3,589 publications found
Sort by
Machine learning models to detect and predict patient safety events using electronic health records: A systematic review.

Identifying patient safety events using electronic health records (EHRs) and automated machine learning-based detection methods can help improve the efficiency and quality of healthcare service provision. This study aimed to systematically review machine learning-based methods and techniques, as well as their results for patient safety event management using EHRs. We reviewed the studies that focused on machine learning techniques, including automatic prediction and detection of patient safety events and medical errors through EHR analysis to manage patient safety events. The data were collected by searching Scopus, PubMed (Medline), Web of Science, EMBASE, and IEEE Xplore databases. After screening, 41 papers were reviewed. Support vector machine (SVM), random forest, conditional random field (CRF), and bidirectional long short-term memory with conditional random field (BiLSTM-CRF) algorithms were mostly applied to predict, identify, and classify patient safety events using EHRs; however, they had different performances. BiLSTM-CRF was employed in most of the studies to extract and identify concepts, e.g., adverse drug events (ADEs) and adverse drug reactions (ADRs), as well as relationships between drug and severity, drug and ADEs, drug and ADRs. Recurrent neural networks (RNN) and BiLSTM-CRF had the best results in detecting ADEs compared to other patient safety events. Linear classifiers and Naive Bayes (NB) had the highest performance for ADR detection. Logistic regression had the best results in detecting surgical site infections. According to the findings, the quality of articles has non-significantly improved in recent years, but they had low average scores. Machine learning can be useful in automatic detection and prediction of patient safety events. However, most of these algorithms have not yet been externally validated or prospectively tested. Therefore, further studies are required to improve the performance of these automated systems.

Just Published
Relevant
Patient and staff experiences of using technology-enabled and analogue models of remote home monitoring for COVID-19 in England: A mixed-method evaluation.

To evaluate patient and staff experiences of using technology-enabled ('tech-enabled') and analogue remote home monitoring models for COVID-19, implemented in England during the pandemic. Twenty-eight sites were selected for diversity in a range of criteria (e.g. pre-hospital or early discharge service, mode of patient data submission). Between February and May 2021, we conducted quantitative surveys with patients, carers and staff delivering the service, and interviewed patients, carers, and staff from 17 of the 28 services. Quantitative data were analysed using descriptive statistics and both univariate and multivariate analyses. Qualitative data were interpreted using thematic analysis. Twenty-one sites adopted mixed models whereby patients could submit their symptoms using either tech-enabled (app, weblink, or automated phone calls) or analogue (phone calls with a health professional) options; seven sites offered analogue-only data submission (phone calls or face-to-face visits with a health professional). Sixty-two patients and carers were interviewed, and 1069 survey responses were received (18% response rate). Fifty-eight staff were interviewed, and 292 survey responses were received (39% response rate). Patients who used tech-enabled modes tended to be younger (p=0.005), have a higher level of education (p=0.011), and more likely to identify as White British (p=0.043). Most patients found relaying symptoms easy, regardless of modality, though many received assistance from family or friends. Staff considered the adoption of mixed delivery models beneficial, enabling them to manage large patient numbers and contact patients for further assessment as needed; however, they suggested improvements to the functionality of systems to better fit clinical and operational needs. Human contact was important in all remote home monitoring options. Organisations implementing tech-enabled remote home monitoring at scale should consider adopting mixed models which can accommodate patients with different needs; focus on the usability and interoperability of tech-enabled platforms; and encourage digital inclusivity for patients.

Open Access
Relevant
Mobile health interventions for cancer patient education: A scoping review.

Mobile health (mHealth) is using mobile devices and applications to deliver health information and services. mHealth has been increasingly applied in cancer care to support patients in various aspects of their disease journey. This scoping review aimed to explore the current evidence on the use of mHealth interventions for cancer patient education. This scoping review followed the Joanna Briggs Institute (JBI) methodology for scoping reviews. We searched four electronic databases (PubMed, Web of Science, CINAHL, and Cochrane) using a combination of keywords related to mHealth, cancer, and education. After finding articles at the initial search the screening has been done based on the inclusion and exclusion criteria. We included only original research articles and excluded all other types of publications, such as review papers, reports, editorials, letters to the editor, book reviews, short communications, conference proceedings, graduate dissertations, protocols, and commentaries. We extracted data on the characteristics and outcomes of the included studies using a standardized form. We conducted a narrative synthesis and inductive content analysis to summarize and categorize the evidence. Out of 2131 records found in the initial search, 28 full-text articles reported on the use of mHealth educational interventions for cancer patients. The majority of the studies focused on breast cancer patients (n=21, 75%). The most common type of mHealth intervention was exercise-based education delivered through various media such as text messages, videos, audio, images, and social networks. The main objectives of mHealth educational interventions were to enhance self-management skills, improve psychological well-being, and promote healthy lifestyle behaviors among cancer patients. The reported outcomes of mHealth interventions included reduced chemotherapy-related side effects, improved mental health, improved quality of life and lifestyle, and better pain management. This scoping review showed that mHealth is a promising and feasible modality for delivering educational interventions to cancer patients. However, more rigorous and diverse studies are needed to evaluate the effectiveness and cost-effectiveness of mHealth interventions for different types of cancers, stages, and settings.

Relevant
Machine learning prediction models for clinical management of blood-borne viral infections: a systematic review of current applications and future impact.

Machine learning (ML) prediction models to support clinical management of blood-borne viral infections are becoming increasingly abundant in medical literature, with a number of competing models being developed for the same outcome or target population. However, evidence on the quality of these ML prediction models are limited. This study aimed to evaluate the development and quality of reporting of ML prediction models that could facilitate timely clinical management of blood-borne viral infections. We conducted narrative evidence synthesis following the synthesis without meta-analysis guidelines. We searched PubMed and Cochrane Central Register of Controlled Trials for all studies applying ML models for predicting clinical outcomes associated with hepatitis B virus (HBV), human immunodeficiency virus (HIV), or hepatitis C virus (HCV). We found 33 unique ML prediction models aiming to support clinical decision making. Overall, 12 (36.4%) focused on HBV, 10 (30.3%) on HCV, 10 on HIV (30.3%) and two (6.1%) on co-infection. Among these, six (18.2%) addressed the diagnosis of infection, 16 (48.5%) the prognosis of infection, eight (24.2%) the prediction of treatment response, two (6.1%) progression through a cascade of care, and one (3.03%) focused on the choice of antiretroviral therapy (ART). Nineteen prediction models (57.6%) were developed using data from high-income countries. Evaluation of prediction models was limited to measures of performance. Detailed information on software code accessibility was often missing. Independent validation on new datasets and/or in other institutions was rarely done. Promising approaches for ML prediction models in blood-borne viral infections were identified, but the lack of robust validation, interpretability/explainability, and poor quality of reporting hampered their clinical relevance. Our findings highlight important considerations that can inform standard reporting guidelines for ML prediction models in the future (e.g., TRIPOD-AI), and provides critical data to inform robust evaluation of the models.

Open Access
Relevant
A multilayered graph-based framework to explore behavioural phenomena in social media conversations.

Social media is part of current health communications. This research aims to delve into the effects of social contagion, biased assimilation, and homophily in building and changing health opinions on social media. Conversations about COVID-19 vaccination on English and Spanish Twitter are the case studies. A new multilayered graph-based framework supports the integrated analysis of content similarity within and across posts, users, and conversations to interpret contrasting and confluent user stances. Deep learning models are applied to infer stance. Graph centrality and homophily scores support the interpretation of information reproduction. The results show that semantically related English posts tend to present a similar stance about COVID-19 vaccination (rstance=0.51) whereas Spanish posts are more heterophilic (rstance=0.38). Neither case showed evidence of homophily regarding user influence or vaccine hashtags. Graph filters for Pfizer and Astrazeneca with a similarity threshold of 0.85 show stance homophily in English scenarios (i.e. rstance=0.45 and rstance=0.58, respectively) and small homophily in Spanish scenarios (i.e. r=0.12 and r=0.3, respectively). Highly connected users are a minority and are not socially influential. Spanish conversations showed stance homophily, i.e. most of the connected conversations promote vaccination (rstance=0.42), whereas English conversations are more likely to offer contrasting stances. The methodology proposed for quantifying the impact of natural and intentional social behaviours in health information reproduction can be applied to any of the main social platforms and any given topic of conversation. Its effectiveness was demonstrated by two case studies describing English and Spanish demographic and sociocultural scenarios.

Open Access
Relevant
A comprehensive evaluation of ChatGPT consultation quality for augmentation mammoplasty: A comparative analysis between plastic surgeons and laypersons.

ChatGPT has gained significant popularity as a source of healthcare information among the general population. Evaluating the quality of chatbot responses is crucial, requiring comprehensive and qualitative analysis. This study aims to assess the answers provided by ChatGPT during hypothetical breast augmentation consultations across various categories and depths. The evaluation involves the utilization of validated tools and a comparison of scores between plastic surgeons and laypersons. A panel consisting of five plastic surgeons and five laypersons evaluated ChatGPT's responses to 25 questions spanning consultation, procedure, recovery, and sentiment categories. The DISCERN and PEMAT tools were employed to assess the responses, while emotional context was examined through ten specific questions. Additionally, readability was measured using the Flesch Reading Ease score. Qualitative analysis was performed to identify the overall strengths and weaknesses. Plastic surgeons generally scored lower than laypersons across most domains. Scores for each evaluation domain varied by category, with the consultation category demonstrating lower scores in terms of DISCERN reliability, information quality, and DISCERN score. Plastic surgeons assigned significantly lower overall quality ratings to the procedure category compared to other question categories. They also gave lower emotion scores in the procedure category compared to laypersons. The depth of the questions did not impact the scoring. Existing health information evaluation tools may not be entirely suitable for comprehensively evaluating the quality of individual responses generated by ChatGPT. Consequently, the development and implementation of appropriate evaluation tools to assess the appropriateness and quality of AI consultations are necessary.

Open Access
Relevant
Predicting COVID-19 severity: Challenges in reproducibility and deployment of machine learning methods.

The increasing use of electronic health records (EHR) based computable phenotypes in clinical research is providing new opportunities for development of data-driven medical applications. Adopted widely in the United States and globally, EHRs facilitate systematic collection of patients' longitudinal information, which serves as one of the important foundations for artificial intelligence applications in medicine. Harmonization of input variables and outcome definitions is critically important for wider clinical applicability of artificial intelligence (AI) methodologies. In this review, we focused on Coronavirus Disease 2019 (COVID-19) severity machine learning prediction models and explored the pipeline for standardizing future disease severity model development using EHR information. We identified 2,967 studies published between 01/01/2020 and 02/15/2022 and selected 135 independent studies that had built machine learning prediction models to predict severity related outcomes of COVID-19 patients based on EHR data for the final review. These 135 studies spanning across 27 counties covered a broad range of severity related prediction outcomes. We observed substantial inconsistency in COVID-19 severity phenotype definitions among models in these studies. Moreover, there was a gap between the outcome of these models and clinician-recognized clinical concepts. Accordingly, we recommend that robust clinical input metrics, with outcome definitions which eliminate ambiguity in interpretation, to reduce algorithmic bias, mitigate model brittleness and improve generalizability of a universal model for COVID-19 severity. This framework can potentially be extended to broader clinical application.

Relevant
On the inter-dataset generalization of machine learning approaches to Parkinson's disease detection from voice.

Parkinson's disease is the second-most-common neurodegenerative disorder that affects motor skills, cognitive processes, mood, and everyday tasks such as speaking and walking. The voices of people with Parkinson's disease may become weak, breathy, or hoarse and may sound emotionless, with slurred words and mumbling. Algorithms for computerized voice analysis have been proposed and have shown highly accurate results. However, these algorithms were developed on single, limited datasets, with participants possessing similar demographics. Such models are prone to overfitting and are unsuitable for generalization, which is essential in real-world applications. We evaluated the computerized Parkinson's disease diagnosis performance of various machine learning models and showed that these models degraded rapidly when used on different datasets. We evaluated two mainstream state-of-the-art approaches, one based on deep convolutional neural networks and another based on voice feature extraction followed by a shallow classifier (i.e., extreme gradient boosting (XGBoost)). An investigation with four datasets (CzechPD, PC-GITA, ITA, and RMIT-PD) proved that even if the algorithms yielded excellent performance on a single dataset, the results obtained on new data or even a mix of datasets were very unsatisfactory. More work needs to be done to make computerized voice analysis methods for Parkinson's disease diagnosis suitable for real-world applications.

Relevant
Change agent's role in facilitating use of technology in residential aged care: A systematic review.

To synthesise the evidence on the roles and outcomes of change agents in facilitating the use of powered technology systems and devices for staff end-users in residential aged care workplaces. Systematic review and narrative synthesis. CINAHL, MEDLINE and EMBASE databases were searched for articles published in English between January 2010 and July 2022. Two of three reviewers independently screened each title and abstract, and subsequently the full texts of selected records. The Mixed Method Appraisal Tool was used to assess the quality of the included articles. Of 3,680 records identified, ten articles reporting nine studies were included. In all the studies, the change agent role was a minor component embedded within implementation processes. Three key change agent roles were identified: 1) providing guidance, expertise, and support with implementing a new technology; 2) delivering training to others, and 3) troubleshooting and responding to issues. The key outcome of change agent roles related to achieving project implementation milestones and higher levels of implementation of technology. Change agent processes, however, were compromised when the designated change agent role was included late in the implementation process, or was not supported, recognised, embraced, or when roles or responsibilities were unclear. The direct contribution of change agents was difficult to elucidate because the roles and outcomes of change agents were embedded in multi-faceted implementation strategies. The change agent can play an important role in facilitating technology implementation by providing support, training, and troubleshooting. Challenges with the change agent role included unclear role expectations and appointment late in the implementation process. Overall, there was limited evidence specific to the role and outcome of the change agent role to inform ideal approaches for their role in technology facilitation for end-users in residential aged care workplaces.

Open Access
Relevant