In the paper, Prevalence of Anaesthesia Information Management Systems in University-Affiliated Hospitals in Europe, presented in the February issue of the Journal, Balust et al.1 report the result of a survey of use or intended use of anaesthesia information management systems (AIMS) in university-affiliated departments in Europe. As acknowledged in the paper, the survey had a very high nonresponse rate (71%) and the methodology did not guarantee that all European university-affiliated hospitals were contacted, especially those in eastern Europe where the language barrier made web-search challenging. It is consequently difficult to comment on whether their results give an accurate picture of AIMS take-up in Europe.
The present editorial will instead focus on how the authors might have deepened their analysis to reflect the complex arguments of why hospitals might or might not have introduced an AIMS, moving beyond the singular, although relevant, factor of price that they propose. Highlighting some of the data that the authors present but do not analyse, we will question the authors' initial assumption that AIMS are a necessity and their conclusion that lack of funding is the most appropriate issue to address. It is our goal to open the debate from the need to fund more AIMS as they exist now to what would be the best way to build and fund software to support anaesthesia and intensive care.
The paper by Balust et al. (referred throughout this editorial as the paper) begins with the statement that ‘The adoption of computer and computer software to convert, store, protect, process, transmit and securely retrieve information is fundamentally changing the practice of medicine’. We believe what they mean by this statement is that the computational power to manipulate data gives the possibility of providing clinicians with information that could change clinical practice. It is important to distinguish between data and information – the latter means a transformation of data into something that gives insight. If software makes it difficult to collect data or does not present data in a useful format at appropriate times, data do not become information that clinicians can use.
The paper also briefly mentions a number of reasons why uptake of AIMS has been slower than expected, including high acquisition price and maintenance costs, no clear return on investment for the hospital, necessary customization, lack of interoperability with other hospital systems, immature software and no proven benefit. A number of these are barriers that make it difficult for AIMS to turn data into information and others the result of data not being turned into information. We would suggest that an AIMS is a necessity only if it turns data into information for the clinicians using it, which to a large extent depends on whether the workflow built into the AIMS matches that in the anaesthesia or intensive care department.
If digital data in an AIMS are turned into information, clinical practice will evolve as new techniques and possibilities in patient care arise. An AIMS must be able to evolve with the changing clinical practice that it inspires. The rigid framework of most AIMS does not allow evolution of the system and therefore defeats the purpose of having digital data. As high initial costs of AIMS make it difficult to change or recustomize systems, AIMS that cannot evolve are more likely to slow down, rather than speed up, the rate at which information technology can support innovation in patient care. Although AIMS are generally considered useful because they contain and manipulate digital data, their power comes from being flexible enough to evolve.
At what cost
The paper mentions that the 44 hospitals using (or intending to use) AIMS have systems from nine different commercial companies. Such a high number of companies for a small and relatively young market suggests that the companies consider health technologies to be a lucrative market. In what economists term a ‘perfect market’, competition stimulates innovation. However, the high initial costs of purchasing hardware and software, customizing the software and training staff, make the AIMS market far from perfect because it is difficult to switch suppliers. It is unlikely then that the high number of suppliers will drive costs down; if anything, it will drive costs up because the number of users is small and the amount of duplicated work high.
A further issue with such a large number of suppliers is interoperability. It is likely to be difficult to move data from an AIMS to another clinical information system in a different department, such as intensive care. Interestingly, 20 out of 28 hospitals bought AIMS from a different supplier than existing systems elsewhere in the hospital. A number of survey responders (number not given) did not know what clinical information systems were being used in other departments of the hospital. It would also be difficult to collate data across hospitals. Although interoperability of systems did not seem to be a priority for many anaesthesia departments, which most likely had choice criteria based on immediate needs of workflow and price, interoperability is an issue for hospitals and health authorities that fund such systems.
Some countries, particularly in the north of Europe, have tried to solve problems of cost and interoperability by building large, comprehensive systems to be used nation-wide, such as Britain's National Program for Information Technology (NPfIT). Such systems are meant to solve interoperability problems so that health authorities are provided with the data that they need to better manage nation healthcare and cost of the information technology is minimized through a reduction in duplicated work. It can be argued that the top-down design approach often taken, however, makes it difficult for systems to have the flexibility they need to suit individual department workflows and to evolve as practice changes. Flexible systems allow information technology in clinical practice to be about clinical vision and adaptation as opposed to data collection.
A third way
A more efficient approach would be, for a health authority, to create open data standards and a customizable database platform that is easily extensible. Such systems, frequently called middleware, are comparable to the iPhone and the Apps store from Apple. They provide a secure platform for data, but allow applications to be built on the data. For example, a ventilator company might build a version of its software that takes data from the patient record, such as age and sex, and uses that to optimize the ventilation of the patient. Or, clinicians using customizing tools can develop ways to better visualize data to suit their clinical practice.
Such a platform will also require an interface. A basic interface should be provided and be easily customizable by the clinical team to suit a particular workflow in a unit. It should also be modular (e.g. trends, orders and results) so that its parts can be introduced slowly, reducing high initial costs, staff training costs and staff anxiety. The platform should also be extensible by information technology professionals who can build novel interfaces and data visualization tools. This solution drives clinical vision by allowing clinicians to customize their own systems as well as industry innovation by allowing companies to compete on the tools that they provide.
If health authorities focused on building middleware platforms, funding of AIMS would change dramatically. Although funding would still be needed, it could be spread out in time as a department built up its system. Departments would have an opportunity to experiment at low cost and demonstrate the value of any system they customized before asking for further funding. Furthermore, funds that would have previously been allocated for initial investment in a system could be reallocated to customization, staff training, and system support and development, which is often underfunded and can lead to nonadoption and ultimately to the system collecting data rather than producing information.
We have argued in this editorial how clinicians and funders might reconsider what an AIMS should be and how best to fund it. We pointed out that data and information are not the same and that the advantages of digital data are only actualized if an AIMS fits the workflow of the department, presenting data in an insightful way as information. Further, the system must be flexible enough to evolve as the information presented changes clinical practice. Only in this way is an AIMS a necessity. We then discussed the problems that arise in both current private and public sector development of AIMS and propose that a middleware platform, similar to the iPhone and App Store model, would be a better way to utilize clinical vision and drive private sector innovation.
1 Balust J, Egger-Halbeis CB, Macario A. Prevalence of anaesthesia information management systems in university-affiliated hospitals in Europe. Eur J Anaesthesiol 2010; 27:202–208.