Explore the vast range of titles available.

Understanding users’ needs is crucial to building and maintaining high quality software. Online software user feedback has been shown to contain large amounts of information useful to requirements engineering (RE). Previous studies have created machine learning classifiers for parsing this feedback for development insight. While these classifiers report generally good performance when evaluated on a test set, questions remain as to how well they extend to unseen data in various forms. This study evaluates machine learning classifiers’ performance on feedback for two common classification tasks (classifying bug reports and feature requests). Using seven datasets from prior research studies, we investigate the performance of classifiers when evaluated on feedback from different apps than those contained in the training set and when evaluated on completely different datasets (coming from different feedback channels and/or labelled by different researchers). We also measure the difference in performance of using channel-specific metadata as a feature in classification. We find that using metadata as features in classifying bug reports and feature requests does not lead to a statistically significant improvement in the majority of datasets tested. We also demonstrate that classification performance is similar on feedback from unseen apps compared to seen apps in the majority of cases tested. However, the classifiers evaluated do not perform well on unseen datasets. We show that multi-dataset training or zero shot classification approaches can somewhat mitigate this performance decrease. We discuss the implications of these results on developing user feedback classification models to analyse and extract software requirements.

We study the problem of predicting the next query to be recommended in interactive data exploratory analysis to guide users to correct content. Current query prediction approaches are based on sequence-to-sequence learning, exploiting past interaction data. However, due to the resource-hungry training process, such approaches fail to adapt to immediate user feedback. Immediate feedback is essential and considered as a signal of the user’s intent. We contribute with a novel query prediction ensemble mechanism, which adapts to immediate feedback relying on multi-armed bandits framework. Our mechanism, an extension to the popular Exp3 algorithm, augments Transformer-based language models for query predictions by combining predictions from experts, thus dynamically building a candidate set during exploration. Immediate feedback is leveraged to choose the appropriate prediction in a probabilistic fashion. We provide comprehensive large-scale experimental and comparative assessment using a popular online literature discovery service, which showcases that our mechanism (i) improves the per-round regret substantially against state-of-the-art Transformer-based models and (ii) shows the superiority of causal language modelling over masked language modelling for query recommendations.

[...]our research aims to implement a method for processing misspelled words in Indonesian language using the Levenshtein distance algorithm, which has been proven to search for similar words [23] in the text preprocessing phase. - [...]two classification models using Naïve Bayes algorithm are built, one of them utilizes the method for preprocessing misspelled words and the other does not. - [...]we test the both ability of Levenshtein distance in processing misspelled words in the dataset and the ability of the models to classify texts from the test set. [...]we collected a portion of the feedback data from previous work which can be found in [11] for training and testing the classification model.

Significance & Background: The use of Closed System Transfer Devices (CSTD) for hazardous drug (HD) administration is an industry standard and required by USP . After adopting use of a CSTD in 2019, nurse leaders at a NCI-designated Comprehensive Cancer center recognized a significant increase in accidental disconnections in home infusions of HDs. These disconnections caused HD exposure, lost medication, and patient distress. While home administration of HDs lasting longer than 24 hours is not uncommon, most CSTD on the market are only labeled for less than 24 hours of use. Purpose: With the rising demand for outpatient treatments, ensuring the safe administration of hazardous drugs in the home setting is critical. The aim of this QI project was to systematically evaluate the usability of CSTD for home infusion based on error rates and staff preference for three devices. The goal was to find a CSTD that could be used for home infusion without causing increased errors. Interventions: Three different devices were trialed, one at a time, and error rates evaluated monthly via incident reports within the Chemotherapy Quality and Safety council. Results: Error rates were calculated using the number of reported errors vs dispenses. End user and patient feedback were also collected and considered. The baseline error rate for home infusions from 6 months prior to first intervention was 0.21%. The first intervention added a securement device to the BD PhaSeal N35-C35 product, yielding an unacceptable error rate of 1.07%. The trial was discontinued and the error rate dropped back down to 0.27%. The 2nd intervention, the N40-C35 locking injector yielded 1.79% error rate. The project team switched to a totally different device Texium-Smart Site system. The error rate with this device was 0.18% and thus was accepted. Discussion: While CSTD provide important protections from HD exposure, most are not indicated for use longer than 24 hours. This presents a challenge for institutions giving HDs via home infusion pump. Through the systematic evaluation of products and careful attention to end-user feedback, this system was able to identify a product that maintained protections of the CSTD without increasing the error rate. Institutions need to be aware of the limitations of these devices and apply similar methodologies to evaluate their use in unique cases such as home infusion.

IntroductionResponsiveness to service users’ views is a widely recognised objective of health systems. A key component of responsive health systems is effective interaction between users and service providers. Despite a growing literature on patient feedback from high-income settings, less is known about effectiveness of such systems in low-income and middle-income countries.Methodology and analysisThis paper disseminates the protocol for an 18-month ‘RESPOND’ project that aims to evaluate the system of collecting and responding to user feedback in Bangladesh. This mixed-method study uses a realist evaluation approach to examine user feedback systems at two Upazila health complexes in Comilla District of Bangladesh, and comprises three steps: (1) initial theory development; (2) theory validation; and (3) theory refinement and development of lessons learnt. The project also uses (1) process evaluation to understand causal mechanisms and contexts of implementation; (2) statistical analysis of patient feedback to clarify the nature of issues reported; (3) social science methods to illuminate feedback processes and user and provider experiences; and (4) health policy and systems research to clarify issues related to integration of feedback systems with quality assurance and human resource management. During data analysis, qualitative and quantitative findings will be integrated throughout to help achieve study objectives. Analysis of qualitative and quantitative data will be done using a convergent mixed-methods model, involving continuous triangulation of multiple data sets to facilitate greater understanding of the context of user feedback systems including the links with relevant policies, practices and programmes.Ethics and disseminationEthics approvals were obtained from the University of Leeds and the Bangladesh Medical Research Council. All data collected for this study will be anonymised, and identifying characteristics of respondents will not appear in a final manuscript or reports. The study findings will be presented at scientific conferences and published in peer-reviewed journals.

This article provides updated guidance on methods for conducting rapid reviews of effectiveness, targeted at Cochrane and other stakeholders interested in the methodology of rapid reviews. The guidance, developed by the Cochrane Rapid Reviews Methods Group, builds upon previous interim guidance, and incorporates changes based on an evaluation of its application, a scope of the literature on rapid review methodology, and input from a diverse group of experts in rapid review methods. The guidance consists of 24 specific recommendations supporting the conduct of rapid reviews, applicable both within and outside Cochrane. It underscores the importance of considering the appropriateness of undertaking rapid reviews and advocates for a tailored, iterative approach to each review. Key defining features of rapid reviews, such as restricted methods, how the dimension of timelines factors into rapid reviews, and the involvement of knowledge users (eg, patient and public partners, healthcare providers, policy makers), are outlined. The paper presents a definition of a Cochrane rapid review and additional considerations for rapid reviews of effectiveness to enhance the efficiency of the review process. In conclusion, the Cochrane Rapid Review Methods Group’s updated guidance, complemented by examples, seeks to guide methodological decisions in the design and conduct of rapid reviews, facilitating timely decision making in healthcare.

Purpose The purpose of this paper is to describe a small scale pilot study undertaken in Northern Ireland to gather service user feedback from individuals who have been subject to adult safeguarding procedures. Design/methodology/approach The aims, methods and findings of the “Adult Safeguarding: 10,000 Voices” pilot project are presented. Findings The pilot project highlighted how an initiative which captures the experiences of patients, service users, carers and staff in the health and social care sector (10,000 Voices) could be successfully adapted for use in adult safeguarding, facilitating the collation of complex experiences and enabling insights to be gleaned and shared. Research limitations/implications The pilot study is limited by the small number of participants. The findings are preliminary. Practical implications For the first time in Northern Ireland the 10,000 Voices model was utilised in the context of a non-health related service, namely, adult safeguarding. Social implications This outline of the model and methodology for obtaining service user feedback can inform user involvement in other contexts. Originality/value This paper provides an accessible overview of an innovative approach to engaging service users in adult safeguarding, such approaches, to date have been limited.

IntroductionA novel, mobile ecological momentary assessment (EMA) platform, Recovering Concussion Update on Progression of Symptoms (ReCoUPS) allows clinicians to remotely monitor concussion symptoms in real-time. This study evaluated the psychometric properties and usability of a concussion symptom checklist and psychological health-related quality of life (PHRQoL) inventories administered by ReCoUPS in healthy individuals.Materials and MethodsSeventy-nine healthy individuals (female=54; μ age=20.85+2.49 years) enrolled in our test-retest reliability study using ReCoUPS. Thirty survey questions from the Sport Concussion Assessment Tool6 (SCAT6) symptom checklist and PROMIS Emotional Distress Short Forms Anxiety and Depression (PROMISAnx and PROMISDepress), were administered at random times, daily, for 7 days via text messages. After 8 days, participants completed recalled PHRQoL (rPHRQoL) inventories (‘within the past 7 days’) and a user feedback survey. Cronbach’s alpha (α) determined the internal consistency of mPHRQoL. Intraclass correlation coefficients (ICC) assessed agreement between rHRQoL and mPHRQoL. Spearman’s rho correlation coefficients evaluated relationships between mPHRQoL and SCAT6 symptom clusters.ResultsInternal consistency was strong for all mPHRQoL items (PROMISAnx α=0.82; PROMISDepress α=0.94). Excellent agreement was demonstrated between rPHRQoL and mPHRQoL (PROMISAnxICC=0.99, 95% CI:0.98–0.99; PROMISDepress ICC=0.99, 95% CI: −0.99–0.99). Both mPROMISAnx (affective-rs=0.79, p<0.001; cognitive-fatigue-rs=0.66, p<0.001; migraine-rs=0.45; p<0.001) and mPROMISDepress (affective-rs=0.67, p<0.001; cognitive-fatigue-rs= 0.58, p<0.001; migraine-rs=0.25; p<0.001) were significantly correlated with all symptom clusters. Twenty-six (33%) participants completed the user feedback survey, with >73% rating ReCoUPS highly for usability and health benefits.ConclusionsThese findings ensure the robustness and usability of ReCoUPS to remotely monitor concussion symptoms and PHRQoL in healthy participants.

The significant advancements in applying artificial intelligence (AI) to healthcare decision-making, medical diagnosis, and other domains have simultaneously raised concerns about the fairness and bias of AI systems. This is particularly critical in areas like healthcare, employment, criminal justice, credit scoring, and increasingly, in generative AI models (GenAI) that produce synthetic media. Such systems can lead to unfair outcomes and perpetuate existing inequalities, including generative biases that affect the representation of individuals in synthetic data. This survey study offers a succinct, comprehensive overview of fairness and bias in AI, addressing their sources, impacts, and mitigation strategies. We review sources of bias, such as data, algorithm, and human decision biases—highlighting the emergent issue of generative AI bias, where models may reproduce and amplify societal stereotypes. We assess the societal impact of biased AI systems, focusing on perpetuating inequalities and reinforcing harmful stereotypes, especially as generative AI becomes more prevalent in creating content that influences public perception. We explore various proposed mitigation strategies, discuss the ethical considerations of their implementation, and emphasize the need for interdisciplinary collaboration to ensure effectiveness. Through a systematic literature review spanning multiple academic disciplines, we present definitions of AI bias and its different types, including a detailed look at generative AI bias. We discuss the negative impacts of AI bias on individuals and society and provide an overview of current approaches to mitigate AI bias, including data pre-processing, model selection, and post-processing. We emphasize the unique challenges presented by generative AI models and the importance of strategies specifically tailored to address these. Addressing bias in AI requires a holistic approach involving diverse and representative datasets, enhanced transparency and accountability in AI systems, and the exploration of alternative AI paradigms that prioritize fairness and ethical considerations. This survey contributes to the ongoing discussion on developing fair and unbiased AI systems by providing an overview of the sources, impacts, and mitigation strategies related to AI bias, with a particular focus on the emerging field of generative AI.

Background Systematic review with meta-analysis integrates findings from multiple studies, offering robust conclusions on treatment effects and guiding evidence-based medicine. However, the process is often hampered by challenges such as inconsistent data reporting, complex calculations, and time constraints. Researchers must convert various statistical measures into a common format, which can be error-prone and labor-intensive without the right tools. Implementation Meta-Analysis Accelerator was developed to address these challenges. The tool offers 21 different statistical conversions, including median & interquartile range (IQR) to mean & standard deviation (SD), standard error of the mean (SEM) to SD, and confidence interval (CI) to SD for one and two groups, among others. It is designed with an intuitive interface, ensuring that users can navigate the tool easily and perform conversions accurately and efficiently. The website structure includes a home page, conversion page, request a conversion feature, about page, articles page, and privacy policy page. This comprehensive design supports the tool’s primary goal of simplifying the meta-analysis process. Results Since its initial release in October 2023 as Meta Converter and subsequent renaming to Meta-Analysis Accelerator, the tool has gained widespread use globally. From March 2024 to May 2024, it received 12,236 visits from countries such as Egypt, France, Indonesia, and the USA, indicating its international appeal and utility. Approximately 46% of the visits were direct, reflecting its popularity and trust among users. Conclusions Meta-Analysis Accelerator significantly enhances the efficiency and accuracy of meta-analysis of systematic reviews by providing a reliable platform for statistical data conversion. Its comprehensive variety of conversions, user-friendly interface, and continuous improvements make it an indispensable resource for researchers. The tool’s ability to streamline data transformation ensures that researchers can focus more on data interpretation and less on manual calculations, thus advancing the quality and ease of conducting systematic reviews and meta-analyses.