Project objectives

Problem: The need for living guidelines

Guidelines

As defined by the U.S. Institute of Medicine [Field and Lohr, 1990], clinical guidelines are "systematically developed statements to assist practitioner and patient decisions about appropriate health care for specific clinical circumstances". This definition stresses three aspects: a systematic approach to their development; a method to support clinical decision making; and coverage of a specific clinical problem.

Such guidelines are based on the best empirical evidence available at the moment [Hanka et al., 1999]. A guideline thus describes the optimal care for patients and therefore, when properly applied, improves the quality of patient care. Evidence-based guidelines are becoming an important and indispensable part of quality healthcare because of their potentials. At the same time, guidelines are more and more embedded in coherent quality systems in healthcare organisations. In this way, they facilitate the spreading of high standard practices that otherwise would have much less impact [Grimshaw and Russell, 1993]. Guidelines can be produced at any level, international, national, regional or local.

Protocols

Protocols are local tools that set out specifically what should happen, when and by whom in the care process. They can be seen as the local definition of a particular care process derived from a more discretionary guideline. Protocols reflect local circumstances, and variation will due to the differing types of local provision.

The need for living guidelines

It has been proved that adherence to guidelines and protocols may reduce health-care costs up to a 25% [Clayton and Hripsak, 1995]. Nevertheless, in the current scenario of guideline development, dissemination and deployment, there is a major problem with clinical guidelines:

• Recommendations can be outdated or not applicable in practice, because most guidelines are only revised every 5 years. In contrast with this, scientific and pragmatic knowledge is growing faster every year. At this moment, a guideline is a static document which cannot be modified easily.

• Update of the guidelines on a more continuous basis: clinical guidelines have to become flexible, adaptable documents. The aim is to develop guidelines which present up-to-date and state-of-the-art knowledge to practitioners. To make this possible, guidelines have to be modular in structure, so that only part of a guideline can be adjusted and not the whole document needs revision.

To make the approach of living guidelines possible, there must be some major changes in the guideline development process. Most guidelines are authored in an unstructured narrative form. Computer-based support depends on a more formal, structured representation, and also faces a number of other challenges at all stages of the guideline life-cycle: modelling, authoring, dissemination, implementation and update. At this moment, guidelines are often multi-interpretable [Marcos et al., 2001]. Also, different guidelines can include the same modules, which can be in conflict with other contents of the guidelines. Guidelines are complex documents. As a result, guidelines can be ambiguous, incomplete and even inconsistent. In the modelling phase of guidelines, methods have to be developed to support this process. Also, terminology is a problem here; precise, abstract definitions of core notions for medical management are necessary. In summary: to enable living guidelines, they must be developed in a more structured way. Formal methods can be of help here. This will be an important first step to enhance further computer-based support of guidelines and protocols.

Building on the success of Protocure I

During the Protocure I pilot phase of this project, the focus of interest has been the use of formal methods in quality improvement of medical protocols and guidelines. In this one-year pilot phase, we have shown that formal verification of medical protocols is possible not only in principle but also in practice: two full scale high quality medical protocols used in daily medical practice have been formalised in a two stage process first in Asbru [Seyfang et al., 2002a] and then in KIV [Reif, 1992c] and subjected to machine-verification [Seyfang et al., 2002b], [Kosara et al., 2002], [Marcos et al., 2002b], [Marcos et al. 2003]. This process revealed a number of anomalies even in these high quality protocols, and a subset of these anomalies were judged to be of medical relevance by a panel of independent medical experts.

The approach of Protocure I can be summarised in the metaphor "protocol ~ program". Just as a program, a protocol is a set of instructions of how to behave in certain circumstances. Just as formal methods in Software Engineering are known to help improve the quality of programs, we have shown how these methods can also be successfully employed to improve the quality of protocols. We have shown how informal medical protocols can be modelled and formalised in a two-stage process. Required properties can be formulated on each of these increasingly precise versions of the protocols, enabling the formal verification of such properties.

The Protocure I project focussed only on already existing protocols, and did not take into account the development process of these protocols. In order to meet the future challenge of turning guidelines into "living guidelines", Protocure II will extend the static metaphor of "protocol ~ program" into a dynamic metaphor:

As a result, the focus of attention in Protocure II will be the incorporation of formal methods in the life-cycle of medical guidelines. Instead of aiming at developing methods to analyse existing guidelines and protocols, we will concentrate on techniques and tools to support the whole guideline development process.

Summary of objectives

The overall objective of Protocure II project is:

We will achieve this by incorporating methods and processes from software engineering into the life-cycle of medical guidelines and protocols. This overall objective can be translated into the following measurable and verifiable concrete objectives:

1. A prescriptive process model for the integrated and iterative development of formal and informal versions of guidelines.
2. A set of software tools which supports this development process.
3. A library of guideline components, in the form of re-usable patterns.
4. A demonstration guideline, which has been developed according to the above process model, using the above tools, and based on the contents of the library.