Using preference learning for detecting inconsistencies in clinical practice guidelines: Methods and application to antibiotherapy

Tsopra, Rosy; Lamy, Jean-Baptiste; Sedki, Karima

doi:10.1016/j.artmed.2018.04.013

Cited by 27 publications

(33 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In previous studies [20,21,19,37], we built a knowledge base describing antibiotics in terms of 11 features used by experts to establish recommendations (Table 1). Each feature is Boolean, and its value depends on the antibiotic, the patient profile (e.g.…”

Section: Contextmentioning

confidence: 99%

“…This knowledge base was formalized as an OWL 2.0 ontology [19]. It contains 144,038 RDF triples describing 5,696 classes, 19 properties and 34,483 axioms, and it belongs to the ALC(D) family 2 of description logics (DL).…”

Section: Contextmentioning

confidence: 99%

“…The proposed preference model had an error rate of 3.5% (115 errors in 3,300 prescriptions). In our previous study [19], we used preference learning to identify candidate inconsistencies in CPGs. Our previous preference model had slightly higher error rates (3.8% for first-line treatment and 4.0% for any line), and about half these errors were manually classified as being due to inconsistencies in CPGs.…”

Section: Preference Learningmentioning

confidence: 99%

“…In a recent study [19], we used preference learning to detect inconsistencies in CPGs for antibiotic treatment. We described the learning of a simple preference model from an ontology of antibiotic treatments [20,21] including CPG recommendations.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Explainable decision support through the learning and visualization of preferences from a formal ontology of antibiotic treatments

Lamy

Sedki

Tsopra

2020

Journal of Biomedical Informatics

Self Cite

View full text Add to dashboard Cite

The aim of eXplainable Artificial Intelligence (XAI) is to design intelligent systems that can explain their predictions or recommendations to humans. Such systems are particularly desirable for therapeutic decision support, because physicians need to understand recommendations to have confidence in their application and to adapt them if required, e.g. in case of patient contraindication. We propose here an explainable and visual approach for decision support in antibiotic treatment, based on an ontology. There were three steps to our method. We first generated a tabular dataset from the ontology, containing features defined on various domains and n-ary features. A preference model was then learned from patient profiles, antibiotic features and expert recommendations found in clinical practice guidelines. This model made the implicit rationale of the expert explicit, including the way in which missing data was treated. We then visualized the preference model and its application to all antibiotics available on the market for a given clinical situation, using rainbow boxes, a recently developed technique for set visualization. The resulting preference model had an error rate of 3.5% on the learning data, and 5.2% on test data (10-fold validation). These findings suggest that our system can help physicians to prescribe antibiotics correctly, even for clinical situations not present in the guidelines (e.g. due to allergies or contraindications for the recommended treatment).

show abstract

Section: Contextmentioning

confidence: 99%

Section: Contextmentioning

confidence: 99%

Section: Preference Learningmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Explainable decision support through the learning and visualization of preferences from a formal ontology of antibiotic treatments

Lamy

Sedki

Tsopra

2020

Journal of Biomedical Informatics

Self Cite

View full text Add to dashboard Cite

show abstract

“…In a previous work [16], we built a knowledge base on antibiotic properties. From this knowledge base, we extracted 9 rules for the prescription of antibiotics for urinary infections (Table I).…”

Section: A Designing and Learning The Neural Networkmentioning

confidence: 99%

Visual Explanation of Simple Neural Networks using Interactive Rainbow Boxes

Lamy¹,

Tsopra

2019

2019 23rd International Conference Information Visualisation (IV)

Self Cite

View full text Add to dashboard Cite

How Useful? Fish-Friendly Irrigation Guidelines for the Lower Mekong Lack Definition in Five Key Areas

Duncan

Bond

Conallin

et al. 2023

Environmental Management

View full text Add to dashboard Cite

A proliferation of irrigation infrastructure throughout the Mekong River has impacted the ability of certain fish species to migrate to fulfil their lifecycle. In response, fishways, a type of fish-friendly irrigation structure, have been developed to provide passage for these fish. In recent years, several guidelines documents providing guidance on fish-friendly irrigation structures and their construction have been published. The development process from guideline inception to publication is unclear, while their purpose, audience, and contribution to fishway practice are vague. This study is the first to review the development of three fish-friendly guideline documents, using structural criteria analysis, combined with qualitative data from 27 key informant interviews. It aimed to understand document elements such as purpose, audience, scope, and framing. The results showed reviewed guideline utility and impact could be improved by attention to five key aspects, namely: definition of target audience; engagement of target audience in guideline design; definition of guideline scope; specificity of recommendations; and evaluation. Attention to these 5 aspects may result in guidelines that are perceived as more useful by their target audience and have greater impact on water management practice.

show abstract

Using preference learning for detecting inconsistencies in clinical practice guidelines: Methods and application to antibiotherapy

Cited by 27 publications

References 42 publications

Explainable decision support through the learning and visualization of preferences from a formal ontology of antibiotic treatments

Explainable decision support through the learning and visualization of preferences from a formal ontology of antibiotic treatments

Visual Explanation of Simple Neural Networks using Interactive Rainbow Boxes

How Useful? Fish-Friendly Irrigation Guidelines for the Lower Mekong Lack Definition in Five Key Areas

Contact Info

Product

Resources

About