APPENDIX 9
Memorandum by Dr Jean Chapple (MS 14)
BIOGRAPHICAL DETAILS
Jean Chapple is a consultant in public health
medicine with a special interest in obstetric and perinatal epidemiology.
She qualified as a doctor from Liverpool University and worked
in clinical posts in obstetrics and paediatrics before moving
to London. Jean is the director of the North Thames Perinatal
Public Health Unit, which runs a malformation register, a maternity
care denominator database and the Confidential Enquiry in Stillbirths
and Deaths in Infancy and the Confidential Enquiry into Maternal
Deaths.
Jean has had a career long interest in maternity
data. Together with clinicians from St Mary's Hospital, Paddington,
in the 1980s she helped to develop a clinical maternity database
that also supports audit, epidemiology and management of maternity
care. A large number of units in the former North West Thames
Region and some units outside the region use this system.
In 1997, Jean was given a two-year research
and development contract from North Thames R&D Directorate
"Maternal and child health systems—validation and demonstration
project (MACHIS)". This was set up to investigate whether
routine maternity and community child health information systems
could be used to provide reliable population based clinical data
on mothers and children for public health and epidemiological
research purposes. The objectives were to:
— assess existing maternity and child
health information systems in the Thames regions for content and
comparability of data, and to validate a proportion of these data;
— create a regional maternity database
by pooling maternity data from each system in use;
— demonstrate the utility of the maternity
database through a study of the geographical variation in the
incidence of low birth weight; and
— assess the feasibility of linking
the child health data in one inner London community trust with
information from maternity systems.
The outcomes of this study are available as
a report and as the proceedings of a conference.
Jean's work on maternity data at a national
level includes—
— Member of NHS Information Authority
Board for NHS Numbers for Babies project 1999-2003.
— Member of NHS Information Authority
quality assurance group for Maternity Data Project 1999-2000.
— Member of NHS Number Caldicott and
Babies Project Board 1998.
— Membership of the working group to
develop a national patient held pregnancy health record (funded
through "Changing Childbirth" development funds) 1995-2000.
— Member of normal pregnancy outcome
indicators working group commissioned by the Department of Health
1997-98.
— Member of Audit Sub-committee, Royal
College of Obstetrics and Gynaecology 1990-1994.
— Member of working group to develop
audit standards for incorporation into contracts for maternity
services.
— Member of steering group for the
national audit of neonatal (Guthrie) screening, organised by the
RCP 1994-96.
— Member of Working Group of the Registrar
General's Medical Advisory Committee on the OPCS Monitoring Scheme
for Congenital Malformations 1993-94.
SUMMARY
1. Maternity data are important as a measure
of maternal and child health. Data items are relatively easy to
define and count.
2. The best way to collect reliable, validated
high quality data is for the data to be captured in the course
of clinical care by clinicians. Incentives for clinicians to collect
high quality standardised data include easy access to their own
data for analysis to review personal practice, audit, evaluation
and research and useful hard copy outputs from computer systems
to file in records and pass information on to primary care.
3. Collecting written data in a standardised
patient held record has been shown to be effective in at least
one English region. A national record would minimise the resources
needed to develop, produce, provide training and implement written
records. It would also minimise costs in translating patient focussed
records.
4. Scotland aspires to produce a national
maternity record. Experience of trying to introduce a national
record in England suggests that there needs to be a common data
set but two or three versions differing in presentation and format
to allow some degree of local choice across units serving a birth
population of 50,000 to 100,000 births a year. Training in the
use of a record is also essential.
5. Clinical computerised information systems
collect high quality data in accessible and analysable forms,
but are rarely integrated into hospital computer systems or link
with child health systems and remain stand alone. This means that
double data entry and transcription errors are inevitable and
that data collected for management purposes are inaccurate and
not passed on to national information systems such as HES.
6. A consensus agreed by national experts
at a 1999 conference in London concluded that the way forward
to produce high quality maternity dated linked to child health
outcomes was:
1. Investment in the maternal and child health
information infrastructure.
2. The provision of the NHS number at birth
rather than at registration.
3. Ensure the early development of a core
essential dataset for mothers, babies and children.
4. Establish electronic linkage of maternal,
neonatal and child health information.
5. Establish an organisational infra structure
at national and regional level.
These recommendations are still the key to developing
integrated high quality maternity data collection. Unfortunately,
for a variety of reasons, little progress has been made apart
from the successful NHS Numbers for Babies project which went
live on 29 October 2002.
7. A standardised birth notification minimum
maternity data set which can be transmitted electronically to
child health systems is urgently needed to maximise the outcome
of the NHS Numbers for babies project and to lead the way to electronic
patient records for mothers and babies.
1. BACKGROUND
1.1 Data from maternity units are important
as they give a measure of the health of not only pregnant women
but also their babies. Data on its mother's health play an important
part in a child's health record and maternal and child health
data should be linked.
1.2 Maternity services are fortunate in
that they involve discrete and countable events. There are no
problems in defining whether pregnancy is present—unlike
diabetes or hypertension, a woman cannot be "a little bit
pregnant". Similarly, there are no problems in defining when
a pregnancy has ended. There are UK national laws on registering
and notifying births so some vital pieces of information for each
pregnancy could be collected on a national scale. Why then is
the collection of data from maternity units a cause for concern?
2. WHO COLLECTS
GOOD QUALITY
DATA?
2.1 The only way to collect high quality
maternity data is to ensure that it is captured in the course
of routine clinical contacts. This affects both the quality of
data collection and interventions offered. For example, data are
needed to monitor progress against Local Delivery Plan targets
for Primary Care Trusts (PCTs). The Planning and Priorities guidance
on reducing health inequalities (http://www.doh.gov.uk/planning2003-2006/appa.htm)
requires PCTs to "deliver a one percentage point reduction
per year in the proportion of women continuing to smoke throughout
pregnancy." Little thought has been given as to how to collect
data to see if these targets have been met. We need a standard
definition and time scale of "continuing to smoke" and
the time at which a woman is asked. If a midwife asks a woman
about whether she smokes, not only can the midwife collect the
data required for monitoring the target, but she can offer interventions
to make the target achievable. This opportunity is lost if data
are collected by lay clerks or by tick boxes on forms.
2.2 The best quality data are therefore
captured in the course of care by clinical staff who understand
why they are seeking the data items and the importance of collecting
data correctly using agreed definitions. The best national data
come from a "bottom up" approach using data collected
in the course of care to build up information on practice and
outcomes in the maternity unit, in the surrounding locality and
across regions. These data should be built on individual mothers
and babies—"episodes of care" have little meaning
in a clinical context and cannot be used for audit or evaluation
of clinical services.
2.3 In return, clinicians must have local
access to the data they produce. The "carrot" of having
access to pooled data to look at their own individual case mix
and practice, help with audits and research encourages clinicians
to take pride and care in data collection and makes up for the
"stick" of writing and entering data items. These incentives
are missing from data coded by clerks for management purposes.
2.4 Computerised maternity data systems
can also offer carrots in the form of easy and quick transfer
of data to others who need to know about the health of mother
and baby—for example, computerised maternity systems offer
printouts for community midwives, health visitors and staff in
primary care, or other acute units. If such transfers can be effected
electronically, this prevents duplication of effort in data entry.
3. HOW IS
GOOD QUALITY
DATA COLLECTED?
3.1 Written records
3.1.1 "Changing Childbirth" led
to wide spread adoption of patient held records so that written
data are readily available to both clinicians and the women and
families they care for. In NW Thames, as many as 25% of women
booking at each unit do not finally deliver there. Some women
miscarry and many change address and place of care during their
pregnancy. Transfer of written records from one unit to another
via the woman ensures that tests are not needlessly repeated.
3.1.2 Some regions, such as the West Midlands,
adopted the same patient held record successfully. This enabled
clinicians to know where to look for information and standardised
the data recorded.
3.1.3 The "Changing Childbirth"
initiative funded a national Maternity Record Project in 1995.
The unique feature of these notes is that they are directed to
the woman ("What is your name?" rather than "Name")
and explain clinical terminology and give sources for further
information. Scotland proposes to have a national maternity record
in its recently published national service framework for maternity
care.
3.1.4 The other advantages of a standardised
tool for recording maternity data are:
— Cost—printing large quantities
reduces costs and allows up dated editions to be published readily.
— Translation costs—evaluation
of the antenatal module showed that women whose first language
was not English liked the notes, as it was easy to take them home
and go through them with someone who could read and write English.
This gave them a good idea of the information about their health
and past and present pregnancies that was needed. If developed
on a large scale, translated guidance on the notes could be readily
available.
— Training on completing the notes—videos
and written guidance on how to keep accurate maternity records
can be developed and disseminated to staff.
3.1.5 The Maternity Record project has been
difficult to implement nationally across England and Wales for
a variety of reasons. The project has developed only an antenatal
module to date and many units chose to wait for implementation
until a full antenatal, delivery and postnatal set of notes were
developed. There was also much discussion about the format of
the notes—some clinicians felt the notes had become too
skewed towards the consumer of care and said that they found it
difficult to find clinical data. In retrospect, more emphasis
should also have been given to training about how to use the notes.
Many units also complained that the format of the notes did not
match the data entry screens on their computerised maternity data
systems and so involved constant turning of pages to input data.
3.1.6 The success of developing a standardised
maternity record in a region and the proposal to implement a record
in Scotland suggest that such a project succeeds best at a birth
population of 50,000 to 100,000 births a year rather than at a
national level of 600,000 births a year. Consideration should
be given to developing a standard set of data to be kept in written
(and, in time, electronic) records but with two or three different
formats to allow some degree of clinical choice locally.
3.2 Electronically recorded data
3.2.1 A survey done by the NHS Information
Authority for the NHS numbers for babies project (www.nhsia.nhs.uk/nn4b/)
showed that two thirds of the 300 or so maternity units in the
UK had a computerised maternity data collection system. There
were ten main suppliers of systems while 13 units had in house
systems. Few of these connect to hospital wide patient administration
systems (PAS), so there is often duplication of data entry within
a unit for management data. Not surprisingly, clinical staff give
greater effort to collecting clinical data for the maternity system
than to ensuring management data are available. This may explain
why national data collection through HES is poor.
3.2.2 Electronic data collection should
in theory allow ready transfer of information across the NHS.
However, the Maternal and Child Health Information System (MACHIS)
project in London showed that despite computerisation in parts
of London, there was little direct transfer of data direct from
maternity systems to child health systems. Midwives printed out
birth notifications from their computer system and sent hard copy
to the child health department, where the data were re-keyed.
This leads to duplication of effort and transcription errors.
Experience across the NHS shows similar regional patterns of development
of computerised maternity data as with written records. It appears
easier to get agreement and commitment to developing and using
a standardised computer system across 20 to 30 units and a birth
population of 50,000 to 100,000 births year than to try to develop
national systems.
4. WAYS FORWARD
4.1 The researchers on the London wide MACHIS
project held a national conference in January 1999 for an invited
audience of experts on maternal and child health data. The conference
proceedings (Chapple J, Golightly S, Charles Z, editors 2000,
Linking data for better health in pregnancy and childhood. Proceedings
of a conference held at the King's Fund, 12 and 13 January 1999.
London: CASPE Research ISBN 1 898845 15 8) ended with a consensus
statement agreed with over 60 people attending and is quoted in
full here.
4.2 "There is a well-documented lack
of high quality clinical information held nationally on maternity
and child health. However, at local level considerable time is
devoted to recording data about women, babies and children.
4.3 There is a long-standing but never the
less urgent need to prioritise information in regard to maternal
and infant health in order to best serve the health of the public.
The importance of maternal and child health is well recognised
internationally. The need for linked information on mothers and
babies readily available to clinicians for care or audit and to
health planners has been long recognised in both government policy
and documents from professional organisations in England and Wales.
Despite this, there is a well-documented lack of available clinical
information on both maternity, neonatal and child health service
activity and subsequent child health outcomes in England and Wales.
4.4 Though Information for Health: the
Information Strategy for the New NHS is welcome and, when
implemented, will lead to a great improvement in the ability of
the NHS to deliver and monitor health, there is no mention of
the importance of or the unique issues regarding maternal, infant
and later child health information. As a result, this area may
fail to gain the priority it deserves. Moreover with the rapid
changes occurring in the reorganisation of community services,
the maintenance of comprehensive population based child health
information including that on special needs is under threat.
4.5 The unique aspect of maternity and child
health information is that mother and baby are linked in life,
as they should be in the provision of information. A central outcome
for maternity services is the health of the baby. Those providing
or monitoring maternal care need to know health outcomes in the
children. Those providing infant or child care need to know aspects
of maternal and family health and about maternity services. It
is well recognised in research that key influences for child health
arise before birth. The need for information applies equally to
the individual mother and baby as to populations.
4.6 The benefits to the NHS from ensuring
linked maternal and child information would be immediate and significant.
These include:
— Improved delivery of care to mothers
and babies across organisational boundaries.
— More effective and efficient clinical
audit.
— Valid national statistics not currently
available.
— Monitoring long term outcomes of
care such as obstetric and neonatal care.
— More efficient epidemiological, economic
and clinical research.
— Improved monitoring of the health
of whole communities.
4.7 Implementing these recommendations will
enhance the ability of the NHS to deliver key priorities of current
health policy such as:
— The effectiveness of the National
Institute of Clinical Effectiveness (NICE) and the Commission
for Health Improvement (CHIMP).
— Health improvement programmes.
— Primary care group developments.
— Reduction of health inequalities
(where maternal and child health is a recognised priority).
— The implementation of an electronic
patient record and an electronic health record.
— Health Action Zones (HAZ) and Education
Action Zones (EAZ).
— Confidential Enquiry into Stillbirth
and Death in Infancy (CESDI).
4.8 Recommendations:
1. Investment in the maternal and child
health information infrastructure. This is an essential prerequisite
to the collection and linkage of high quality clinical information
on maternity and child health. Nationally only two-thirds of maternity
units are currently computerised and although all child health
departments use some sort of computer systems to assist with service
delivery some of these systems are outdated and unsupportable.
2. The provision of the NHS number at birth
rather than at registration. Essential to the ability to provide
comprehensive health information on new-borns and to link it to
maternal or later child health is the provision of the NHS number
at birth. The current NHS Number Caldicott and Babies Project
is likely to recommend this for early implementation. This will
immediately facilitate the delivery of care to babies and enable
the linkage of maternal and infant health information called for
in this document. Under no circumstances should this project be
delayed.
3. Ensure the early development of a core
essential dataset for mothers, babies and children with a set
of nationally agreed definitions, collected electronically with
on line validation checks. Datasets are in the process of being
agreed by the Royal College of Obstetricians and Gynaecologists
(RCOG) (Obstetric data), the British Association for Perinatal
Medicine (BAPM) (obstetric and neonatal data) and the Child Health
Informatics Consortium (CHIC) (community child health). Some of
these sets are larger than the essential core of data needed by
each of the three clinical groups, but there is a core of data
which is common to all of them, which should be agreed as soon
as possible.
4. Establish electronic linkage of maternal,
neonatal and child health information. This requires:
— The NHS number issued to babies at
birth.
— Integration of existing policies
on data collection.
— Dedicated funding for pilot studies,
links, validation and professional involvement.
— Dedicated and named lead professionals,
both regionally and nationally.
— Clear accountability for the linkage
process.
5. Establish an organisational infra structure
at national and regional level. This will act as a local clearing
house for routinely collected maternal, neonatal and child health
data as a linkage centre for quality reviews such as the Confidential
Enquiry into Stillbirths and Deaths in Infancy (CESDI) and congenital
anomaly registers. A national network of regional perinatal public
health units could provide such a service."
5. CURRENT PROGRESS
TOWARDS INTEGRATED
QUALITY DATA
COLLECTION IN
MATERNITY CARE
Almost exactly four years later, what progress
has been made?
5.1 Investment in the maternal and child
health information infrastructure.
The situation remains almost unchanged, with one
third of maternity units still not having any computerised maternity
system apart from the NHS Numbers for Babies Interim NHS Numbering
System (INNS).Some of the delays have been caused by uncertainty
in Trusts for future requirements for systems to deliver electronic
patient records and lack of integration of clinical information
systems into hospital administration systems.
5.2 The provision of the NHS number at birth
rather than at registration. This NHS Information Authority project
reached a successful conclusion on 29 October 2002 when systems
to issue NHS numbers at birth as part of the statutory midwifery
birth notification system were put into action. Further work needs
to be done to ensure that the NHS number is available for NHS
use—anecdotal examples suggest that the NHS number is not
currently available on Guthrie cards for neonatal screening as
it cannot be printed as a bar code. Neonatalogists also report
problems in obtaining the NHS number.
5.3 Ensure the early development of a core
essential dataset for mothers, babies and children.
5.3.1 An early NHS Information Authority
project set out to develop a core essential data set for maternity.
Clinicians hoped that this would lead to a small data set of 20
to 40 crucial clinical items which every clinician would collect
using a nationally specified definition. Unfortunately, this did
not occur and the only outcome of this project was a maternity
care data dictionary (www.nhsia.nhs.uk/datastandards/pages/mcdp/asp
). Further progress on this project seems to have stalled.
5.3.2 The clinicians advising the NHS Numbers
for Babies project had hoped that a national birth notification
core data set would be developed as part of the initiative and
had made plans to consult on a data set. However, consultation
with the NHS Information Authority Caldicott guardian led to an
opinion given to the IA that this aim could not be achieved, as
it would break confidentiality rules if clinical data were collected
without consent and sent to the Central Issuing System. As birth
notification has been required by law for nearly a century, the
clinicians queried the need for consent—unfortunately, the
early 20th century legislators had failed to specify the information
need for a birth notification and had merely stated that the local
medical officer of health must be told of a birth. This explains
the myriad different birth notification forms across the country
and prevented the development of a core data set as part of the
successful NHS Numbers for Babies project.
5.4 Establish electronic linkage of maternal,
neonatal and child health information. The NHS Numbers for Babies
project has provided a means for transfer of NHS numbers attached
to each baby's details to every child health system. The lack
of a standard birth notification minimum data set has severely
limited the use of the opportunity to transfer clinically relevant
data.
5.5 Establish an organisational infra structure
at national and regional level. Many regions used funds from the
Confidential Enquiry into Stillbirths and Deaths in infancy (CESDI)
to support denominator data collection as a background to pregnancies
in which mothers lost their babies. In some regions, health authorities
made additional funds available for maternity initiatives. The
results in regions where these investments have been made are
standardised data collection systems backed by enthusiastic local
clinicians. However, progress made so far is threatened by the
reconfiguration of regions, the difficulty of engaging new Primary
Care Trusts into investing in projects crossing PCT boundaries
and the assimilation of CESDI into NICE. Reconfiguration of CESDI
has resulted in a large cut in budget and the first casualties
of this are the denominator data "add ons" that some
regions developed.
6. CONCLUSION
Clinicians and epidemiologists are agreed on
appropriate ways forward for maternity data collection. There
will inevitably be some debate about the finer points of a national
minimum data set but this is not insurmountable. A standardised
birth notification is urgently needed to maximise the effects
of the NHS Numbers for Babies project and future developments
in electronic data collection.
| 
