Computerised Diagnostic Decision Support System for the Classification of Pre-Invasive Cervical Squamous Lesions.

Previous studies have revealed considerable interobserver and intraobserver variation in the histological classification of preinvasive cervical squamous lesions. The aim of the present study was to develop a decision support system (DSS) for the histological interpretation of these lesions. Knowledge and uncertainty were represented in the form of a Bayesian belief network that permitted the storage of diagnostic knowledge and, for a given case, the collection of evidence in a cumulative manner that provided a final probability for the possible diagnostic outcomes. The network comprised 8 diagnostic histological features (evidence nodes) that were each independently linked to the diagnosis (decision node) by a conditional probability matrix. Diagnostic outcomes comprised normal; koilocytosis; and cervical intraepithelial neoplasia (CIN) 1, CIN II, and CIN M. For each evidence feature, a set of images was recorded that represented the full spectrum of change for that feature. The system was designed to be interactive in that the histopathologist was prompted to enter evidence into the network via a specifically designed graphical user interface (i-Path Diagnostics, Belfast, Northern Ireland). Membership functions were used to derive the relative likelihoods for the alternative feature outcomes, the likelihood vector was entered into the network, and the updated diagnostic belief was computed for the diagnostic outcomes and displayed. A cumulative probability graph was generated throughout the diagnostic process and presented on screen. The network was tested on 50 cervical colposcopic biopsy specimens, comprising 10 cases each of normal, koilocytosis, CIN 1, CIN H, and CIN III. These had been preselected by a consultant gynecological pathologist. Using conventional morphological assessment, the cases were classified on 2 separate occasions by 2 consultant and 2 junior pathologists. The cases were also then classified using the DSS on 2 occasions by the 4 pathologists and by 2 medical students with no experience in cervical histology. Interobserver and intraobserver agreement using morphology and using the DSS was calculated with K statistics. Intraobserver reproducibility using conventional unaided diagnosis was reasonably good (kappa range, 0.688 to 0.861), but interobserver agreement was poor (kappa range, 0.347 to 0.747). Using the DSS improved overall reproducibility between individuals. Using the DSS, however, did not enhance the diagnostic performance of junior pathologists when comparing their DSS-based diagnosis against an experienced consultant. However, the generation of a cumulative probability graph also allowed a comparison of individual performance, how individual features were assessed in the same case, and how this contributed to diagnostic disagreement between individuals. Diagnostic features such as nuclear pleomorphism were shown to be particularly problematic and poorly reproducible. DSSs such as this therefore not only have a role to play in enhancing decision making but also in the study of diagnostic protocol, education, self-assessment, and quality control. (C) 2003 Elsevier Inc. All rights reserved.

Effect on prescribing of the limited list in a computerised group practice

The prescribing of drugs in the therapeutic classes that are affected by the government's limited list was investigated in a computerised group practice of just over 3,000 patients. Prescribable drugs in categories that are affected by the list were identified for two consecutive six month periods before and one six month period after the introduction of the list. A significant decrease in the prescribing of cough and cold remedies, vitamins, and antacids occurred after the list was introduced, whereas no change occurred in the prescribing of laxatives, benzodiazepines, or analgesics. The prescribing of iron and penicillin increased significantly after the list was introduced, whereas the use of H2 antagonists and non-steroidal anti-inflammatory drugs showed no significant change.

Comparison between laser scanning tomography and computerised image analysis of the optic disc

Aims - To study the interchangeability of the measurements of the optic disc topography obtained by one computerised image analyser and one confocal laser tomographic scanner. Methods - One eye of 28 patients with glaucoma or glaucoma suspects was studied. All cases had simultaneous stereoscopic disc photographs taken with the fundus camera Topcon TRC-SS and optic disc examination with the Heidelberg retina tomograph (HRT) during the same visit. The optic disc photographs were digitised and analysed with the Topcon ImageNet (TI) system. Three variables of the optic disc topography provided by the TI and the HRT were compared - cup volume (CV), rim area (RA), and cup area to disc area ratio (CA/DA). Results - The mean values of CV and RA provided by the TI (0.52 (SD 0.32) mm and 1.58 (0.39) mm , respectively) were greater (p <0.01) than the mean values of CV and RA determined by the HRT (0.32 (0.25) mm , and 1.33 (0.47) mm , respectively). The mean value of CA/DA provided by the TI (0.42 (0.14)) and the HRT (0.42 (0.18)) was similar (p = 0.93). Correlation coefficients between measurements obtained by the two methods ranged from 0.53 to 0.73. Conclusion - There was a significant discrepancy in the measurements of rim area and cup volume of the optic disc obtained by a computerised image analyser and a laser scanning tomograph.

The problem with law in books and law in computers - the oral nature of law

It is becoming clear that, contrary to earlier expectations, the application of AI techniques to law is not as easy nor as effective as some claimed. Unfortunately, for most AI researchers, there seems to be little understanding of just why this is. In this paper I argue, from empirical study of lawyers in action, just why there is a mismatch between the AI view of law, and law in practice. While this is important and novel, it also - if my arguments are accepted - demonstrates just why AI will never have success in producing the computerised lawyer.

Development of a risk model for adverse drug events in the elderly

The aim of this study was to develop a predictive model for adverse drug events (ADEs) in elderly patients. Socio-demographic and medical data were collected from chart reviews, computerised information and a patient interview, for a population of 929 elderly patients (aged greater than or equal to 65 years) whose admission to the Waveney/B raid Valley Hospital in Northern Ireland was not scheduled. A further 204 patients formed a validation group. An ADE score was assigned to each patient using a modified Naranjo algorithm scoring system. The ADE scores ranged from 0 to 8. For the purposes of developing a risk model, scores of 4 or more were considered to constitute a high risk of an ADE.