PLOS digital health最新文献

Background: Diabetic retinopathy (DR) is a leading cause of blindness globally. The gold standard for DR screening is stereoscopic colour fundus photography with tabletop cameras. VistaView is a novel smartphone-based retinal camera which offers mydriatic retinal imaging. This study compares the diagnostic accuracy of the smartphone-based VistaView camera compared to a traditional desk mounted fundus camera (Triton Topcon). We also compare the agreement between graders for DR screening between VistaView images and Topcon images.

Methodology: This prospective study took place between December 2021 and June 2022 in Pakistan. Consecutive diabetic patients were imaged following mydriasis using both VistaView and Topcon cameras at the same sitting. All images were graded independently by two graders based on the International Classification of Diabetic Retinopathy (ICDR) criteria. Individual grades were assigned for severity of DR and maculopathy in each image. Diagnostic accuracy was calculated using the Topcon camera as the gold standard. Agreement between graders for each device was calculated as intraclass correlation coefficient (ICC) (95% CI) and Cohen's weighted kappa (k).

Principal findings: A total of 1428 images were available from 371 patients with both cameras. After excluding ungradable images, a total of 1231 images were graded. The sensitivity of VistaView for any DR was 69.9% (95% CI 62.2-76.6%) while the specificity was 92.9% (95% CI 89.9-95.1%), and PPV and NPV were 80.5% (95% CI 73-86.4%) and 88.1% (95% CI 84.5-90.9) respectively. The sensitivity of VistaView for RDR was 69.7% (95% CI 61.7-76.8%) while the specificity was 94.2% (95% CI 91.3-96.1%), and PPV and NPV were 81.5% (95% CI 73.6-87.6%) and 89.4% (95% CI 86-92%) respectively. The sensitivity for detecting maculopathy in VistaView was 71.2% (95% CI 62.8-78.4%), while the specificity was 86.4% (82.6-89.4%). The PPV and NPV of detecting maculopathy were 63% (95% CI 54.9-70.5%) and 90.1% (95% CI 86.8-92.9%) respectively. For VistaView, the ICC of DR grades was 78% (95% CI, 75-82%) between the two graders and that of maculopathy grades was 66% (95% CI, 59-71%). The Cohen's kappa for retinopathy grades of VistaView images was 0.61 (95% CI, 0.55-0.67, p<0.001), while that for maculopathy grades was 0.49 (95% CI 0.42-0.57, p<0.001). For images from the Topcon desktop camera, the ICC of DR grades was 85% (95% CI, 83-87%), while that of maculopathy grades was 79% (95% CI, 75-82%). The Cohen's kappa for retinopathy grades of Topcon images was 0.68 (95% CI, 0.63-0.74, p<0.001), while that for maculopathy grades was 0.65 (95% CI, 0.58-0.72, p<0.001).

Conclusion: The VistaView offers moderate diagnostic accuracy for DR screening and may be used as a screening tool in LMIC.

Signs and symptoms of Attention-Deficit/Hyperactivity Disorder (ADHD) are present at preschool ages and often not identified for early intervention. We aimed to use machine learning to detect ADHD early among kindergarten-aged children using population-level administrative health data and a childhood developmental vulnerability surveillance tool: Early Development Instrument (EDI). The study cohort consists of 23,494 children born in Alberta, Canada, who attended kindergarten in 2016 without a diagnosis of ADHD. In a four-year follow-up period, 1,680 children were later identified with ADHD using case definition. We trained and tested machine learning models to predict ADHD prospectively. The best-performing model using administrative and EDI data could reliably predict ADHD and achieved an Area Under the Curve (AUC) of 0.811 during cross-validation. Key predictive factors included EDI subdomain scores, sex, and socioeconomic status. Our findings suggest that machine learning algorithms that use population-level surveillance data could be a valuable tool for early identification of ADHD.

Community health programs are gaining relevance within national health systems and becoming inherently more complex. To ensure that community health programs lead to equitable geographic access to care, the WHO recommends adapting the target population and workload of community health workers (CHWs) according to the local geographic context and population size of the communities they serve. Geographic optimization could be particularly beneficial for those activities that require CHWs to visit households door-to-door for last mile delivery of care. The goal of this study was to demonstrate how geographic optimization can be applied to inform community health programs in rural areas of the developing world. We developed a decision-making tool based on OpenStreetMap mapping and route optimization algorithms in order to inform the micro-planning and implementation of two kinds of community health interventions requiring door-to-door delivery: mass distribution campaigns and proactive community case management (proCCM) programs. We applied the Vehicle Routing Problem with Time Windows (VRPTW) algorithm to optimize the on-foot routes that CHWs take to visit households in their catchment, using a geographic dataset obtained from mapping on OpenStreetMap comprising over 100,000 buildings and 20,000 km of footpaths in the rural district of Ifanadiana, Madagascar. We found that personnel-day requirements ranged from less than 15 to over 60 per CHW catchment for mass distribution campaigns, and from less than 5 to over 20 for proCCM programs, assuming 1 visit per month. To illustrate how these VRPTW algorithms can be used by operational teams, we developed an "e-health" platform to visualize resource requirements, CHW optimal schedules and itineraries according to customizable intervention designs and hypotheses. Further development and scale-up of these tools could help optimize community health programs and other last mile delivery activities, in line with WHO recommendations, linking a new era of big data analytics with the most basic forms of frontline care in resource poor areas.

Biases in medical artificial intelligence (AI) arise and compound throughout the AI lifecycle. These biases can have significant clinical consequences, especially in applications that involve clinical decision-making. Left unaddressed, biased medical AI can lead to substandard clinical decisions and the perpetuation and exacerbation of longstanding healthcare disparities. We discuss potential biases that can arise at different stages in the AI development pipeline and how they can affect AI algorithms and clinical decision-making. Bias can occur in data features and labels, model development and evaluation, deployment, and publication. Insufficient sample sizes for certain patient groups can result in suboptimal performance, algorithm underestimation, and clinically unmeaningful predictions. Missing patient findings can also produce biased model behavior, including capturable but nonrandomly missing data, such as diagnosis codes, and data that is not usually or not easily captured, such as social determinants of health. Expertly annotated labels used to train supervised learning models may reflect implicit cognitive biases or substandard care practices. Overreliance on performance metrics during model development may obscure bias and diminish a model's clinical utility. When applied to data outside the training cohort, model performance can deteriorate from previous validation and can do so differentially across subgroups. How end users interact with deployed solutions can introduce bias. Finally, where models are developed and published, and by whom, impacts the trajectories and priorities of future medical AI development. Solutions to mitigate bias must be implemented with care, which include the collection of large and diverse data sets, statistical debiasing methods, thorough model evaluation, emphasis on model interpretability, and standardized bias reporting and transparency requirements. Prior to real-world implementation in clinical settings, rigorous validation through clinical trials is critical to demonstrate unbiased application. Addressing biases across model development stages is crucial for ensuring all patients benefit equitably from the future of medical AI.

Remote longitudinal studies are on the rise and promise to increase reach and reduce participation barriers in chronic disease research. However, maintaining long-term retention in these studies remains challenging. Early identification of participants with different patterns of long-term retention offers the opportunity for tailored survey adaptations. Using data from the online arm of the Swiss Multiple Sclerosis Registry (SMSR), we assessed sociodemographic, health-related, and daily-life related baseline variables against measures of long-term retention in the follow-up surveys through multivariable logistic regressions and unsupervised clustering analyses. We further explored follow-up survey completion measures against survey requirements to inform future survey designs. Our analysis included data from 1,757 participants who completed a median of 4 (IQR 2-8) follow-up surveys after baseline with a maximum of 13 possible surveys. Survey start year, age, citizenship, MS type, symptom burden and independent driving were significant predictors of long-term retention at baseline. Three clusters of participants emerged, with no differences in long-term retention outcomes revealed across the clusters. Exploratory assessments of follow-up surveys suggest possible trends in increased survey complexity with lower rates of survey completion. Our findings offer insights into characteristics associated with long-term retention in remote longitudinal studies, yet they also highlight the possible influence of various unexplored factors on retention outcomes. Future studies should incorporate additional objective measures that reflect participants' individual contexts to understand their ability to remain engaged long-term and inform survey adaptations accordingly.

The availability of quality and timely data for routine monitoring of mental, neurological and substance use (MNS) disorders is a challenge, particularly in Africa. We assessed the feasibility of using an open-source data science technology (R Shiny) to improve health data reporting in Nairobi City County, Kenya. Based on a previously used manual tool, in June 2022, we developed a digital online data capture and reporting tool using the open-source Kobo toolbox. Primary mental health care providers (nurses and physicians) working in primary healthcare facilities in Nairobi were trained to use the tool to report cases of MNS disorders diagnosed in their facilities in real-time. The digital tool covered MNS disorders listed in the World Health Organization's (WHO) Mental Health Gap Action Program Intervention Guide (mhGAP-IG). In the digital system, data were disaggregated as new or repeat visits. We linked the data to a live dynamic reproducible dashboard created using R Shiny, summarising the data in tables and figures. Between January and August 2023, 9064 cases of MNS disorders (4454 newly diagnosed, 4591 revisits and 19 referrals) were reported using the digital system compared to 5321 using the manual system in a similar period in 2022. Reporting in the digital system was real-time compared to the manual system, where reports were aggregated and submitted monthly. The system improved data quality by providing timely and complete reports. Open-source applications to report health data is feasible and acceptable to primary health care providers. The technology improved real-time data capture, reporting, and monitoring, providing invaluable information on the burden of MNS disorders and which services can be planned and used for advocacy. The fast and efficient system can be scaled up and integrated with national and sub-national health information systems to reduce manual data reporting and decrease the likelihood of errors and inconsistencies.

The Home Health Project, set on Clare Island, five kilometres off the Irish Atlantic coast, is a pilot exploration of ways in which various forms of technology can be utilised to improve healthcare for individuals living in isolated communities. The integration of digital health technologies presents enormous potential to revolutionise the accessibility of healthcare systems for those living in remote communities, allowing patient care to function outside of traditional healthcare settings. This study aims to explore the personal experiences and perspectives of participants who are using digital technologies in the delivery of their healthcare as part of the Home Health Project. Individual semi-structured interviews were conducted with nine members of the Clare Island community participating in the Home Health Project. Interviews took place in-person, in June 2023. Interviews were audio-recorded and transcribed verbatim. The data were analysed inductively using reflexive thematic analysis. To identify determinants of engagement with the Home Health Project, the data was then deductively coded to the Theoretical Domains Framework (TDF) and organised into themes. Seven of the possible 14 TDF domains were supported by the interview data as influences on engagement with the Project: Knowledge, Beliefs about capabilities, Optimism, Intentions, Environmental context and resources, Social influences and Emotion. Overall, participants evaluated the Home Health Project as being of high quality which contributed to self-reported increases in health literacy, autonomy, and feeling well supported in having their health concerns addressed. There was some apprehension related to data protection, coupled with a desire for extended training to address aspects of digital illiteracy. Future iterations can capitalise on the findings of this study by refining the technologies to reflect tailored health information, personalised to the individual user.

Food and beverage marketing on social media contributes to poor diet quality and health outcomes for youth, given their vulnerability to marketing's effects and frequent use of social media. This study benchmarked the reach and frequency of earned and paid media posts, an understudied social media marketing strategy, of food brands frequently targeting Canadian youth. The 40 food brands with the highest brand shares in Canada between 2015 and 2020 from frequently marketed food categories were determined using Euromonitor data. Digital media engagement data from 2020 were licensed from Brandwatch, a social intelligence platform, to analyze the frequency and reach of brand-related posts on Twitter, Reddit, Tumblr, and YouTube. The 40 food brands were mentioned on Twitter, Reddit, Tumblr, and YouTube a total of 16.85M times, reaching an estimated 42.24B users in 2020. The food categories with the most posts and reach were fast food restaurants (60.5% of posts, 58.1% of total reach) and sugar sweetened beverages (29.3% of posts, 37.9% of total reach). More men mentioned (2.77M posts) and were reached (6.88B users) by the food brands compared to women (2.47M posts, 5.51B users reached). The food and beverage brands (anonymized), with the most posts were fast food restaurant 2 (26.5% of the total posts), soft drink 2 (10.4% of the total posts), and fast food restaurant 6 (10.1% of the total posts). In terms of reach, the top brands were fast food restaurant 2 (33.1% of the total reach), soft drink 1 (18.1% of the total reach), and fast food restaurant 6 (12.2% of the total reach). There is a high number of posts on social media related to food and beverage brands that are popular among children and adolescents, primarily for unhealthy food categories and certain brands. The conversations online surrounding these brands contribute to the normalization of unhealthy food and beverage intake. Given the popularity of social media use amongst of children and adolescents, policies aiming to protect these vulnerable groups need to include the digital food environment.

Recent estimates report a high incidence and prevalence of HIV among men who have sex with men (MSM) in Malaysia. Mobile apps are a promising and cost-effective intervention modality to reach stigmatized and hard-to-reach populations to link them to HIV prevention services (e.g., HIV testing, pre-exposure prophylaxis, PrEP). This study assessed attitudes and preferences toward the format, content, and features of a mobile app designed to increase HIV testing and PrEP uptake among Malaysian MSM. We conducted six online focus groups between August and September 2021 with 20 MSM and 16 stakeholders (e.g., doctors, nurses, pharmacists, and NGO staff) to query. Transcripts were analyzed in Dedoose software to identify thematic content. Key themes in terms of app functions related to stylistic preferences (e.g., design, user interface), engagement strategies (e.g., reward systems, reminders), recommendations for new functions (e.g., enhanced communication options via chat, discussion forum), cost of services (e.g., PrEP), and legal considerations concerning certain features (e.g., telehealth, patient identification), minimizing privacy and confidentiality risks. Our data suggest that a tailored HIV prevention app would be acceptable among MSM in Malaysia. The findings further provide detailed recommendations for successfully developing a mobile app to improve access to HIV prevention services (e.g., HIV testing, PrEP) for optimal use among MSM in Malaysia.