Requirements for Managing Clinical Information in Integrated Care

• Introduction 
• EHRs in Integrated Care
• The Need for an EHR Architecture
• The GEHR Architecture
• Data Communication and the Distributed EHR
• Data Security
• Examples of Integrated Care IM/IT Implementations
  • 1. Paradigm IPA, Hawke’s Bay, New Zealand
  • 2. Australian Co-ordinated Care Trials
• Summary
• References

Introduction

The first attempts at computerising patient records occurred in the 1960s but this was mainly limited to the recording of demographic and other administrative information about patients. Computerised clinical records began to be introduced in the seventies and became widespread in the eighties. This paper will concentrate on the clinical information management requirements of integrated care.

The primary care sector has generally been ahead of hospitals in maintaining patient-centred clinical notes in electronic form. In some European countries such as Norway, the Netherlands and the UK, almost all GPs now have electronic health records (EHRs). Hospital clinical records are still largely oriented towards departmental systems (eg pathology, radiology, ICU, A&E) and despite much recent publicity about clinical data repository projects in hospitals, they are rarely in the form of an integrated patient-centred record. The emphasis in the few “integrated” hospital clinical computing systems which do exist has been on tracking and recording of the clinical process (critical/clinical pathways) rather than on recording clinical details such as history, examination, management plan, and progress notes which form the “raw material” for computerised decision support.

During the period of the evolution from paper-based to EHRs, there has been a dramatic change in the organisation and delivery of health care. Over the past fifty years there has been an enormous increase in the degree of specialisation within the medical profession, due mainly to the explosion of knowledge in every branch of medicine. There has also been a huge increase in the sophistication and complexity of medical technology for diagnosis, monitoring, and treatment of patients. This has meant that patients are now routinely managed by a team of doctors from different specialties, especially in the acute hospital setting. The trend to super specialisation and high technology medicine has, however, had some adverse side effects. The cost of health care in some developed countries, most notably the US, has spiralled almost out of control. There has also been a recognition that over-specialisation has often lead to poor co-ordination of care and sometimes to adverse outcomes by seeing the patient as an organ system rather than a holistic being.

The era of super-specialisation probably peaked in the eighties and there has been an evolving trend in many countries over the last decade towards a more holistic and integrated approach to health care. This has precipitated a change in the balance between primary and secondary care. Hospitals around the world have been “down-sized”, there is a renewed interest in preventive care strategies, and there is now talk in many countries of a “primary care lead” health system. This model usually involves integrated multi-disciplinary patient care with the general practitioner at the hub of the system. The terminology, organisation, and emphasis has been somewhat different between countries – “shared care” in the UK, “managed care” in the US, “co-ordinated care” in Australia, “integrated care” in New Zealand – but the ultimate goal is the same: to improve the overall quality of well-being and patient care while at the same time stabilising or reducing the cost of health care to the community. The term “integrated care” will be used in this paper to broadly encompass all of these other terms.

EHRs in Integrated Care

The greatest challenge in this new paradigm of health care delivery is to provide comprehensive, reliable, relevant, accessible, and timely patient information to each member of the health-care team whether in primary or secondary care and whether a doctor, nurse, allied health professional or patient. With the complexity and volume of information to be managed, these requirements can really only be met through the use of electronic records and “connected” computerised systems that enable clinicians to share and exchange patient information.

The problem at the moment is that the sharing and exchange of clinical data are impeded by the lack of standards for electronic health record structures and the lack of harmonisation between different clinical computing systems. This has been called the “vertical silo” approach to computerised clinical systems development. Currently, most software suppliers build monolithic applications that sit on top of their own proprietary systems’ architectures and data formats. It is often difficult or impossible in primary care even to share clinical data between two different applications on the same PC let alone exchange data between practices or with hospitals and other health care institutions. The situation is no better in the secondary care sector where hospitals spend large amounts of money trying to link disparate departmental systems using interface “engines”, often resulting in poor system performance, poor quality data, and disgruntled clinicians.

The various members of a multi-disciplinary integrated care team will often require different types of patient information, different levels of detail, and different ways of entering, retrieving, and displaying this information. Having said this, there will also usually be a significant degree of commonality in the core clinical and administrative patient information requirements and a commonality of purpose within the team, hopefully co-ordinated by a nominated care team member. The care co-ordinator will most often be the patient’s GP but this may not always be the case depending on the particular model of integrated care for a given country or region.

The key component of the integrated care information management system is the electronic health record. This can be broadly defined as “a repository of information in computer readable format regarding the health of a subject of care”.  ¹   There are a number of terms used as synonyms for the EHR. These are largely country specific. Examples include the American CPR (Computerised Patient Record), the European EHCR (Electronic Health Care Record), the New Zealand EMR (Electronic Medical Record) and the British EPR (Electronic Patient Record). The recent NHS Information Strategy document  ²   makes a distinction between the EPR, which it defines as “the record of the periodic care provided mainly by one institution”, and the EHR, which it defines as “the concept of a longitudinal record of patient’s health and health care – from cradle to grave”. The EPR by these definitions becomes a logical or physical subset of the EHR. This is a useful conceptual distinction in the context of information management for integrated care. The NHS document confirms its view of the centrality of primary care when it states that the "majority of patient contacts are with primary care, and the information contained within the EHR is essential to support primary care teams. It therefore follows that the creation and maintenance of the EHR is best undertaken within the primary care setting.”

The Need for an EHR Architecture

The integrated care requirements for shareability and exchangeability of patient information can best be met by standardisation of the EHR architecture. The architecture in this context means the structure or framework of the EHR rather than the specific format of the data within the EHR. Thus, the architecture is a model of the generic features of the EHR and does not restrict or prescribe the particular type of data that can be stored in the record. There may be some situations in which it is desirable or necessary to be prescriptive about the particular data elements and field sizes within a record (this is called the data format). However, standardisation at the higher level of the architecture will give much greater flexibility and will accommodate the variability of requirements of the different health care disciplines.

It is important to emphasise the difference between a record architecture and a data format as this has been the source of much misunderstanding and confusion, particularly in relation to exchange formats like HL7 and Edifact. There are two types of data formats: storage formats and exchange formats or protocols. An HL7 message, as specified in the HL7 protocol, specifies in great detail the segments, data elements, and field sizes of every clinical message just as a storage format specifies the data elements and field sizes in a health record database.

It is also important to understand the key differences between a message and a record. A message has a specific purpose, is inherently transient, and is concerned with transferring data between information management systems. A record on the other hand has a generic purpose, is persistent (ie “permanent”), and is contained within an information management system (although it can obviously can also be exchanged between systems). Exchange or message standards such as HL7 are an important component of an integrated care information management system but they are insufficient on their own to provide all the necessary attributes and functions.

A standardised EHR architecture can provide a number of other essential and desirable functions that cannot be achieved at the data format level. These include the requirements to provide and ensure comprehensiveness; portability; ethical, medico-legal, security, and educational functions; code set support; language independence; audit control; and transaction version control.

The GEHR Architecture

There is, as yet, only one comprehensive EHR architecture that is available in the public domain and is therefore a candidate for national and international standardisation: the GEHR (Good Electronic Health Record) architecture, formerly known as the Good European Health Record. It is the result of a three-and-a-half year European Commission project that ran from 1992 to 1995. The objective of the project was “to develop a comprehensive and widely applicable common data structure (architecture) for using and sharing electronic health care records in Europe.”  ³   The GEHR project consortium involved 21 participating organisations in eight European countries and included clinicians from different professions and disciplines, computer scientists in commercial and academic institutions, and major multi-national companies. More than 6 million ECUs (approximately $NZ 10,000,000) were spent on the project. All of the documentation and other deliverables, including a formal object model, are in the public domain.  ⁴  

GEHR is not currently a standard in its own right in Europe or anywhere else. It is however, the basis of the European Standards Committee (CEN) pre-standard for the Electronic Healthcare Record Architecture.  ⁵   GEHR was adopted last year as the recommended EHR architecture for primary care computing in Australia and it is also starting to attract considerable interest in New Zealand. Much work has been done over the past two years in both Europe and Australia on updating and extending the GEHR model and it is expected that the first GEHR compliant systems will be available in Australia and New Zealand in early 1999.

Data Communication and the Distributed EHR

The final key component for integrated care information management is connectivity. The particular communication method chosen is probably not crucial although the ubiquity of the Internet (and its close cousins, the intranets) is making this medium the obvious choice for most if not all health care communications. Since almost all people will have used more than one health care provider and have health care records at more than one institution, electronic communication is essential for effective and timely integrated care. There are two main views on the form of the ideal electronic health record.

The first may be termed the centralised EHR approach. This is the model espoused in the new UK NHS IT strategy where a comprehensive EHR is compiled and maintained by the patient’s GP. This lifelong EHR will contain summaries of each encounter and episode with other health care professionals and institutions. Under this model it will still be possible for the GP or other authorised user to access more detailed information contained in distributed subsets of the EHR (these are termed EPRs – cf above). This model will work best in a capitated health system where every patient has a single allocated primary care clinician, as in the UK and as proposed for New Zealand.

The second model is the fully distributed or virtual patient record system. In this model there is no one pre-eminent location or version of the EHR. Each user accessing the patient’s record will build a logical view of the record from data obtained from one or more physically distributed components of the total EHR. [Few people would seriously advocate a fully centralised model where literally every clinical item for a lifetime is stored in a single data repository.] This model relies on a very high degree of technological and communication capability which is not yet fully achievable although there have been some early attempts such as the “Health Data Network” developed by IBM in Dayton, Ohio.

A third model, which is really a subset of the first, is the patient-held electronic record, usually called a smart card. This model has been widely deployed in a few European countries such as Germany, France, and Belgium. The smart card, at least in its current form, is not a replacement for the more traditional EHR storage devices since it can only hold a relatively small amount of data and is therefore not suitable as a lifelong record. It can, however, be very useful for storing a health summary and details of the more recent encounters. This gives the smart card a definite potential role in an integrated health care system. The smart card can also serve a valuable function in increasing the patient’s feeling of ownership and involvement in their own health care. Its main disadvantages are the need for widespread deployment of card readers and the likelihood of the patient loosing or forgetting to bring their card. The need for smart cards is questionable in the longer term should either of the first two models be widely implemented. These models, of course, would also potentially give the patient access to their own records at home via the Internet or other computer network. It should be noted that the GEHR architecture with its strong transaction version control, audit control and language independence would ably support all of these models.

Data Security

Security (or lack of it) is a major concern with any type of EHR system and has been the single major concern about using the Internet for transmitting patient information. Fortunately, the advent of intranets, firewalls, strong encryption and digital certificate authentication is rapidly overcoming most security concerns. There is currently debate about the relative merits of using an intranet rather than the open Internet for health data transmission. An intranet should, in theory, offer stronger security protection but the advocates of the open Internet argue that Internet security is now more than adequate and that intranets tend to discourage legitimate exchange of information between different groups of professionals. This is an important question in the context of integrated care. For example, in Australia there is a GP intranet and a trial retail pharmacy intranet but there is no capability to link these two networks.

A high degree of security for confidential patient clinical data is, of course, a legitimate requirement; but probably no more so than for other personal data such as social security, taxation and banking records, which are now routinely transmitted electronically. Moreover, a number of studies done in these other industries have shown that far more breaches of security and privacy occur via authorised users (“insiders”) rather than unauthorised users (“outsiders”); ie, it is often human rather than technology failures which lead to breaches of security. It should also be remembered that there is an inverse relationship between security and usability; the more “locks” there are on the system, the more difficult it is for authorised personnel to use the system for legitimate purposes.

Technology can, however, offer assistance to minimise even this type of security breach. Beth Israel hospital in Boston has one of the best and most widely used computerised clinical systems in the US. It has a policy that any authorised user can access any part of any patient’s electronic health record. Over a long period, the number of inappropriate ‘accesses’ is almost negligible despite the hospital having a large number of “VIP” patients. This has been achieved by the Beth Israel system recording every access to every record so that there is a complete audit trail at all times. Furthermore, users are reminded regularly by random messages on the system that their accessing of records is being monitored. It is worth noting that the GEHR architecture supports this high level of security and audit control.

Examples of Integrated Care IM/IT Implementations

1. Paradigm IPA, Hawke’s Bay, New Zealand
Paradigm IPA (Independent Practitioners’ Association) commenced an integrated care project in March 1998, electronically linking 50 GPs in 25 practices to a central server in the IPA. The main objective was to enable IPA-wide integrated care population health projects. The connectivity is based on software from Trident Health, and a central data repository in the IPA is populated from each of the GP’s EHR systems. The central data repository was not initially GEHR compliant but will be upgraded to a GEHR system early next year. Electronic messaging for all GPs for pathology, radiology and hospital transactions will be added in the near future.

The first project undertaken was for population diabetes management. An asthma project is also planned to commence in the near future. The degree of enthusiasm and sense of purpose among the participating GPs has been high and the project is already seen as giving the IPA a new raison d’être. It is important to note that the IPA central data repository is not a dump of all the clinical data from the contributing GPs but rather is a carefully selected and quality controlled set of the relevant clinical data only.

2. Australian Co-ordinated Care Trials
A series of eight two-year, co-ordinated care trials is currently in progress around Australia. The trials are being financed and supported jointly by the federal and state governments. They involve mainly the integrated/co-ordinated management of complex and costly chronic diseases such as diabetes, chronic obstructive pulmonary disease, and congestive heart failure although some of the trials are more broadly based around cohorts such as the frail elderly. The purpose of the trials is to improve service delivery, achieve better patient outcomes and deliver significant cost savings, mainly through reductions in hospitalisation. The degree of computerised information management varies widely between the trials but two trials stand out because of the success and innovation of their IM/IT systems. The first is the Illawarra region “CareNet” trial in NSW, which involves over 100 GPs and 1,800 patients who are 65 or older and have complex health needs. All GP practices are computerised and receive electronically pathology, radiology and hospital attendance and discharge details. They also work with a multi-disciplinary care team to develop and monitor electronically patient care plans. Early indications are that hospitalisations (and therefore total health costs) have been substantially reduced by the integrated care methodology and its supporting IM/IT systems.

The second trial with innovative IM/IT is the South Australian “HealthPlus” trial. This is really a series of eight sub-trials that collectively comprise nearly 5,000 patients making it the largest of the Australian trials. The IM/IT emphasis in HealthPlus has been on a combination of a centralised data collection system and a computerised care planning system for use by GP care co-ordinators. The centralised data repository holds patient information from a variety of sources on prescribing, tests, hospitalisations, etc. The care planning system has been developed as part of a separate research project on improved user interfaces and knowledge engineering for decision support. The resulting Care Planning On Line (CPOL) program contains a number of innovative features including multi-level guideline support to assist the GP in developing or modifying the patient care plan. The communications network is based on intranet technology.

Summary

Good information management is essential for effective integrated care, and the quantity and complexity of clinical information in today’s health care environment means that computerised systems are the only practical way to provide this management. The timely sharing and exchange of computerised patient information among the multi-disciplinary integrated care team depends on good communication systems and on standards for EHR and clinical messaging. The Internet, with adequate security safeguards, is the obvious communications network to provide widespread accessibility to patient information. This needs to be underpinned by a standardised EHR architecture like GEHR and a clinical data exchange standard like HL7. Many of the integrated care information requirements will be common to any model of health care delivery. Effective integrated care will, however, have a heightened dependence on computerised care planning and co-ordination tools and the availability of applications which can rapidly access and display clinical information in a variety of formats to suit the diverse needs of the multi-disciplinary integrated care team members.

References