regions according to local priorities. A more appropriate comparison with Scotland would be an English region of similar size and with a similar rate of uptake of GAP, such as the East Midlands, which is the second smallest English region; in 2018, it had 50 177 births12 and GAP was implemented in 8 of 10 hospitals, while, in the same year, Scotland had 51 308 births11 and GAP was implemented in 12 of 15 hospitals. Figure 1 illustrates a similar decline in stillbirth rate in both areas. As a devolved nation, Scotland benefited from a centrally coordinated Patient Safety Programme (SPSP) and the Maternity and Children Quality Improvement Collaborative (MCQIC) established in 201313, with its top two priorities being fetal growth and fetal movement. Due to a high postmortem rate, the 2012 Scottish Perinatal and Infant Mortality Report14 was able to highlight the predominance of placental causes of stillbirth. The report (section 10.4) also referred to our West Midlands program of standardized case reviews, in which several Scottish maternity units participated before 2013. These reviews often highlighted the association between stillbirth and fetal growth restriction and its potential avoidability, which helped raise clinical awareness. The Scottish government commissioned GAP nationally, and the rollout of ‘SCOT-GAP’ commenced in May 2014 through a series of ‘train-the-trainer’ workshops. This led to gradual implementation in 13 of the 15 health boards, supported by an e-learning and accreditation program with over 1500 registered Scottish midwives, doctors and ultrasonographers. An audit conducted by Edinburgh University between December 2014 and February 2015 found that the three maternity units that had implemented GAP by that time had significantly higher SGA detection rates than did the 12 other Scottish units, and called for standardization of care in Scotland15. A parallel initiative of relevance was the 2014–2016 AFFIRM cluster randomized trial16 which investigated the effectiveness of a decreased fetal movement care package that included all Scottish maternity units. While the results of the trial showed no significant effect on stillbirth rate due to the care package itself, the protocol led to additional third-trimester growth scans and Doppler studies, and resulted in significantly fewer SGA deliveries at or after 40 weeks' gestation16. These factors (national programs with general raising of awareness, informed mortality reports, case reviews highlighting the importance of growth restriction and the potential avoidability of stillbirth, rollout of GAP training, and a national trial resulting in increased growth scans and Doppler examinations) specific to Scotland are all much more plausible contributing factors to any reduction in stillbirths during 2010–2015 than are the effects of the 2006/2007 smoking cessation policy and the 2008 NICE antenatal care guidelines, as speculated by the authors1. Such special initiatives also limit the suitability of using Scotland as a control for any pre-2015 comparisons in stillbirth rate. According to the latest NRS data11, the decline in stillbirth rate in Scotland continued in 2018, at which time 12 of the 15 Scottish health boards were in the SCOT-GAP program. To account for small number variation, the MCQIC presents this sustained drop in stillbirth rate as a ‘rephased median’ from 2014, showing a 22.6% improvement13 (Figure 2). This is similar to the 25% reduction in stillbirths in England, from an average rate of 5.35 per 1000 births in 2000–2009 to 4.01 per 1000 births in 201812, resulting in 842 fewer deaths per year. We have outlined a very different picture of GAP from that portrayed by Iliodromiti et al.1, by: (1) reviewing the underlying evidence and performance of the program; (2) evaluating the full range of data including those from the last 3 years which the authors omitted; (3) performing like-for-like comparisons of areas of similar size and with similar uptake of the GAP program; and (4) acknowledging specific national efforts in Scotland which helped to raise clinical awareness, even before GAP was implemented. We hope that these additional facts, which are all in the public domain yet were omitted from the study in question1, will help in providing a more objective assessment and allow readers to reach their own conclusion.