Background: Perinatal depression (PD) and postpartum depression (PPD) are leading causes of morbidity in the United States (U.S.). Asian American women, the fastest-growing racial groups in U.S., are disproportionately affected by cultural stigma, language barriers, and limited access to culturally responsive healthcare. This review examines the current methods, evidence gaps, and opportunities to address perinatal and PPD through mobile health (mHealth) applications among Asian American women. Methods: A scoping review was conducted using PubMed, EBSCOhost, Google Scholar searches following the principles of systematic and rapid review methodology. Articles were included if they addressed Asian American women with PD or PPD, focused on mHealth or telehealth interventions, and peer-reviewed publications from last ten years. A total of 246 articles were identified, from which 25 studies were selected for inclusion. Data were synthesized thematically across six domains. Results: The sample included observational (24%), qualitative (16%), pilot (12%), RCTs (12%), reviews (24%), protocols (8%), and mixed (4%); Edinburgh Postnatal Depression Scale (EPDS) was most frequently used (68%). mHealth tools with hybrid approaches were widely used regardless of location and systemic barriers. Mindfulness and cognitive behavioral therapy (CBT)-based interventions were effective in reducing depressive symptoms, improving maternal self-efficacy, and enhancing psychosocial outcomes. However, engagement was lower among women with severe depressive symptoms, and mental illness stigmatization limited access to digital tools. Key motivators for uptake included connectivity, feasibility, and adaptability. Discussion: mHealth interventions demonstrate considerable potential to improve depressive symptoms during perinatal and postpartum periods. However, their implementation and evaluation among Asian American women remain limited. Future interventions should prioritize culturally and linguistically tailored digital platforms, integrate peer and professional support, and align with existing maternal healthcare systems to improve accessibility, engagement, and equity while addressing persistent disparities in maternal mental health.
Background: The convergence of multi-omics technologies and artificial intelligence (AI) has opened new frontiers in precision medicine; however, the complexity and opacity of advanced AI models remain a major barrier to clinical adoption. This systematic review aims to critically evaluate explainable AI (XAI) strategies for multi-omics integration and their role in bridging the translational gap between computational innovation and clinical utility. Methods: A systematic literature search was conducted across PubMed/MEDLINE, Scopus, and Web of Science databases for studies published between 2020 and 2025, following PRISMA 2020 guidelines. Studies addressing multi-omics integration using explainable or interpretable AI methods in precision medicine were included. Data extraction and narrative synthesis were performed due to methodological heterogeneity. Results: A total of 116 studies were included in the final analysis. Computational approaches ranged from classical machine learning and deep learning to graph-based and transformer architectures. XAI techniques, including SHAP (SHapley Additive exPlanations), attention mechanisms, and saliency maps, enabled interpretable predictions across gene, pathway, and network levels. Applications were most prominent in cancer subtyping, biomarker discovery, drug response prediction, and prognosis modeling. Despite promising performance, key challenges persist, including data heterogeneity, high dimensionality, batch effects, overfitting, limited reproducibility, and insufficient clinical validation. Discussion: XAI enhances transparency, trust, and biological interpretability in multi-omics models, facilitating their integration into clinical workflows. Emerging directions such as federated learning, causal AI, foundation models, digital twins, and human-in-the-loop systems offer potential solutions to current limitations. Standardized evaluation frameworks and robust clinical validation are essential to advance real-world implementation. This review provides a comprehensive roadmap for developing reliable and clinically actionable XAI-driven multi-omics systems in precision medicine.
Aim: The global promotion of digital health is accelerating the transformation of healthcare systems. Consequently, in contemporary healthcare environments characterized by information overload, nurses are increasingly demanded to possess advanced information-processing abilities to appropriately search for, critically appraise, apply, and disseminate health information. Therefore, in this study, we aimed to investigate the current state of digital health literacy among hospital nurses and examine its association with nursing informatics competency. Methods: We conducted this cross-sectional, web-based survey between May and August 2025. We recruited participants from 50 randomly selected large hospitals (≥ 400 beds) in the Kansai region of Japan. We measured digital health literacy and nursing informatics competency using the validated Japanese versions of the Digital Health Literacy Instrument and the Nursing Informatics Competency Scale, respectively, and then described the total digital health literacy score (mean of all items) and its subscale scores. We applied Pearson correlation and multiple regression analyses to evaluate the association between these variables, adjusting for potential confounders. Results: We included 113 nurses in the final analysis. The overall mean score for digital health literacy was 2.8. While operational skills and privacy protection scored the highest, the evaluation of reliability and addition of self-generated content were the lowest-scoring domains. Digital health literacy was positively correlated with nursing informatics competency. In the multivariable model, digital health literacy was independently and positively associated with nursing informatics competency, indicating the strongest association among all examined factors. Conclusions: Nurses displayed moderate digital health literacy, with proficiency largely limited to basic information-access skills. Beyond demographic and occupational factors, individual digital health literacy may represent an important enabling factor for professional nursing informatics competency. Future research is needed to clarify how digital health literacy is related to nursing informatics competency and to examine the broader mechanisms and contextual factors underlying this association.
Aim: Evaluate the associations between usage of the standalone SmartMoms Canada mHealth intervention, and gestational weight gain (GWG) guideline adherence and lifestyle improvements in pregnant individuals in the context of a pragmatic study. Methods: Participants (18–40 years, BMI 18.5–39.9 kg/m2) were recruited into a single-arm trial conducted in Winnipeg and Ottawa, Canada. All participants were provided with the app, a Fitbit® tracker, and a smart scale. Participants were assessed in early, mid-, and late pregnancy. Physical activity was measured with the Godin Leisure Time Exercise score, and the Fitbit® tracker (steps and time in physical activity). Fitbit® app was used to measure dietary intake. App usage and GWG were monitored. GWG guideline adherence was compared with data from the Statistics Canada Maternal Experiences Survey (MES). GWG adherence and lifestyle changes were compared between app usage groups (≥ median weekly app usage vs. < median) with multinomial logistic regressions or t tests. Trajectories in lifestyle changes were compared between groups with repeated measure analyses. Results: Of the 75 participants recruited in early pregnancy, 51 were followed through pregnancy (32% drop out). Overall app usage was low (median 1.30 min/wk). Adequate GWG was achieved by 35.7% (95% CI: 23.2–48.2) of participants vs. 32.6% in the MES; while excessive GWG occurred in 50.0% (95% CI: 36.9–63.1) vs. 48.7%. GWG adherence was not different between usage groups (P = 0.399), but a higher mean weekly app usage (continuous) was associated with lower odds of insufficient GWG (OR = 0.01, P = 0.035). There were no significant associations between app usage and changes in physical activity, but a lower increase in carbohydrate intake was observed in the higher usage group. Conclusions: Few associations were found between app usage and GWG or lifestyle outcomes. Lack of significant results could relate to low protocol and intervention adherence (Trial registration: http://www.isrctn.com/ISRCTN16254958).
Aim: Planning orthopedic tumor surgery requires substantial cognitive effort to interpret 3D plans derived from 2D preoperative images and translate them into the patients’ actual anatomy. Mixed Reality (MR) 3D holograms overlaid on patients may help surgeons visualize surgical steps more intuitively before making skin incisions. This study evaluated the use of MR for preoperative assessment in 72 patients with primary or revision orthopedic oncology conditions, as well as the technical issues encountered during clinical implementation, between July 2021 and November 2025. Methods: 3D Slicer or MIMICS software was used to generate tumor models and support surgical planning. A proprietary MR platform (versions 1 and 2) was developed to integrate patients’ medical images and 3D models into digital asset bundles, which were then downloaded to the MR headset in the operating room via the hospital’s Wi-Fi network. The surgeon examined each patient preoperatively using the conventional 2D method first, and then applied the MR 3D hologram method. Results: A Likert-scale questionnaire showed that the MR 3D hologram group outperformed the 2D group across all aspects of spatial awareness of the patient’s pathoanatomy and was viewed as a more effective tool for preoperative planning. Regarding NASA-TLX scores, the overall cognitive workload during preoperative assessment was lower in the MR 3D hologram group. Since December 2024, generating cinematic-rendered 3D models with the upgraded MR software platform (version 2) has taken an average of 61 minutes (49–156). Engineer intervention was needed in 4 of 36 cases (11.1%). All cases were wirelessly accessible and completed an MR assessment. The average time to perform hologram-to-patient registration for the last 26 cases was 2.3 minutes (0.95–5.17). Conclusions: Our results suggest that MR technology could enhance surgeons’ 3D spatial awareness in various orthopedic tumor surgeries and reduce cognitive load during the translation of surgical plans.
Aim: Telemonitoring apps are increasingly prescribed as part of self-management for patients with Chronic Obstructive Pulmonary Disease (COPD), yet patients still make minimal use of these apps. This research investigates explanatory factors associated with the behavioral intention to use and actual use of COPD telemonitoring apps among users and non-users. Methods: A cross-sectional study was conducted among 200 COPD patients from two Dutch hospitals. Eligible participants (≥ 18 years, diagnosed with COPD, ≥ 2 outpatient pulmonology visits in 2023) were identified through the electronic health record and invited by mail. Participants completed a self-administered questionnaire assessing demographics, disease severity, literacy, facilitating conditions, and app-related factors, based on the Unified Theory of Acceptance and Use of Technology 2 (UTAUT2), the Technology Acceptance Model (TAM), and the Reasoned Action Approach (RAA). Behavioral intention was analyzed using hierarchical multiple regression, and use was analyzed using binomial logistic regression. Results: Intention was explained by performance expectancy (coefficient = 0.760, p ≤ 0.001), self-efficacy (coefficient = 0.207, p = 0.009), and alignment with personal norms and values (coefficient = 0.163, p = 0.006). Use was explained by self-efficacy (OR = 1.992, p = 0.023), social influence (OR = 1.642, p = 0.039), personalization (OR = 0.628, p = 0.039), and intention to use (OR = 3.459, p ≤ 0.001). App users showed significantly higher digital literacy, performance expectancy, and fewer symptoms compared to non-users. Users also experienced significantly higher importance of social influence and alignment with norms and values than non-users. Demographic variables and disease severity were no significant predictors of behavioral intention and use. Conclusions: Optimizing the app and the supportive role of the healthcare professional, enhancing digital and health literacy, and hybrid care ensures that patients can benefit from both traditional care and the advantages of remote monitoring.
Aim: Cerebral palsy (CP) is one of the most common motor neurodevelopmental disorders, affecting approximately three in every thousand live births in North America. The study aims to investigate and identify the factors influencing manual dexterity performance among children with CP and typically developing (TD) children according to the Manual Ability Classification System (MACS) levels. Methods: A total of 100 children aged 4 to 12 years were enrolled, including 50 diagnosed with CP and 50 TD children. Manual dexterity performance was assessed across MACS levels. A Bayesian seemingly unrelated regression (BayesSUR) framework was applied to identify influential factors, explicitly accounting for interrelationships among multiple response variables. This probabilistic approach allowed for robust estimation under uncertainty while incorporating correlations across outcomes. Results: The BayesSUR analysis revealed distinct factor influences MACS levels. For children with mild CP (MACS level 1), object type had the strongest effect on response time. For moderately affected children (MACS level 2), direction most strongly influenced movement error, while age impacted both error and success rate. Among severely affected children (MACS level 3) and TD children, gender emerged as the dominant factor influencing response time. However, the low inclusion probabilities of other factors suggest that additional data and validation are warranted. Conclusions: The findings highlight the importance of considering both individual characteristics and task-specific factors when designing interventions to improve manual dexterity in children with CP. These results contribute to a better understanding of the key determinants influencing motor performance and may guide the development of more effective therapeutic and rehabilitation strategies. The Trial Registration Number: CTRI/2018/07/014900.
Aim: To examine the behavioral signature of the “Algorithmic Self,” characterizing how users adapt their identity and behaviors in response to algorithmic reinforcement among active digital media users in Pakistan. Methods: A cross-sectional quantitative design was employed with 422 adults aged 18–45 years across five major cities. Participants completed a structured online questionnaire capturing demographic data, digital usage patterns, the Algorithmic Exposure Score (AES), and Algorithmic Self Behavioral Signature Scale (ASBSS). Validated instruments assessed social comparison, Fear of Missing Out (FoMO), self-esteem, and digital stress. Data were analyzed using descriptive statistics, Pearson correlations, and multiple linear regression in SPSS version 26, with significance set at p < 0.05. Results: Participants demonstrated moderate-to-high levels of Algorithmic Self formation, with 39.8% classified in the high category. Higher daily screen time, greater platform diversity, stronger algorithmic trust, and elevated social comparison were associated with higher Algorithmic Self Scores. In multiple linear regression analysis, daily screen time (β = 0.34), social comparison (β = 0.31), algorithmic trust (β = 0.29), and algorithmic exposure (β = 0.28) emerged as significant predictors of Algorithmic Self formation, while FoMO was not a significant predictor (β = 0.11, p = 0.09). The final model explained 56% of the variance in Algorithmic Self formation (R2 = 0.56, adjusted R2 = 0.54, p < 0.001). Conclusions: AI-driven digital environments are associated with self-presentation, identity adaptation, and behavioral regulation among Pakistani users. These findings highlight the importance of enhancing digital literacy, improving awareness of algorithmic influence, and further investigating the psychological and societal implications of Algorithmic Self formation in digitally mediated environments.
Aim: To benchmark three deep learning-based retinal image registration methods RetinaRegNet, EyeLiner, and GeoFormer on the Fundus Image Registration (FIRE) dataset to compare registration accuracy and computational efficiency using mean landmark error (MLE) as the primary outcome measure. Methods: The three image registration approaches were evaluated using the FIRE dataset under consistent conditions across varying image overlap conditions (Classes S, A, and P). These included: (a) RetinaRegNet, which incorporates diffusion features, dual keypoint sampling through Scale-Invariant Feature Transform (SIFT) and random, two-stage outlier removal, and a multilevel registration hierarchy progressing from homography to polynomial transforms; (b) EyeLiner, which integrates anatomical segmentation with SuperPoint feature extraction, LightGlue matching, and thin-plate spline warping; (c) GeoFormer, which builds on Local Feature Transformers (LoFTR) through cross-attention mechanisms and Random Sampling Consensus (RANSAC)-based refinement. Registration performance was quantified using MLE. Results: Across all 134 FIRE image pairs, RetinaRegNet achieved the lowest overall MLE (3.12 pixels), outperforming EyeLiner (3.81 pixels) and GeoFormer (6.06 pixels). Class-specific analysis showed that RetinaRegNet delivered the highest accuracy in Class S images (1.70 pixels), competitive performance in Class A (5.24 pixels), and the strongest results in the most challenging Class P cases (4.57 pixels). GeoFormer demonstrated the shortest processing time at 0.32 seconds per image pair, compared with 4.92 seconds for EyeLiner and 31.23 seconds for RetinaRegNet. In Class P, RetinaRegNet achieved a 59.2% improvement in accuracy relative to GeoFormer (4.57 vs 11.20 pixels). The code is available at: https://github.com/ThenukaDharmaseelan/image_Registration. Conclusions: Overall, the evaluation reveals a clear trade-off between registration precision and computational speed. RetinaRegNet achieves the lowest MLE for complex clinical cases despite higher computational cost. EyeLiner balances precision and speed for routine use, while GeoFormer prioritizes rapid throughput where processing speed is critical.
Telepsychiatry has transitioned from a supplementary modality to a sustained component of contemporary mental healthcare, driven by technological advancement, workforce shortages, and the COVID-19 pandemic. This narrative review synthesizes current evidence on clinical effectiveness, service models, technological integration, and ethical–legal considerations, and contextualizes these domains through institutional implementation experience in Türkiye. Across major diagnostic groups, including mood, anxiety, psychotic, neurodevelopmental, and substance use disorders, published studies generally indicate comparable outcomes and patient satisfaction to face-to-face care when delivered within structured clinical frameworks. We further articulate the theoretical foundations of clinical equivalence, emphasizing language-mediated therapeutic mechanisms, alliance formation in video-based settings, and behavioral factors influencing adherence. The manuscript introduces a system-level perspective for Türkiye, positioning telepsychiatry as a capacity-extending model within geographically uneven workforce distribution. Institutional applications, including disaster response, postpartum screening pathways, and hybrid specialty clinics, illustrate context-sensitive implementation strategies. Emerging innovations such as digital phenotyping, artificial intelligence, and virtual reality are discussed alongside regulatory, equity, and data governance considerations. We conclude that telepsychiatry represents not merely an emergency substitute but an increasingly integrated and policy-relevant model of care.
Multicenter imaging studies are increasingly critical in epidemiology, yet variability across scanners, acquisition protocols, and reconstruction algorithms introduces systematic biases that threaten reproducibility and comparability of quantitative biomarkers. This paper reviews the major sources of heterogeneity in MRI, CT, and PET-CT data, highlighting their impact on epidemiologic inference, including misclassification, reduced statistical power, and compromised generalizability. We outline harmonization strategies spanning pre-acquisition standardization, phantom-based calibration, post-acquisition intensity normalization, and advanced statistical and machine learning methods such as ComBat and domain adaptation. Illustrative examples from MRI flow quantification and radiomic feature extraction demonstrate how harmonization can mitigate site effects and enable robust large-scale analyses.
Background: Sepsis is a major cause of disease worldwide. Mobile applications (apps) have been developed to assist clinical practice. Current evidence evaluating such apps is diverse. This scoping review aimed to map currently available literature investigating the usage of mobile apps for sepsis-related healthcare. This will highlight evidence gaps, and areas for future innovation and app development. Methods: Databases MEDLINE, Embase, CINAHL, Cochrane, Scopus, and Web of Science were searched in June 2023 (updated in July 2024). Studies containing original research investigating mobile apps for sepsis-related healthcare were included and analysed in three categories identified from the primary purpose of the app: (1) education and awareness, (2) clinical assistance, and (3) biomarker or pathogen detection. Results: A total of 1,755 studies were identified and 27 included following screening, of which 19 (70%) were published in 2020 or later. Most of the 27 studies investigated apps for clinical assistance (70%, n = 19). These apps were diverse, acting as digital solutions for data collection (n = 2), triage (n = 6), clinical guideline access (n = 5), alert delivery (n = 1), and outcome prediction (n = 5). There were five apps (19%) used to assist biomarker or pathogen detection. Of these, most (80%, n = 4) mobile apps were used to detect and quantify colorimetric signals in combination with assays, and all five apps had attachments necessary for laboratory processes. Lastly, three apps (11%) were designed to enhance education and awareness, two targeting medical education and one targeting public awareness. Discussion: Mobile applications offer innovative and exciting digital solutions for biomarker detection, education, and clinical support in sepsis-related healthcare. Current literature is highly heterogenous and rapidly developing.
Aim: This study aims to evaluate the outreach achieved by psychiatry-related posts using the hashtag #YouthMentalHealth, highlighting how social media platforms can shape public discourse on adolescent mental health. Methods: We utilized the Fedica research analytics tool to characterize posts containing #YouthMentalHealth from January 10, 2018, to January 10, 2023. This analysis examined the #YouthMentalHealth activity timeline, identifying the number of posts containing the hashtag and the geographical distribution to assess the effectiveness of hashtag campaigns. Results: The #YouthMentalHealth movement resulted in 58,000 posts shared by around 25,000 X users, generating 292.7 million impressions (views). The top three countries from which most posts containing #YouthMentalHealth were shared included the United States (35.14%), Canada (29.15%), and the United Kingdom (14.37%). The three largest contributor groups were management companies (20.6%), educational advocacy organizations (17.5%), and social advocacy groups (14%). Conclusions: This first-of-its-kind study explores the impact and utilization of #YouthMentalHealth globally, reporting trends and patterns from digital media platforms. By mapping the hashtag’s global footprint, the study offers novel insights into how digital advocacy can amplify youth mental health awareness and connect multidisciplinary stakeholders. These findings contribute to emerging frameworks in digital psychiatry by underscoring the role of social media as a complementary tool for mental health promotion and community engagement, while illuminating diverse strategies to aid the psychiatric community in effectively addressing the mental health needs of adolescents.
Background: To synthesize evidence on how medical thermography, integrated with artificial intelligence (AI), blockchain, 5G (5th Generation mobile networks), and Internet of Things (IoT), enhances diagnostics, fraud prevention, and personalized health insurance in emerging markets, addressing cost escalation and access gaps. Methods: This systematic review followed AMSTAR 2 and PRISMA guidelines, synthesizing 25 sources (22 peer-reviewed articles, 3 industry reports) from a pre-analyzed dataset. Inclusion focused on relevance to thermography, insurance, or synergistic technologies; exclusions included non-peer-reviewed or irrelevant items. Data extraction via Microsoft Excel (version 2409) covered diagnostics, applications, synergies, and contexts. Quality appraisal used the Mixed Methods Appraisal Tool (MMAT) to assess methodological rigor. Narrative synthesis addressed heterogeneity, without meta-analysis due to design diversity and resource limits. Results: Thermography achieves 83–98% sensitivities for breast cancer (asymmetries > 3.0°C), diabetic foot ulcers (DFUs; 96.71% with AI), and rheumatoid arthritis (RA; inflammation > 0.5°C), reducing triage times by 25% and costs by 30% in mobile settings. Blockchain’s six-layer architecture, with Practical Byzantine Fault Tolerance and InterPlanetary File System, secures data at US$0.028 per transaction, potentially reducing fraud through enhanced verification. In emerging markets like India and Brazil, portable thermography with 5G supports screening, aligned with standards like T/ZADT 005-2002. Discussion: These integrations enable early detection (saving US$8,000–12,000 per DFU), fraud mitigation, and equitable access, though protocol variances and biases require attention. Recommendations include standardization, pilots in rural areas, and bias-mitigating AI frameworks to optimize health insurance outcomes.
Aim: The aim of this study is to compare the accuracy, reliability, and educational quality of YouTube videos on osteochondritis dissecans based on their YouTube Health verification status. Methods: The term “osteochondritis dissecans” was searched on June 3, 2024. The first 50 videos found on YouTube after searching “osteochondritis dissecans” were evaluated. The Journal of the American Medical Association (JAMA) benchmark criteria was used to score video reliability and accuracy (0–4 points), the Global Quality Score (GQS) was used to score nonspecific educational content (0–5 points), and the osteochondritis dissecans specific score (OCDSS) was used to score specific educational content (0–11 points). Three independent reviewers scored all videos, and interrater reliability was assessed with intraclass correlation coefficients (ICC). Group differences were analyzed with one-way analysis of variance (ANOVA) and independent sample t-tests, and multivariable linear regression was used to identify independent predictors of JAMA, GQS, and OCDSS scores. Results: A total of 50 videos were analyzed with a cumulative 326,851 views. The mean JAMA score was 2.28 ± 0.64, the mean GQS score was 2.60 ± 1.36, and the mean OCDSS was 5.02 ± 3.16. The mean JAMA score for YouTube Health verified videos was 2.44 ± 0.34, GQS was 2.72 ± 1.22, and OCDSS was 5.72 ± 2.69. The mean JAMA score for videos not verified by YouTube Health was 2.29 ± 0.65, GQS score was 2.61 ± 1.44, and OCDSS was 4.95 ± 3.37. These differences were not statistically significant: JAMA p = 0.380, GQS p = 0.837, OCDSS p = 0.546. Conclusions: There were no significant differences in reliability, educational content, and comprehensiveness between videos that were verified by YouTube Health and videos that were not verified.
Vaccines have eliminated once-deadly diseases, yet rising vaccine hesitancy threatens these gains. Human papillomavirus (HPV) illustrates this crisis: Although it is one of the few vaccines that directly prevents cancer, uptake remains low in the United States and globally, particularly in regions with high cervical cancer incidence. This persistent gap undermines both individual and public health. This paper examines how digital health technologies, aligned with policy frameworks and community engagement, can address HPV vaccine hesitancy. We propose the Digital Vaccine Advocacy Toolkit, a structured, HPV-focused framework that integrates electronic health record (EHR)-based clinical decision support, personalized reminders, population dashboards, AI-driven misinformation surveillance, and culturally tailored education. As a conceptual model, it draws on secondary evidence and policy recommendations rather than original empirical data, emphasizing interoperability, privacy safeguards, equity-driven design, and stakeholder engagement to support feasibility across diverse health systems. The Toolkit is organized into illustrative workflows that demonstrate how technical features could be combined with policy mechanisms and financing models to strengthen HPV vaccination. By situating HPV within the World Health Organization’s 90-70-90 elimination targets and the recent adoption of single-dose schedules, the framework highlights both translational relevance and global applicability, though its recommendations require pilot testing and empirical validation. Overall, the Digital Vaccine Advocacy Toolkit offers a practical roadmap for improving HPV vaccine uptake through the integration of technology, policy, and ethics, and provides a transferable model for advancing digital health strategies to increase vaccine confidence and equity in immunization programs worldwide.
Aim: A comprehensive understanding of current digital literacy and perspectives of the psychiatric workforce is important to introduce appropriate digital psychiatry interventions and implement contextually relevant measures in Pakistan. This study aims to address a gap in the existing literature by assessing psychiatrists’ knowledge, attitudes, perceived barriers, and willingness to integrate digital psychiatry into their clinical practice. Methods: A cross-sectional online survey was conducted from January 2023 to June 2023 across psychiatric departments of 18 public hospitals in Pakistan. The study included psychiatry residents, fellows, and consultants. A 48-item questionnaire, internally and externally validated, assessed knowledge, perceptions, and willingness to adopt digital psychiatry tools—telepsychiatry, artificial intelligence, mental health applications, and virtual reality. Data were analyzed using Statistical Package for the Social Sciences (version 26) for descriptive statistics, correlation, and regression analyses, while thematic analysis of open-ended responses was performed using Quirkos. Results: A total of 200 participants (56.0% aged 20–30 years, n = 112; 55.5% male, n = 111) were part of this study. 68.5% (n = 137) understood the applications of telepsychiatry, while 72.5% (n = 145) agreed that it is time-efficient and cost-effective. Only 39.5% (n = 79) of participants had received relevant artificial intelligence training to incorporate it in their psychiatric clinical practice. 62.0% (n = 124) of respondents reported unfamiliarity with the use of mental health applications. Regarding virtual reality, 32.5% (n = 65) were familiar with the technology, but only 42.5% (n = 85) were aware of its applications in psychiatric care. Thematic reflexive analysis revealed major challenges, including a ‘lack of infrastructure/resources’ (44.5%, n = 89) and a ‘lack of education/awareness’ (21.5%, n = 43). Conclusions: This study represents the first cross-sectional examination of digital psychiatric literacy in Pakistan’s healthcare system, which revealed significant gaps in digital health competencies among psychiatrists. Given the vast potential of emerging technologies in addressing mental health challenges, there is an urgent need for mental health professionals in Pakistan to integrate digitization in psychiatric practice.
Artificial intelligence (AI) is transforming healthcare by equipping clinicians and patients with tools that support more efficient, patient-centered care. In pediatrics, however, the implementation of AI demands a higher threshold for responsibility, transparency, and family-centered engagement. This perspective explores the opportunities and challenges of AI in pediatric healthcare, highlighting the unique ethical and developmental considerations that distinguish children’s care from adult medicine. Drawing on Kaiser Permanente’s seven principles for responsible AI, the article emphasizes the importance of augmentation over automation, the need for pediatric-specific validation, and the necessity of trustworthiness and fairness in clinical deployment. It outlines how AI can support primary care providers through enhanced decision support, early screening for developmental and behavioral disorders, including the potential for AI to create personalized developmental trajectories, moving beyond static population norms to provide earlier, more precise insights into a child’s neurodevelopmental progress, improved electronic health record usability, and risk prediction models. However, without careful governance, AI poses risks of bias, inequity, and erosion of clinician judgment. Policy recommendations include redesigning family consent models, ensuring robust clinician training, and mandating pediatric-specific testing of AI systems with diverse, representative datasets. Ultimately, AI should function as a supportive tool that strengthens, not replaces, human empathy, clinical expertise, and family-centered values. Responsible innovation is essential to ensure that children benefit equitably from AI while maintaining trust, safety, and compassion in pediatric healthcare.
Digital twin technology is emerging as a transformative paradigm in healthcare, shifting practice from provider-centered models toward more personalized forms of medicine. As dynamic virtual representations of the human body, digital twins integrate biometric data, lifestyle patterns, and clinical records to simulate, monitor, and predict health trajectories in real time. Their growing use raises not only technical possibilities but also important questions about how patients relate to these data-driven counterparts, particularly when twins inform everyday health decisions in chronic care, such as diabetes or oncology. This perspective examines these relational dynamics and their ethical, cultural, and experiential implications for autonomy, decision-making, and the lived experience of being represented in data. To guide this analysis, we introduce a scale framework with three intersecting lenses: time, distinguishing asynchronous from synchronous updating; twining, ranging from close mirroring to more augmentative forms of representation; and control, spanning human-led to twin-driven decision authority. Using this framework, we position four common types of digital twins: mirror, shadow, intelligent, and simulacra as an evolution from basic representation to transformative modeling. We argue that future healthcare and public health policy must go beyond technical innovation to address patients’ lived experiences, ensuring that digital twins enhance rather than diminish autonomy, trust, and equity. This perspective thus calls for a patient-centered approach in designing and implementing digital twin technologies.
Aim: Diagnosing and treating major depressive disorder (MDD) remains a pressing global health challenge. Generative-AI tools, by lowering technical barriers and offering rapid visual feedback, may open new avenues for art-based assessment and intervention. Methods: In this exploratory qualitative pilot, we conducted reflexive thematic analysis of semi-structured interviews with N = 10 young adults at elevated risk for depression who generated self-representative images in Midjourney during a 45-minute session. Participants were selected from a larger cohort described elsewhere; no quantitative analyses were conducted in the present paper. Results: Qualitative findings suggested therapeutic-like mechanisms that mirror—and in some cases amplify—those reported for traditional art therapy, including the experience of flow and spontaneity, a heightened sense of creative agency, and the safe externalization of difficult or extreme emotions. Some participants described abrupt “sentiment switches,” where joyful imagery was immediately followed by scenes of sudden, intrusive self-criticism. Importantly, the generative process also surfaced idiosyncratic “resource images” (e.g., nature motifs, hobbies, values, loved ones) that participants experienced as calming or empowering, hinting at personalised anchors for future interventions. Conclusions: In line with prior quantitative work showing that more negative prompt sentiment statistically relates to higher BDI scores, the present qualitative narratives offer an interpretive account of how such negativity may emerge during AI-assisted self-representation. However, the current study does not integrate datasets or perform mixed-methods triangulation and uses those prior findings solely for contextualization. We conclude that, with appropriate ethical safeguards, generative-AI image making may serve as a flexible, low-cost adjunct to existing diagnostic and art-therapeutic practices, offering clients and clinicians a shared visual language for exploring the multi-layered experience of depression.
