Objective:Acknowledging study limitations in a scientific publication is a crucial element in scientific transparency and progress. However, limitation reporting is often inadequate. Natural language processing (NLP) methods could support automated reporting checks, improving research transparency. In this study, our objective was to develop a dataset and NLP methods to detect and categorize self-acknowledged limitations (e.g., sample size, blinding) reported in randomized controlled trial (RCT) publications.Methods:We created a data model of limitation types in RCT studies and annotated a corpus of 200 full-text RCT publications using this data model. We fine-tuned BERT-based sentence classification models to recognize the limitation sentences and their types. To address the small size of the annotated corpus, we experimented with data augmentation approaches, including Easy Data Augmentation (EDA) and Prompt-Based Data Augmentation (PromDA). We applied the best-performing model to a set of about 12K RCT publications to characterize self-acknowledged limitations at larger scale.Results:Our data model consists of 15 categories and 24 sub-categories (e.g., Population and its sub-category DiagnosticCriteria). We annotated 1090 instances of limitation types in 952 sentences (4.8 limitation sentences and 5.5 limitation types per article). A fine-tuned PubMedBERT model for limitation sentence classification improved upon our earlier model by about 1.5 absolute percentage points in F1 score (0.821 vs. 0.8) with statistical significance (). Our best-performing limitation type classification model, PubMedBERT fine-tuning with PromDA (Output View), achieved an F1 score of 0.7, improving upon the vanilla PubMedBERT model by 2.7 percentage points, with statistical significance ().Conclusion:The model could support automated screening tools which can be used by journals to draw the authors’ attention to reporting issues. Automatic extraction of limitations from RCT publications could benefit peer review and evidence synthesis, and support advanced methods to search and aggregate the evidence from the clinical trial literature.
MULTIFILE
ObjectivesOsteoarthritis (OA) of the foot-ankle complex is understudied. Understanding determinants of pain and activity limitations is necessary to improve management of foot OA. The aim of the present study was to investigate demographic, foot-specific and comorbidity-related factors associated with pain and activity limitations in patients with foot OA.MethodsThis exploratory cross-sectional study included 75 patients with OA of the foot and/or ankle joints. Demographic and clinical data were collected with questionnaires and by clinical examination. The outcome variables of pain and activity limitations were measured using the Foot Function Index (FFI). Potential determinants were categorized into demographic factors (e.g., age, sex), foot-specific factors (e.g., plantar pressure and gait parameters), and comorbidity-related factors (e.g., type and amount of comorbid diseases). Multivariable regression analyses with backward selection (p-out≥0.05) were performed in two steps, leading to a final model.ResultsOf all potential determinants, nine factors were selected in the first step. Five of these factors were retained in the second step (final model): female sex, pain located in the hindfoot, higher body mass index (BMI), neurological comorbidity, and Hospital Anxiety and Depression Scale (HADS) score were positively associated with the FFI score. The explained variance (R2) for the final model was 0.580 (adjusted R2 = 0.549).ConclusionFemale sex, pain located in the hindfoot, higher BMI, neurological comorbidity and greater psychological distress were independently associated with a higher level of foot-related pain and activity limitations. By addressing these factors in the management of foot OA, pain and activity limitations may be reduced.
Objective: To automatically recognize self-acknowledged limitations in clinical research publications to support efforts in improving research transparency.Methods: To develop our recognition methods, we used a set of 8431 sentences from 1197 PubMed Central articles. A subset of these sentences was manually annotated for training/testing, and inter-annotator agreement was calculated. We cast the recognition problem as a binary classification task, in which we determine whether a given sentence from a publication discusses self-acknowledged limitations or not. We experimented with three methods: a rule-based approach based on document structure, supervised machine learning, and a semi-supervised method that uses self-training to expand the training set in order to improve classification performance. The machine learning algorithms used were logistic regression (LR) and support vector machines (SVM).Results: Annotators had good agreement in labeling limitation sentences (Krippendorff's α = 0.781). Of the three methods used, the rule-based method yielded the best performance with 91.5% accuracy (95% CI [90.1-92.9]), while self-training with SVM led to a small improvement over fully supervised learning (89.9%, 95% CI [88.4-91.4] vs 89.6%, 95% CI [88.1-91.1]).Conclusions: The approach presented can be incorporated into the workflows of stakeholders focusing on research transparency to improve reporting of limitations in clinical studies.
Everyone has the right to participate in society to the best of their ability. This right also applies to people with a visual impairment, in combination with a severe or profound intellectual and possibly motor disability (VISPIMD). However, due to their limitations, for their participation these people are often highly dependent on those around them, such as family members andhealthcare professionals. They determine how people with VISPIMD participate and to what extent. To optimize this support, they must have a good understanding of what people with disabilities can still do with their remaining vision.It is currently difficult to gain insight into the visual abilities of people with disabilities, especially those with VISPIMD. As a professional said, "Everything we can think of or develop to assess the functional vision of this vulnerable group will help improve our understanding and thus our ability to support them. Now, we are more or less guessing about what they can see.Moreover, what little we know about their vision is hard to communicate to other professionals”. Therefore, there is a need for methods that can provide insight into the functional vision of people with VISPIMD, in order to predict their options in daily life situations. This is crucial knowledge to ensure that these people can participate in society to their fullest extent.What makes it so difficult to get this insight at the moment? Visual impairments can be caused by a range of eye or brain disorders and can manifest in various ways. While we understand fairly well how low vision affects a person's abilities on relatively simple visual tasks, it is much more difficult to predict this in more complex dynamic everyday situations such asfinding your way or moving around during daily activities. This is because, among other things, conventional ophthalmic tests provide little information about what people can do with their remaining vision in everyday life (i.e., their functional vision).An additional problem in assessing vision in people with intellectual disabilities is that many conventional tests are difficult to perform or are too fatiguing, resulting in either no or the wrong information. In addition to their visual impairment, there is also a very serious intellectual disability (possibly combined with a motor impairment), which makes it even more complex to assesstheir functional vision. Due to the interplay between their visual, intellectual, and motor disabilities, it is almost impossible to determine whether persons are unable to perform an activity because they do not see it, do not notice it, do not understand it, cannot communicate about it, or are not able to move their head towards the stimulus due to motor disabilities.Although an expert professional can make a reasonable estimate of the functional possibilities through long-term and careful observation, the time and correct measurement data are usually lacking to find out the required information. So far, it is insufficiently clear what people with VZEVMB provoke to see and what they see exactly.Our goal with this project is to improve the understanding of the visual capabilities of people with VISPIMD. This then makes it possible to also improve the support for participation of the target group. We want to achieve this goal by developing and, in pilot form, testing a new combination of measurement and analysis methods - primarily based on eye movement registration -to determine the functional vision of people with VISPIMD. Our goal is to systematically determine what someone is responding to (“what”), where it may be (“where”), and how much time that response will take (“when”). When developing methods, we take the possibilities and preferences of the person in question as a starting point in relation to the technological possibilities.Because existing technological methods were originally developed for a different purpose, this partly requires adaptation to the possibilities of the target group.The concrete end product of our pilot will be a manual with an overview of available technological methods (as well as the methods themselves) for assessing functional vision, linked to the specific characteristics of the target group in the cognitive, motor area: 'Given that a client has this (estimated) combination of limitations (cognitive, motor and attention, time in whichsomeone can concentrate), the order of assessments is as follows:' followed by a description of the methods. We will also report on our findings in a workshop for professionals, a Dutch-language article and at least two scientific articles. This project is executed in the line: “I am seen; with all my strengths and limitations”. During the project, we closely collaborate with relevant stakeholders, i.e. the professionals with specific expertise working with the target group, family members of the persons with VISPIMD, and persons experiencing a visual impairment (‘experience experts’).
In the last decade, the automotive industry has seen significant advancements in technology (Advanced Driver Assistance Systems (ADAS) and autonomous vehicles) that presents the opportunity to improve traffic safety, efficiency, and comfort. However, the lack of drivers’ knowledge (such as risks, benefits, capabilities, limitations, and components) and confusion (i.e., multiple systems that have similar but not identical functions with different names) concerning the vehicle technology still prevails and thus, limiting the safety potential. The usual sources (such as the owner’s manual, instructions from a sales representative, online forums, and post-purchase training) do not provide adequate and sustainable knowledge to drivers concerning ADAS. Additionally, existing driving training and examinations focus mainly on unassisted driving and are practically unchanged for 30 years. Therefore, where and how drivers should obtain the necessary skills and knowledge for safely and effectively using ADAS? The proposed KIEM project AMIGO aims to create a training framework for learner drivers by combining classroom, online/virtual, and on-the-road training modules for imparting adequate knowledge and skills (such as risk assessment, handling in safety-critical and take-over transitions, and self-evaluation). AMIGO will also develop an assessment procedure to evaluate the impact of ADAS training on drivers’ skills and knowledge by defining key performance indicators (KPIs) using in-vehicle data, eye-tracking data, and subjective measures. For practical reasons, AMIGO will focus on either lane-keeping assistance (LKA) or adaptive cruise control (ACC) for framework development and testing, depending on the system availability. The insights obtained from this project will serve as a foundation for a subsequent research project, which will expand the AMIGO framework to other ADAS systems (e.g., mandatory ADAS systems in new cars from 2020 onwards) and specific driver target groups, such as the elderly and novice.
The maritime transport industry is facing a series of challenges due to the phasing out of fossil fuels and the challenges from decarbonization. The proposal of proper alternatives is not a straightforward process. While the current generation of ship design software offers results, there is a clear missed potential in new software technologies like machine learning and data science. This leads to the question: how can we use modern computational technologies like data analysis and machine learning to enhance the ship design process, considering the tools from the wider industry and the industry’s readiness to embrace new technologies and solutions? The obbjective of this PD project is to bridge the critical gap between the maritime industry's pressing need for innovative solutions for a more agile Ship Design Process; and the current limitations in software tools and methodologies available via the implementation into Ship Design specific software of the new generation of computational technologies available, as big data science and machine learning.
