PLOS digital health最新文献

{"title":"Design and implementation of digital literacy training programme: Findings of a quasi- experimental study from rural India.","authors":"Aparajita Gogoi, Mercy Manoranjini, Mamta Gupta","doi":"10.1371/journal.pdig.0000617","DOIUrl":"https://doi.org/10.1371/journal.pdig.0000617","url":null,"abstract":"<p><p>In India, digital divide is evident among rural sections of the society among women. For bridging this gap, the physical access to information and communications technology (ICT) along with training to harness these skills is important. This would empower the girls to gain more control over their health, finances, and safety, and can more freely assert their voice and agency. The project aimed to measure the knowledge and skills about the digital tools and its impact in different spheres of life (status of schooling, uptake of government schemes and mobility) for the adolescent girls before and after the implementation of the digital literacy training programme (DLTP) in rural India. The project was implemented in eight blocks of District Gumla of Jharkhand for three years from 2017-2020. The evaluations were conducted in three terms that is, baseline, midline and endline by using quasi-experimental research design. A quantitative questionnaire was developed to collect data from school dropout girls of 10-19 years. The sample coverage for intervention and comparison arm was 314 and 272 at the baseline, 318 and 260 at the midline, and 402 vs 202 in the endline. A state-of-the-art training curriculum was designed covering various components of digital literacy (hardware, software, applications, internet, emailing, social media, cyber security, education and career opportunities in the digital space). A team of facilitators were provided a laptop and a pico projector to conduct trainings in the intervention blocks. Majority of the participants were unmarried, lived along their parents and had ever attended school in both intervention and comparison arm. There was no significant difference in the proportion of girls having digital literacy score above median between intervention and comparison arm at the baseline. At the midline, the proportion of these girls with above median score was significantly higher in the intervention arm [n=203, 63.8%] as compared to the comparison arm [n=107, 41.2%]. Similarly, at the endline, the proportion of girls with higher median score for digital literacy increased significantly in the intervention arm [n=362, 90%] as compared to comparison arm [n=84, 41.6%]. The proportion of girls continuing education, physical access to ICT devices at home, exposure to mass media and perceived physical security increased in the intervention arm at the midline and endline as compared to comparison arm in the baseline. A significantly higher proportion of girls had knowledge about government schemes in the intervention arm as compared to comparison arm at the endline (p<0.05). We conclude that such tailor-made training programs which are weaved within cultural contexts can effectively bridge the digital gaps in resource scare settings. Along with this, they have the potential to bridge the literacy and economic gaps also within the community.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"4 4","pages":"e0000617"},"PeriodicalIF":0.0,"publicationDate":"2025-04-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11984724/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144045529","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":"Chronobiologically-informed features from CGM data provide unique information for XGBoost prediction of longer-term glycemic dysregulation in 8,000 individuals with type-2 diabetes.","authors":"Jamison H Burks, Leslie Joe, Karina Kanjaria, Carlos Monsivais, Kate O'laughlin, Benjamin L Smarr","doi":"10.1371/journal.pdig.0000815","DOIUrl":"https://doi.org/10.1371/journal.pdig.0000815","url":null,"abstract":"<p><p>Type 2 Diabetes causes dysregulation of blood glucose, which leads to long-term, multi-tissue damage. Continuous glucose monitoring devices are commercially available and used to track glucose at high temporal resolution so that individuals can make informed decisions about their metabolic health. Algorithms processing these continuous data have also been developed that can predict glycemic excursion in the near future. These data might also support prediction of glycemic stability over longer time horizons. In this work, we leverage longitudinal Dexcom continuous glucose monitoring data to test the hypothesis that additional information about glycemic stability comes from chronobiologically-informed features. We develop a computationally efficient multi-timescale complexity index, and find that inclusion of time-of-day complexity features increases the performance of an out-of-the-box XGBoost model in predicting the change in glucose across days. These findings support the use of chronobiologically-inspired and explainable features to improve glucose prediction algorithms with relatively long time-horizons.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"4 4","pages":"e0000815"},"PeriodicalIF":0.0,"publicationDate":"2025-04-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11981153/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144001611","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":"Deep neural network modeling for brain tumor classification using magnetic resonance spectroscopic imaging.","authors":"Erin B Bjørkeli, Knut Johannessen, Jonn Terje Geitung, Anna Karlberg, Live Eikenes, Morteza Esmaeili","doi":"10.1371/journal.pdig.0000784","DOIUrl":"https://doi.org/10.1371/journal.pdig.0000784","url":null,"abstract":"<p><p>This study is driven by the complex and specialized nature of magnetic resonance spectroscopy imaging (MRSI) data processing, particularly within the scope of brain tumor assessments. Traditional methods often involve intricate manual procedures that demand considerable expertise. In response, we investigate the application of deep neural networks directly to raw MRSI data in the time domain. Given the significant health risks associated with brain tumors, the necessity for early and accurate detection is crucial for effective treatment. While conventional MRI techniques encounter limitations in the rapid and precise spatial evaluation of diffuse gliomas, both accuracy and efficiency are often compromised. MRSI presents a promising alternative by providing detailed insights into tissue chemical composition and metabolic changes. Our proposed model, which utilizes deep neural networks, is specifically designed for the analysis and classification of spectral time series data. Trained on a dataset that includes both synthetic and real MRSI data from brain tumor patients, the model aims to distinguish MRSI voxels that indicate pathological conditions from healthy ones. Our findings demonstrate the model's robustness in classifying glioma-related MRSI voxels from those of healthy tissue, achieving an area under the receiver operating characteristic curve of 0.95. Overall, these results highlight the potential of deep learning approaches to harness raw MR data for clinical applications, signaling a transformative impact on diagnostic and prognostic assessments in brain tumor examinations. Ongoing research is focused on validating these approaches across larger datasets, to establish standardized guidelines and enhance their clinical utility.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"4 4","pages":"e0000784"},"PeriodicalIF":0.0,"publicationDate":"2025-04-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11981170/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144045255","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":"Using machine learning and single nucleotide polymorphisms for improving rheumatoid arthritis risk Prediction in postmenopausal women.","authors":"Yingke Xu, Qing Wu","doi":"10.1371/journal.pdig.0000790","DOIUrl":"https://doi.org/10.1371/journal.pdig.0000790","url":null,"abstract":"<p><p>Genetic factors contribute to 60-70% of the variability in rheumatoid arthritis (RA). However, few studies have used genetic variants to predict RA risk. This study aimed to enhance RA risk prediction by leveraging single nucleotide polymorphisms (SNPs) through machine-learning algorithms, utilizing Women's Health Initiative data. We developed four predictive models: 1) based on common RA risk factors, 2) model 1 incorporating polygenic risk scores (PRS) with principal components, 3) model 1 and SNPs after feature reduction, and 4) model 1 and SNPs with kernel principal component analysis. Each model was assessed using logistic regression (LR), random forest (RF), eXtreme Gradient Boosting (XGBoost), and support vector machine (SVM). Performance metrics included the area under the receiver operating characteristic curve (AUC), sensitivity, specificity, positive and negative predictive values (PPV and NPV), and F1-score. The fourth model, integrating SNPs with XGBoost, outperformed all other models. In addition, the XGBoost model that combines genomic data with conventional phenotypic predictors significantly enhanced predictive accuracy, achieving the highest AUC of 0.90 and an F1 score of 0.83. The DeLong test confirmed significant differences in AUC between this model and the others (p-values < 0.0001), particularly highlighting its efficacy in utilizing complex genetic information. These findings emphasize the advantage of combining in-depth genomic data with advanced machine learning for RA risk prediction. The most robust performance of the XGBoost model, which integrated both conventional risk factors and individual SNPs, demonstrates its potential as a tool in personalized medicine for complex diseases like RA. This approach offers a more nuanced and effective RA risk assessment strategy, underscoring the need for further studies to extend broader applications.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"4 4","pages":"e0000790"},"PeriodicalIF":0.0,"publicationDate":"2025-04-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11981130/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144060201","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":"Feasibility and preliminary efficacy of an online mindful walking intervention among COVID-19 long haulers: A mixed methods study including daily diary surveys.","authors":"Abhishek Aggarwal, Shan Qiao, Chih-Hsiang Yang, Slone Taylor, Cheuk Chi Tam, Xiaoming Li","doi":"10.1371/journal.pdig.0000794","DOIUrl":"10.1371/journal.pdig.0000794","url":null,"abstract":"<p><p>COVID-19 long haulers face profound psychosocial stressors (e.g., depression, anxiety, PTSD) and physical health challenges (e.g., brain fog, fatigue). This study tests the feasibility and initial impact of a digitally delivered mindful-walking (MW) intervention for improving the physical and psychosocial wellbeing of COVID-19 long haulers. We recruited 23 participants via Facebook groups, between March and November 2021, for a 4-week online MW intervention (i.e., 2 mindfulness practice sessions per week), that was delivered entirely through the study Facebook group. The intervention was assessed using mixed methods. Quantitative data were collected through brief daily evening surveys (i.e., 28 days) over the 4-week intervention period, and measured affect, cognition, mindfulness, physical activity, and MW engagement. Qualitative data were extracted from the Facebook group's Paradata (i.e., participant feedback, engagement metrics, and all social media interactions). Multilevel modeling was employed for the statistical analysis and a pragmatic approach was used for the qualitative analysis. The participants reported a high feasibility score (mean=4.93/7, SD=1.88), which was comprised of perceived usefulness, satisfaction, and ease of use. Those who engaged in MW, on any given day, frequently reported better psychosocial moods with more positive affect (β=0.89, p<0.01), less negative affect (β=-0.83, p<0.01), higher perceived cognitive ability (β=0.52, p<0.05), and more physical activity (β=0.41, p<0.05). Additionally, participants who practiced MW more consistently during the study reported higher levels of momentary mindfulness (β=0.3 p<0.01). Participants expressed satisfaction with the intervention, reporting benefits such as better symptom management and an overall improvement in wellbeing. Despite the small sample size, the digital delivery of our MW intervention via Facebook showed high acceptability. Preliminary efficacy findings indicate improved mental wellbeing and physical activity among long haulers. Larger-scale RCTs are needed in the future to improve the robustness and applicability of findings.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"4 4","pages":"e0000794"},"PeriodicalIF":0.0,"publicationDate":"2025-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11978064/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143813080","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":"Evaluating the feasibility, adoption, cost-effectiveness, and sustainability of telemedicine interventions in managing COVID-19 within low-and-middle-income countries (LMICs): A systematic review.","authors":"Nonye M Okafor, Imani Thompson, Vandana Venkat, Courtney Robinson, Aishwarya Rao, Sumedha Kulkarni, Leah Frerichs, Khady Ndiaye, Deborah Adenikinju, Chukwuemeka Iloegbu, John Pateña, Hope Lappen, Dorice Vieira, Joyce Gyamfi, Emmanuel Peprah","doi":"10.1371/journal.pdig.0000771","DOIUrl":"10.1371/journal.pdig.0000771","url":null,"abstract":"<p><p>COVID-19 has tragically taken the lives of more than 6.5 million people globally, significantly challenging healthcare systems and service delivery, especially in low-and middle-income countries (LMICs). This systematic review aims to: (1) evaluate the feasibility of telemedicine interventions for COVID-19 management; (2) assess the adoption of telemedicine interventions during the COVID-19 pandemic; (3) examine the cost-effectiveness of telemedicine implementation efforts and (4) analyze the sustainability of telemedicine interventions for COVID-19 disease management within LMIC service settings. We reviewed studies from selected public health and health science databases, focusing on those conducted in countries classified as low and middle-income by the World Bank, using telemedicine for confirmed COVID-19 cases, and adhering to Proctor's framework for implementation outcomes. Of the 766 articles identified and 642 screened, only 3 met all inclusion criteria. These studies showed reduced reliance on antibiotics, prescription drugs, and emergency department referrals among telemedicine patients. Statistical parity was observed in the length of stay, diagnostic test ordering rates, and International Classification of Diseases (ICD)-10 diagnoses between telemedicine and in-person visits. Telemedicine interventions designed for post-COVID physical rehabilitation demonstrated safety, sustainability, and enhanced quality of life for patients without requiring specialized equipment, proving adaptable across contexts with appropriate technology. These interventions were also economically sustainable and cost-effective for healthcare systems as a whole. Proposed strategies to bridge implementation gaps include community-level assessments, strategic planning, multisectoral partnerships of local hospital administration and lawmakers, legal consultations, and healthcare informatics improvements. Increased investment in telemedicine research focusing on infectious disease management is crucial for the continued development and refinement of effective strategies tailored to resource-constrained regions.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"4 4","pages":"e0000771"},"PeriodicalIF":0.0,"publicationDate":"2025-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11978082/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143813119","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":"Telehealth or in-person HIV care? Qualitative study findings on decision-making from people with HIV and HIV care providers in South Carolina during the COVID-19 pandemic.","authors":"Valerie Yelverton, Salome-Joelle Gass, Daniel Amoatika, Christopher Cooke, Jan Ostermann, Nabil Natafgi, Nicole L Hair, Bankole Olatosi, Otis L Owens, Shan Qiao, Xiaoming Li, Caroline Derrick, Sharon Weissman, Helmut Albrecht","doi":"10.1371/journal.pdig.0000812","DOIUrl":"10.1371/journal.pdig.0000812","url":null,"abstract":"<p><p>The COVID-19 pandemic disrupted HIV care services across the United States. Telehealth was rapidly implemented to ensure HIV care continuity. Despite the evidence of unequal telehealth uptake among some people with HIV (PWH), the decision-making processes to determine who received telehealth or in-person care are under-researched. This study assessed which decision criteria and processes determined which HIV care visit type was used by PWH and HIV care providers during the COVID-19 pandemic. Qualitative in-depth interviews with 18 PWH and 10 HIV care providers from South Carolina assessed PWHs' and HIV care providers' decision-making criteria and processes for telehealth HIV care during the COVID-19 pandemic. Interviews were analyzed using thematic analysis. Most PWH (11 out of 18) and all providers had used telehealth for HIV care. To guide visit type decisions, interviewees reported decision-making criteria across four domains: patient-related criteria, clinical criteria, provider preference, and HIV care continuity. Patient-related criteria included patient preference, convenience, fear of COVID-19 exposure and stigma, and transportation barriers. Clinical criteria included the need for a physical exam, a person's care history and health status. While all identified decision criteria were important, we found a hierarchical structure: care continuity superseded other criteria. Some clinical criteria were reported as decision-relevant criteria by providers but not PWH. Most PWH reported that they were included or took the lead in the visit type decision process. Decision-making processes to determine PWHs' HIV care visit types considered criteria across multiple domains. The superseding criteria was to sustain HIV care continuity. To guide future telehealth use, shared decision-making is needed to weigh patient-related, provider-related, and clinical decision criteria and maintain care continuity, and to comprehensively include all relevant decision criteria.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"4 4","pages":"e0000812"},"PeriodicalIF":0.0,"publicationDate":"2025-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11977962/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143813190","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":"Deep learning-based assessment of pulp involvement in primary molars using YOLO v8.","authors":"Aydin Sohrabi, Nazila Ameli, Masoud Mirimoghaddam, Yuli Berlin-Broner, Hollis Lai, Maryam Amin","doi":"10.1371/journal.pdig.0000816","DOIUrl":"10.1371/journal.pdig.0000816","url":null,"abstract":"<p><p>Dental caries is a major global public health problem, especially among young children. Rapid decay progression often necessitates pulp treatment, making accurate pulp condition assessment crucial. Despite advances in pulp management techniques, diagnostic methods for assessing pulp involvement have not significantly improved. This study aimed to develop a machine learning (ML) model to diagnose pulp involvement using radiographs of carious primary molars. Clinical charts and bitewing radiographs of 900 children treated from 2018-2022 at the University of Alberta dental clinic were reviewed, yielding a sample of 482 teeth. images were preprocessed, standardized, and labeled based on clinical diagnoses. Data were split into training, validation, and test sets, with data augmentation applied to classify 2 categories of outcomes. The YOLOv8m-cls model architecture included convolutional and classification layers, and performance was evaluated using top-1 and top-5 accuracy metrics. The YOLOv8m-cls model achieved a top-1 accuracy of 78.7% for upper primary molars and 87.8% for lower primary molars. Validation datasets showed higher accuracy for lower primary teeth. Performance on new test images demonstrated precision, recall, accuracy, and F1-scores, highlighting the model's effectiveness in diagnosing pulp involvement, with lower primary molars showing superior results. This study developed a promising CNN model for diagnosing pulp involvement in primary teeth using bitewing radiographs, showing promise for clinical application in pediatric dentistry. Future research should explore whole bitewing images, include clinical variables, and integrate heat maps to enhance the model. This tool could streamline clinical practice, improve informed consent, and assist in dental student training.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"4 4","pages":"e0000816"},"PeriodicalIF":0.0,"publicationDate":"2025-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11977986/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143813106","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 effect of software and hardware version on Apple Watch activity measurement: A secondary analysis of the COVFIT retrospective cohort study.","authors":"Shelby L Sturrock, Rahim Moineddin, Dionne Gesink, Sarah Woodruff, Daniel Fuller","doi":"10.1371/journal.pdig.0000727","DOIUrl":"10.1371/journal.pdig.0000727","url":null,"abstract":"<p><p>The objective of this study was to estimate the impact of software and hardware version on Apple Watch activity measurement using data from the COVFIT retrospective cohort study. We estimated the impact of software and hardware versions on activity measurement by comparing daily active calories and daily exercise minutes in the 7 days before and 7 days after upgrading from watchOS 5 to 6, 6 to 7, 7 to 8, 8 to 9 or between two hardware versions. For each transition, we fit mixed effect negative binomial regression models to estimate the effect of the upgrade on daily (a) exercise minutes and (b) active calories, overall and stratified by sex, with and without adjusting for weekday. We also calculated and plotted the mean person-level change in average activity levels between the two weeks. As a control, we repeated the entire analysis comparing activity data two weeks before vs. one week before each upgrade. 253 participants contributed data about at least one transition (software = 250, hardware = 74). Hardware upgrades were not associated with either outcome; however, some software upgrades were. Upgrading from watchOS 7 to 8 was associated with a large, statistically significant increase in daily exercise minutes (unadjusted rate ratio (RR) = 1.13, 95% CI: 1.06, 1.20). WatchOS 6 to 7 and 8 to 9 transitions were associated with statistically significant decreases in daily exercise minutes (6 to 7: unadjusted RR = 0.92, 95% CI: 0.86, 0.99; 8 to 9: unadjusted RR = 0.91, 95% CI: 0.86, 0.96) and active calories (6 to 7: RR = 0.96, 95% CI: 0.94, 0.99); 8 to 9: RR = 0.97, 95% CI: 0.94, 0.99). There was no significant change in either outcome during in the two-week control period for most transitions. Differences in software version over time or between people may confound physical activity analyses using Apple Watch data.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"4 4","pages":"e0000727"},"PeriodicalIF":0.0,"publicationDate":"2025-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11977988/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143813206","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":"Exploring the diagnostic accuracy of an HIV self-test optimized by a digital app-based solution: Results from a secondary data analysis of a field trial in South Africa.","authors":"Ashlyn Beecroft, Aliasgar Esmail, Olivia Vaikla, Thomas Duchaine, Nora Engel, Chen Liang, Qihuang Zhang, Keertan Dheda, Nitika Pant Pai","doi":"10.1371/journal.pdig.0000791","DOIUrl":"10.1371/journal.pdig.0000791","url":null,"abstract":"<p><strong>Background: </strong>To reach UNAIDS 95-95-95 targets, digital HIV self-testing (HIVST) strategy aided by applications, platforms, and readers can engage young people and adults living with undetected HIV infection. Evidence on its acceptability, feasibility, impact exists, yet accuracy data are limited.</p><p><strong>Methods: </strong>A secondary data analysis of a quasi-RCT of digital HIVST in South Africa was performed. We hypothesized app-guided digital interpretation of oral self-test enhanced test accuracy. We compared accuracy between digital HIVST supervised vs. unsupervised (with/without healthcare worker). Self-test results were interpreted and uploaded by participants, compared using computer vision technology, against lab reference standard by trained healthcare professionals.</p><p><strong>Results: </strong>1513 digital HIVST participants reported pooled Sensitivity (Sn) = 95.52% (95% CI, 94.48%-96.56%); Specificity (Sp): 99.93% (95% CI, 99.79%-100.06%); Positive predictive value (PPV): 99.22% (95% CI, 98.78%-99.67%); Negative Predictive Value (NPV): 99.57% (95% CI, 99.24%-99.90%). 565 participants on supervised digital HIVST, reported a pooled Sn: 93.65% (95% CI, 91.64-95.66); Sp: 100.00% (95% CI, 100.00-100.00); PPV: 100.00% (95% CI, 100.00-100.00); NPV: 99.21% (95% CI, 98.48-99.94). 968 unsupervised digital HIVST participants, reported a pooled Sn: 97.18% (95% CI, 96.13-98.24); Sp: 99.89% (95% CI, 99.67-100.10); PPV: 98.57% (95% CI, 97.82-99.33); NPV: 99.77% (95% CI, 99.47-100.08). Non-digital HIVST vs. study digital HIVST data at 5% significance level - Sn: chi = 0.6495, p-value = 0.4203, Sp: chi = 0.3831, p-value = 0.5259. Supervised vs. unsupervised HIVST at 5% significance level - Sn: chi = 0.973, p-value = 0.3237, Sp: chi = 0.527, p-value = 0.4449.</p><p><strong>Conclusions: </strong>Digital HIVST improved interpretation of test results, increased accuracy and predictive value estimations (upper limit 98%-100%), removing subjectivity. Unsupervised digital HIVST users performed better than supervised. Digital HIVST results can potentially signal a rapid triage to therapy or prevention pathways, while awaiting lab confirmation. Findings have implications for scale up of digital HIVST initiatives in global settings.</p>","PeriodicalId":74465,"journal":{"name":"PLOS digital health","volume":"4 4","pages":"e0000791"},"PeriodicalIF":0.0,"publicationDate":"2025-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11977960/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143813120","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}