-
Scientific Reports Nov 2023Crop disease detection and management is critical to improving productivity, reducing costs, and promoting environmentally friendly crop treatment methods. Modern...
Crop disease detection and management is critical to improving productivity, reducing costs, and promoting environmentally friendly crop treatment methods. Modern technologies, such as data mining and machine learning algorithms, have been used to develop automated crop disease detection systems. However, centralized approach to data collection and model training induces challenges in terms of data privacy, availability, and transfer costs. To address these challenges, federated learning appears to be a promising solution. In this paper, we explored the application of federated learning for crop disease classification using image analysis. We developed and studied convolutional neural network (CNN) models and those based on attention mechanisms, in this case vision transformers (ViT), using federated learning, leveraging an open access image dataset from the "PlantVillage" platform. Experiments conducted concluded that the performance of models trained by federated learning is influenced by the number of learners involved, the number of communication rounds, the number of local iterations and the quality of the data. With the objective of highlighting the potential of federated learning in crop disease classification, among the CNN models tested, ResNet50 performed better in several experiments than the other models, and proved to be an optimal choice, but also the most suitable for a federated learning scenario. The ViT_B16 and ViT_B32 Vision Transformers require more computational time, making them less suitable in a federated learning scenario, where computational time and communication costs are key parameters. The paper provides a state-of-the-art analysis, presents our methodology and experimental results, and concludes with ideas and future directions for our research on using federated learning in the context of crop disease classification.
Topics: Algorithms; Communication; Data Collection; Data Mining; Electric Power Supplies
PubMed: 37932344
DOI: 10.1038/s41598-023-46218-5 -
Science Translational Medicine Nov 2023Clinical trials for central nervous system disorders often enroll patients with unrecognized heterogeneous diseases, leading to costly trials that have high failure... (Review)
Review
Clinical trials for central nervous system disorders often enroll patients with unrecognized heterogeneous diseases, leading to costly trials that have high failure rates. Here, we discuss the potential of emerging technologies and datasets to elucidate disease mechanisms and identify biomarkers to improve patient stratification and monitoring of disease progression in clinical trials for neuropsychiatric disorders. Greater efforts must be centered on rigorously standardizing data collection and sharing of methods, datasets, and analytical tools across sectors. To address health care disparities in clinical trials, diversity of genetic ancestries and environmental exposures of research participants and associated biological samples must be prioritized.
Topics: Humans; Mental Disorders; Data Collection; Disease Progression; Environmental Exposure
PubMed: 38190501
DOI: 10.1126/scitranslmed.adg4775 -
Canadian Journal of Public Health =... Aug 2023Tuberculosis incidence in Canada has remained essentially unchanged over the past decade. A strategic plan to reduce the burden of disease, underpinned by high-quality...
Tuberculosis incidence in Canada has remained essentially unchanged over the past decade. A strategic plan to reduce the burden of disease, underpinned by high-quality surveillance data, is sorely needed. However, tuberculosis surveillance data are lacking in Canada for multiple reasons. There is no single entity responsible for coordinating a tuberculosis response, including strategies for surveillance, thus inhibiting effective solutions. This in turn affects the timeliness and comprehensiveness of national tuberculosis surveillance reporting: between 2000 and 2020, there was an average 25-month delay to publication of annual surveillance data and the comprehensiveness of reports has precipitously fallen over time. Compounding these issues are case report forms for tuberculosis surveillance data which have not been updated since 2011, failing to keep up with the changing tuberculosis epidemiology and to provide information required for strategic planning. Common-sense steps can be taken to vastly improve the utility of collected tuberculosis surveillance data, and the development of a strategic plan for tuberculosis elimination. These include initiating a country-wide consultation on surveillance needs; allocating resources for data collection and analysis and data sharing; setting precise, measurable goals; and, importantly, establishing an oversight committee with representation from all provincial/territorial tuberculosis program leads who are held to account for performance.
Topics: Humans; Tuberculosis; Data Collection; Data Accuracy; Information Dissemination; Canada
PubMed: 37014575
DOI: 10.17269/s41997-023-00767-4 -
The American Journal of Cardiology Sep 2023The electronic health record (EHR) represents a rich source of patient information, increasingly being leveraged for cardiovascular research. Although its primary use... (Review)
Review
The electronic health record (EHR) represents a rich source of patient information, increasingly being leveraged for cardiovascular research. Although its primary use remains the seamless delivery of health care, the various longitudinally aggregated structured and unstructured data elements for each patient within the EHR can define the computational phenotypes of disease and care signatures and their association with outcomes. Although structured data elements, such as demographic characteristics, laboratory measurements, problem lists, and medications, are easily extracted, unstructured data are underused. The latter include free text in clinical narratives, documentation of procedures, and reports of imaging and pathology. Rapid scaling up of data storage and rapid innovation in natural language processing and computer vision can power insights from unstructured data streams. However, despite an array of opportunities for research using the EHR, specific expertise is necessary to adequately address confidentiality, accuracy, completeness, and heterogeneity challenges in EHR-based research. These often require methodological innovation and best practices to design and conduct successful research studies. Our review discusses these challenges and their proposed solutions. In addition, we highlight the ongoing innovations in federated learning in the EHR through a greater focus on common data models and discuss ongoing work that defines such an approach to large-scale, multicenter, federated studies. Such parallel improvements in technology and research methods enable innovative care and optimization of patient outcomes.
Topics: Electronic Health Records; Information Storage and Retrieval; Delivery of Health Care; Data Collection; Natural Language Processing; Multicenter Studies as Topic
PubMed: 37499593
DOI: 10.1016/j.amjcard.2023.06.104 -
Health Information Management : Journal... Sep 2023With increasing implementation of enhanced recovery programs (ERPs) in clinical practice, standardised data collection and reporting have become critical in addressing... (Review)
Review
With increasing implementation of enhanced recovery programs (ERPs) in clinical practice, standardised data collection and reporting have become critical in addressing the heterogeneity of metrics used for reporting outcomes. Opportunities exist to leverage electronic health record (EHR) systems to collect, analyse, and disseminate ERP data. (i) To consolidate relevant ERP variables into a singular data universe; (ii) To create an accessible and intuitive query tool for rapid data retrieval. We reviewed nine established individual team databases to identify common variables to create one standard ERP data dictionary. To address data automation, we used a third-party business intelligence tool to map identified variables within the EHR system, consolidating variables into a single ERP universe. To determine efficacy, we compared times for four experienced research coordinators to use manual, five-universe, and ERP Universe processes to retrieve ERP data for 10 randomly selected surgery patients. The total times to process data variables for all 10 patients for the manual, five universe, and ERP Universe processes were 510, 111, and 76 min, respectively. Shifting from the five-universe or manual process to the ERP Universe resulted in decreases in time of 32% and 85%, respectively. The ERP Universe improves time spent collecting, analysing, and reporting ERP elements without increasing operational costs or interrupting workflow. Manual data abstraction places significant burden on resources. The creation of a singular instrument dedicated to ERP data abstraction greatly increases the efficiency in which clinicians and supporting staff can query adherence to an ERP protocol.
Topics: Humans; Data Collection; Costs and Cost Analysis
PubMed: 35695132
DOI: 10.1177/18333583221095139 -
Journal of Public Health Policy Dec 2023Patient surgical registries are essential tools for public health specialists, creating research opportunities through linkage of registry data with healthcare outcomes.... (Review)
Review
Patient surgical registries are essential tools for public health specialists, creating research opportunities through linkage of registry data with healthcare outcomes. However, little is known regarding data error sources in the management of surgical registries. In June 2022, we undertook a scoping study of the empirical literature including publications selected from the PUBMED and EMBASE databases. We selected 48 studies focussing on shared experiences centred around developing surgical patient registries. We identified seven types of data specific challenges, grouped in three categories- data capture, data analysis and result dissemination. Most studies underlined the risk for a high volume of missing data, non-uniform geographic representation, inclusion biases, inappropriate coding, as well as variations in analysis reporting and limitations related to the statistical analysis. Finally, to expand data usability, we discussed cost-effective ways of addressing these limitations, by citing aspects from the protocols followed by established exemplary registries.
Topics: Humans; Registries; Surgical Procedures, Operative; Patients
PubMed: 37726394
DOI: 10.1057/s41271-023-00442-5 -
Topics in Spinal Cord Injury... 2023Assessment of aerobic exercise (AE) and lipid profiles among individuals with spinal cord injury or disease (SCI/D) is critical for cardiometabolic disease (CMD) risk...
BACKGROUND
Assessment of aerobic exercise (AE) and lipid profiles among individuals with spinal cord injury or disease (SCI/D) is critical for cardiometabolic disease (CMD) risk estimation.
OBJECTIVES
To utilize an artificial intelligence (AI) tool for extracting indicator data and education tools to enable routine CMD indicator data collection in inpatient/outpatient settings, and to describe and evaluate the recall of AE levels and lipid profile assessment completion rates across care settings among adults with subacute and chronic SCI/D.
METHODS
A cross-sectional convenience sample of patients affiliated with University Health Network's SCI/D rehabilitation program and outpatients affiliated with SCI Ontario participated. The SCI-HIGH CMD intermediary outcome (IO) and final outcome (FO) indicator surveys were administered, using an AI tool to extract responses. Practice gaps were prospectively identified, and implementation tools were created to address gaps. Univariate and bivariate descriptive analyses were used.
RESULTS
The AI tool had <2% error rate for data extraction. Adults with SCI/D ( = 251; 124 IO, mean age 61; 127 FO, mean age 55; = .004) completed the surveys. Fourteen percent of inpatients versus 48% of outpatients reported being taught AE. Fifteen percent of inpatients and 51% of outpatients recalled a lipid assessment ( < .01). Algorithms and education tools were developed to address identified knowledge gaps in patient AE and lipid assessments.
CONCLUSION
Compelling CMD health service gaps warrant immediate attention to achieve AE and lipid assessment guideline adherence. AI indicator extraction paired with implementation tools may facilitate indicator deployment and modify CMD risk.
Topics: Adult; Humans; Middle Aged; Spinal Cord Injuries; Cross-Sectional Studies; Artificial Intelligence; Quality Indicators, Health Care; Data Collection; Cardiovascular Diseases; Lipids
PubMed: 38174138
DOI: 10.46292/sci23-00018S -
F1000Research 2023Research suggests that gamification can increase work engagement by providing employees with a sense of autonomy, competence, and relatedness, and by creating a fun and...
Research suggests that gamification can increase work engagement by providing employees with a sense of autonomy, competence, and relatedness, and by creating a fun and engaging work environment. Gamification is designed to increase consumer and employee engagement and see that they holistically collaborate to achieve a shared vision. The concept of gamification is as old as learning itself, just that the use of the terminology "Gamification" is of a recent origin. This article focuses on the impact of gamification in various organizations and simultaneously sees its relationship with job engagement and productivity. A primary investigation was done to determine the nexus between the various variables and data collection from 400 respondents working in various fraternities of the economy from both public and private domains from countries in the Gulf region. The structural equation model and SPSS has been inferred to analyse the results. The study results show that variable such as perceived adoption and usefulness in the gamified system is significantly associated with job engagement. Similarly, employee's recognition and perceived motivation have a positive impact on productivity. The study identified job engagement mediating factor to enhance organisational productivity in a gamified system. The effectiveness of gamification in enhancing work engagement may depend on factors such as the design of the gamification system, the preferences and motivations of individual employees, and the organizational culture and goals. The findings have significant implications for insight into how employees in the service sector are aware of the gamified working environment and react to the system through work engagement and productivity.
Topics: Humans; Gamification; Efficiency, Organizational; Data Collection; Learning; Motivation
PubMed: 38434668
DOI: 10.12688/f1000research.131579.2 -
Expert Review of Pharmacoeconomics &... Jul 2024
Topics: Humans; Biomarkers; Data Collection; Time Factors; Endpoint Determination
PubMed: 38362754
DOI: 10.1080/14737167.2024.2320233 -
Journal of Pharmacy & Pharmaceutical... 2023Patient support programs (PSPs) offer a unique opportunity to collect real-world data that can contribute to improving patient care and informing healthcare decision...
Patient support programs (PSPs) offer a unique opportunity to collect real-world data that can contribute to improving patient care and informing healthcare decision making. In this perspective article, we explore the collection of data through PSPs in Canada, current advances in data collection methods, and the potential for generating acceptable real-world evidence (RWE). With PSP infrastructure already in place for most specialized drugs in Canada, adding and strengthening data collection capacities has been a focus in recent years. However, limitations in PSP data, including challenges related to quality, bias, and trust, need to be acknowledged and addressed. Forward-thinking PSP developers have been taking steps to strengthen the PSP datasphere, such as engaging third parties for data analysis, publishing peer-reviewed studies that utilize PSPs as a data source and incorporating quality controls into data collection processes. This article illustrates the current state of PSP data collection by examining six PSP RWE studies and outlining their data characteristics and the health outcomes collected from the PSP. A framework for collecting real-world data within a PSP and a checklist to address issues of trust and bias in PSP data collection is also provided. Collaboration between drug manufacturers, PSP vendors, and data specialists will be crucial in elevating PSP data to a level acceptable to healthcare decision makers, including health technology assessors and payers, with the ultimate beneficiary being patients.
Topics: Humans; Data Collection; Delivery of Health Care; Canada
PubMed: 37901362
DOI: 10.3389/jpps.2023.11877