PLOS digital health最新文献

{"title":"Boosting efficiency in a clinical literature surveillance system with LightGBM.","authors":"Cynthia Lokker, Wael Abdelkader, Elham Bagheri, Rick Parrish, Chris Cotoi, Tamara Navarro, Federico Germini, Lori-Ann Linkins, R Brian Haynes, Lingyang Chu, Muhammad Afzal, Alfonso Iorio","doi":"10.1371/journal.pdig.0000299","DOIUrl":"10.1371/journal.pdig.0000299","url":null,"abstract":"<p><p>Given the suboptimal performance of Boolean searching to identify methodologically sound and clinically relevant studies in large bibliographic databases, exploring machine learning (ML) to efficiently classify studies is warranted. To boost the efficiency of a literature surveillance program, we used a large internationally recognized dataset of articles tagged for methodological rigor and applied an automated ML approach to train and test binary classification models to predict the probability of clinical research articles being of high methodologic quality. We trained over 12,000 models on a dataset of titles and abstracts of 97,805 articles indexed in PubMed from 2012-2018 which were manually appraised for rigor by highly trained research associates and rated for clinical relevancy by practicing clinicians. As the dataset is unbalanced, with more articles that do not meet the criteria for rigor, we used the unbalanced dataset and over- and under-sampled datasets. Models that maintained sensitivity for high rigor at 99% and maximized specificity were selected and tested in a retrospective set of 30,424 articles from 2020 and validated prospectively in a blinded study of 5253 articles. The final selected algorithm, combining a LightGBM (gradient boosting machine) model trained in each dataset, maintained high sensitivity and achieved 57% specificity in the retrospective validation test and 53% in the prospective study. The number of articles needed to read to find one that met appraisal criteria was 3.68 (95% CI 3.52 to 3.85) in the prospective study, compared with 4.63 (95% CI 4.50 to 4.77) when relying only on Boolean searching. Gradient-boosting ML models reduced the work required to classify high quality clinical research studies by 45%, improving the efficiency of literature surveillance and subsequent dissemination to clinicians and other evidence users.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"3 9","pages":"e0000299"},"PeriodicalIF":0.0,"publicationDate":"2024-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11419392/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142309311","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Google's new AI Chatbot produces fake health-related evidence-then self-corrects.","authors":"Gary M Franklin","doi":"10.1371/journal.pdig.0000619","DOIUrl":"10.1371/journal.pdig.0000619","url":null,"abstract":"","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"3 9","pages":"e0000619"},"PeriodicalIF":0.0,"publicationDate":"2024-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11419366/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142309312","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Low-cost and convenient screening of disease using analysis of physical measurements and recordings.","authors":"Jay Chandra, Raymond Lin, Devin Kancherla, Sophia Scott, Daniel Sul, Daniela Andrade, Sammer Marzouk, Jay M Iyer, William Wasswa, Cleva Villanueva, Leo Anthony Celi","doi":"10.1371/journal.pdig.0000574","DOIUrl":"https://doi.org/10.1371/journal.pdig.0000574","url":null,"abstract":"<p><p>In recent years, there has been substantial work in low-cost medical diagnostics based on the physical manifestations of disease. This is due to advancements in data analysis techniques and classification algorithms and the increased availability of computing power through smart devices. Smartphones and their ability to interface with simple sensors such as inertial measurement units (IMUs), microphones, piezoelectric sensors, etc., or with convenient attachments such as lenses have revolutionized the ability collect medically relevant data easily. Even if the data has relatively low resolution or signal to noise ratio, newer algorithms have made it possible to identify disease with this data. Many low-cost diagnostic tools have been created in medical fields spanning from neurology to dermatology to obstetrics. These tools are particularly useful in low-resource areas where access to expensive diagnostic equipment may not be possible. The ultimate goal would be the creation of a \"diagnostic toolkit\" consisting of a smartphone and a set of sensors and attachments that can be used to screen for a wide set of diseases in a community healthcare setting. However, there are a few concerns that still need to be overcome in low-cost diagnostics: lack of incentives to bring these devices to market, algorithmic bias, \"black box\" nature of the algorithms, and data storage/transfer concerns.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"3 9","pages":"e0000574"},"PeriodicalIF":0.0,"publicationDate":"2024-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11412657/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142302859","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Ethical, legal, and social issues (ELSI) and reporting guidelines of AI research in healthcare.","authors":"Junko Kameyama, Satoshi Kodera, Yusuke Inoue","doi":"10.1371/journal.pdig.0000607","DOIUrl":"https://doi.org/10.1371/journal.pdig.0000607","url":null,"abstract":"","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"3 9","pages":"e0000607"},"PeriodicalIF":0.0,"publicationDate":"2024-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11412540/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142302856","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Performance of Publicly Available Large Language Models on Internal Medicine Board-style Questions.","authors":"Constantine Tarabanis, Sohail Zahid, Marios Mamalis, Kevin Zhang, Evangelos Kalampokis, Lior Jankelson","doi":"10.1371/journal.pdig.0000604","DOIUrl":"https://doi.org/10.1371/journal.pdig.0000604","url":null,"abstract":"<p><p>Ongoing research attempts to benchmark large language models (LLM) against physicians' fund of knowledge by assessing LLM performance on medical examinations. No prior study has assessed LLM performance on internal medicine (IM) board examination questions. Limited data exists on how knowledge supplied to the models, derived from medical texts improves LLM performance. The performance of GPT-3.5, GPT-4.0, LaMDA and Llama 2, with and without additional model input augmentation, was assessed on 240 randomly selected IM board-style questions. Questions were sourced from the Medical Knowledge Self-Assessment Program released by the American College of Physicians with each question serving as part of the LLM prompt. When available, LLMs were accessed both through their application programming interface (API) and their corresponding chatbot. Mode inputs were augmented with Harrison's Principles of Internal Medicine using the method of Retrieval Augmented Generation. LLM-generated explanations to 25 correctly answered questions were presented in a blinded fashion alongside the MKSAP explanation to an IM board-certified physician tasked with selecting the human generated response. GPT-4.0, accessed either through Bing Chat or its API, scored 77.5-80.7% outperforming GPT-3.5, human respondents, LaMDA and Llama 2 in that order. GPT-4.0 outperformed human MKSAP users on every tested IM subject with its highest and lowest percentile scores in Infectious Disease (80th) and Rheumatology (99.7th), respectively. There is a 3.2-5.3% decrease in performance of both GPT-3.5 and GPT-4.0 when accessing the LLM through its API instead of its online chatbot. There is 4.5-7.5% increase in performance of both GPT-3.5 and GPT-4.0 accessed through their APIs after additional input augmentation. The blinded reviewer correctly identified the human generated MKSAP response in 72% of the 25-question sample set. GPT-4.0 performed best on IM board-style questions outperforming human respondents. Augmenting with domain-specific information improved performance rendering Retrieval Augmented Generation a possible technique for improving accuracy in medical examination LLM responses.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"3 9","pages":"e0000604"},"PeriodicalIF":0.0,"publicationDate":"2024-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11407633/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142302860","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Six-year (2016-2022) longitudinal patterns of mental health service utilization rates among children developmentally vulnerable in kindergarten and the COVID-19 pandemic disruption.","authors":"Fernanda Talarico, Dan Metes, Mengzhe Wang, Jake Hayward, Yang S Liu, Julie Tian, Yanbo Zhang, Andrew J Greenshaw, Ashley Gaskin, Magdalena Janus, Bo Cao","doi":"10.1371/journal.pdig.0000611","DOIUrl":"https://doi.org/10.1371/journal.pdig.0000611","url":null,"abstract":"<p><strong>Introduction: </strong>In the context of the COVID-19 pandemic, it becomes important to comprehend service utilization patterns and evaluate disparities in mental health-related service access among children.</p><p><strong>Objective: </strong>This study uses administrative health records to investigate the association between early developmental vulnerability and healthcare utilization among children in Alberta, Canada from 2016 to 2022.</p><p><strong>Methods: </strong>Children who participated in the 2016 Early Development Instrument (EDI) assessment and were covered by public Alberta health insurance were included (N = 23 494). Linear regression models were employed to investigate the association between service utilization and vulnerability and biological sex. Separate models were used to assess vulnerability specific to each developmental domain and vulnerability across multiple domains. The service utilization was compared between pre- and post-pandemic onset periods.</p><p><strong>Results: </strong>The analysis reveals a significant decrease in all health services utilization from 2016 to 2019, followed by an increase until 2022. Vulnerable children had, on average, more events than non-vulnerable children. There was a consistent linear increase in mental health-related utilization from 2016 to 2022, with male children consistently experiencing higher utilization rates than females, particularly among vulnerable children. Specifically, there was a consistent linear increase in the utilization of anxiety-related services by children from 2016 to 2022, with females having, on average, 25 more events than males. The utilization of ADHD-related services showed different patterns for each group, with vulnerable male children having more utilization than their peers.</p><p><strong>Conclusion: </strong>Utilizing population-wide data, our study reveals sex specific developmental vulnerabilities and its impact on children's mental health service utilization during the COVID-19 pandemic, contributing to the existing literature. With data from kindergarten, we emphasize the need for early and targeted intervention strategies, especially for at-risk children, offering a path to reduce the burden of childhood mental health disorders.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"3 9","pages":"e0000611"},"PeriodicalIF":0.0,"publicationDate":"2024-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11407640/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142302862","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The role of Open Access Data in democratizing healthcare AI: A pathway to research enhancement, patient well-being and treatment equity in Andalusia, Spain.","authors":"Álvaro Ritoré, Claudia M Jiménez, Juan Luis González, Juan Carlos Rejón-Parrilla, Pablo Hervás, Esteban Toro, Carlos Luis Parra-Calderón, Leo Anthony Celi, Isaac Túnez, Miguel Ángel Armengol de la Hoz","doi":"10.1371/journal.pdig.0000599","DOIUrl":"https://doi.org/10.1371/journal.pdig.0000599","url":null,"abstract":"","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"3 9","pages":"e0000599"},"PeriodicalIF":0.0,"publicationDate":"2024-09-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11404816/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142302863","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Impact of electronic medical records on healthcare delivery in Nigeria: A review.","authors":"Sarah Oreoluwa Olukorode, Oluwakorede Joshua Adedeji, Adetayo Adetokun, Ajibola Ibraheem Abioye","doi":"10.1371/journal.pdig.0000420","DOIUrl":"https://doi.org/10.1371/journal.pdig.0000420","url":null,"abstract":"<p><p>Electronic medical records (EMRs) have great potential to improve healthcare processes and outcomes. They are increasingly available in Nigeria, as in many developing countries. The impact of their introduction has not been well studied. We sought to synthesize the evidence from primary studies of the effect of EMRs on data quality, patient-relevant outcomes and patient satisfaction. We identified and examined five original research articles published up to May 2023 in the following medical literature databases: PUBMED/Medline, EMBASE, Web of Science, African Journals Online and Google Scholar. Four studies examined the influence of the introduction of or improvements in the EMR on data collection and documentation. The pooled percentage difference in data quality after introducing or improving the EMR was 142% (95% CI: 82% to 203%, p-value < 0.001). There was limited heterogeneity in the estimates (I2 = 0%, p-heterogeneity = 0.93) and no evidence suggestive of publication bias. The 5th study assessed patient satisfaction with pharmacy services following the introduction of the EMR but neither had a comparison group nor assessed patient satisfaction before EMR was introduced. We conclude that the introduction of EMR in Nigerian healthcare facilities meaningfully increased the quality of the data.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"3 9","pages":"e0000420"},"PeriodicalIF":0.0,"publicationDate":"2024-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11398640/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142302858","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Contextualised digital health communication infrastructure standards for resource-constrained settings: Perception of digital health stakeholders regarding suitability for Uganda's health system.","authors":"Andrew Egwar Alunyu, Mercy Rebekah Amiyo, Josephine Nabukenya","doi":"10.1371/journal.pdig.0000603","DOIUrl":"https://doi.org/10.1371/journal.pdig.0000603","url":null,"abstract":"<p><p>Ignoring the need to contextualise international standards has caused low-resourced countries to implement digital health systems on the ad-hoc, thereby often failing to meet the local needs or scale up. Authors have recommended adapting standards to a country's context. However, to date, most resources constrained countries like Uganda have not done so, affecting their success in attaining the full benefits of using ICT to support their health systems. They apply the standards 'as is' with little regard for their fitness for potential use and ability to fulfil the country's digital health needs. A design science approach was followed to elicit digital health communication infrastructure (DHCI) requirements and develop the contextual DHCI standards for Uganda. The design science methodology's design cycle supported DHCI standards' construction and evaluation activities. Whereas two workgroup sessions were held to craft the standards, three cycles of evaluation and refinement were performed. The final refinement produces the contextualised DHCI standards approved by Uganda's DH stakeholders through summative evaluation. Results of the summative evaluation show that DH stakeholders agree that the statement of the standards and the requirements specification are suitable to guide DHCI standards implementation in Uganda. Stakeholders agreed that the standards are complete, have the potential to realise DHCI requirements in Uganda, that have been well structured and follow international style for standards, and finally, that the standards are fit to realise their intended use in Uganda. Having been endorsed by DH stakeholders in Uganda's health system, the standards should be piloted to establish their potency to improve health information exchange and healthcare outcomes. Also, we recommend other low middle income countries (LMICs) with similar challenges to those in Uganda adopt the same set of contextualised DHCI standards.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"3 9","pages":"e0000603"},"PeriodicalIF":0.0,"publicationDate":"2024-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11392385/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142302854","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Artificial intelligence and social accountability in the Canadian health care landscape: A rapid literature review.","authors":"Alex Anawati, Holly Fleming, Megan Mertz, Jillian Bertrand, Jennifer Dumond, Sophia Myles, Joseph Leblanc, Brian Ross, Daniel Lamoureux, Div Patel, Renald Carrier, Erin Cameron","doi":"10.1371/journal.pdig.0000597","DOIUrl":"https://doi.org/10.1371/journal.pdig.0000597","url":null,"abstract":"<p><strong>Background: </strong>Situated within a larger project entitled \"Exploring the Need for a Uniquely Different Approach in Northern Ontario: A Study of Socially Accountable Artificial Intelligence,\" this rapid review provides a broad look into how social accountability as an equity-oriented health policy strategy is guiding artificial intelligence (AI) across the Canadian health care landscape, particularly for marginalized regions and populations. This review synthesizes existing literature to answer the question: How is AI present and impacted by social accountability across the health care landscape in Canada?</p><p><strong>Methodology: </strong>A multidisciplinary expert panel with experience in diverse health care roles and computer sciences was assembled from multiple institutions in Northern Ontario to guide the study design and research team. A search strategy was developed that broadly reflected the concepts of social accountability, AI and health care in Canada. EMBASE and Medline databases were searched for articles, which were reviewed for inclusion by 2 independent reviewers. Search results, a description of the studies, and a thematic analysis of the included studies were reported as the primary outcome.</p><p><strong>Principal findings: </strong>The search strategy yielded 679 articles of which 36 relevant studies were included. There were no studies identified that were guided by a comprehensive, equity-oriented social accountability strategy. Three major themes emerged from the thematic analysis: (1) designing equity into AI; (2) policies and regulations for AI; and (3) the inclusion of community voices in the implementation of AI in health care. Across the 3 main themes, equity, marginalized populations, and the need for community and partner engagement were frequently referenced, which are key concepts of a social accountability strategy.</p><p><strong>Conclusion: </strong>The findings suggest that unless there is a course correction, AI in the Canadian health care landscape will worsen the digital divide and health inequity. Social accountability as an equity-oriented strategy for AI could catalyze many of the changes required to prevent a worsening of the digital divide caused by the AI revolution in health care in Canada and should raise concerns for other global contexts.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"3 9","pages":"e0000597"},"PeriodicalIF":0.0,"publicationDate":"2024-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11392241/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142302844","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}