Previous Volume Next Volume
Volume 3. Maternal and Neonatal Datasets
Volume Editors: Rupert Fawdry and Helga Perry
If interested in the possibility of editing this volume Click here
To see a sample page Click here
Discussion Link: http://eepd.org.uk/?page_id=499
Help us all to have Open Access to any relevant perinatal datasets.
Send copies (As a ".pdf" or ".doc" to email@example.com
"In order to create a future-proof joined-up computer system,
it is essential to start with a comprehensive, open-access analysis
of ALL the existing relevant paperwork and datasets.
Intermittent short-term, unpaid expert advisors cannot take adequate account of the needs of all stakeholders"
Reason for the EEPD
“On both technical and financial grounds, integrating (or even adequately linking) complex independent flow-patterned databases is virtually always impossible. For this simple reason, structured Electronic Patient Records (EPRs) will only attain their true potential to improve the quality of patient care and reduce the risk of human errors, without unsustainable data re-entry overload, when - following intense, open, web-based discussions, their detailed, logically and chronologically arranged, flow-patterned questions and the full range of all allowable answer-options - [unlike traditional coding, always including whenever needed “Unknown (Free Text)” and “Other (Free Text)”] - are, by stages, taking account of as many interested parties as possible, individual question by individual question, internationally standardised.”
Computers will never, ever, cost-effectively cope with the following chaos of incompatible paper and electronic datasets. What follows is my best attempt so far at some kind of classification:
"This Volume of the EEPD is just the start of this process of open-access documentation! Do not be put off by its current patchy nature - If we are to achieve the quality of perinatal software that we need it will in time need to be properly updated by all of us."
Cost-effective perinatal computer systems will never be able to cope with the number and variations of data items which I started to document in Nightmare Findings. With a maximum U.K. market of about 200 such systems, the cost of the detailed programming to cope with incompatible data will always be prohibitive.
The logical next step has been gradually to build up a collection of official paper-based datasets, either proposed or currently required, to see how all these variations might be slowly reconciled into a single universal set of electronic questions and all allowable answers. This essential foundation set of documents has after 30 years led to what is now Volume 3 of the EEPD.
Few tasks in my whole life have proved to be more difficult or more time consuming than this Volume of the EEPD, mainly because the huge variety of relevant material makes it extremely difficult to create any logical classification. The paper documents I have collected almost always contain data items in fairly random order that overlap between different Clinical, Statistical, Audit, Management and Individual Patient Care requirements and between “Mother” (One set per Mother’s Birth Event), “Baby” (One set per Baby) and “Special Neonatal Care” Items (One set per Neonate requiring Intensive Care) etc.
What follows is my best attempt so far at some kind of classification.
This massive task has also turned out to be the best way to try to take full account of the many stakeholders wishing to use maternity and neonatal data; and should really have been done in detail before the creation of the very earliest Electronic Maternity and Neonatal IT systems! A task which, in my own case should have been done, had I had the time and resources, before I designed the original Protos/iSoft Maternity IT system back in 1990!
Although many of these datasets are now no longer in use, they remain valuable, both because they were the original basis for what have become Volume 4 (The Resource Document) and Volume 5 (The Logical Priority Core Data Definitions); but, more importantly, because even outdated versions provide permanent insights into the different perspectives of those needing to use the data.
Earlier versions of what is now the EEPD formed the “knowledge engineering” basis for (a) the Protos/iSoft maternity computer system (the most successful commercially marketed U.K. maternity computer system), (b) the West Midlands Perinatal Institute MANNERS Maternity data set and (c) the Government funded Maternity Care Data Project, as well as several other initiatives. (See EEPD. Appendix A. Initiatives)
(Updated 17 Sep 2010)
Request for any Missing or Outdated Datasets.
The future of cost-effective electronic patient records will eventually depend on high quality international data input standardisation. Open access world-wide, through the EEPD, to as many current incompatible paper or electronic datasets relevant to perinatal care as possible, will greatly facilitate this unavoidable process.
If you can provide an electronic or paper version of any updated or missing datasets; especially the full documentation for any complex, detailed, chronologically arranged, flow-patterned commercial maternity systems, please get in touch urgently.
The wording found in
still occurs far too often.
(Updated 17 Sep 2010)
At first, collecting datasets seemed like a relatively simple task, and indeed was an essential basis for the creation of any good maternity system such as the design for the original Protos maternity system. My collection started with documenting what was required for Birth Notification, Birth Registration and the original Korner committee recommendations.
As soon as Protos sold maternity computer systems in Scotland, note needed to be taken of the Scottish SMR datasets.
But since then the task has again grown and grown.
Confusion became greater and greater, with more and more different datasets being documented.
And the U.K. taxpayer was nearly always ultimately required to provide funding for the data entry work of the majority of these different datasets!
Yet almost all of these datasets are virtually independent in their data selection and definitions and therefore required fresh data entry onto each different computer system!
(see www.fawdry.info/eepd/00_ima/ohps/Mountain.pdf and a more detailed discussion in www.fawdry.info/eepd/01_ess/b_eprs/B01_Railways.pdf)
Even more worrying was the fact that many of these data collections, (e.g. The Cardiff Births Survey, the Oxford maternity database) have been running for many years and have required major human efforts and taxpayers’ funding to maintain, yet it has proved almost impossible for me to find out
a) When did that database start collecting data?
b) Is it still running?
c) What data items form part of that database?
d) Is the data still accessible?
e) If still accessible, how and by whom?
Despite working first in Cardiff for six months and later near Oxford for ten months, it proved impossible to find all the answers to the above crucial questions. The originators of the data collection project had often retired or moved on, and funding had been a problem, or the old database had been superseded by a new computer system that required different questions and answers. Again, the selection and wording of each question and each answer seemed universally to depend on the particular views of those directly involved the creation of each dataset. (For a proposed solution see the Introduction to EEPD Volume 5. “Logical Priority”)
It may be objected that several of the datasets included in the EEPD are either outdated or were only proposals; but despite the fact that the item selection and wording in such datasets seems almost arbitrary, they remain useful (a) in providing some insight into the kind of data required by each special interest group and (b) more importantly, they have provided the essential material for deciding on the best content and wording both for the Resource Document (Volume 4) and the Logical Priority document (Volume 5).
Any offers to add to this personal, and somewhat random, selection are very welcome (in MS Word “.doc”, Mac “.pages”, “.pdf” “.rtf” or “.txt” if possible).
Whenever it is intended that a dataset should ever have any reasonable relationship to an electronic patient record or to be compared with any other similar database, it must first be re-organised until it is:
i) Logically Divided. e.g. Every Mother/Birth Event (once for each mother), separate from Every Baby (once for every baby or fetus), separate from Every Special Care Baby, separate from Every Anaesthetic etc
ii) Set out Chronologically and
If a dataset is presented without these three crucial characteristics, then
a) It seems that no-one involved in creating the dataset has bothered to think about its relationship to other obligatory datasets; and to any subsequent Electronic Patient Record Database
b) As a result it is almost impossible for clinicians to do any reasonably efficient quality checks.
How sad, not to have done such re-sorting when each dataset is first proposed; and not to have taken any adequate account of the very simple principles set out in: Credible Perinatal Datasets.
So far, apart from my “Logical Prioritisation” dataset, i.e. the EEPD Volume 5, I have not come across a single open source perinatal dataset anywhere, which is thus chronologically arranged and flow-patterned. Yet all complex maternity systems such as iSoft/Protos or Euroking or Terranova or TRAX depend on such flow-patterned datasets in order to function but their versions are not yet openly available; and may even be specifically described as commercially confidential.
In future all health service purchasers of any clinical IT system must demand that such systems will be considered for purchase only if all the questions, and all allowable answer options, in flow-patterned, logical and chronological order, are in the public domain; and are available on the internet and in a printed out format for full discussion by all the front-line users, before any contracts are signed.
The present chaos will inevitably continue until this becomes mandatory.
A typical example of a total failure to consider the needs of other datasets is provided by the “New” Birth Notification which is presented both in its original form and then as re-sorted chronologically.
Rupert Fawdry. (Updated 10 Nov 2010)
For an overview of our current chaos see General Message and Summary.
Next Volume Next Page