Studies in Health Technology and Informatics

Evidence-based clinical decision making in oncology is challenging. Multi-disciplinary team (MDTs) meetings are organized to consider different diagnostic and treatment options. MDT advice are often based on clinical practice guideline recommendations which can be extensive and ambiguous, making it difficult to implement in clinical practice. To address this issue, guideline-based algorithms have been developed. These are applicable in clinical practice and enable accurate guideline adherence evaluation. This ongoing study aims to determine the optimal decision-making approach for different subpopulations of patients with high-incidence gynecological cancers...

Understanding the aspects of progression for atherosclerotic cardiovascular disease and treatment is key to building reliable clinical decision-support systems. To promote system trust, one step is to make the machine learning models (used by the decision support systems) explainable for clinicians, developers, and researchers. Recently, working with longitudinal clinical trajectories using Graph Neural Networks (GNNs) has attracted attention among machine learning researchers. Although GNNs are seen as black-box methods, promising explainable AI (XAI) methods for GNNs have lately been proposed...

In pharmacovigilance, signal assessment of a medicinal product and adverse event can involve reviewing prohibitively large numbers of case reports. A prototype of a decision support tool guided by a needs assessment was developed to help manual review of many reports. In a preliminary qualitative evaluation, users said the tool was easy to use, improved efficiency and provided new insights.

The implementation process in the routine clinical care of a new predictive tool based on machine learning algorithms has been investigated using the RE-AIM framework. Semi-structured qualitative interviews have been conducted with a broad range of clinicians to elucidate potential barriers and facilitators of the implementation process across five major domains: Reach, Efficacy, Adoption, Implementation, and Maintenance. The analysis of 23 clinician interviews demonstrated a limited reach and adoption of the new tool and identified areas for improvement in implementation and maintenance...

The search strategy of a literature review is of utmost importance as it impacts the validity of its findings. In order to build the best query to guide the literature search on clinical decision support systems applied to nursing clinical practice, we developed an iterative process capitalizing on previous systematic reviews published on similar topics. Three reviews were analyzed relatively to their detection performance. Errors in the choice of keywords and terms used in title and abstract (missing MeSH terms, failure to use common terms), may make relevant articles invisible...

Risk of bias (RoB) assessment of randomized clinical trials (RCTs) is vital to conducting systematic reviews. Manual RoB assessment for hundreds of RCTs is a cognitively demanding, lengthy process and is prone to subjective judgment. Supervised machine learning (ML) can help to accelerate this process but requires a hand-labelled corpus. There are currently no RoB annotation guidelines for randomized clinical trials or annotated corpora. In this pilot project, we test the practicality of directly using the revised Cochrane RoB 2...

Glaucoma is one of the leading causes of blindness worldwide. Therefore, early detection and diagnosis are key to preserve full vision in patients. As part of the SALUS study, we create a blood vessel segmentation model based on U-Net. We trained U-Net on three different loss functions and used hyperparameter tuning to find their optimal hyperparameters for each loss function. The best models for each of the loss functions achieved an accuracy of over 93%, Dice scores around 83% and Intersection over Union scores over 70%...

The objective of this study was to compare different convolutional neural networks (CNNs), as employed in a Python-produced deep learning process, used on white light images of colorectal polyps acquired during the process of a colonoscopy, in order to estimate the accuracy of the optical recognition of particular histologic types of polyps. The TensorFlow framework was used for Inception V3, ResNet50, DenseNet121, and NasNetLarge, which were trained with 924 images, drawn from 86 patients.

Preterm birth (PTB) is defined as delivery occurring before 37 weeks of gestation. In this paper, Artificial Intelligence (AI)-based predictive models are adapted to accurately estimate the probability of PTB. In doing so, pregnant women' objective results and variables extracted from the screening procedure in combination with demographics, medical history, social history, and other medical data are used. A dataset consisting of 375 pregnant women is used and a number of alternative Machine Learning (ML) algorithms are applied to predict PTB...

Finding the right time for weaning from ventilator is a difficult clinical decision. Several systems based on machine or deep learning are reported in literature. However, the results of these applications are not completely satisfactory and may be improved. An important aspect is represented by the features used as input of these systems. In this paper we present the results of the application of genetic algorithms to perform feature selection on a dataset containing 13688 patients under mechanical ventilation characterizing by 58 variables, extracted from the MIMIC III database...

Machine learning methods are becoming increasingly popular to anticipate critical risks in patients under surveillance reducing the burden on caregivers. In this paper, we propose an original modeling that benefits of recent developments in Graph Convolutional Networks: a patient's journey is seen as a graph, where each node is an event and temporal proximities are represented by weighted directed edges. We evaluated this model to predict death at 24 hours on a real dataset and successfully compared our results with the state of the art...

The evolution of clinical decision support (CDS) tools has been improved by usage of new technologies, yet there is an increased need to develop user-friendly, evidence-based, and expert-curated CDS solutions. In this paper, we show with a use-case how interdisciplinary expertise can be combined to develop CDS tool for hospital readmission prediction of heart failure patients. We also discuss how to make the tool integrated in clinical workflow by understanding end-user needs and have clinicians-in-the-loop during the different development stages...

Adverse Drug Reactions (ADRs) are an important public health issue as they can impose significant health and monetary burdens. This paper presents the engineering and use case of a Knowledge Graph, supporting the prevention of ADRs as part of a Clinical Decision Support System (CDSS) developed in the context of the PrescIT project. The presented PrescIT Knowledge Graph is built upon Semantic Web technologies namely the Resource Description Framework (RDF), and integrates widely relevant data sources and ontologies, i...

Association rules are one of the most used data mining techniques. The first proposals have considered relations over time in different ways, resulting in the so-called Temporal Association Rules (TAR). Although there are some proposals to extract association rules in OLAP systems, to the best of our knowledge, there is no method proposed to extract temporal association rules over multidimensional models in these kinds of systems. In this paper we study the adaptation of TAR to multidimensional structures, identifying the dimension that establishes the number of transactions and how to find time relative correlations between the other dimensions...

The use and shareability of Clinical Quality Language (CQL) artefacts is an important aspect in enabling the exchange and interoperability of clinical data to support both clinical decisions and research in the medical informatics field. This paper, while basing on use cases and synthetic data, developed purposeful CQL reusable libraries to showcase the possibilities of multidisciplinary teams and how CQLs could be best used to support clinical decision making.

Since its emergence, the COVID-19 pandemic still poses a major global health threat. In this setting, a number of useful machine learning applications have been explored to assist clinical decision-making, predict the severity of disease and admission to the intensive care unit, and also to estimate future demand for hospital beds, equipment, and staff. The present study examined demographic data, hematological and biochemical markers routinely measured in Covid-19 patients admitted to the intensive care unit (ICU) of a public tertiary hospital, in relation to the ICU outcome, during the second and third Covid-19 waves, from October 2020 until February 2022...

VAR Healthcare is a clinical decision support system for nurses that aspires to become even more advanced. By applying The Five Rights model, we have evaluated the status and direction of its development to bring potential lacks or barriers into the fore. The evaluation shows that ensuring APIs that will allow the nurses to combine the assets of VAR Healthcare with information on individual patients from EPRs would bring advanced decision support to nurses. This would adhere to all the principles of the five rights model...

This paper presents the results of a study performed on Parallel Convolutional Neural Network (PCNN) toward detecting heart abnormalities from the heart sound signals. The PCNN preserves dynamic contents of the signal in a parallel combination of the recurrent neural network and a Convolutional Neural Network (CNN). The performance of the PCNN is evaluated and compared to the one obtained from a Serial form of the Convolutional Neural Network (SCNN) as well as two other baseline studies: a Long- and Short-Term Memory (LSTM) neural network and a Conventional CNN (CCNN)...

With the advent of SARS-CoV-2, several studies have shown that there is a higher mortality rate in patients with diabetes and, in some cases, it is one of the side effects of overcoming the disease. However, there is no clinical decision support tool or specific treatment protocols for these patients. To tackle this issue, in this paper we present a Pharmacological Decision Support System (PDSS) providing intelligent decision support for COVID-19 diabetic patient treatment selection, based on an analysis of risk factors with data from electronic medical records using Cox regression...

The application of machine learning (ML) algorithms to electronic health records (EHR) data allows the achievement of data-driven insights on various clinical problems and the development of clinical decision support (CDS) systems to improve patient care. However, data governance and privacy barriers hinder the use of data from multiple sources, especially in the medical field due to the sensitivity of data. Federated learning (FL) is an attractive data privacy-preserving solution in this context by enabling the training of ML models with data from multiple sources without any data sharing, using distributed remotely hosted datasets...