5.1 Introduction

In a previous chapter in Key Health Data (Key Health Data for the West Midlands 2001), we have shown that

• West Midlands perinatal mortality is consistently one of the highest of all government regions;


• stillbirths are the largest contributor to perinatal mortality, and with current classification systems, most are described as ‘unexplained’;


• a new classification is being applied which demonstrates that many of these deaths are associated with fetal growth restriction at relatively mature gestations and are therefore potentially avoidable;


• there is a significant link between perinatal mortality and social deprivation.

The purpose of this chapter is to

1. summarise a recently completed analysis of stillbirth and infant mortality trends from 1998-2003 in the West Midlands;
2. assess the main categories contributing to these deaths;
3. analyse the trends of mortality associated with inequalities and the association with deprivation within different mortality subgroups.

The analysis presented here refers to the whole of the West Midlands. Sub-analyses for the three strategic health authorities are presented on www.perinatal.nhs.uk/pnm/trends. For PCTs, there are limitations on the detail in which potentially identifiable data can be published on the web or in print. The results for each PCT are therefore being sent directly to the respective director of Public Health, in line with data protection and confidentiality guidelines.

5.2 Methods

Definitions:

Stillbirth rate
Number of stillbirths > 24 weeks (168 days) per 1,000 live births and stillbirths.

Early neonatal mortality rate
Number of deaths at ages under 1 week, per 1,000 live births.

Neonatal mortality rate
Number of deaths at ages under 4 weeks, per 1,000 live births.

Postneonatal mortality rate
Number of deaths at ages 28 days and over but under one year, per 1,000 live births.

Perinatal mortality rate
Number of stillbirths plus number of deaths at ages under 1week, per 1,000 live births and stillbirths.

Infant mortality rate
Number of deaths at ages under one year, per 1,000 live births.

5.3 Data source

The data was derived from Rapid Report Forms submitted to the Institute from maternity units in the West Midlands and the ONS Birth Tapes.

5.4 Classification

The two classifications in this report are listed in Tables 5.1 A and 5.1 B. They map to those applied by ONS in terms of the chronology of the event. However, they also use a previously developed categorisation which aims to elucidate underlying conditions or causality.

• Classification for Stillbirth

The standard classifications such as Wigglesworth are summarised in the CESDI national reports. However, most of these result in a large ‘unexplained’ category which confirms the need to improve the classification system (CESDI 2001).

The Perinatal Institute developed a classification for stillbirth, ReCoDe (http://www.perinatal.nhs.uk/pnm/recode). For the purpose of this aggregated analysis, stillbirth rates are summarised in the following ReCoDe categories:

• Classification for Neonatal and Infant Death

This was based on the Fetal and Neonatal Classification (Hey 1986) and amended to include cot deaths as a separate category.

5.5 Deprivation Scores

For the purpose of the current analysis, we applied the widely used Index of Multiple Deprivation (IMD2000). This is a ward-based measure with seven sub domains (income, employment, health, education, housing and services, environment and crime). Larger values indicate higher degrees of deprivation.

For comparative analysis, we defined as cut-off the 197 (8%) most deprived wards of the 798 wards in the West Midlands. These wards had 93,962 of the total of 377,744 births (25%) over the period of the study. We thus compared the 25% of births from the 8% most deprived wards (‘IMD 8/25’) with 75% of births from the remaining 92% of the wards in the West Midlands (‘IMD 92/75’).

It should be noted that these births, although from the most deprived wards, were still from a range of social class strata, a consequence of this measure of inequality being ward and not patient based. We currently do not have social class indicators for each birth in the West Midlands, although we expect data from the ONS which will allow us to use these for further analysis in the near future.

5.6 Rates and Trends in Perinatal and Infant Mortality

Figure 05.01 Stillbirths, Early, Late and Postneonatal deaths West Midlands 1998-2003.

The PSA targets on reducing inequalities use infant mortality as a key indicator (Health Inequalities DoH 2004). However, as Figure 5.1 shows,

• most infant deaths (birth to 12 months) occur in the first week of life;
• stillbirths are the largest component of adverse outcome, and these are not contained in infant mortality statistics.

It is well established that many instances of adverse outcome have antenatal origins. Therefore, we need to include stillbirths, not just in their own right, but also when seeking to understand and reduce infant mortality. The analysis presented in this chapter will include perinatal (stillbirth and early neonatal) as well as infant (early, late and post neonatal) data.

	Table 05.02 Trends for perinatal mortality rates in the West Midlands and England & Wales.
	Figure 05.02 Trends for perinatal mortality rates in the West Midlands and England & Wales.

There was a non-significant increase in perinatal mortality in England & Wales (p=0.11; linear trend analysis) and West Midlands (p=0.21) This appears to be associated with a rise in stillbirth rates over the last 2 years (Figure 5.3).

	Table 05.03 Stillbirth rates in the West Midlands and England and Wales 1998-2003.
	Figure 05.03 Stillbirth rates in the West Midlands and England and Wales 1998-2003.

Linear trend analysis shows an increase in stillbirth rates in the West Midlands over the last two years, which mirrored those in England and Wales over the same period. Both were significant (p<0.01).

	Table 05.04 Wigglesworth classification of stillbirths in the West Midlands 1998-2003.
	Figure 05.04 Wigglesworth classification of stillbirths in the West Midlands 1998-2003.

Consistent with national statistics (CESDI 2001), the Wigglesworth classification used by CESDI / CEMACH leaves consistently two-thirds of stillbirths in the ‘Unexplained’ Category.

Fetal growth restriction was the main category in each year, and represented an average 42% of all stillbirths. There was a slight, non-significant downward trend (p=0.65).

	Figure 05.05 Stillbirths by major ReCoDe groups 1998-2003, West Midlands.
	Table 05.06 Early Neonatal Death Rates in the West Midlands and England & Wales 1998-2003.
	Figure 05.06 Early Neonatal Death Rates in the West Midlands and England & Wales 1998-2003.

Early Neonatal Deaths showed a gradual downward trend in England and Wales (p=0.02) but not in the West Midlands.

	Table 05.07 Early neonatal deaths in West Midlands 1998-2003 - Neonatal and Infant Death classification.
	Figure 05.07 Early neonatal deaths in West Midlands 1998-2003 - Neonatal and Infant Death classification.

There was a significant upward trend in the category Severe Pulmonary Immaturity (p=0.04), while the late drop in congenital anomalies did not reach significance (p=0.22).

5.7 Neonatal Deaths and Gestational Age

While stillbirths are recorded from 24 weeks only, neonatal deaths have no lower gestational age limit, and a baby is considered live born if it has demonstrated any signs of life.

The deaths classified as ‘Severe Pulmonary Immaturity’ were analysed according to gestational age at birth (Table 5.8)

The table shows an increase in deaths reported at very early gestations. The increasing trend in reports of ENND at gestations < 22 wks is highly significant (p<0.01). This is further illustrated in Figure 5.8.

Figure 05.08 Early Neonatal Deaths <22 weeks due to severe pulmonary immaturity West Midlands 1998-2003.

The Perinatal Institute intends to discuss the issue of gestational age and the recording of neonatal deaths with regional stakeholders. One possible explanation for the observed rise the tendency in recent years to move the management of late miscarriages from a gynaecological to an obstetric ward, where perhaps ‘signs of life’ are more actively looked for, even at pre-viable gestations.

	Table 05.09 Late Neonatal Death Rates in the West Midlands and England & Wales 1998-2003.
	Figure 05.09 Late Neonatal Death Rates in the West Midlands and England & Wales 1998-2003.

There was a gradual drop in late neonatal death rates in England and Wales, (p=0.01) which was not reflected in the West Midlands data.

	Table 05.10 Late neonatal deaths by Neonatal & Infant classification, West Midlands 1998-2003.
	Figure 05.10 Late neonatal deaths by Neonatal & Infant classification, West Midlands 1998-2003.

It should be noted that numbers in the 'Late Neonatal' age range are much smaller, which explains some of the variation.

The downward trend in congenital anomalies was statistically significant (p=0.04) but the drop in 2003 was accompanied by an increase in the unclassified/unknown category, which is likely due to delays in diagnosis and classifications of some anomalies.

	Table 05.11 Post Neonatal Death Rates in the West Midlands and England & Wales 1998-2002. NB - Some post neonatal deaths in 2003 have not yet been notified. Therefore the rates are presented up to 2002 only.
	Figure 05.11 Post Neonatal Death Rates in the West Midlands and England & Wales 1998-2002.

There appears to be a slow downward trend in both data series. This achieves significance in the figures for England and Wales (p<0.01) but not in the West Midlands (p=0.38).

	Table 05.12 Post neonatal deaths analysed using major Neonatal and Infant Death classification groups 1998-2002. NB - A proportion of these deaths in 2003 have not yet been notified, are coroners cases, and /or their final diagnoses have yet to be entered. Therefore the data is presented to 2002 only.
	Figure 05.12 Post neonatal deaths using Neonatal and Infant Classification - West Midlands 1998-2002.

There was a significant drop in cot deaths (p<0.01) The apparent downward trend in congenital anomalies is not statistically significant (p=0.56).

	Table 05.13 Infant Death Rates in the West Midlands and England & Wales 1998-2002.
	Figure 05.13 Infant Death Rates in the West Midlands and England & Wales 1998-2002.

There was a downward trend in infant mortality rates in England and Wales (p<0.01) which was not reflected in the West Midlands.

Table 5.14 and Figure 5.14 show an analysis of Infant deaths by the Neonatal and Infant classification.

	Table 05.14 Infant deaths by Fetal and Neonatal classification groups, West Midlands 1998-2002.
	Figure 05.14 Infant deaths by Neonatal and Infant classification groups, West Midlands 1998-2002.

The two largest categories are 'severe pulmonary immaturity' and 'congenital anomalies'.

There was a fall in cot deaths (p<0.01). The apparent late drop of congenital anomaly related deaths did not reach statistical significance (p=0.39).

5.8 Stillbirth and Infant Mortality and Deprivation

To analyse the effect of deprivation, a cut–off was selected which identified the 8% most deprived wards in the West Midlands, which accounted for 25% of all births (IMD 8/25). This is used to compare with the other 92% of wards, representing 75% of births (IMD 92/75).

	Table 05.15 Stillbirth Rates by Index of Multiple Deprivation, West Midlands Region 1998-2003.
	Figure 05.15 Stillbirth Rates by Index of Multiple Deprivation, West Midlands Region 1998-2003.

This analysis shows the well known association between deprivation and stillbirths. However, it is also apparent that the gap is not reducing but in fact getting larger, due to a significant increase in stillbirths for the most deprived areas from 6.9 in 1998 to 8.4 in 2003 (p=0.04).

	Table 05.16 Infant Death Rates by Index of Multiple Deprivation, West Midlands Region 1998-2002.
	Figure 05.16 Infant Death Rates by Index of Multiple Deprivation, West Midlands Region 1998-2002.

This analysis confirms the well known association between deprivation and infant mortality. There is no decrease in the size of the gap between the most deprived and others. There was even a mild but nonsignificant (p=0.34) increase in infant deaths in the most deprived areas.

	Table 05.17 Stillbirth, early neonatal and perinatal mortality rates in most deprived areas (IMD 8/25) compared to the rest of the population, West Midlands 1998-2003.
	Figure 05.17 Stillbirth and Deprivation: Effect of deprivation (IMD 8/25) on main categories of stillbirth ReCoDe Groups, West Midlands 1998-2003. Proportion of deaths and significant Odds Ratios and Confidence Intervals are shown.
	Table 05.18 Infant mortality rates in most deprived areas (IMD 8/25) compared to the rest of the population, West Midlands 1998-2002.
	Figure 05.18 Infant Mortality and Deprivation, Effect of deprivation (IMD 8/25) on main categories of Infant Mortality, Fetal and Neonatal Classification Groups, West Midlands 1998-2003. Proportion of deaths and significant Odds Ratios and Confidence Intervals are shown.

5.9 Summary Conclusions

1. Stillbirths are the highest contributor (57%) to perinatal mortality, and most infant deaths (58p%) occur in the first week of life.
2. Intrauterine growth restriction is the most frequent condition associated with stillbirth.
3. Severe immaturity and congenital anomalies are the largest contributors to infant mortality. There was a recent increase in the proportion of neonatal death registrations at extremely early gestations which the Perinatal Institute will investigate further and consult on with stakeholders.
4. There is a significant gap in stillbirth and infant mortality rates between the most deprived wards and the rest of the West Midlands population, which has been increasing over the period of this analysis.
5. The increased risk applies in all main sub-categories, including deaths associated with congenital anomalies, fetal growth restriction and immaturity.

References

Health Inequalities: Life Expectancy and Infant Mortality DoH, 2004

Hey EN, LLoyd DJ, Wigglesworth JS. Classifying perinatal death: fetal and neonatal factors. British Journal Obstetrics & Gynaecology 1986;93:1213-1223.

Key Health Data for the West Midlands 2001. University of Birmingham DPHE 2002

Maternal and Child Health Consortium. CESDI 8th Annual Report: Confidential Enquiry of Stillbirths and Deaths in Infancy, 2001. Commentary: Clinical Implications of ‘Unexplained Stillbirths’, J Gardosi, pp 40-47

ReCoDe: Stillbirth classification denoting the ‘Relevant Condition at Death’. http://www.perinatal.nhs.uk/pnm/recode accessed 7th May 2005