Jacques Bouaud

How to translate therapeutic recommendations in clinical practice guidelines into rules for critiquing physician prescriptions? Methods and application to five guidelines

BackgroundClinical practice guidelines give recommendations about what to do in various medical situations, including therapeutical recommendations for drug prescription. An effective way to computerize these recommendations is to design critiquing decision support systems, i.e. systems that criticize the physicians prescription when it does not conform to the guidelines. These systems are commonly based on a list of if conditions then criticism rules. However, writing these rules from the guidelines is not a trivial task. The objective of this article is to propose methods that (1) simplify the implementation of guidelines therapeutical recommendations in critiquing systems by automatically translating structured therapeutical recommendations into a list of if conditions then criticize rules, and (2) can generate an appropriate textual label to explain to the physician why his/her prescription is not recommended.MethodsWe worked on the therapeutic recommendations in five clinical practice guidelines concerning chronic diseases related to the management of cardiovascular risk. We evaluated the system using a test base of more than 2000 cases.ResultsAlgorithms for automatically translating therapeutical recommendations into if conditions then criticize rules are presented. Eight generic recommendations are also proposed; they are guideline-independent, and can be used as default behaviour for handling various situations that are usually implicit in the guidelines, such as decreasing the dose of a poorly tolerated drug. Finally, we provide models and methods for generating a human-readable textual critique. The system was successfully evaluated on the test base.ConclusionWe show that it is possible to criticize physicians prescriptions starting from a structured clinical guideline, and to provide clear explanations. We are now planning a randomized clinical trial to evaluate the impact of the system on practices.

Formalisation de la démarche diagnostique des pneumopathies médicamenteuses : le système PneumoDoc

Resume La pneumopathie medicamenteuse (PM), souvent evoquee, reste un diagnostic difficile car les medicaments pneumotoxiques sont nombreux, la clinique est heterogene, et il n’y a pas de critere diagnostique formel. PneumoDoc est un systeme informatise qui propose une formalisation de la demarche diagnostique du pneumologue devant un tableau evocateur sur la base d’arguments chronologiques, semiologiques, d’imagerie et cytologiques (lavage broncho-alveolaire). Ces criteres dits « intrinseques » sont croises avec les criteres extrinseques tires de la litterature, en particulier Pneumotox. Ils sont renseignes sous forme de questions/reponses successives a choix ferme. Un recapitulatif final permet de visualiser l’ensemble des donnees caracterisant la situation clinique observee. Le logiciel estime la probabilite de PM selon une des cinq modalites suivantes : incompatible, douteux, compatible, suggestif, tres suggestif. La multiplicite des drogues, l’intrication a une pathologie cardiopulmonaire et l’absence de cas rapporte constituent les limites du systeme.Establishing the diagnosis of drug-related pulmonary disease (DRPD) remains a difficult task because of the large number of drug-related toxic situations and the variety of clinical presentations. PneumoDoc is a computer-based support system designed to facilitate the diagnosis of lung disease using chronological, clinical, imaging, and cytological (alveolar lavage) input. These intrinsic items are crosschecked against extrinsic items reported in the literature (Pneumotox). Data input is in the form of yes-no questions. The final output displays the characteristic features of the observed clinical situation and calculates the probability of DRPD defined in five categories: incompatible, doubtful, compatible, suggestive, and highly suggestive. Use of multiple drugs, interaction with cardiopulmonary disease, and the absence of reported cases are limitations of the system.

A Reconstruction of Conceptual Graphs on Top of a Production System

In this paper, we study how several aspects of the Conceptual Graph theory can be implemented using the pattern-matching mechanisms of production systems. Usually, standard pattern matching applies to arbitrary data that, unlike CGs, do not rely on a particular theory. Reconstructions of Conceptual Graphs in terms of basic graphs have been proposed in the literature. We show that K, a graph representation language with “high-level” (rule-based) graph manipulation facilities, allows an elegant implementation of these proposals. We show how the CG projection is reconstructed from standard pattern matching. Such a mechanism provides the user with graph retrieval facilities. Moreover, Ks inherent features, such as forward reasoning rules, are gracefully transferred to the resulting CG implementation with no further effort. The result is a production system that operates within the CG theory thus providing the basis for a flexible CG processor.

Using formal concept analysis to discover patterns of non-compliance with clinical practice guidelines: a case study in the management of breast cancer

Clinical decision support systems (CDSSs) may be appropriate tools to promote the use of clinical practice guidelines (CPGs). However, compliance with CPGs is a multifactorial process that relies on the CPGs to be implemented, the physician(s) in charge of the decision, and the patient to manage. Formal concept analysis (FCA) allows to derive implicit relationships from a set of objects described by their attributes, based on the principle of attribute sharing between objects.We used FCA to elicit patient-based formal concepts related to the non-conformity of multidisciplinary staff meetings (MSMs) decisions with CPGs in the domain of breast cancer management. We developed a strategy for selecting attributes and make lattices manageable. We found that when not using the guideline-based CDSS OncoDoc2, patients with bad prognostic factors were associated with non-compliant decisions. This was corrected when the system was used during MSMs.

Automatic derivation of a decision tree to represent guideline-based therapeutic strategies for the management of chronic diseases

In the management of chronic diseases, therapeutic decisions depend on previously administered therapies as well as patient answers to these prior treatments. To take into account the specific management of chronic diseases, the knowledge base of the guided mode of the system ASTI has been structured as a double level decision tree, a clinical level to characterize the clinical profile of a patient, and a therapeutic level to identify the new recommended treatment when taking into account the patients therapeutic history. We propose to automatically derive the therapeutic level of the decision tree from the formal expression of guideline-based therapeutic strategies. The method has been developed using Augmented Transition Networks. Preliminary results obtained with additive therapeutic strategies such as (A, A+ B, A + B + C) where A, B, and C are therapeutic classes which can be substituted respectively by (A′, A′′), (B′, B′′), and (C′, C′′) are promising. However, the method needs to be extended to take into account more complex patterns.

Computerized Drug Prescription Decision Support

Drug prescription has to satisfy three quality criteria. Orders have to be adapted to the patient state, be compatible with all the other drugs of the prescription, and in compliance with the recommendations described in clinical practice guidelines (CPGs). Computer provider order entry systems (CPOEs) have been developed to secure drug orders and they address the first two criteria. Clinical decision support systems (CDSSs) have been developed to improve the implementation of CPGs and promote evidence-based medicine. This chapter first introduces the different medication errors. Then, the general architecture of CPOEs (user interface, drug database, interface with electronic medical records (EMRs) and inference engine) is presented. The main modalities of entering drug orders are described. Alert generation for contra-indications, or drug-drug interactions, are detailed. CDSSs are tools to provide patient-specific recommended treatments. They rely on a knowledge base embedding CPGs. The translation process of CPGs from their original narrative format to a structured formalized representation is described. The difficulty of text translation is emphasized and documentary tools such as GEM that help formalize guideline content are described. The main guideline representation formalisms, Arden Syntax, decision trees, EON and GLIF, are presented. Then, ways of operating CDSSs are described, from the totally automated alert-based mode, to various documentary approaches where the user navigates through a structured knowledge base. Finally, examples of clinical decision support systems currently routinely used are given.

Place de la «réactance» des médecins dans la e-iatrogénie: Une étude de cas avec ASTI mode guidé sur la prise en charge de l’hypertension artérielle

Clinical decision support systems (CDSSs) have the potential to increase physician’s adherence to clinical practice guidelines (CPGs), but factors of success are not well understood. ASTI-GM is an on-demand guideline-based CDSS where the user interactively describes patient characteristics while browsing the system knowledge base to obtain the recommended treatment. We performed a web-based evaluation of ASTI-GM, carried out as a before-after study, where general practitioners (GPs) were asked to solve 5 clinical cases, first without ASTI-GM, then using the system. During a 2-month period, 266 GPs participated and 1,981 prescription orders were collected. Of the 136 GPs that solved case #2 on the management of hypertension, compliance with best practices increased from 69.1% to 80.9% with ASTI-GM. When the navigation matched the set of patient’s parameters described in the clinical case, the increase was even higher and reached 92.9%. E-iatrogenesis has been measured at 19.1%, with 5.1% of commission errors, 8.1% of negative reactance, and 5.9% of neutral reactance.

Caractérisation de l’évolution d’une base de connaissances suite à la réactualisation de recommandations de pratiques cliniques: une étude de cas sur la prise en charge des tumeurs de la vessie

The constant evolution of medical knowledge requires clinical practice guidelines (CPGs) to be regularly updated Clinical decision support systems based on computerized CPGs have thus to be revised as their knowledge sources evolve. We propose a formal characterization of recommendations evolution between two successive CPG versions. Each atomic recommendation is represented as a rule linking a clinical situation with a recommended action plan. Using subsumption functions on recommendation components, we identified 7 evolution patterns of recommendations. The method has been applied on the knowledge bases of the 2002 version of the French bladder cancer guidelines and its 2004 revision. Surprisingly, despite similar coverage, only 54% of the 2002 version was formally reused in 2004 and 36% of the 2004 version were considered new recommendations. Evaluation of CPG update on actual patient medical records from the Urology Department of the Bichat-Claude Bernard Hospital showed a significant fall of physician decisions compliance with guidelines as a side-effect of the CPG revision.

Computer-Aided Assessment of Drug-Induced Lung Disease Plausibility

Drug-induced lung disease (DILD), often suspected in pneumology, is still a diagnostic challenge because of the ever increasing number of pneumotoxic drugs and the large diversity of observed clinical patterns. As a result, DILD can only be evoked as a plausible diagnosis after the exclusion of all other possible causes. PneumoDoc is a computer-based decision support that formalises the evaluation process of the drug-imputability of a lung disease. The knowledge base has been structured as a two-level decision tree. Patient-specific chronological and semiological criteria are first examined leading to the assessment of a qualitative intrinsic DILD plausibility score. Then literature-based data including the frequency of DILD with a given drug and the frequency of the observed clinical situation among the clinical patterns reported with the same drug are evaluated to compute a qualitative extrinsic DILD plausibility score. Based on a simple multimodal qualitative model, extrinsic and intrinsic scores are combined to yield an overall DILD plausibility score.