If you're a New South Welshman, the next time you or a loved one go to a hospital you might just find the admission process smoother than it's ever been before. You also might get better continuity of care from the health professionals there, and while you may not notice, those professionals could be using information across a range of different systems to support their clinical decision making and health care planning. You might even find yourself waiting for shorter periods in Casualty, and if you're coming in for overnight elective surgery you might not have to wait as long for that service.
All because NSW Health has recognised that information is a resource that has value and must be managed like any other asset. Moreover it knows that information management strategies, while not in themselves core business strategies, are crucial to effectively developing an organisation's information infrastructure. And in capitalising on that recognition, NSW Health is developing an information management system (IMS) to underpin its ability to achieve better health outcomes. The aim is to provide a framework of definitions and business rules to enable information about health needs; health service delivery; the resources used to deliver that service; and, health outcomes from providing that service to be related in a way never before possible.
According to information architecture manager John Lewis, before developing its IMS strategy, NSW Health had enterprise-level statutory and management information needs that were poorly coordinated, complex and a burden on its 17 separate Area Health Service units. Within an Area Health Service, the data owned by an individual application was like an "information island". When viewed from "above" these islands of information showed duplication and inconsistency, and that inconsistency had to be managed - to some extent - through development of expensive interfaces.
There was overlap in data content. Differences in versions of data often meant difficulties in reconciling the basis for decision making. Worse still, the reliability of much of the data was questionable, because different people relied on different definitions for the same information. An individual might be classified as a same-day patient in one collection, counted in another collection as an overnight patient and in another collection altogether might be counted as just a patient.
"I think at one point we had 14 different definitions of what a full-time equivalent employee was," Lewis says. "At last count we had something like 87 separate data collections which ranged from our large in-patient statistics collections through human resources and financial performance information to very small collections.
"We are now trying to rationalise those collections so that the burden doesn't fall as much as it does now on our operational people, our clinicians and our providers of care. Because if you analyse the information that they are collecting, it breaks down into just three broad areas: information about the services we provide, information about the people we provide those services to and information about the resources we use to provide those services."Across the State, the application sets supporting operational needs were different. The implementation of even the same system at different sites sometimes resulted in differences in the code sets used. In the past this was inevitable. After all, NSW Health must always balance the need to provide a standard set of definitions for State-wide reporting purposes and State-wide performance management against the fact that the areas of NSW served by the NSW Health system are very different in their needs. This diversity sees some parts of the State needing a particular code set not required in other areas.
"Providing a service to a sparsely populated, large geographical area is very different to providing a service to a highly populated urban area," Lewis says.
"We need to centralise and tailor our services to meet the needs of particular areas and population groups, but on the other hand, we still need to get some sort of standard measure of how we are going. It's that kind of dilemma that we face."Added together, these factors made the cost of providing consistent, comparable data across the State very high. To Lewis, a data warehouse seemed the best possible answer. "We realised a data warehouse strategy that aimed to collect that information as a by-product of normal operational systems and to integrate it into one database would overcome, to a large extent, those kinds of problems. First of all the burden of the actual collection and secondly the need to individually, with each collection, integrate and cross tabulate the information you get from it."The first step was to accept that information was a resource that had value and that must be managed like any other asset. Thus emerged the NSW Health Information Vision. It states: "Information should be person-centred and available in the clinical setting to manage episodes of illness and episodes of care, useful for clinical review and as a management tool - timely, accurate, relevant and presented in a usable form." In other words, information should be used to improve clinical productivity, facilitate continuity of care, enhance decision-making processes and improve quality measurement and management.
"The Vision is a major step forward as it is written expression of the value placed on information (and information management) by the top management of NSW Health," Lewis says. "It is used at a conceptual level to guide the development of appropriate IM & T strategies. Expenditure on IT can be measured against how it implements the goals of the information policy."NSW Health next adopted an information management strategy (NIMS) outlining both an information and applications architecture. Lewis' Information Architecture Group (IAG) is responsible for developing and maintaining the information architecture, which aims to be accessible to all users. As a direct outcome of that work, the Health Department has been able to begin work on prioritising and implementing key aspects of the information policy.
Initially, the department drew on NIMS to make a number of recommendations, including development and adoption of the enterprise information model (EIM) and associated enterprise-wide data standards. Developed by consultants Simsion Bowles and Associates in conjunction with NSW Health, the EIM provides a framework and common language on which to hang the information needs of the NSW Public Health System.
As a "living", layered model of the information requirements of an organisation, the EIM is derived from an inventory process that continues to evolve as business requirements change. That evolution entails development of a series of integrated subject databases, rationalisation of data collections and implementation of standard code sets across NSW Health with the aim of improving ease of data integration, data quality, and data comparability.
"The EIM was not developed to be implemented as a single application," Lewis says. "Its purpose was to provide a framework that allows business rules and needs to be identified so that they can be incorporated in specific developments, not only in IT but also in the way the business operates, or to reflect changes in the way the business operates.
"We're trying to describe in a standard way our business rules, such as that a medical health professional (doctor) may prescribe zero or more prescriptions and that a prescription must be prescribed by a registered medical practitioner." According to Lewis, by going through each information item of interest, the department can relate each one to other items and in that way define the rules by which it does business. "You can then take that business-level or conceptual model and bring it down to a greater level of detail," he says. And the work had other pay-offs, as well.
"If you're an application systems developer developing a computer system, it will save you a lot of work in having to go through and identify and map those business rules. Anyone who develops a computer system needs to do that if they want the computer system to reflect the business of the organisation," Lewis says.
"The other area where it is useful is when we buy pre-packaged software. We can look at the software package and see how well it matches the information items we need to keep about that particular subject area and how well it matches the business rules that we have. Then we'd obviously be trying to choose the one that best fits our business need and our information needs."By mapping them to the EIM, specific context models of the business of health, including patient administration, clinical services, finances, providers and clients, can be related to one another. This makes it easy to recognise information overlaps and sources of data and identify areas of interest that have inadequate information resources. Doing so requires capture of the detailed structure of components of the bigger picture in subsidiary models such as complaint management, clinical cancer, ambulatory care, mental health and Aboriginal health.
Lewis says the architecture operates first at a conceptual (or business) level then at a more detailed logical level that specifies information required in a computer system. Finally it operates at the physical level, where it is implemented within a computer database structure.
"The use of standards has critical importance in the health management reform process," says Lewis. "Data standards allow aggregation of data from disparate sources and comparisons to be made. These are fundamental to evaluation of the efficiency of service delivery."The next step will be to establish a Health Information Resources Directory, an Internet/intranet-enabled database containing details of the NSW Health System's information resources, how they relate to each other, who manages the information (acts as the custodian) and their contact details. Lewis says the directory will provide URL hotkey access to other sources of data.
Since the team saw "data warehouse" as too esoteric a concept for a widely variant user community, it coined the term Health Information Exchange (HIE) to convey a simpler message: information would be stored from different parts of the organisation but made available across the organisation was well.
"The HIE data warehouse project evolved out of the IMS strategy recommendation for subject databases," says Lewis "The HIE will physically implement significant aspects of the information management strategy by drawing information together from separate operational systems, using a standard data model to validate and integrate the information. It will also be the backend database for efficient distribution and interaction through various reporting tools and retailing systems like the NSW Health EIS, and local data marts."The aim is to rationalise the large number of data collections and reduce the areas of redundant data needing to be stored. "If you've got a [data] collection about in-patient statistics and there's a lot of information about patient demographics there, and you've got a collection about ambulatory care statistics where a lot of the same patients have gone through emergency and out-patient departments, you're keeping that information twice about the same people," Lewis says.
During the pilot release of the warehouse this will mean taking extracts of information from emergency department information systems, in-patient patient administration systems and waiting list information systems and integrating those patient records in order to link the episodes of treatment or service.
"Say you wanted to know how many people had spent more than six months on a waiting list for cardio-surgery and then had serious heart attacks that ended up with them being treated in the emergency department or admitted to hospital.
"To find that information out now you'd have to go to the waiting list system to see who was on the waiting list. You'd have to go to the emergency department and then see if you could find that same person having come in for an emergency treatment for cardiovascular illness. Then you'd have to go and look in the in-patient system to see whether that person had been admitted for treatment.
"With a system like the information warehouse, you can get that kind of information, because the information warehouse stores it as a matter of course and you've got it all together in the one database," Lewis says.
The HIE is in fact a set of databases linked by a common data model, kept up to date and synchronised across the organisation by periodic updates. Local users are free to extend the data structure by adding local classifications or organisation units. There is a series of 10 physical databases with a logical data warehouse for each of the 17 Area Health Services plus the Health Department. There are common data model links between each warehouse, making it easy to exchange comparable information such as performance management.
Detailed data about individual patient treatment or service events is kept in the Area Health Service where the patient is treated, while a summary and aggregate data that is de-identified is provided to the department. Lewis says taken together, the set-up makes it easy for the department to support key performance indicators and State-wide clinical indicators, while supporting local decision making for managers and clinicians. At the same time, it reduces the data collection burden and improves consistency of data across systems, hospitals and areas.
Collecting integrated information organised by subject area, as a series of snapshots, "is of significantly more value than having the same information stored in different forms within multiple operational systems", Lewis insists.
"It allows a dynamic reporting capability and the ability to obtain longitudinal views of historical data, something which is extremely difficult to achieve in operational systems," he says.
Initially set up to reduce the number of data collections across NSW Health, HIE will in fact address a range of issues for the organisation. These include the difficulty of linking inputs with activities and clinical outcomes, the need for adherence to standards, the time consuming and fragmented nature of data reporting and the resources required for collection of Department of Health data.
It also makes it possible to make links between health needs as expressed in demand for services, population health surveys and the services NSW Health provides, to the effectiveness of those services and to the resources used to provide those services. And it makes it possible to provide information that will support patient/client care across a variety of settings, from hospital in-patient through GP to Community Health, accident and emergency departments.
"The IMS underpins the ability to manage information to achieve better health outcomes," Lewis says. "It aims to provide a framework of definitions and business rules that will enable information about health needs, health service delivery - along with the resources used to deliver that service - and health outcomes from providing those services to be interrelated in a way not before possible.
"While much of the benefit is difficult to quantify, savings in data collections alone, without considering the improved infrastructure for decision support, justify the IMS strategy," Lewis says. He adds that realisation of benefits will be an iterative process and will become more tangible with each releaseWith release two scheduled for early next year, and the pilot already established at the departmental level and in two Area Health Services, Lewis says other recipients will have to work to realise the benefits of the HIE.
And, while the benefits of the HIE are already proving tangible, from now on each Area Health Service will be required to put up a business case for the HIE, to ensure their total commitment to the exercise.
"The days are long gone when we had this cargo cult mentality that users got what the IT department thought was good for them," says Lewis.
Join the CIO Australia group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.