Born but not properly counted

Whitehall is better armed with data on children who no longer exist than on those that can expect to live for the next 80 years.

While abortions are carefully counted and categorised by ethnicity, live births in the NHS are the subject of one of the most slapdash data-gathering systems it is possible to imagine. In last year's tables, one hospital even managed to count 99 per cent of its births as stillbirths – and still had the nerve to submit the data.
 
In the week in which the Prime Minister has discovered that immigration may not be an unmixed blessing, it is worth taking a more detailed look at these figures, because they illustrate a disregard for the very data that ought to illuminate the changing ethnic make-up of the UK.  
 
Governments need to know the ethnic mix of the population. Fill in any government form, apply for a grant, respond to a questionnaire, write after a job vacancy and, as like as not, you’ll be asked about your ethnicity. Every Town Hall has a breakdown of employees by gender and by ethnicity.
 
The objective, a proper one, is to be sure that we are providing the same level of services to all those in the community, of whatever ethnic origin, and to establish well in advance the likely composition of any future society by age, sex and race. Governments need to know that schools are prepared for the annual intakes by size and cultural mix and that lessons are sympathetic to cultural diversity and sensitivities.
 
Why, then, are greater efforts not made to record births accurately? Of course, they must all be registered, but registration data are slow to appear and limited in the information they contain. They record the country of birth of the mother but not ethnicity. That the NHS is capable of counting is shown by the abortion data, which record every termination in precise detail – a much higher standard than is applied to the recording of live births. 
 
 

 
Abortions are properly quantified and costs allocated either to the NHS or to private agencies. We have a good idea of the abortion rates, what age is the patient and from what ethic group. Of the 185,415 ‘terminations’ in 2004, 111,807 (60 per cent) were to white mothers and only 10,937 to Asian mothers.
Fig 1 (above), shows only those women who have had legal abortions and who are resident in this country. Women who pay for their own abortion or who are resident overseas do not appear in this table.
 
But in matters of new born babies we have no accurate idea of ethnicity or even of total births. The data that does exist has a wide tolerance and is collected on a voluntary basis, so that it is only indicative and could be misleading. Fig 2 (below) shows the approximate numbers of new born babies in 2007-08 by ethnic group. Not surprisingly babies of White parents predominate, at 282,109. Asian babies, that is, Indian and Pakistani, make up 38,963 of all new-borns and Afro-Caribbean or British African 20,717. What might be unexpected is the number of Chinese and other births from that region of the world, including Vietnam and the Philippines, at 14,081.
 
  
     
This table comes with a lengthy explanation,  which could be paraphrased as admitting that it is virtually useless. Why? Because a number of hospital trusts “persistently submit poor quality data” the Information Centre admits.
 
Among the problems are trusts submitting more “delivery episodes” than “birth episodes”, or failing to submit delivery data at all. Some trusts says that they have lots of beds available but send no data on deliveries or births; others say that they have no beds available, but still manage to record a high number of deliveries or births.
 
As a result, the total number of births in this table is far below the actual number: 217,465 births (more than a third of the total) are excluded because their birth weight was not known. In the past, ONS data on birth registrations was used to “gross up” these tables and conceal the gaping holes, but that practice was abandoned before the 2006-07 birth data were published,. Even within this truncated sample, for a large number of births ethnicity is either not stated (33,929)  or not known (10,012).
 
So one might fairly conclude that Whitehall is very good  at counting those babies whose lives have been terminated and who therefore need not be considered  with regard future policy decisions, but is all at sea with those children who do make it past the delivery room.
 
The prize for the craziest entry in this birth lottery goes to Heatherwood and Wexham Park Hospitals NHS Trust which recorded 5,459 stillbirths in its return for 2007-08. Clearly this was an error of data entry, rather than an epidemic of stillbirths, but as a result, all the hospital’s data was excluded.
 
HESonline is aware of these shortcomings but either can’t find a workable solution or can’t find the time to address it. There seems little point in publishing statistics so slipshod as these.
 
The latest maternity publication has recently appeared, covering 2007-08, though only a few of the tables have so far been released. Are they any better? Alas, no. “The coverage of 2007-08 HES maternity data has deteriorated across a range of maternity fields” the Information Centre admits. But it has gone ahead with publication in order to share the information available, and help trusts improve the quality of the data for future years.
 
With the second of these objectives we can all concur.