Evaluation as a tool to increase knowledge in healthcare informatics.

The evaluation of information systems is an important topic in Clinical Informatics. It is argued that past evaluations have not been particularly informative in progressing the effective use of IT in healthcare due to their narrow focus. The different roles of evaluation in Clinical Informatics are examined, and the breadth and diversity of the available methodological tool kit highlighted. The aim is to stimulate a greater awareness of the roles and methods of evaluation. Challenges in evaluation which face the Clinical Informatics community are discussed and finally some comments made concerning the way in which evaluation might be made more effective in order to improve our knowledge of how to deliver useful systems into healthcare.

Evaluating large scale health information systems: from practice towards theory.

With the introduction of large scale health information systems which are incrementally developed from legacy systems, evaluators are faced with difficult methodological and practical problems. Some of the problems involved in multidisciplinary multi-method evaluations care discussed. It is argued that the development of a framework for evaluation is necessary in order to successfully plan an evaluation, understand the implications of the results and make future predictions based upon them. Some suggestions for arriving at such a framework are put forward.

Data entry in computer-based care planning.

If computer-based care planning is to be acceptable to nurses, they must be able to enter care plans in a flexible and intuitive manner that does not distort their observations about the patient. The limitations of existing computer-based data entry techniques are discussed. A novel method of data entry called Predictive Data Entry is presented, and the underlying model of nursing terminology which supports it is described. The results of user evaluations are discussed.

Computer-based medical clerking.

All admissions to hospital begin with medical clerking. Thus, it is an essential element of any electronic patient record system. This paper describes some initial experiments to develop a computer-based clerking tool with which doctors can record their clinical observations. The clerking tool uses a structured data entry technique which is based upon an underlying semantic network of clinical terminology. This enables data to be captured and stored in a semantically structured manner, facilitating its use in computer-based record systems and also for audit, management, and decision support purposes. The prototype tool is described, and the evaluation and preliminary results obtained are reported. Initial results indicate that structured data entry is feasible for hospital-based clerking.

The PEN & PAD medical record model: development of a nursing record for hospital-based care of the elderly.

The PEN & PAD Medical Record model describes a framework for an information model, designed to meet the requirements of an electronic medical record. This model has been successfully tested in a computer-based record system for General Practitioners as part of the PEN & PAD (GP) Project. Experiences of using the model for developing computer-based nursing records are reported. Results show that there are some problems with directly applying the model to the nursing domain. Whilst the main purpose of the nursing record is to document and communicate a patient's care, it has several other, possibly incompatible, roles. Furthermore, the structure and content of the information contained within the nursing record is heavily influenced by the need for the nursing profession to visibly demonstrate the philosophical frameworks underlying their work. By providing new insights into the professional background of nursing records, this work had highlighted the need for nurses to clarify and make explicit their uses of information, and also provided them with some tools to assist in this task.

Development of Design-a-Trial, a knowledge-based critiquing system for authors of clinical trial protocols.

Many published clinical trials are poorly designed, suggesting that the protocol was incomplete, disorganised or contained errors. This fact, doctors' limited statistical skills and the shortage of medical statisticians, prompted us to develop a knowledge-based aid, Design-a-Trial, for authors of clinical trial protocols. This interviews a physician, prompts them with suitable design options, comments on the statistical rigour and feasibility of their proposed design and generates a 6-page draft protocol document. This paper outlines the process used to develop Design-a-Trial, presents preliminary evaluation results, and discusses lessons we learned which may apply to the developed of other medical decision-aids.

Using the PEN&PAD information model to support hospital-based clinical care.

The PEN&PAD model for clinical record systems has been successfully used for medical records in both General Practice and hospital-based care. This paper describes experiences of using the model for developing computer-based nursing records. Results from this work show that there are some problems with directly applying the model to the nursing domain. Whilst the main purpose of the nursing record is to document and communicate a patient's care, it has several other, possibly incompatible, roles. Furthermore, the structure and content of the information contained within the nursing record is heavily influenced by the need for the nursing profession to visibly demonstrate the philosophical frameworks underlying their work. By providing new insights into the professional background of nursing records, this work has highlighted the need for nurses to clarify and make explicit, their uses of information, and also provided them with some tools to assist in this task.

Philosophies for the design and development of clinical decision-support systems.

Little significance is attached by medical informatics workers to the many practical issues which affect the development of clinical decision-support systems. We examine the current state of research in clinical decision-support, the characteristics and motivations of developers, and the perceptions of intended end-users. Factors which adversely affect the success of systems are highlighted and pointers to good practice discussed. We then propose a coherent approach to system development, consisting of requirements analysis, software design, implementation, testing, evaluation and maintenance.

PEN&PAD (Elderly Care): designing a patient record system for elderly care.

Practical experiences in the work analysis and design of a computer-based patient record system for use by doctors on an elderly care ward are discussed. A diagrammatic technique named "Task Oriented Flow Diagrams" is briefly described. This technique enables designers to formalise their understanding of users work tasks in manner that is comprehensible to users, and thus can form the basis of a dialogue between designer and users. A prototype patient record system is described and the results of its evaluation with users presented.

Decision support system for the differential diagnosis of breast disease.

The histopathological diagnosis of breast disease is representative of many problems of differential diagnosis encountered in the medical domain. It requires highly trained and experienced experts and is characterized by a large number of features whose presence or absence involves much uncertainty. Computer-based decision support systems intended to function in a consultative capacity during differential diagnosis have had limited success for two fundamental reasons. Firstly, they take an autonomous role and assume that the user has no contribution to make to the problem-solving process. Secondly, the established techniques for representing and reasoning with medical knowledge are of limited suitability in this domain. Such systems are unable to reach a correct diagnosis quickly and often subject the user to a cumbersome dialogue. These are not tolerated by pathologists working under severe time constraints. We first look at the problem-solving methods employed by pathologists in this domain and examine the functionality of traditional expert system methodologies. We then present a cooperative design which allows the pathologists to express his or her ideas within a decision support system whilst gaining assistance in required areas. A novel inference technique based upon the set partitioning technique in hypergraphs is also described. This mathematical method has the ability to cope with the incomplete or inadequate knowledge which is a characteristic of breast disease, whilst directing data gathering in a meaningful manner. In particular this approach can significantly reduce the amount of irrelevant data which the pathologist must enter before a conclusion is reached. Thus it can potentially improve the efficiency and user acceptability of medical expert systems.

A menu-driven knowledge base browsing tool.

Conventional computer-assisted medical decision-making systems have had limited impact on routine clinical practice. This has stimulated an alternative approach to the utilization of medical knowledge bases. Centering on the storage and retrieval of medical information, it aims to provide clinicians with computerized medical reference systems. In this paper we describe the development of a prototype menu-driven browsing tool, which allows clinicians to browse through the contents of a knowledge base in a number of ways. Operations include interrogation via disease classes, names or attributes; hierarchical display of all or part of a disease profile; printing of a disease profile; construction of differential diagnosis lists and comparison of two diseases. We discuss how the use of a menu-driven interface can help to overcome some of the problems encountered with previous designs of medical reference systems.

Computer assisted diagnosis of fine needle aspirate of the breast.

The development of the National Breast Screening Programme has created a demand for the widespread availability of fine needle aspiration cytology services. To meet this demand there must be a rapid increase in the number of pathologists and laboratories able to offer this service. In turn there is a need for improved training methods. The technique of fine needle aspiration cytology is not inherently complicated. The number of possible conclusions is essentially limited to four: unsatisfactory, benign, suspicious and malignant. A computer based expert system, designed to assist pathologists in the diagnosis of fine needle aspirates of the breast, has been developed. The system prompts pathologists to categorize a number of variables in the aspirate including nuclear and cytoplasmic features, and the degree of cellular cohesion, and uses these data to reason about possible conclusions. The final diagnosis is displayed with a detailed explanation listing the factors supporting it. Initial trials with this system have been encouraging and it is envisaged that this system will be of value both in training and as an aid to routine diagnosis.

Computer-assisted breast cancer grading.

Morphological tumour differentiation has been shown in numerous studies to give a good prognosis in breast cancer, but as histological grading is based upon a subjective assessment of microscopical appearances, difficulties in consistency and reproducibility are inevitable. A review of the many conventional methods served to highlight a common limitation in their approach; lack of structure. We introduce a new approach which seeks to overcome the problem, by formalizing the methods and identifying aspects which are well suited to computer aided analysis, these being incorporated into a microcomputer system facilitating the collection and appraisal of morphometric data. Within the Information Technology Institute (ITRI) at Brighton Polytechnic a research team is carrying out multidisciplinary work into the elucidation of biological systems. This programme, entitled 'Intelligent Medical Systems', used methods of mathematical signal processing and artificial intelligence, applied to a number of areas, one of which is described in this paper. The aim has been to utilize the inherent skill exercised by the histopathologist in interpreting microscopical images, whilst making quantitization more accurate and reproducible. the system has been developed within a highly structured framework and will have applications in teaching and routine histological analysis. The value of artificial intelligence techniques in the wider issues of this area is discussed.