PLOS digital health最新文献

{"title":"ChatGPT and large language models (LLMs) awareness and use. A prospective cross-sectional survey of U.S. medical students.","authors":"Conner Ganjavi, Michael Eppler, Devon O'Brien, Lorenzo Storino Ramacciotti, Muhammad Shabbeer Ghauri, Issac Anderson, Jae Choi, Darby Dwyer, Claudia Stephens, Victoria Shi, Madeline Ebert, Michaela Derby, Bayan Yazdi, Giovanni E Cacciamani","doi":"10.1371/journal.pdig.0000596","DOIUrl":"10.1371/journal.pdig.0000596","url":null,"abstract":"<p><p>Generative-AI (GAI) models like ChatGPT are becoming widely discussed and utilized tools in medical education. For example, it can be used to assist with studying for exams, shown capable of passing the USMLE board exams. However, there have been concerns expressed regarding its fair and ethical use. We designed an electronic survey for students across North American medical colleges to gauge their views on and current use of ChatGPT and similar technologies in May, 2023. Overall, 415 students from at least 28 medical schools completed the questionnaire and 96% of respondents had heard of ChatGPT and 52% had used it for medical school coursework. The most common use in pre-clerkship and clerkship phase was asking for explanations of medical concepts and assisting with diagnosis/treatment plans, respectively. The most common use in academic research was for proof reading and grammar edits. Respondents recognized the potential limitations of ChatGPT, including inaccurate responses, patient privacy, and plagiarism. Students recognized the importance of regulations to ensure proper use of this novel technology. Understanding the views of students is essential to crafting workable instructional courses, guidelines, and regulations that ensure the safe, productive use of generative-AI in medical school.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"3 9","pages":"e0000596"},"PeriodicalIF":0.0,"publicationDate":"2024-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11376538/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142141930","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":"Systematic review of adherence to technology-based falls prevention programs for community-dwelling older adults: Reimagining future interventions.","authors":"Maureen C Ashe, Isis Kelly Dos Santos, Jefferson Erome, Jared Grant, Juliana Mollins, Sze-Ee Soh","doi":"10.1371/journal.pdig.0000579","DOIUrl":"10.1371/journal.pdig.0000579","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;Prevention programs, and specifically exercise, can reduce falls among community-dwelling older adults, but low adherence limits the benefits of effective interventions. Technology may overcome some barriers to improve uptake and engagement in prevention programs, although less is known on adherence for providing them via this delivery mode. We aimed to synthesize evidence for adherence to technology-based falls prevention programs in community-dwelling older adults 60 years and older. We conducted a systematic review following standard guidelines to identify randomized controlled trials for remote delivered (i.e., no or limited in-person sessions) technology-based falls prevention programs for community-dwelling older adults. We searched nine sources using Medical Subject Headings (MeSH) terms and keywords (2007-present). The initial search was conducted in June 2023 and updated in December 2023. We also conducted a forward and backward citation search of included studies. Two reviewers independently conducted screening and study assessment; one author extracted data and a second author confirmed findings. We conducted a random effects meta-analysis for adherence, operationalized as participants' completion of program components, and aimed to conduct meta-regressions to examine factors related to program adherence and the association between adherence and functional mobility. We included 11 studies with 569 intervention participants (average mean age 74.5 years). Studies used a variety of technology, such as apps, exergames, or virtual synchronous classes. Risk of bias was low for eight studies. Five interventions automatically collected data for monitoring and completion of exercise sessions, two studies collected participants' online attendance, and four studies used self-reported diaries or attendance sheets. Studies included some behavior change techniques or strategies alongside the technology. There was substantial variability in the way adherence data were reported. The mean (range) percent of participants who did not complete planned sessions (i.e., dropped out or lost to follow-up) was 14% (0-32%). The pooled estimate of the proportion of participants who were adherent to a technology-based falls prevention program was 0.82 (95% CI 0.68, 0.93) for studies that reported the mean number of completed exercise sessions. Many studies needed to provide access to the internet, training, and/or resources (e.g., tablets) to support participants to take part in the intervention. We were unable to conduct the meta-regression for adherence and functional mobility due to an insufficient number of studies. There were no serious adverse events for studies reporting this information (n = 8). The use of technology may confer some benefits for program delivery and data collection. But better reporting of adherence data is needed, as well as routine integration and measurement of training and skill development to use techno","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"3 9","pages":"e0000579"},"PeriodicalIF":0.0,"publicationDate":"2024-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11371225/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142127499","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":"Towards artificial intelligence-based disease prediction algorithms that comprehensively leverage and continuously learn from real-world clinical tabular data systems.","authors":"Terrence J Lee-St John, Oshin Kanwar, Emna Abidi, Wasim El Nekidy, Bartlomiej Piechowski-Jozwiak","doi":"10.1371/journal.pdig.0000589","DOIUrl":"10.1371/journal.pdig.0000589","url":null,"abstract":"<p><p>This manuscript presents a proof-of-concept for a generalizable strategy, the full algorithm, designed to estimate disease risk using real-world clinical tabular data systems, such as electronic health records (EHR) or claims databases. By integrating classic statistical methods and modern artificial intelligence techniques, this strategy automates the production of a disease prediction model that comprehensively reflects the dynamics contained within the underlying data system. Specifically, the full algorithm parses through every facet of the data (e.g., encounters, diagnoses, procedures, medications, labs, chief complaints, flowsheets, vital signs, demographics, etc.), selects which factors to retain as predictor variables by evaluating the data empirically against statistical criteria, structures and formats the retained data into time-series, trains a neural network-based prediction model, then subsequently applies this model to current patients to generate risk estimates. A distinguishing feature of the proposed strategy is that it produces a self-adaptive prediction system, capable of evolving the prediction mechanism in response to changes within the data: as newly collected data expand/modify the dataset organically, the prediction mechanism automatically evolves to reflect these changes. Moreover, the full algorithm operates without the need for a-priori data curation and aims to harness all informative risk and protective factors within the real-world data. This stands in contrast to traditional approaches, which often rely on highly curated datasets and domain expertise to build static prediction models based solely on well-known risk factors. As a proof-of-concept, we codified the full algorithm and tasked it with estimating 12-month risk of initial stroke or myocardial infarction using our hospital's real-world EHR. A 66-month pseudo-prospective validation was conducted using records from 558,105 patients spanning April 2015 to September 2023, totalling 3,424,060 patient-months. Area under the receiver operating characteristic curve (AUROC) values ranged from .830 to .909, with an improving trend over time. Odds ratios describing model precision for patients 1-100 and 101-200 (when ranked by estimated risk) ranged from 15.3 to 48.1 and 7.2 to 45.0, respectively, with both groups showing improving trends over time. Findings suggest the feasibility of developing high-performing disease risk calculators in the proposed manner.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"3 9","pages":"e0000589"},"PeriodicalIF":0.0,"publicationDate":"2024-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11371204/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142127560","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":"Comparison of professional and everyday wearable technology at different body positions in terms of recording gait perturbations.","authors":"Lea Feld, Lena Schell-Majoor, Sandra Hellmers, Jessica Koschate, Andreas Hein, Tania Zieschang, Birger Kollmeier","doi":"10.1371/journal.pdig.0000553","DOIUrl":"https://doi.org/10.1371/journal.pdig.0000553","url":null,"abstract":"<p><p>Falls are a significant health problem in older people, so preventing them is essential. Since falls are often a consequence of improper reaction to gait disturbances, such as slips and trips, their detection is gaining attention in research. However there are no studies to date that investigated perturbation detection, using everyday wearable devices like hearing aids or smartphones at different body positions. Sixty-six study participants were perturbed on a split-belt treadmill while recording data with hearing aids, smartphones, and professional inertial measurement units (IMUs) at various positions (left/right ear, jacket pocket, shoulder bag, pants pocket, left/right foot, left/right wrist, lumbar, sternum). The data were visually inspected and median maximum cross-correlations were calculated for whole trials and different perturbation conditions. The results show that the hearing aids and IMUs perform equally in measuring acceleration data (correlation coefficient of 0.93 for the left hearing aid and 0.99 for the right hearing aid), which emphasizes the potential of utilizing sensors in hearing aids for head acceleration measurements. Additionally, the data implicate that measurement with a single hearing aid is sufficient and a second hearing aid provides no added value. Furthermore, the acceleration patterns were similar for the ear position, the jacket pocket position, and the lumbar (correlation coefficient of about 0.8) or sternal position (correlation coefficient of about 0.9). The correlations were found to be more or less independent of the type of perturbation. Data obtained from everyday wearable devices appears to represent the movements of the human body during perturbations similar to that of professional devices. The results suggest that IMUs in hearing aids and smartphones, placed at the trunk, could be well suited for an automatic detection of gait perturbations.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"3 8","pages":"e0000553"},"PeriodicalIF":0.0,"publicationDate":"2024-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11364241/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142115733","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":"Digital assessment of cognitive-affective biases related to mental health.","authors":"Sang-Eon Park, Jisu Chung, Jeonghyun Lee, Minwoo Jb Kim, Jinhee Kim, Hong Jin Jeon, Hyungsook Kim, Choongwan Woo, Hackjin Kim, Sang Ah Lee","doi":"10.1371/journal.pdig.0000595","DOIUrl":"https://doi.org/10.1371/journal.pdig.0000595","url":null,"abstract":"<p><p>With an increasing societal need for digital therapy solutions for poor mental health, we face a corresponding rise in demand for scientifically validated digital contents. In this study we aimed to lay a sound scientific foundation for the development of brain-based digital therapeutics to assess and monitor cognitive effects of social and emotional bias across diverse populations and age-ranges. First, we developed three computerized cognitive tasks using animated graphics: 1) an emotional flanker task designed to test attentional bias, 2) an emotional go-no-go task to measure bias in memory and executive function, and 3) an emotional social evaluation task to measure sensitivity to social judgments. Then, we confirmed the generalizability of our results in a wide range of samples (children (N = 50), young adults (N = 172), older adults (N = 39), online young adults (N=93), and depression patients (N = 41)) using touchscreen and online computer-based tasks, and devised a spontaneous thought generation task that was strongly associated with, and therefore could potentially serve as an alternative to, self-report scales. Using PCA, we extracted five components that represented different aspects of cognitive-affective function (emotional bias, emotional sensitivity, general accuracy, and general/social attention). Next, a gamified version of the above tasks was developed to test the feasibility of digital cognitive training over a 2-week period. A pilot training study utilizing this application showed decreases in emotional bias in the training group (that were not observed in the control group), which was correlated with a reduction in anxiety symptoms. Using a 2-channel wearable EEG system, we found that frontal alpha and gamma power were associated with both emotional bias and its reduction across the 2-week training period.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"3 8","pages":"e0000595"},"PeriodicalIF":0.0,"publicationDate":"2024-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11361731/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142115734","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":"A health equity monitoring framework based on process mining.","authors":"Jan Niklas Adams, Jennifer Ziegler, Matthew McDermott, Molly J Douglas, René Eber, Judy Wawira Gichoya, Deirdre Goode, Swami Sankaranarayanan, Ziyue Chen, Wil M P van der Aalst, Leo Anthony Celi","doi":"10.1371/journal.pdig.0000575","DOIUrl":"10.1371/journal.pdig.0000575","url":null,"abstract":"<p><p>In the United States, there is a proposal to link hospital Medicare payments with health equity measures, signaling a need to precisely measure equity in healthcare delivery. Despite significant research demonstrating disparities in health care outcomes and access, there is a noticeable gap in tools available to assess health equity across various health conditions and treatments. The available tools often focus on a single area of patient care, such as medication delivery, but fail to examine the entire health care process. The objective of this study is to propose a process mining framework to provide a comprehensive view of health equity. Using event logs which track all actions during patient care, this method allows us to look at disparities in single and multiple treatment steps, but also in the broader strategy of treatment delivery. We have applied this framework to the management of patients with sepsis in the Intensive Care Unit (ICU), focusing on sex and English language proficiency. We found no significant differences between treatments of male and female patients. However, for patients who don't speak English, there was a notable delay in starting their treatment, even though their illness was just as severe and subsequent treatments were similar. This framework subsumes existing individual approaches to measure health inequities and offers a comprehensive approach to pinpoint and delve into healthcare disparities, providing a valuable tool for research and policy-making aiming at more equitable healthcare.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"3 8","pages":"e0000575"},"PeriodicalIF":0.0,"publicationDate":"2024-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11355534/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142086445","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 application of Borg scale in cardiopulmonary resuscitation: An integrative review.","authors":"Lucia Tobase, Sandra Helena Cardoso, Renata Tavares Franco Rodrigues, Dhieizom Rodrigo de Souza, Debora Gugelmin-Almeida, Thatiane Facholi Polastri, Heloisa Helena Ciqueto Peres, Sergio Timerman","doi":"10.1371/journal.pdig.0000592","DOIUrl":"10.1371/journal.pdig.0000592","url":null,"abstract":"<p><p>The study of human performance and perception of exertion constitutes a fundamental aspect for monitoring health implications and enhancing training outcomes such as cardiopulmonary resuscitation (CPR). It involves gaining insights into the varied responses and tolerance levels exhibited by individuals engaging in physical activities. To measure perception of exertion, many tools are available, including the Borg scale. In order to evaluate how the Borg scale is being used during CPR attempts, this integrative review was carried out between October/2020 and December/2023, with searches from PubMed, CINAHL, Web of Science, Embase, PsycINFO and VHL. Full publications relevant to the PICO strategy were included and letters, editorials, abstracts, and unpublished studies were excluded. In total, 34 articles were selected and categorised into three themes: a) CPR performed in different contexts; b) CPR performed in different cycles, positions, and techniques; c) CPR performed with additional technological resources. Because CPR performance is considered a strenuous physical activity, the Borg scale was used in each study to evaluate perception of exertion. The results identified that the Borg scale has been used during CPR in different contexts. It is a quick, low-cost, and easy-to-apply tool that provides important indicators that may affect CPR quality, such as perception of exertion, likely improving performance and potentially increasing the chances of survival.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"3 8","pages":"e0000592"},"PeriodicalIF":0.0,"publicationDate":"2024-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11355535/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142086446","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":"Self-supervised contrastive learning improves machine learning discrimination of full thickness macular holes from epiretinal membranes in retinal OCT scans.","authors":"Timothy William Wheeler, Kaitlyn Hunter, Patricia Anne Garcia, Henry Li, Andrew Clark Thomson, Allan Hunter, Courosh Mehanian","doi":"10.1371/journal.pdig.0000411","DOIUrl":"10.1371/journal.pdig.0000411","url":null,"abstract":"<p><p>There is a growing interest in using computer-assisted models for the detection of macular conditions using optical coherence tomography (OCT) data. As the quantity of clinical scan data of specific conditions is limited, these models are typically developed by fine-tuning a generalized network to classify specific macular conditions of interest. Full thickness macular holes (FTMH) present a condition requiring urgent surgical repair to prevent vision loss. Other works on automated FTMH classification have tended to use supervised ImageNet pre-trained networks with good results but leave room for improvement. In this paper, we develop a model for FTMH classification using OCT B-scans around the central foveal region to pre-train a naïve network using contrastive self-supervised learning. We found that self-supervised pre-trained networks outperform ImageNet pre-trained networks despite a small training set size (284 eyes total, 51 FTMH+ eyes, 3 B-scans from each eye). On three replicate data splits, 3D spatial contrast pre-training yields a model with an average F1-score of 1.0 on holdout data (50 eyes total, 10 FTMH+), compared to an average F1-score of 0.831 for FTMH detection by ImageNet pre-trained models. These results demonstrate that even limited data may be applied toward self-supervised pre-training to substantially improve performance for FTMH classification, indicating applicability toward other OCT-based problems.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"3 8","pages":"e0000411"},"PeriodicalIF":0.0,"publicationDate":"2024-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11346922/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142074669","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":"Balance confidence classification in people with a lower limb amputation using six minute walk test smartphone sensor signals.","authors":"Pascale Juneau, Natalie Baddour, Helena Burger, Edward D Lemaire","doi":"10.1371/journal.pdig.0000570","DOIUrl":"10.1371/journal.pdig.0000570","url":null,"abstract":"<p><p>The activities-specific balance confidence scale (ABC) assesses balance confidence during common activities. While low balance confidence can result in activity avoidance, excess confidence can increase fall risk. People with lower limb amputations can present with inconsistent gait, adversely affecting their balance confidence. Previous research demonstrated that clinical outcomes in this population (e.g., stride parameters, fall risk) can be determined from smartphone signals collected during walk tests, but this has not been evaluated for balance confidence. Fifty-eight (58) individuals with lower limb amputation completed a six-minute walk test (6MWT) while a smartphone at the posterior pelvis was used for signal collection. Participant ABC scores were categorized as low confidence or high confidence. A random forest classified ABC groups using features from each step, calculated from smartphone signals. The random forest correctly classified the confidence level of 47 of 58 participants (accuracy 81.0%, sensitivity 63.2%, specificity 89.7%). This research demonstrated that smartphone signal data can classify people with lower limb amputations into balance confidence groups after completing a 6MWT. Integration of this model into the TOHRC Walk Test app would provide balance confidence classification, in addition to previously demonstrated clinical outcomes, after completing a single assessment and could inform individualized rehabilitation programs to improve confidence and prevent activity avoidance.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"3 8","pages":"e0000570"},"PeriodicalIF":0.0,"publicationDate":"2024-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11346636/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142074668","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":"Community-engaged artificial intelligence research: A scoping review.","authors":"Tyler J Loftus, Jeremy A Balch, Kenneth L Abbott, Die Hu, Matthew M Ruppert, Benjamin Shickel, Tezcan Ozrazgat-Baslanti, Philip A Efron, Patrick J Tighe, William R Hogan, Parisa Rashidi, Michelle I Cardel, Gilbert R Upchurch, Azra Bihorac","doi":"10.1371/journal.pdig.0000561","DOIUrl":"10.1371/journal.pdig.0000561","url":null,"abstract":"<p><p>The degree to which artificial intelligence healthcare research is informed by data and stakeholders from community settings has not been previously described. As communities are the principal location of healthcare delivery, engaging them could represent an important opportunity to improve scientific quality. This scoping review systematically maps what is known and unknown about community-engaged artificial intelligence research and identifies opportunities to optimize the generalizability of these applications through involvement of community stakeholders and data throughout model development, validation, and implementation. Embase, PubMed, and MEDLINE databases were searched for articles describing artificial intelligence or machine learning healthcare applications with community involvement in model development, validation, or implementation. Model architecture and performance, the nature of community engagement, and barriers or facilitators to community engagement were reported according to PRISMA extension for Scoping Reviews guidelines. Of approximately 10,880 articles describing artificial intelligence healthcare applications, 21 (0.2%) described community involvement. All articles derived data from community settings, most commonly by leveraging existing datasets and sources that included community subjects, and often bolstered by internet-based data acquisition and subject recruitment. Only one article described inclusion of community stakeholders in designing an application-a natural language processing model that detected cases of likely child abuse with 90% accuracy using harmonized electronic health record notes from both hospital and community practice settings. The primary barrier to including community-derived data was small sample sizes, which may have affected 11 of the 21 studies (53%), introducing substantial risk for overfitting that threatens generalizability. Community engagement in artificial intelligence healthcare application development, validation, or implementation is rare. As healthcare delivery occurs primarily in community settings, investigators should consider engaging community stakeholders in user-centered design, usability, and clinical implementation studies to optimize generalizability.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"3 8","pages":"e0000561"},"PeriodicalIF":0.0,"publicationDate":"2024-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11343451/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142044219","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}