An 'Issue' in Primary Health Care
Dr Graeme Miller & Dr Helena Britt
Family Medicine Research Unit, University of Sydney and the Royal Australian College of General Practitioners
This paper describes the development of the Australian Primary and Community Health Services Information Model and the derived data concepts and code sets.
The Australian health system is administratively complicated by the existence of dual responsibility (split between the Federal Government and the State Governments) for the delivery of health care. Hospital services and non-medical community-based services are managed by the State Governments, supported by Federal Government grants. General practitioners and most specialists are reimbursed for the large bulk of their activity on a fee-for-service basis through Medicare, a Federally-run compulsory health insurance scheme. The major initiatives in health information technology have been developed by State health departments and have been concentrated in the hospital sector. Hence most of the developments have an obvious institutional bias. There is also a considerable variance in the implementation of similar projects between States. As a result of concern regarding the variance and the perceived need for management data at a Federal level, the Australian Health Ministers' Advisory Council initiated the National Health Data Agreement. This has been implemented by the National Health Information Management Group (NHIMG) and the National Health Data Committee.
One of the major outputs of the NHIMG has been the National Health Information Model which aims to provide:
The National Health Information Model therefore represents two important initiatives:
Following the development of the 'high-level' National Health Information model, the NHIMG initiated a number of second-level projects, including the Primary and Community Health Services National Information Management (PACHSNIM) Project. This aims to develop an information model for primary and community health services and provide national community-based data definitions and code sets. The scope of the project was to cover all community-based 'health' services, which in the Australian context also includes community-based social and welfare workers. The scope was eventually expanded to a default option of covering all services which were provided to any client, patient or population group other than to hospital inpatients, outpatients or emergency department attendees. Thus general practice, allied health, community nursing, health promotion, health education and social work were all included.
The wide variety of professional disciplines involved and the multitude of target individuals and groups made it clear early in the project that the institutional framework of earlier work had resulted in deficiencies in the national model which could not be addressed unless its conceptual base was broadened.
The hospital one-patient biomedical paradigm needed expansion into a dual patient/ population biopsychosocial model if it was to address the data needs of all the players in the primary care and community services sector. Continued relevance to the institutional sector was ensured by locking the model into the framework of the National Health Data Model.
While the modelling process used in the PACHSNIM project was essentially 'top-down', it involved a wide cross-section of services providers. The project was directed by the New South Wales Health Department on behalf of the National Health Data Committee. The Department was advised by a National Reference Group of thirty people representing all State Health Departments, the Federal Health Department, professional organisations, non-government service provider organisations, and a wide range of health and community professionals. The formal modelling process has been conducted by consultants experienced in the discipline, and has followed a pattern of focus groups and workshops which has become the standard method for developing subsets of the National Health Data Model.
Two new high-level data concepts were developed to broaden the conventional biomedical concepts of diagnosis and treatment.
The first was the concept of 'issue', which was introduced to encompass not only 'diagnosis' or 'problem', but also the wider variety of factors which may influence the state of wellbeing of any party - be they an individual, a family, a corporation, a population group or a community. This may include any factor which triggers an interaction between health service provider and party. 'Issue' has been split into several domains to facilitate the development of code sets. To the conventional medical domains of physical, mental and functional 'issues' has been added a wide range of social 'issues' such as environmental, cultural, relationship, legal and lifeskill factors.
The definition of the high level data element of 'issue' has led to a search for an appropriate classification system for its sub-elements which would be acceptable to all disciplines and allow the development of a computerised data collection system at the service level. No classification system available in Australia was comprehensive enough to cover all the possible sub-types of 'issue' without some enhancement. However, the basic structure of the International Classification of Primary Care (ICPC) was judged to be appropriate for primary and community health services in Australia. An enhanced version of ICPC (ICPC PLUS) had been developed by the University of Sydney for general practice computerised patient records. This was regarded as a logical starting point for the development of a universal system for all community health services. A consortium of State health departments has commissioned a feasibility study of ICPC PLUS to determine its usefulness in the broader community context. This will involve testing the addition of terms used by community-based services which fit within the broad definition of 'issue'. ICPC PLUS is being mapped to ICD-9 CM(A), the standard Australian hospital coding system, to standardise data transfer between the primary and secondary health sectors.
The second concept is that of 'service type' which was introduced to include both personal and population based activities of primary care and community health providers. As with all process codes, 'service type' has implications for the funding of services and providers. It was therefore important to cover all the possible activities of service providers in order to demonstrate productivity and improve accountability to service purchasers. There is little international work of relevance to the development of process codes and classifications in Australia. However the ICPC framework was again seen as relevant in the development of a community-based process coding system.
There is a significant overlap between 'issue' and 'service type'. For example, 'immunisation' may be an 'issue' or a 'service type' depending on the context. Some service providers see delivery of a 'service type' as being the only 'issue', for example, the provision of 'meals on wheels'. As with all health data systems, the classification of a 'term' will depend not only on the concept described but also on the context of its use.
These concepts of 'issue' and 'service type' are being used, with other data elements, to develop a primary and community health services information model which is applicable to community-based health and welfare services. For this to be compatible with the National Health Information Model, there needs to be a review of some basic concepts included in the higher-level model to broaden its scope. Indications are that the National Health Data Committee is amenable to such an approach.
The New South Wales Government has now developed a detailed community health data model. This has been used to develop tender specifications for a clinical computer system for community health services. A three-State consortium is developing this project for use in Australia and South East Asia.
Following completion of this project, it is proposed that a General Practice Data Modelling project be initiated to develop a detailed model down to the clinical level which would provide a suitable standard for developing clinical computer systems.
The evolution of appropriate information systems for primary care requires a disciplined approach to the development of data models as a precursor to the refinement of data definitions and code sets. This requires close collaboration with potential users of the information systems and a broad consensus regarding the validity and applicability of the model across the spectrum of primary health care services. The Australian experience suggests that this can be achieved, given the will of governments to consult, and the enthusiastic co-operation of service providers.