BJA Advance Access originally published online on May 26, 2006
British Journal of Anaesthesia 2006 97(1):69-76; doi:10.1093/bja/ael138
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Integrating anaesthesia and intensive care into the National Care Record
Department of Anaesthesia, Torbay Hospital Torquay, Devon TQ2 7AA, UK
*E-mail: roger.tackley{at}nhs.net
 |
Abstract |
|---|
 Top Abstract IntroductionThe NPfIT in the...The SCREPR in secondary careIntegration in the NCR...Methods of integrating clinical...Anaesthesia and intensive careDiscussionReferences |
|---|
The National Care Record for England is planned to be delivered as part of the National Programme for Information Technology (NPfIT) by the National Health Service Connecting for Health. It will be made up from a National Summary Care Record, Local Detailed Care Records and from images in Picture Archiving and Communication Systems. Full benefits for clinical care will only come when there is true integration of the clinical records systems which enables rapid clinical decision support, a consistent user interface, single entry of data items and analysis of information across the full spectrum of clinical care. Currently there are few hospitals with fully electronic anaesthetic or critical care systems, and these are only partly linked to the hospital systems. This limits their benefit to patient care and health care staff. As NPfIT is being mandated for all hospitals in England it is essential to consider now how the next generation of anaesthetic and intensive care systems will integrate with it.
Keywords: National Programme for IT; records, anaesthesia, computerized; records, integration of computerized clinical records; records, intensive care record; records, National Care Record
|
Introduction |
|---|
|
TopAbstractIntroductionThe NPfIT in the...The SCREPR in secondary careIntegration in the NCR...Methods of integrating clinical...Anaesthesia and intensive careDiscussionReferences |
|---|
The National Care Record (NCR) is planned to be delivered across England as part of the National Programme for Information Technology (NPfIT)1 by the National Health Service (NHS) Connecting for Health (CfH). It will be made up from a National Summary Care Record (SCR), Local Detailed Care Records and images in picture archiving and communication systems (PACS).
The reasons for moving towards electronic health records were laid out in the 1998 NHS Information Strategy2 and described as ‘compelling’. This is now reflected in plans worldwide. In the USA, the president made a decision for widespread adoption of electronic health care records in 2004. In Europe, an action plan was adopted by the European Commission in April 2004 to make use of e-Health technologies such as electronic records, patient identifiers and health cards, to support better quality health care.
The SCR will hold limited information about every person (if they allow it) on the NHS ‘Spine’. An initial version of this will be populated later in 2006 but it is unlikely to be fully populated for another 3 yr.
A detailed care record will be held in every NHS organization but will only document the local clinical details of the patient's care. Some of this will be shared either with the Spine and other organizations. In this article, the detailed care record will be referred to as an electronic patient record (EPR) which was originally defined in the NHS Information Strategy of 1998.2 The original target was for all trusts to have a ‘level 3’ EPR by 2005. However, there are still very few EPRs in England, and many of the community projects were stopped when the NPfIT was announced such as the Shires EPR project in the South West.3 In most cases the implementation will be staged to allow the organizations to adapt to new working practices and some local service providers (LSPs) have already installed early versions which usually consist mainly of a new patient administration system and electronic ordering and results for pathology.
PACS is now being introduced as a core part of NPfIT. It will only support radiology imaging to start with, but eventually all clinical images, such as clinical photographs, clinical diagrams and fibreoptic endoscopy images will be available.
Within the EPR there will be ways of recording anaesthesia and intensive care unit (ICU) information but this will not necessarily look similar to an electronic version of the paper record. Most of the NPfIT systems are still in the early stages of development of specialist clinical areas but there are some suggested designs and principles. With the current timescales it is unlikely that more than 25% of hospitals will have fully integrated anaesthesia and intensive care systems before 2008.
This article discusses how the NCR service4 (NCRS) plans to achieve integration of clinical systems. It explains the need for integration at many different levels ranging from the physical connection of medical devices through to the structure and presentation of the clinical record. All of these factors impact on how to integrate anaesthesia and intensive care records into the NCR. Some principles of how to proceed towards integration are suggested at the end.
|
The NPfIT in the NHS |
|---|
|
TopAbstractIntroductionThe NPfIT in the...The SCREPR in secondary careIntegration in the NCR...Methods of integrating clinical...Anaesthesia and intensive careDiscussionReferences |
|---|
NPfIT, the largest single IT project ever undertaken, was launched in 2002. So far about £18bn worth of contracts have been let, to run over 10 yr. Many see this large scale as a risk, but the reasons for running it as one project are to be able to take advantage of the economy of scale with computer system suppliers, and to encourage a consistent approach to health information across England. Scotland, Wales and Ireland are planning their own strategies, probably to be influenced by the success of NPfIT.
Apart from the NCRS there are three other main components to NPfIT (Table 1). High speed internet connections have now been made to the majority of hospitals and general practice surgeries in England and Scotland. This has a firewall connection to the Internet allowing more secure communication between health care sites and also the provision of some NHS only websites for business purposes. There is now an NHS e-mail service which allows NHS employees in England and Scotland to have one permanent e-mail address wherever they are based. ‘Choose and Book’ is the name given to a new electronic booking system for patient appointments. Although only in its infancy, it will enable authorized users such as general practitioners and eventually patients to book appointments for clinics online. Patients will soon be able to benefit from the Electronic Transmission of Prescriptions service which will replace more than 300 million paper prescriptions with electronic versions sent from a doctor direct to the dispensing pharmacy.
|
The consistent approach should mean that whatever system is being used there will be similar presentations of clinical records using the same underlying clinical information standards across health care organizations. This will improve communication between clinical systems, reduce the burden of training staff on new and different systems, and allow comparable analyses of information for clinical governance, epidemiology and management reports.
The NPfIT is divided into five clusters (Fig. 1), each with its own (LSP) (Table 2) to manage local computer systems in hospitals, GP surgeries, NHS community organizations and mental health. To reduce costs and enable consistency, LSPs are encouraging most organizations in a cluster to use the same systems. For secondary care the most important of these will be the EPR. Recently a decision has been made to allow GP surgeries to continue using their existing systems as long as they are made compatible with NPfIT communication standards.
|
|
|
The SCR |
|---|
|
TopAbstractIntroductionThe NPfIT in the...The SCREPR in secondary careIntegration in the NCR...Methods of integrating clinical...Anaesthesia and intensive careDiscussionReferences |
|---|
From later in 2006, the SCR will be populated with major diagnoses, allergies and adverse reactions, procedures performed and medications. This information will be an extract from GP systems. Patients will be asked first if they allow the sharing of this information but there has been much discussion about whether they should have to ‘opt out’ or be expected to ‘opt in’. The Care Record Development Board had previously recommended opting out.5
The structure of the SCR will be based on care record elements (CREs) (Table 3) explained in the internal CfH document—NPFIT-NCR-DES-0135.05 Version 3.0 April 2006. It is planned that more CREs will be added after each care event. These are defined as ‘... a discrete period of time, usefully identifiable by the patient or Care Professional, during which a Care Professional, and/or the patient, and/or their carer, makes a material contribution to the care of the patient, resulting in a change to a patient's NHS Care Record’. This can be a single consultation as with a GP or a whole episode in theatre or intensive care.
|
It is essential that the SCR will have a national standard format, because it is planned that it will be accessed by any authorized user of an NPfIT computer system, and also by the public via the internet site ‘Healthspace’ (https://www.healthspace.nhs.uk/).
|
EPR in secondary care |
|---|
|
TopAbstractIntroductionThe NPfIT in the...The SCREPR in secondary careIntegration in the NCR...Methods of integrating clinical...Anaesthesia and intensive careDiscussionReferences |
|---|
The 1998 information strategy described in some detail the core components of an EPR (Table 4). These have not changed in the NPfIT and the integrated care record service (ICRS) output based specification (OBS) of 2003 outlined these in more detail.6 These generic functions of a health record apply to all specialties in some form or another and can form the basis of all health records.
|
|
Integration in the NCR service |
|---|
|
TopAbstractIntroductionThe NPfIT in the...The SCREPR in secondary careIntegration in the NCR...Methods of integrating clinical...Anaesthesia and intensive careDiscussionReferences |
|---|
The ‘ICRS OBS 2’ recommends that systems should be integrated. It states ‘... it [the LSP] must describe how its approach reduces the total number of systems in the Authority, provides greater degrees of integration, and allows for controlled migration away from today's silo-based systems...’ An example of this would be enterprisewide scheduling and when booking a patient for a radiology appointment the user should see similar screens and functions as when scheduling any outpatient appointment. The system should also be integrated so that multiple appointments for the same patient and in different areas (such as phlebotomy, pre-assessment and orthotics) can be seen together, and managed so they do not cause conflicts or unnecessary patient attendances.
Clinical decision support can be ‘passive’, for example providing relevant, context-sensitive, guidance-like access to a local protocol or guideline, or it can be ‘active’. Active clinical decision support (CDS) is usually triggered by a patient process in the EPR such as ordering a treatment and then invokes a previously agreed rule for computation. The rule then provides advice after accessing patient information already held in the system. An example might be checking that a heart rate (beats min–1) had recently been checked, and that it was not too low, before prescribing digoxin.
For active decision support to work correctly the system must have rapid access to patient information of many different sorts—investigations, drugs, diagnoses, allergies, vital signs—wherever they have been entered in the system. Sophisticated decision support systems already exist in some ICU systems as reviewed by Martich and colleagues7 which highlights the importance of integration of information.
This is just one reason why an EPR should be a fully integrated system, and not consist of multiple separate modules which have to communicate using complex, relatively slow, messaging interfaces.
|
Methods of integrating clinical systems |
|---|
|
TopAbstractIntroductionThe NPfIT in the...The SCREPR in secondary careIntegration in the NCR...Methods of integrating clinical...Anaesthesia and intensive careDiscussionReferences |
|---|
Traditionally hospitals have a large number of computer systems, some of which are linked using complex interfaces (Fig. 2). Each of these systems will contain its own patient index which has to be kept up to date with the master patient index. To do this specific messages are sent from one system to another every time there is a change in either list. This often entails hundreds of messages per minute but has been used successfully for many years for data such as patient characteristic data and laboratory tests.
|
As systems become more complex and need to exchange all aspects of clinical information, the load on the interface engine will exponentially increase. However, the need for a nearly instant response will remain. It is likely therefore that a system using large numbers of complex interfaces will become inadequate for integration of clinical information.
Integration implies that clinical information needs only to be entered once, and is only stored in one place. This is best achieved by having one database. For instance, when a stored diagnosis is edited, the system only has to look in one place, and it only has to update one place. The risk of data being out of synchronization with its copy in another part of the system is eradicated.
As far as the user is concerned, integration will mean only logging on once, and then being able to access all clinical areas. To ensure security and confidentiality each user needs to have two factor authentication such as a unique card and PIN number. Once logged in and identified, the system can be configured to only allow access to particular items in the database. The NPfIT is specifying a complex system of ‘role-based access’ rules to limit users to specific parts of the clinical system. An example would be limiting ward nurses to be able to see laboratory results but not be able to request them. In addition there will be items of information held in ‘sealed envelopes’5 which can only be accessed by whoever the patient has nominated. For instance a patient may have asked that only the consultant in gynaecology can see the information about a past termination of pregnancy. Every clinical system will have to implement these highly complex rules for their database. If there are multiple interfaced systems, and the user is only expecting to log on once, then the rules will have to be kept in step and activated for each system. This again is a massive overhead in computing transactions.
For these reasons and many others, EPRs often use one database to store all information.
Even though the information might be stored in one place, the presentation of the information depends on the purpose for which it is being used. Figure 3 shows an example of a single database, with examples of common information categories, and from this varying views can be generated. For example, a theatre system view will need to access patient characteristic data, scheduling and probably administration details, whereas in outpatients the user will need to see views of patient characteristic data, scheduling, clinical correspondence and probably electronic prescribing.
|
EPR systems have not yet developed sufficiently sophisticated views of the information to support all clinical specialties and situations. Anaesthetic record keeping systems have been developing over many years and use very carefully designed screens to allow rapid data entry and optimal presentation of information. It may therefore be necessary for some functions to link to an external system to optimally present and manipulate the data (Fig. 3).
Interfacing between clinical systems will therefore still be needed for:
- linking to specialist systems until EPRs have developed more complex modules;
- linking to medical devices;
- linking to the NHS spine. Links to other organizations will probably also be via the NHS Spine (Fig. 4).
|
Interfacing requirements
Communication between computer systems requires defined standards. Figure 5 shows examples of where agreed standards are needed. The International Standards Organization (ISO) published a model for interconnection called Open System Interconnection (OSI)8 (Table 5) which has seven levels. It starts with the physical connection between the devices, for example specifying the voltages and the pattern of bits, then describes networking, and ultimately how the application will communicate the message. ‘Health Level 7’9 (HL7) messaging standard refers to the seventh level of this model.
|
|
The ICRS OBS mandates many standards including Systematized Nomenclature of Medicine (SNOMED) Clinical Terms,10 HL7, DICOM v3 and IEEE 1073. IEEE 1073 specifies ‘plug and play’ standards for medical devices. This allows the user to connect and disconnect devices at any time and still expect the communication to continue. In Figure 5 it can be seen that medical devices may be connected directly to a local computer (dotted line) or into the EPR via a network hub which allows translation in other messages such as HL7.
DICOM is a standard for Digital Imaging and Communication in Medicine. It is needed to exchange pictures such as X-rays.
SNOMED Clinical Terms is the terminology to be used for communicating all structured clinical information in NHS CfH computer systems. It is the most comprehensive international terminology available, has more than 900 000 terms and is on course for becoming an ISO standard for health care. There are more than 3000 specific anaesthesia and intensive care concepts, including procedures, diagnoses, drugs, critical incidents and measurements.
HL7 v2 is a standard which is widely used for clinical administration and laboratory messages between computer systems. Version 3 includes definitions for describing more complex clinical communications. Part of this is the clinical document architecture (CDA).9 Work has been done by Gardner and Peachey11 supported by the Society for Computing and Technology Clinical Information Standards Group to define an XML (eXtensible Markup Language) schema to describe anaesthetic records and this could then be used in a message with CDA.
Common clinical record structure
One of the key standards needed to allow unambiguous communication of clinical information between systems is a common clinical record structure. The only standard that CfH have proposed is that of the CREs. This will be useful to ensure that the right type of information is passed from a section in one system to the same section in another for example diagnoses, allergies or medication (Fig. 6). Diagnoses recorded in the ICU episode should be stored and related to diagnoses in the rest of that spell or the next. They should also be available to be sent to the SCR as a diagnosis.
|
However, the CREs do not help in defining what content should be in a specialty record, such as for intensive care, or preoperative assessment. Professional standards need to clarify the headings and data items in clinical records. The Royal College of Anaesthetists (RCA) proposed some years ago the recommended anaesthetic record set12 (Table 6). These headings have stood the test of time and most appear in some form in the new XML schema for anaesthetic records.
|
More content for record structures will need to be agreed to support different purposes. For instance what information about an anaesthetic should be sent to the SCR? The Association of Anaesthetists of Great Britain and Ireland information strategy group has considered this and proposed the following as a minimum.
- Name of anaesthetist(s) involved in the care record event.
- Procedures performed, for example sedation, general anaesthetic and neuraxial block.
- Risks and warnings, for example difficult intubations, abnormal pseudocholinesterase, malignant hyperpyrexia status, allergies.
- Findings of adverse events; for example pneumothorax, crown damaged, persistent loss of sensation after epidural.
|
Anaesthesia and intensive care |
|---|
|
TopAbstractIntroductionThe NPfIT in the...The SCREPR in secondary careIntegration in the NCR...Methods of integrating clinical...Anaesthesia and intensive careDiscussionReferences |
|---|
Only a minority of English hospitals have fully functional anaesthesia or intensive care record keeping systems in spite of the benefits that have been described for over 10 yr.13 The majority of these are standalone but some have interfaces to the master patient index, laboratory systems and other medical devices.
The lack of integration with the rest of the EPR has the following types of implications. Electronic prescribing and medicine administration recording is usually limited to that given within theatre or intensive care. For this reason CDS to warn about drug interactions or allergies has not been available as these rely on the system having a full list of the patient's allergies, all medical conditions and the knowledge of what other drugs the patient is receiving. Diagnoses recorded in these systems would probably not be in the context of seeing all the previous diagnoses for that episode and previous episodes. This leads to duplication and possibly errors. Fluid balance would be started afresh as would vital signs recording.
What is the difference between the information requirements for a ward or for critical care areas? They both share the same care processes, such as patient assessments, giving drugs, recording vital signs and performing procedures. These are all covered by the generic parts of the core EPR in Table 4.
One of the differences between generic EPR information and a specialist anaesthetic record keeping system is the way it is presented. Figure 7 shows a medicine administration screen from Cerner, but this is very different from the list of drugs in the Recall anaesthetic screen (Fig. 8).
|
|
Presentation of information in intensive care has the same problems and even with large computer screens it is not possible to display all the information seen on a paper A1 size chart in the same fashion.
Another difference is the amount of detail to record. Vital signs may be recorded every second or in the case of arterial blood pressure continuously by some devices, and for best analysis this detail needs to be retained or very carefully compressed. The EPR database will almost certainly not be configured to contain that high level of detail, or to provide suitable algorithms for data compression, and so some raw data may have to stay in a separate anaesthesia or critical care repository. It will still be available for analysis by the local critical care team but not available to every user of the EPR.
The last main difference is analysis of information. Critical care and anaesthesia need complex real time interpretation of multiple variables—for trends, alarms, artifact rejection, decision support and reports. Scoring systems, Intensive Care National Audit and Research Centre (ICNARC) and the critical care minimum data set all require complex algorithms to interpret the data correctly. Some of these can be supplied by EPR systems (Fig. 9) but many need years of development. In this case the presentation may need to be changed to enhance confidentiality of mortality prediction scores.
|
|
Discussion |
|---|
|
TopAbstractIntroductionThe NPfIT in the...The SCREPR in secondary careIntegration in the NCR...Methods of integrating clinical...Anaesthesia and intensive careDiscussionReferences |
|---|
The NCR is being introduced over the next 6 yr although there are many similar to Cross who question whether CfH will deliver its promises.14 Having negotiated the LSP contracts, the NPfIT has fallen behind on many of its milestone targets and the NCR is the most complex part to get right. One area of tension is that between the standard representation of clinical concepts that computer systems need to store and communicate information unambiguously, and the clinicians' perceived need to record clinical information in a personalized way, embellished by free text. The compromise between these extremes is essential for the success of the NCR and requires an in depth involvement from specialist clinicians who have an interest in health informatics. As there are few incentives for clinicians to be involved in the NPfIT, these solutions are only slowly being achieved. It is the author's view, however, that implementation of even very basic integrated clinical systems will be of great benefit, and lay a good foundation for more complex aspects of clinical record keeping.
The purpose of the NCR is to provide clinicians with access to relevant patient information wherever and whenever it is needed. Full integration of the record improves data quality, presentation of the whole clinical picture, reduces potential system errors, allows faster complex transactions and enables a complete analysis of a patient's health over time and across organizational boundaries.
However, the majority of hospitals still have multiple clinical systems, such as for administration, laboratories, pharmacy and theatres, which are linked by interfaces. This will take many years to change but the direction of travel should be to reduce the overall number of separate systems. It is easier for suppliers to develop standalone or single specialty systems but in future it is important that any new ones become part of the whole IT programme so that, for instance, there is only one type of ICU system per cluster, and it conforms to all the ICRS connectivity standards.
The timescales for achieving full integration are probably past 2013, the end date for the NPfIT procurements but it is essential to consider now the principles for integration in the future.
Proposed integration principles for hospital EPRs
- Create one core of clinical information that is shared in all disciplines. This includes patient characteristic data, investigations, diagnoses, procedures, drugs, fluids and basic vital signs.
- Specific areas such as ICU may need more detail but this can remain local to the specialist area. For example, the full detail of all vital signs traces from monitors would not need to be shared, only an extract.
- Analysis of core information must be possible across all clinical areas. Analysis of detailed information, in conjunction with the core information, should be available to specialist areas.
- Presentation of information must conform to general standards (in the same way as a suite such as Microsoft Office does) but screens must be tailored to the needs of the area. For instance entering information about performing an epidural in the ICU should bring up the same entry screen as in obstetrics, asking for level of block, depth of puncture, etc., but the detail surrounding the need for the epidural and the presentation of vital signs after the epidural will almost certainly be different.
If departments either already have an anaesthetic record keeping system or need to get one within 2 yr, then it is worth considering how to move from this to full integration with an EPR. Before any sort of integration it is important to develop the core EPR first. Ensure that the EPR core information structure can be built on with the detail needed to support critical care areas. If necessary use an anaesthetic or ICU record keeping system in parallel with the EPR to start with. Ensure that the record keeping system conforms to interoperable standards including SNOMEDCT, HL7 and IEEE 11073. The importance of consistent terminology in anaesthesia has been recognized for many years15 and so now there is a rich set of terms within SNOMEDCT to support anaesthesia and critical care. Over time link increasing amounts of anaesthetic clinical information to the core system for example drugs, so that the user of the anaesthetic system begins to use the core EPR information system for most purposes. Gradually, enhance the core EPR system until the presentation and analysis of data is sophisticated enough to be used instead of the original clinical system.
|
References |
|---|
|
TopAbstractIntroductionThe NPfIT in the...The SCREPR in secondary careIntegration in the NCR...Methods of integrating clinical...Anaesthesia and intensive careDiscussionReferences |
|---|
1 National Programme for Information Technology in the NHS. Available from http://www.connectingforhealth.nhs.uk/
2 Information for Health Burns F. An Information Strategy for the Modern NHS 1998–2005 NHS Executive September 1998. Available from http://www.nhsia.nhs.uk/informatics/pages/resource_informatics/Information%20for%20Health.pdf
3 Tackley RM, Jones S, Madden A, Dunnill R. Making the most of the National Programme for IT in the NHS—learning from experience. Br J Healthc Comput Inform Manage 2003; 20:25–7
4 National Care Record Service. Available from http://www.connectingforhealth.nhs.uk/delivery/programmes/nhscrs
5 Care Record Development Board. Available from http://www.connectingforhealth.nhs.uk/crdb/
6 Output Based Specification for Integrated Care Record Service August 2003. Available from http://www.dh.gov.uk/PublicationsAndStatistics/Publications/PublicationsPolicyAndGuidance/PublicationsPolicyAndGuidanceArticle/fs/en?CONTENT_ID=4118312&chk=Qp6rff
7 Martich GD, Waldermann CS, Imhoff M. Clinical Informatics in critical care. J Intensive Care Med 2004; 19:154–63
8 Information processing systems: Open System Interconnection. Basic reference model ISO/IEC 7498:1984
9 Health Level Seven. Available from http://www.hl7.org/ http://www.hl7.org.uk/
10 SNOMED Clinical Terms. Available from http://www.connectingforhealth.nhs.uk/technical/standards/snomed/
11 Gardner M and Peachey T. A standard XML schema for computerised anaesthetic records. Anaesthesia 2002; 57:1174–82[CrossRef][Web of Science][Medline]
12 Lack JA. New guidelines for anaesthetic records. RCA Newsl 1996; 27:8–9
13 Fisher MF. Clinical experiences of an in-house derived system for automated anaesthetic records. In: Kenny GNC, ed. Automated anaesthetic records. Clinical Anaesthesiology 1990. London: Bailliere Tindall 4:1:: pp. 47–65
14 Cross M. Will connecting for health deliver its promises? Br Med J 2006; 332:599–601
15 Tackley RM, Stuart-Taylor ME, Hurrell M. Why do anaesthetists need codes? Br J Anaesth 1993; 71:602–6
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||