House of Commons - Health

Select Committee on Health Fourth Report

2 COLLECTION OF DATA FROM MATERNITY UNITS

HOW ARE DATA ON MATERNITY CARE COLLECTED AT NATIONAL LEVEL?

How does the current system operate?

18. Reliable data on the care provided for pregnant women and new mothers are essential to the auditing and development of maternity services at local and national level. There are three main ways in which data are collected following the birth of a baby. The first is through birth notification. All live births and stillbirths at 24 or more completed weeks of gestation must be notified by the midwife or other professional attending at the birth to the local Director of Public Health within 36 hours. This is now being achieved electronically through the NHS Numbers for Babies scheme which started in October 2002.

19. Under rules governing the civil registration of births and deaths, births, stillbirths at 24 or more completed weeks of gestation, and deaths must be registered at the local register office. Births must be notified within six weeks, live stillbirths within three months, and deaths within five days. The cause of stillbirth or death must be medically certified. These data are anonymised for statistical analysis by the Office for National Statistics which also maintains separate registers of identifiable information for legal purposes.

20. The third major source of data about maternity care given at birth is the Department's system of Hospital Episode Statistics (HES). Statistics on maternity care are collected via items in an appendix to the standard record of care for patients admitted to hospital, known as the 'maternity tail'.[29] These data are sent by trusts from their Patient Administration Systems to the Department (there are now arrangements for maternity units whose systems are not linked to their hospital's patient system to send maternity data directly to the Department).[30]

21. Data are also collected from maternity units via the Confidential Enquiry into Maternal Deaths (a review of individual maternal deaths) and the Confidential Enquiry into Stillbirths and Deaths in Infancy (which reviews samples of events), now merged as the Confidential Enquiry into Maternal and Child Health and run by the National Institute for Clinical Excellence.

WHAT ARE THE PROBLEMS WITH THE CURRENT SYSTEM?

22. We heard throughout our inquiry, however, that there were major problems with maternity care data collection systems. One is incompleteness. Maternity tail data for nearly a third of deliveries do not reach the Department because of problems with incompatible or non-existent computer systems, or because computer systems are in the process of reconfiguration. According to the Department, since 1992-93, about 95% of deliveries in each year have at least a core record (noting the birth, and mode of delivery of each baby). However, even where HES records are sent, data may be missing. In 2000-01, out of 197 hospitals/trusts with deliveries, 52 did not submit any information for the maternity tail.[31]

23. Perhaps the most significant problem with maternity care data is that different units and different systems may define data items in different ways, which calls into question the validity of the statistics produced. The Maternity Care Data Project, run by the NHS Information Authority, was intended to have tackled this problem by April 2003. One of the main strands of the project was the compilation of a 'data dictionary'. This would provide standard definitions for all items of data, including the subset required by the Maternity HES. The aim was to have standardised and consistent recording of data relating to maternity and childbirth (for women and infants), within electronic patient record systems in all affected NHS organisations. However, we heard that the development of the Maternity Care Data Project has stalled owing to the lack of a 'champion'[32]

24. We are concerned that the accuracy of maternity care statistics is adversely affected not only by missing data but by data submitted according to different interpretations of the terms used to define the data required by the Maternity Hospital Episode Statistics. We recommend that the NHS Information Authority clarify the progress made to date on the Maternity Care Data Project, and in particular on the compilation of the 'data dictionary'. We further recommend that work on this important area continue, overseen by a 'national champion' for maternity care data, alongside efforts to ensure that all maternity units submit data to the Maternity HES.

25. We also heard that the data collected failed to reflect the changes in maternity care policy and practice which had taken place over the years. The British Association of Perinatal Medicine (BAPM) criticised the routine data collected in relation to childbirth as "a very limited number of items focusing on whether the mother and her baby survive" and asserted that emphasis on the care of mother and baby in the selection of the data required would allow a much greater understanding of the quality of service, and of the changes in practice in maternity services.[33] In the past, the focus of work on maternity care data has been on the reduction of mortality levels for mothers and babies, although we note that no pregnancy-specific data are collected on care for women who have had a miscarriage.[34]

26. While this is still a vital component of data requirements for maternity care, many of those working in the field argued that pregnancy and childbirth should be seen as normal physiological processes which require low levels of medical intervention in the majority of cases. They felt that the medical profession and even women themselves, had lost a sense of what was normal with regard to birth, and that 'normality' should be defined in order for this sense to be regained. Normal or physiological birth is defined by some units as birth without any medical intervention but by others as birth without caesarean section. Helen Shallow, a consultant midwife from Derby City General Hospital told us that such definitions "could hide a multitude of interventions and things happening to a woman."[35]

27. Data are also needed so that work can be done to reduce health inequalities in maternity care, both in terms of the variations in standards of care across the country and of the barriers to receiving appropriate care that some groups in society confront. The requirements for the Maternity HES do not accommodate the need to monitor inequalities in access to care, or to record normal birth statistics. Dr Soo Downe of the University of Central Lancashire told us that:

28. There is no central collection of data on antenatal and postnatal care outside hospital, although such data are often collected by community and General Practice systems and by the maternity records held by women themselves.[37] Central collection of such data might help the Government to monitor targets such as those on cessation of smoking in pregnancy. In its White Paper Smoking Kills, the Government set out to reduce the percentage of women who smoke during pregnancy from 23% to 15% by 2010, with a fall to 18% by 2005.[38] At present progress towards this target is monitored through the Infant Feeding Survey, which is conducted every five years among women who have recently given birth. The data are retrospective and cannot be validated.[39] Routine collection of data on cessation of smoking during pregnancy might represent a more accurate method of monitoring progress towards targets.[40] Lynne Pacanowski, Head of Midwifery at St Mary's Hospital, Paddington also told us about the problem of defining terms and parameters for data collection on breastfeeding rates. She pointed out that it was impossible to build a statistical base for analysing rates because data were collected at different intervals after delivery at different units.[41]

29. We recommend that data on breastfeeding rates, in terms of initiation and duration, should be standardised and collected at national level.

30. A final, but fundamental problem with the present system of national data collection is that information on child health is collected without adequate electronic links to maternity systems and so it can be very difficult to measure long-term outcomes for pregnancy and childbirth. Dr Jean Chapple, a consultant in perinatal epidemiology told us that:

31. We recommend that the Department take immediate action to ensure that maternity care data systems and population-based child health systems for both sick and healthy babies, should be linked together at national and local level in order that health professionals have all information relevant to mother and baby and in order that the long-term outcomes of pregnancy and childbirth for maternal and child health can be measured.

32. Changing Childbirth recommended that all women should carry their own maternity notes. We are disappointed that ten years later there are still some units where this does not happen. We recommend that the Department should insist that all units support the use of woman-held notes. We further recommend the development of a national format of these notes in preparation for the Electronic Patient Record.

WHAT IS BEING DONE TO IMPROVE THE SYSTEM?

33. The Department has introduced several measures to modify the system and method of data collection on maternity care. Changing Childbirth funded the National Maternity Record Project which set out to design records which would be held by the woman herself, and which explained clinical terms and gave sources for further information. However, some clinicians reported difficulty in finding data and it was later ascertained that the format of the records was not conducive to data entry in computerised maternity data systems.[43]

34. With effect from 29 October 2002 all babies born in England and Wales are now issued with their NHS Number at birth. The 'NHS Numbers for Babies' system is part of the NHS Information Authority's 'NHS Numbers for Life' initiative. Under this initiative, midwives request and receive a newborn baby's NHS Number as soon as possible after birth by connecting to the new NHS Central Issue System. The number is passed on to the Registrar of Births and Deaths. The new system has been welcomed as a way of strengthening the mechanisms necessary to collect comprehensive and accurate statistics[44] but some problems with the implementation of the scheme have been drawn to our attention. Professor Alison Macfarlane from City University, London, told us that technical problems have meant that some birthweights are no longer being passed from child health systems to local registrars of births and deaths, and that implementing the scheme has involved considerable duplication of effort in passing information to child health systems. She also raised the concern that the National Congenital Anomaly System operated by the Office for National Statistics could be undermined, as the minimum dataset for 'NHS Numbers for babies' does not contain details of congenital anomalies.[45]

35. The Department told us that it was acting as an intermediary in negotiations between trusts and the NHS-wide clearing service which received maternity care data, to help reduce incompatibilities between data systems. The Department also told us that the HES system was being recommissioned and that options for the collection of maternity care data were being explored.

36. We welcome the Department's efforts to reduce incompatibilities between data systems and to review policy on the collection of maternity care data. We recommend that this review take account of calls for a renewed focus on normal birth and of the need for accurate data on antenatal and postnatal care in order to monitor progress towards targets and reducing health inequalities. We further recommend that in reviewing the Maternity HES the Department should ensure that the figures compiled for each maternity unit take accurate account of factors such as privately-run units within hospitals, and reflect the configuration of services which take in community midwifery teams and midwifery-led units under the auspices of a hospital unit. The Department should also take steps to ensure that data are collected on births in privately-run units and on home births.

How are data on maternity care collected at local level?

WHAT SYSTEMS ARE USED BY MATERNITY UNITS TO ENTER AND RETRIEVE DATA?

37. The data requirements of the Confidential Enquiries, the Maternity HES and the National Sentinel Audit of Caesarean Sections have led to the development of more sophisticated electronic collections than those used in other areas of health care. Professor Philip Steer, from Imperial College, London put forward this argument in terms of the Government's plans for migration of patient records to electronic format:

The use of computers to collect and store clinical information has been extensively explored in maternity care … electronic collection of data in the maternity services is probably nearer to being a true electronic patient record than in any other medical discipline.[

38. Data collected on maternity care are important not only as a means of monitoring trends at national level but also as a source of guidance for maternity units in developing local services. This means that a good maternity data system will allow retrieval as well as collection and transmission of data.

39. However, the serious problems with collection of data at national level reflect problems at maternity unit level. As part of the preliminary work for the NHS Numbers for Babies project the NHS Information Authority contacted the 184 trusts which provide maternity services (which contain a total of 269 units) and asked them to describe their maternity data systems. Some 34% of the trusts told the NHS Authority that they currently operated manual data systems, and 40% reported that the infrastructure of their current system was not sufficient to support implementation of the NHS Numbers for Babies project.[47]

40. Approximately 60% of the maternity units which did have electronic data systems used one of three commercially available systems but this was no guarantee of uniformity in terms of definitions for data items, or in terms of links to main hospital data systems. Other maternity units used systems developed in-house rather than those from commercial suppliers.[48]

41. We asked each of the maternity units giving evidence to tell us about their data systems, and about how they entered and retrieved data. Although we heard several very different kinds of system described, almost all were reported to be inefficient or inadequate.

42. Staff members from St Mary's Hospital, Paddington told us that their system, developed in-house, crashed on a regular basis and did not allow staff to retrieve useful data. Professor Lesley Regan, Consultant Obstetrician at St Mary's described attempts to use the system for this purpose:

We also spoke to representatives from Edgware Birth Centre, which used the same system. Edgware staff told us that the system could not collect data which would be useful in developing and planning practice at the birth centre, and that other tools had to be devised for this purpose.[50]

43. Professor James Walker, a consultant obstetrician at St James's University Hospital, Leeds identified a key difficulty in retrieving data on maternity care from some computerised systems:

44. Professor Walker described the process used to audit caesarean sections at St James's. The computer system could tell staff which patients had undergone caesareans but any further information had to be sought from case records or the delivery suite book. Another source of difficulty in collating data on mothers and babies was inadequate communication between the maternity system and other hospital computer systems such as that used for pathology.[52] This was another near universal problem in terms of the maternity units which provided evidence for our inquiry.

45. Karen Connolly, Head of Midwifery at St Mary's Hospital for Women and Children, Manchester described her unit's maternity data system, which had been initiated in 1987 to the design of one particular consultant and then implemented across the maternity unit ten years later. It collected specific data from the time that a woman booked her maternity care with the unit through to the postnatal period, and could generate monthly and annual statistics. Although the system itself was "comprehensive", it had only been linked to the hospital's patient administration system in the previous twelve months and even then, as Karen Connolly pointed out, "only in a minimal way, just for the demographic details. We are still having some teething problems … trying to get combined data."[53]

46. Even where maternity systems were linked up with other data systems maternity care staff still struggled to achieve worthwhile returns relative to the time they devoted to entering data. The system at Trafford General Hospital was created in-house as a module of the Hospital's Patient Administration System but as Antony Nysenbaum, Clinical Director of Obstetrics and Gynaecology pointed out, lack of financial investment in maintenance of the system meant that desperate measures were needed in order to retrieve data:

47. In some areas midwives working with pregnant women and new mothers in the community had more fundamental problems of access to data in that they had no way of linking up with hospital computer systems. Cathy Rogers from Barnet and Chase Hospital spoke for midwives working in the community when she emphasised:

48. We were even more dismayed to hear reports from units where computer systems were useless or non-existent. Gill Smethurst, a clinical co-ordinator for midwifery and gynaecology told us about collection of maternity data at Goole Midwifery Centre:

49. We also heard from Worcestershire Royal Hospital and from Shrewsbury Royal Hospital, which lacked any computerised form of maternity information system. Just as at Goole, delivery records at the unit were kept on a ledger in the maternity ward, rendering any effort to generate statistics a time-consuming project: "If you are trying to look at why you are having a high Caesarean rate, if you are trying to do a retrospective audit of that, it means you pull in all the notes for the last twelve months to try and identify common reasons."[57]

50. We were appalled to hear of the burden of work imposed on maternity care staff in units where maternity care data systems were inadequate or nonexistent. The dramatic variation in the reliability and availability of maternity care data systems across the country cannot be rationalised by differences in size or configuration of units. We were struck by the disparity between this unacceptable situation, where staff could not retrieve information about their patients, and in turn where reliable national statistics could not be generated, and the Government's intention to use information technology to "enable NHS professionals to have the information they need both to provide … [the best possible] care and to play their part in improving the public's health."[58]We recommend that the Department of Health Statistics Division 3G liaises with other relevant parts of the Department and the NHS Information Authority to issue a direction to trusts on the provision and maintenance of maternity care data systems, and on links between these systems and other health information systems, so that maternity units can collect and retrieve accurate data in a more efficient way to meet both local and national data needs.

WHO COLLECTS AND ENTERS THE DATA?

51. According to the Centre for Nursing and Midwifery Research at the University of Brighton, responsibility for data collection, data entry and data reporting lies with maternity managers or heads of midwifery, who may not necessarily have the skills needed to work with data on computerised systems.[59] Responsibility for these tasks may be delegated, either to IT staff or, in one third of maternity services across the country, to an IT-specialist midwife. Where the IT-specialist midwife position is a full-time senior post, "the computerised systems function considerably more effectively to produce meaningful clinical data."[60]

52. It became clear to us that those units which had the benefit of a team member interested or skilled in IT were able to make the best use of the computer systems available to them. Responsibility for, and a sense of 'ownership' of, the data collection system were crucial to successful data retrieval, as Donna Ockenden, Head of Midwifery at St Mary's Hospital, Portsmouth told us: "we have an experienced midwife with a big interest in IT and she is responsible for a lot of the good quality detail."[61] Technical support seemed to inspire more confidence in the system. Karen Connolly, from St Mary's, Manchester was able to tell us that:

53. In the majority of units, however, entry of data into electronic systems or ledgers was mostly undertaken by midwives without any particular expertise in IT, and often with insufficient technical support. We heard from midwives that data entry was seen as a burden which could add to the pressure already exerted by a heavy workload. Jennifer Fake, Head of Midwifery at Watford General Hospital, illustrated this point:

54. Jen Ferry from the Rosie Hospital, Cambridge suggested that IT specialists would be better placed to take responsibility for maintaining data collections:

Written evidence from Dr Luk Yun Chan, a consultant paediatrician at Mayday Hospital, Croydon also argued that clerical staff should input data on the grounds that this would produce more accurate results and ease pressure on clinical staff.[65]

55. Given the shortage of midwives in a number of services it can be difficult to provide even the most basic care. If training is inadequate and there is poor support for IT (particularly in relation to prompt assistance when problems arise), entering data can be time consuming and frustrating for midwifery staff and can hamper the provision of good quality care.

56. The process of entering data on maternity care must not compromise the quality of care that pregnant women, and new mothers and babies receive. Adequate managerial and systems support is vital. Maternity care teams should have access to the services of administrative staff who have been trained to use the data system. While clerical staff can help to alleviate some of the pressure on maternity staff in terms of data entry, it is essential that the ultimate responsibility for overseeing the quality and clinical accuracy of data lies with a senior member of the clinical team. We recommend that the Department ensure that maternity units have access to reliable hardware, systems which can support the handling of individual records, to software which can be used for data analysis, and to appropriate statistical and IT support. Provision should be made for midwives who wish to do so to acquire skills in data analysis for monitoring and audit.

How can maternity unit data collection systems be improved?

57. We received evidence to suggest that the implementation of computerised maternity care data systems in some units had been delayed owing to uncertainty about the future requirements of the Government's electronic patient record (EPR) initiative.[66] This suggestion was confirmed by one of our witnesses, Professor David James, Lead Obstetrician at University Hospital, Nottingham. Professor James felt that the drive to deliver the electronic patient record was commendable in principle but was worried that records generated in different parts of the country by different computer systems might not be transferable: "we could be investing a vast amount and developing that locally, and it could be at variance with what is being done nationally."[67]

58. Cathy Rogers from Barnet and Chase Farm Hospital also expressed some anxiety about the development of the EPR. Her answers to our questions reflected a general feeling of dislocation between those who collect data at maternity units and those who formulate policy on data collection at national level:

59. Our witnesses told us that they would value central direction on the collection of maternity care data. Some went as far as to say that all maternity units should use an identical computer system. Christopher Guyer, Clinical Director of Obstetrics, St Mary's Hospital, Portsmouth suggested that such a system "would not only be able to produce records locally but would also be able to produce records nationally across the country."[69] However, other witnesses pointed out that the variation in the kinds of maternity unit which operate across the country meant that exactly the same system might not be appropriate for all units. Professor James Walker from St James's, Leeds told us that:

60. Throughout this part of our inquiry, our attention was drawn to the system of maternity care data collection in Scotland, where data relating to all admissions to maternity units have been collected on the SMR02 form for over twenty years. Several witnesses agreed with the RCOG that the form seemed to be "simple and readily completed" and also noted that annual accounts of perinatal outcomes drawn from the data were published on the website of the Scottish Perinatal Mortality and Morbidity Review.[71]

61. We recommend that in reviewing policy on the collection of maternity care data, the Department consider the merits of the system used in Scotland, not only in terms of the system itself but also in terms of other factors which might contribute to its success, such as the allocation of resources and the existence of a culture which supports staff who collect, enter and analyse data.

62. Most of the midwives and doctors who spoke to us did not recognise the requirements of the Maternity HES as a common data set, because they had not heard of them, or because they felt that definitions of the data required were not clear, or because the Maternity HES did not correspond with the more detailed information they collected independently for the purposes of care for individual mothers and babies, and development of their service. This is an indication of the disparity between national policy on and local knowledge of, collection of maternity care data. If maternity unit data collection systems are to be improved, communications between the Department and individual trusts and maternity units must be strengthened. We recommend that the Department should set out the implications of the electronic patient record initiative for maternity care data systems, including agreement of data definitions for maternity care, and further that it should consult and communicate with trusts on developments relating to the minimum dataset required by the Maternity HES.

63. We believe the current state of maternity care data systems at units across the country to be so grave as to warrant specific attention by PCTs and trusts, and, where needed, the allocation of funds for the purpose of installing and maintaining adequate systems and for recruiting and training appropriate staff to undertake data entry, analysis and system support. We recommend that maternity care data systems should form part of Local Delivery Plans.

29 Maternity tail data items are: first antenatal assessment date, total previous pregnancies, delivery place (actual), delivery place (intended), delivery place change reason, gestation length, labour/delivery onset method, delivery method, status of person conducting delivery, anaesthetic given during labour/delivery/number of babies/sex (baby), birth order, live or still birth, birth weight, resuscitation method, birth date (baby), birth date (mother). Back

30 Ev 144 Back

31 Ev 155 Back

32 Ev 117 (Professor Philip Steer) Back

33 Ev 130 Back

34 Ev 146 Back

35 Q 183 Back

36 Ev 188 Back

37 Ev 146 Back

38 Department of Health, Smoking Kills: a white paper on tobacco, Cm 4177, November 1998, Chapter 9.5 Back

39 Health Development Agency, Cancer Prevention: a resource to support local action in delivering the NHS Cancer Plan, October 2002, p 8 Back

40 Data on smoking cessation are now being collected at local level but concerns remain about methods of collection and duplication of effort. Back