Beatriz Pérez

Authoring and verification of clinical guidelines: A model driven approach

OBJECTIVES The goal of this research is to provide a framework to enable authoring and verification of clinical guidelines. The framework is part of a larger research project aimed at improving the representation, quality and application of clinical guidelines in daily clinical practice. METHODS The verification process of a guideline is based on (1) model checking techniques to verify guidelines against semantic errors and inconsistencies in their definition, (2) combined with Model Driven Development (MDD) techniques, which enable us to automatically process manually created guideline specifications and temporal-logic statements to be checked and verified regarding these specifications, making the verification process faster and cost-effective. Particularly, we use UML statecharts to represent the dynamics of guidelines and, based on this manually defined guideline specifications, we use a MDD-based tool chain to automatically process them to generate the input model of a model checker. The model checker takes the resulted model together with the specific guideline requirements, and verifies whether the guideline fulfils such properties. RESULTS The overall framework has been implemented as an Eclipse plug-in named GBDSSGenerator which, particularly, starting from the UML statechart representing a guideline, allows the verification of the guideline against specific requirements. Additionally, we have established a pattern-based approach for defining commonly occurring types of requirements in guidelines. We have successfully validated our overall approach by verifying properties in different clinical guidelines resulting in the detection of some inconsistencies in their definition. CONCLUSIONS The proposed framework allows (1) the authoring and (2) the verification of clinical guidelines against specific requirements defined based on a set of property specification patterns, enabling non-experts to easily write formal specifications and thus easing the verification process.

A systematic review of code generation proposals from state machine specifications

Context: Model Driven Development (MDD) encourages the use of models for developing complex software systems. Following a MDD approach, modelling languages are used to diagrammatically model the structure and behaviour of object-oriented software, among which state-based languages (including UML state machines, finite state machines and Harel statecharts) constitute the most widely used to specify the dynamic behaviour of a system. However, generating code from state machine models as part of the final system constitutes one of the most challenging tasks due to its dynamic nature and because many state machine concepts are not supported by the object-oriented programming languages. Therefore, it is not surprising that such code generation has received great attention over the years. Objective: The overall objective of this paper is to plot the landscape of published proposals in the field of object oriented code generation from state machine specifications, restricting the search neither to a specific context nor to a particular programming language. Method: We perform a systematic, accurate literature review of published studies focusing on the object oriented implementation of state machine specifications. Results: The systematic review is based on a comprehensive set of 53 resources in all, which we have classified into two groups: pattern-based and not pattern-based. For each proposal, we have analysed both the state machine specification elements they support and the means the authors propose for their implementation. Additionally, the review investigates which proposals take into account desirable features to be considered in software development such as maintenance or reusability. Conclusions: One of the conclusions drawn from the review is that most of the analysed works are based on a software design pattern. Another key finding is that many papers neither support several of the main components of the expressive richness of state machine specifications nor provide an implementation strategy that considers relevant qualitative aspects in software development.

Evolution of XML schemas and documents from stereotyped UML class models: A traceable approach

Context: UML and XML are two of the most commonly used languages in software engineering processes. One of the most critical of these processes is that of model evolution and maintenance. More specifically, when an XML schema is modified, the changes should be propagated to the corresponding XML documents, which must conform with the new, modified schema. Objective: The goal of this paper is to provide an evolution framework by which the XML schema and documents are incrementally updated according to the changes in the conceptual model (expressed as a UML class model). In this framework, we include the transformation and evolution of UML profiles specified in UML class models because they are widely used to capture domain specific semantics. Method: We have followed a metamodeling approach which allowed us to achieve a language independent framework, not tied to the specific case of UML-XML. Besides, our proposal considers a traceability setting as a key aspect of the transformation process which allows changes to be propagated from UML class models to both XML schemas and documents. Results: As a general framework, we propose a Generic Evolution Architecture (GEA) for the model-driven engineering context. Within this architecture and for the particular case of the UML-to-XML setting, our contribution is a UML-to-XML framework that, to our knowledge, is the only approach that incorporates the following four characteristics. Firstly, the evolution tasks are carried out in a conceptual model. Secondly, our approach includes the transformation to XML of UML profiles. Thirdly, the proposal allows stereotyped UML class models to be evolved, propagating changes to XML schemas and documents in such a way that the different elements are kept in synch. Finally, we propose a traceability setting that enables evolution tasks to be performed seamlessly. Conclusions: Generic frameworks such as that proposed in this paper help to reduce the work overload experienced by software engineers in keeping different software artifacts synchronized.

A Model Driven Approach to Automate the Implementation of Clinical Guidelines in Decision Support Systems

Clinical guidelines and protocols are used to aid decision making in health care. This paper presents a model- based approach to the development of decision support systems for clinical guidelines. The guidelines are modelled using the UML Statechart language and the final system is automatically generated using an automated model transformation. Our approach is illustrated with a Spanish clinical guideline developed on the basis of a guideline published by the well-known National Guideline Clearing House (NGC).

Verification of Clinical Guidelines by Model Checking

Clinical guidelines systematically assist practitioners with providing appropriate health care for specific clinical circumstances. However, a significant number of guidelines are lacking in quality. In this paper, we use the UML modeling language to capture guidelines and model checking techniques for their verification. We have established a classification of possible properties to be verified in a guideline and we present an automated approach based on a translation from UML to PROMELA, the input language of the SPIN model checker. Our approach is illustrated with a guideline based on a guideline published by the National Guideline Clearing House (NGC).

Occurrence-Oriented Design Strategy for Developing Business Process Monitoring Systems

In a competitive market, business process improvement is a requirement for any organization. This improvement can only be achieved with the support of comprehensive systems that fully monitor business processes. We propose an occurrence - based approach to business process monitoring that provides a holistic perspective of system dynamics, lending support to evolution aspects. More specifically, we present a three-dimensional artifact, called Occurrence, in which structure, behavior, and guidance are considered simultaneously. Based on this, we define more complex structures, namely Occurrence Base and Occurrence Management System, which serve as scaffolding to develop business process monitoring systems. We also present a specific occurrence-based design strategy that, using an MDE approach, has been applied by our research group for the development of successful monitoring applications.

UML2PROV: Automating Provenance Capture in Software Engineering

In this paper we present UML2PROV, an approach addressing the gap between application design, through UML diagrams, and provenance design, using PROV-Template. PROV-Template is a declarative approach that enables software engineers to develop programs that generate provenance following the PROV standard. The main contributions of this paper are: (i) a mapping strategy from UML diagrams (UML State Machine and Sequence diagrams) to templates, (ii) a code generation technique that creates libraries, which can be deployed in an application by creating suitable artefacts for provenance generation, and (iii) a demonstration of the feasibility of UML2PROV implemented with Java, and a preliminary quantitative evaluation that shows benefits regarding aspects such as design, development and provenance capture.

A systematic review of provenance systems

Provenance refers to the entire amount of information, comprising all the elements and their relationships, that contribute to the existence of a piece of data. The knowledge of provenance data allows a great number of benefits such as verifying a product, result reproductivity, sharing and reuse of knowledge, or assessing data quality and validity. With such tangible benefits, it is no wonder that in recent years, research on provenance has grown exponentially, and has been applied to a wide range of different scientific disciplines. Some years ago, managing and recording provenance information were performed manually. Given the huge volume of information available nowadays, the manual performance of such tasks is no longer an option. The problem of systematically performing tasks such as the understanding, capture and management of provenance has gained significant attention by the research community and industry over the past decades. As a consequence, there has been a huge amount of contributions and proposed provenance systems as solutions for performing such kinds of tasks. The overall objective of this paper is to plot the landscape of published systems in the field of provenance, with two main purposes. First, we seek to evaluate the desired characteristics that provenance systems are expected to have. Second, we aim at identifying a set of representative systems (both early and recent use) to be exhaustively analyzed according to such characteristics. In particular, we have performed a systematic literature review of studies, identifying a comprehensive set of 105 relevant resources in all. The results show that there are common aspects or characteristics of provenance systems thoroughly renowned throughout the literature on the topic. Based on these results, we have defined a six-dimensional taxonomy of provenance characteristics attending to: general aspects, data capture, data access, subject, storage, and non-functional aspects. Additionally, the study has found that there are 25 most referenced provenance systems within the provenance context. This study exhaustively analyzes and compares such systems attending to our taxonomy and pinpoints future directions.

Automating Provenance Capture in Software Engineering with UML2PROV

UML2PROV is an approach to address the gap between application design, through UML diagrams, and provenance design, using PROV-Template. Its original design (i) provides a mapping strategy from UML behavioural diagrams to templates, (ii) defines a code generation technique based on Proxy pattern to deploy suitable artefacts for provenance generation in an application, (iii) is implemented in Java, using XSLT as a first attempt to implement our mapping patterns. In this paper, we complement and improve this original design in three different ways, providing a more complete and accurate solution for provenance generation. First, UML2PROV now supports UML structural diagrams (Class Diagrams), defining a mapping strategy from such diagrams to templates. Second, the UML2PROV prototype is improved by using a Model Driven Development-based approach which not only implements the overall mapping patterns, but also provides a fully automatic way to generate the artefacts for provenance collection, based on Aspect Oriented Programming as a more expressive and compact technique for capturing provenance than the Proxy pattern. Finally, there is an analysis of the potential benefits of our overall approach.

Generating Persistence Structures for the Integration of Data and Control Aspects in Business Process Monitoring.

Today’s organizations have to monitor increasingly complex business processes that handle large amounts of data. In this context, it is essential to design working frameworks that seamlessly integrate both control flow and data perspectives. Such an integration can be eased by automatically generating the infrastructures for storing data and control aspects. Towards this goal, we propose an automatic process for synthesizing persistence structures for control flow and data storage. In particular, based on an approach centered on the concept of Occurrence, in this paper we present a proposal by means of which, after applying several translation patterns to a business process model, we automatically generate the persistence structures that integrate both data and control aspects of such model. The feasibility of this proposal is demonstrated by developing a prototype and evaluating its application to different examples taken from the literature as a benchmark.