Artificial intelligence (AI) is rapidly reshaping clinical education by embedding assessment and feedback into everyday learning activities. Medical students can now use machine learning dashboards, generative AI, large language models, and emerging agentic systems to practice clinical reasoning, communication, and procedural skills while receiving individualized feedback within seconds. However, the availability of more data and more feedback does not necessarily produce better learning. This Viewpoint is intended for clinical educators, assessment leaders, curriculum committees, faculty developers, and institutional leaders who must decide how AI should be used in formative activities without reducing education to automated scoring. AI-assisted formative assessment is defined in this paper as the intentional use of AI tools to generate, organize, and support interpretation of performance information for learning rather than grading. Its distinctive contribution lies in the scale, adaptivity, conversational simulation, pattern detection, and possible autonomy of AI systems. However, AI outputs become formative only when learners and educators interpret them critically, judge their trustworthiness, and translate them into a small number of focused follow-on learning actions. This paper synthesizes the current evidence base while noting that much of it remains early, heterogeneous, and concentrated in short-term or single-setting studies. It examines key risks, including hallucination, automation bias, epistemic overtrust, hidden curricular effects, and broader concerns related to professional identity, power asymmetries, data privacy, and inequitable access. It also presents context-specific implementation examples for preclinical case-based learning, communication and objective structured clinical examination preparation, procedural skill laboratories, clerkship learning, and programmatic assessment portfolios, together with practical implications for faculty development, institutional governance, and phased local implementation. As a Viewpoint rather than an empirical study or systematic review, the framework and examples should be interpreted as evidence-informed design propositions that require local evaluation and validation. Overall, the value of AI-assisted formative assessment depends less on the volume of AI-generated feedback than on educational designs that preserve learner agency, professional judgment, and human accountability.
To evaluate ocular blood flow in school-aged children born prematurely, with or without retinopathy of prematurity (ROP), using laser speckle flowgraphy (LSFG), and to determine factors associated with LSFG-derived parameters. This cross-sectional observational study included 123 school-aged preterm children: 54 without ROP, 20 with untreated type 2/mild ROP, and 49 with treated type 1 ROP. All participants underwent ophthalmic examinations and optic nerve head (ONH) perfusion assessments using LSFG. Mean blur rate indices of the entire ONH (MA), vessel area (MV), and tissue area (MT) of the right eye were analysed. Multivariable generalised linear models were used to identify variables associated with LSFG parameters, and Pearson's correlation was used to examine the association between axial length and MT. Children with type 1 ROP had lower MA, MV, and MT than those with type 2/mild ROP, whereas MV was the highest in the type 2/mild ROP group. Type 1 ROP was independently associated with reduced MA and MV in the full cohort. Among children with ROP, anti-vascular endothelial growth factor (VEGF) treatment was associated with lower MA, MV, and MT. Longer axial length correlated with reduced MT in both the full cohort (r = -0.30, p < 0.01) and the ROP subgroup (r = -0.35, p < 0.01). Type 1 ROP was associated with reduced ocular blood flow in school-aged children born preterm. Among eyes with ROP, a history of anti-VEGF treatment was associated with lower MA, MV, and MT, although this may reflect disease severity.
Myocardial ischemia-reperfusion injury (MIRI) poses diagnostic challenges due to complex histopathological changes. This study aimed to develop an intelligent framework for evaluating MIRI on hematoxylin-eosin-stained slides, to compare major deep learning architectures, and to determine the advantages of transformer models across multiple interventions and time points. A total of 1280 whole-slide images (~62,000 tiles) from public datasets were analyzed across antioxidant, β-blocker, calcium channel blocker, and control groups at 6, 24, and 72 hours. Seven model families (convolutional neural networks, recurrent neural networks, long short-term memory networks, autoencoders, graph convolutional networks, variational autoencoders, and transformers) were trained under unified preprocessing, with generative adversarial networks used exclusively for leakage-free augmentation. Weak supervision used a clustering-constrained attention multiple-instance learning strategy, and segmentation applied a Transformer-UNet. Data were split into 8:1:1 at the subject level, with 5-fold cross-validation. The transformer achieved the best performance (accuracy=0.942; area under the curve=0.982; and F1-score=0.958). Segmentation Dice scores were 0.847 (necrosis) and 0.821 (apoptosis). Predictions strongly agreed with expert measurements (r=0.886; Bland-Altman limits +5% or -5%), and attention maps aligned with necrotic borders and inflammatory foci. Temporal trends matched biological expectations, with the antioxidant group showing the most stable improvement. Transformer-based pathology offers accurate, robust, and interpretable assessment of MIRI and provides a scalable framework for dynamic injury quantification and therapeutic evaluation.
During radiotherapy of non-small-cell lung cancers, changes in radiomics features extracted from cone-beam computed tomography images can serve as unique predictors. For adapting treatment plans or assessing early treatment responses, there is a pressing need to screen cone-beam computed tomography-derived radiomics features for predicting tumour regression volume. Our study enrolled 58 patients with non-small cell lung cancer undergoing radiotherapy, comprising dataset1, and an additional 32 patients who underwent an adaptive treatment approach, forming dataset2. For dataset1 patients, radiomics features were extracted from both the planning computed tomography and the first treatment cone-beam computed tomography. For dataset2 patients, features were also extracted from the replanning computed tomography and the corresponding cone-beam computed tomography after 20 fractions of radiotherapy. By employing correlation coefficient and intraclass correlation coefficient analyses on planning computed tomography and first treatment cone-beam computed tomography features, we identified robust and reproducible features in dataset1 patients. Reproducible delta radiomics features were further selected to characterize feature changes and screen delta radiomics features for tumour regression in dataset2 patients. Lastly, predictive delta radiomics features were screened using least absolute shrinkage and selection operator regression. Based on dataset1, we selected 34 radiomics features. From these, 16 delta radiomics features were further screened using dataset2. The changes observed in these features were consistent across both computed tomography and cone-beam computed tomography images. Specifically, nine features exhibited significant differences (p<0.05) while seven features remained relatively unchanged (p>0.05). Ultimately, four delta radiomics features were identified to classify tumour regression volume with a 30 % threshold for cone-beam computed tomography images. Our study introduces a novel strategy for selecting computed tomography-based cone-beam computed tomography radiomics features, enabling the identification of reliable features for model development. Preliminary findings firstly demonstrate the feasibility of using cone-beam computed tomography-based delta radiomics features to classify tumour regression volume during radiotherapy of non-small-cell lung cancer.
Chimeric antigen receptor T-cell therapy (CAR-T) has revolutionized the treatment of B-cell precursor ALL (B-ALL), but its global availability is limited. This study assessed current access and barriers to CAR-T CD19 therapy for children across Europe. A country questionnaire developed by the European Group for Blood and Marrow Transplantation Pediatric Diseases Working Party, St Jude Children's Research Hospital, and IBFM assessed current access to advanced therapies for B-ALL in Europe using Qualtrics software. Data from 36 WHO-defined European countries (27 high-income, nine upper-middle-income) observed a median of five pediatric hematology-oncology (PHO) centers per country (0.56 PHO centers/1 million inhabitants, range, 0.05-1.83). Hematopoietic stem-cell transplantation (HSCT) facilities were available in 89% of countries (32/36). CAR-T CD19 therapy was available in 72% of countries; however, 25/36 (69%) countries lacked clinical trials or international collaborations for pediatric CAR-T CD19 therapy. Most countries accepted foreign patients, but referrals remained limited, with 1-2 foreign patients treated annually per country. Eighteen countries expressed interest in a referral network, but only six had established mechanisms for domestic or international referrals. Substantial disparities exist in access to advanced therapies for pediatric B-ALL across Europe. Although CAR-T CD19 therapy is available in most countries, gaps in clinical trials, collaborations, and referral systems limit equitable access. Efforts to improve infrastructure and establish referral networks are essential to enhance care for patients with pediatric B-ALL.
Term and preterm newborns have differences in skin structure. Preterm newborns have skin structures that are more susceptible to various disorders. These differences are believed to impact skin pH regulation and bacterial colonization, particularly Staphylococcus spp. This study investigates the acidity difference and skin Staphylococcus spp. colonization in preterm and term newborns. This is a descriptive observational study with a cross-sectional design. The study compared skin pH and Staphylococcus spp. colonization between term (37-41 weeks) and preterm (28-36 weeks) newborns. Skin pH was measured using a skin pH meter, and Staphylococcus spp. colonization was assessed through bacterial identification and colony-forming unit (CFU) testing. Assessments were conducted within the first 24 hours of life. There were 29 participants in this study: 16 preterm newborns and 13 term newborns. The preterm group was dominated by subjects with alkaline skin pH (80%), and the term group was dominated by subjects with acidic skin pH (71%). Positive Staphylococcus spp. colonization was more prevalent in the preterm group (62%) than in the term group (38%). Skin pH differed between preterm and term newborns, with relatively higher skin pH in the preterm group. Staphylococcus spp. was also more prevalent in the preterm group.
Between-individual variability in physiological adaptations to endurance exercise training is often interpreted as reflecting stable differences in training responsiveness. However, whether the capacity for physiological adaptation itself represents a stable, individual-specific trait under repeated exposure to the same training stimulus remains unclear. This study investigated the reproducibility of within-individual adaptations to repeated endurance training. Forty-two middle-aged, untrained men and women completed two similar eight-week endurance training periods (TP1 and TP2), separated by an eight-week detraining period. Assessments spanned multiple physiological and performance outcomes, including hemoglobin mass (Hbmass), skeletal muscle citrate synthase (CS) activity, capillaries∙fibre-1, maximal oxygen uptake (V̇O2max), and 15-minute maximal mean power output (MMPO15-min). Within-individual reproducibility was evaluated using intraclass correlation coefficients with [95% confidence intervals]. Participants completed 23.9 (0.4) and 23.8 (0.7) training sessions in TP1 and TP2, with minimal within-individual variation. Baseline values were highly consistent between periods for V̇O2max (0.98 [0.97, 0.99]), MMPO15-min (0.96 [0.93, 0.97]), and most other outcomes, indicating effective detraining. Nevertheless, reproducibility of within-individual adaptations was poor for Hbmasssub> (0.00 [0.00, 0.30]), CS activity (0.15 [0.00, 0.33]), capillaries∙fibre-1 (0.00 [0.00, 0.36]), V̇O2max (0.19 [0.00, 0.37]), and MMPO15-min (0.20 [0.00, 0.37]). Within-individual variability in hematological and skeletal muscle adaptations accounted for little of the within-individual variability in V̇O2max and MMPO15-min responses. This study demonstrates that individual adaptations to endurance training exhibit limited reproducibility across multiple physiological and performance outcomes. These findings indicate that physiological adaptability to endurance training does not reflect a stable, individual-specific attribute, but instead emerge from dynamic and context-dependent biological processes.
Autism spectrum disorder (ASD) not only affects a person's social communication and behaviors, but also has an impact on their parents, who encounter different challenges during caregiving. Interventions developed for parents of children with ASD often focus on improving child outcomes and seldom consider the well-being of parents and families. Interventions leveraging mindfulness-based approaches have been developed to support parents of children with ASD, but the costs, inflexibility, and scarcity of resources may limit their accessibility. App-based interventions can be an accessible, scalable, and economical way of providing interventions at a primary health care level. The aim of this study was to develop an evidence-based digital intervention that complements existing, overloaded psychiatric services, to provide mindfulness-based psychoeducation for parents of children with ASD to improve their mental well-being. The app development process follows the systematic approach of intervention mapping. Needs assessment was first conducted through semistructured qualitative interviews with health care professionals. Performance and change objectives were specified; theory-based and practical application methods were selected, followed by the design of the curriculum for a structured intervention. A pilot waitlist randomized controlled trial was conducted to evaluate the feasibility, acceptability, and preliminary efficacy of the app with parents of children with ASD recruited from a tertiary child psychiatric service in Hong Kong. The resulting intervention, the TRIP app, is a 6-week structured intervention consisting of 6 sessions per week (each session lasting 15-20 minutes), covering topics on ASD parenting skills and mindfulness practices. The six weekly themes include (1) cultivating curiosity in parenting, (2) mindfulness of the breath and body, (3) management of core and associated features of ASD, (4) managing conflicts and setting boundaries, (5) perspective taking, and (6) cultivating self-compassion. The curriculum was designed to target the determinants of parental stress, including parents' knowledge, skills, emotions, and attitudes. App content and features were designed to incorporate behavioral change techniques, social cognitive theory, and elaboration likelihood model, to enhance efficacy and promote long-term usage. The app was found to be feasible and acceptable in the pilot randomized controlled trial (n=40), with greater long-term usage among parents of children on the waiting list who were yet to receive diagnostic assessment and clinical management, when compared with parents of children who have already been receiving clinical care. The TRIP app was developed based on knowledge and expertise across psychiatry, public health, behavioral science, and implementation science. It caters to the unmet needs for improving caregiver well-being in the holistic care model for families of children with ASD. The clinical efficacy of the TRIP app is yet to be evaluated through clinical trials.
Cognitive impairment (CI) is a significant burden for patients with multiple sclerosis (MS). However crucial its assessment is, longitudinal measurement of cognitive performance is susceptible to learning effect, making the results of repeated evaluations difficult to interpret. Reliable change index (RCI) and standardized regression based norms (SRB) are accepted statistical methods to assess the reliability of a difference score between two observations. Thus, our aims were to provide RCIs and SRBs for all three subtests of the Brief International Cognitive Assessment for MS (BICAMS) battery and to measure the prevalence of true cognitive worsening and improvement. We retrospectively evaluated the first interim analysis data of the longitudinal follow-up or our BICAMS prevalence study-cohort after 1-year. We analyzed the data of 242 MS patients. We calculated both the RCIs and the SRBs for all three subtests of the BICAMS battery. According to the RCI, 5.4%, 6.9% and 14.6% worsened while 12.3%, 34.3% and 10.6% improved on the SDMT, BVMT-R and CVLT-II respectively. In case of SRB method, 3.8%, 8.3% and 19.7% worsened while 3.8%, 7.6% and 0.0% improved. The κ values revealed a mild-to-moderate agreement (κ=0.391-0.540; p<0.001). In case of the BVMT-R and the CVLT-II assessments the baseline scores influenced this outcome significantly (BVMT-R: OR: 1.068; 90%CI: 1.001-1.138; CVLT-II: OR: 1.096; 90%CI: 1.041-1.153). Comparing the methods, RCI seems to be better in cases with already established CI, while SRB, the more complex method, seemingly detects change better in cognitively intact patients.
Recent evidence links adiposity with the risk of severe infections, but whether muscle strength may also be an independent risk factor is less studied. We investigated the association between handgrip strength and risk of common infections and sepsis and explored potential mediation by plasma proteomic biomarkers. We analyzed data from 405,451 UK Biobank participants and replicated the main findings in 4474 Chinese adults from the Hong Kong Osteoporosis Study (HKOS). Baseline handgrip strength was measured by a dynamometer. Cox models were used to estimate its association with incidence of pneumonia, urinary tract infection (UTI), skin infection, and sepsis, adjusting for sociodemographic, lifestyle, and health-related factors. Mediation analyses were performed using 2912 plasma proteins in a UK Biobank subsample (n = 42,414) to identify biological pathways. In UK Biobank (median follow-up 13.6-15.3 years), lower handgrip strength was associated with significantly increased risk of pneumonia (hazard ratio per 5-kg decrement=1.10; 95% CI=1.09-1.11), UTI (1.10; 1.09-1.11), skin infection (1.05; 1.04-1.05), and sepsis (1.08; 1.07-1.10). Associations were largely consistent in HKOS, and the relative risks associated with low grip strength were generally most pronounced in underweight individuals. GDF15 and PLAUR were identified as the most important proteins which mediated 12-14% of these associations. Low handgrip strength is associated with increased risks of common infections and sepsis, particularly in underweight individuals, with partial mediation by proteins related to inflammation and immune-related pathways. Handgrip strength assessment may provide prognostic value beyond body mass index for clinical risk stratification.
Multiple Sclerosis is a neurodegenerative disease frequently associated with gait impairments that can emerge early and progressively worsen, substantially affecting mobility and independence. The Expanded Disability Status Scale (EDSS) is widely used to quantify overall disability in people with multiple sclerosis. However, gait alterations may present as specific spatiotemporal and kinematic changes that are not explicitly described by the EDSS scoring criteria. In this context, quantitative gait analysis can provide complementary information on motor function. This study investigates the feasibility of a markerless, single-camera two-dimensional video-based approach to extract quantitative gait parameters in people with multiple sclerosis and to examine whether these measures scale with overall disability level as indexed by scores on the Expanded Disability Status Scale. Twenty people with multiple sclerosis were recorded while walking at a comfortable pace using a standard video camera. Spatiotemporal parameters and lower-limb elevation angles (thigh, shank, and foot) were extracted using the position of the keypoints obtained with a pose estimation algorithm. Associations between these measures and disability scores were assessed. Results showed a significant reduction in normalized stride length and stride time with increasing disability level. In addition, the range of motion of shank and foot elevation angles exhibited strong associations with disability scores. These findings suggest that the range of motion of lower-limb elevation angles, particularly at distal segments, might provide sensitive indicators of disability-related gait impairment. The proposed two-dimensional video-based method offers a low-cost, non-invasive, and unconstrained tool for objective gait assessment in people with multiple sclerosis.
Leiomyomas remain highly prevalent by midlife, with persistent racial disparities in age at onset and tumor burden. Although they are traditionally expected to regress after menopause, longitudinal data demonstrate variable behavior, with many leiomyomas persisting, some regressing modestly, and a subset continuing to grow. This structured, nonsystematic review of the contemporary literature was conducted using PubMed and Embase, supplemented by guideline and reference review, with the objective of guiding the evaluation and management of uterine leiomyomas across the menopausal transition. In perimenopausal and postmenopausal patients, abnormal uterine bleeding and pelvic pain require systematic evaluation to exclude endometrial pathology and alternative causes. Transvaginal ultrasonography with structured assessment provides the foundation for evaluation, and magnetic resonance imaging serves as an adjunct when findings are indeterminate or clinical concern is elevated. No single imaging feature reliably distinguishes benign leiomyoma from leiomyosarcoma, and the prevalence of unsuspected leiomyosarcoma at surgery remains low, increasing with age. Management should be individualized and symptom driven, incorporating patient age, comorbidities, hormonal exposures, imaging findings, and preferences, with expectant management appropriate for asymptomatic patients and intervention reserved for refractory symptoms or diagnostic uncertainty.
Depression and anxiety are often undetected and untreated in individuals with multiple sclerosis (MS), despite a high prevalence of these conditions in this population. This study evaluated the clinical utility, diagnostic accuracy, and acceptability of depression and anxiety screeners and their integration into routine MS clinic appointments. Participants with MS (N = 207; age M = 47.3 ± 12.7 years, 77.3% female) were enrolled in this cross-sectional study conducted at an MS clinic in Melbourne, Australia. Consenting participants completed the Patient Reported Questionnaire-9 (PHQ-9) and the Depression, Anxiety and Stress Scale-21 (DASS-21) via an electronic tablet in the clinic waiting room or online for telehealth appointments, and underwent a Structured Clinical Interview for DSM-5 major depressive disorder and generalised anxiety disorder. Sensitivity, specificity, internal consistency, and patient comfort were assessed. The PHQ-9 and DASS-21 depression subscales demonstrated excellent internal consistency (α=0.90-.93) and good diagnostic performance. For anxiety, DASS subscales showed moderate validity, with the stress subscale outperforming the anxiety subscale. Participants rated screening as highly acceptable (mean comfort score=7.7/10). Among people with MS, self-administered depression and anxiety screening tools are valid and acceptable for routine use at MS clinic appointments. Tablet and online survey administration provide scalable options for integrating mental health assessment into standard care.
Falls are highly prevalent after stroke. Although trunk control is critical for dynamic stability during walking, the contribution of upper body (head, neck, thorax, spine) kinematics to fall risk discrimination and prediction remains unknown. This study examined whether trunk kinematics during walking differentiate people with stroke (PwS) with high versus low fall-risk and whether these measures improve fall risk classification. Fifty sub-acute PwS from a public 3D gait dataset were analysed. Participants walked barefoot at self-selected speed while full-body kinematics were recorded (Vicon®). Fall risk was classified with the Tinetti Performance-Oriented Mobility Assessment (≤18/28 = high risk). Sagittal and frontal upper body kinematics and thoraco-pelvic coordination (continuous relative phase, CRP) were computed. Group differences were tested using ANCOVA adjusted for age and walking speed. Linear discriminant analysis assessed fall risk prediction using walking speed alone and combined with trunk kinematics. Participants were classified as high (n = 22) and low (n = 28) fall risk. High-risk PwS showed significantly greater mean and maximal sagittal spinal flexion, greater maximal frontal head angle, and lower mean and maximal CRP indicating reduced thoraco-pelvic dissociation. Sagittal head and neck kinematics did not differ. Walking speed alone yielded an AUC of 0.83. Adding maximal sagittal spine angle improved AUC to 0.87 and specificity (0.75-0.88) without affecting sensitivity. High fall risk PwS showed greater trunk flexion, reduced thoraco-pelvic coordination, and greater frontal plane head angle during walking. Walking speed alone strongly predicted fall risk, while adding trunk kinematics improved specificity without affecting sensitivity.
Electroconvulsive therapy (ECT) is an effective treatment for severe and treatment-resistant bipolar depression, yet intracranial vascular malformations such as cerebral cavernous malformations raise safety concerns due to transient ECT-induced increases in blood pressure and intracranial pressure, potentially elevating the risk of hemorrhage. We report a case of a woman in her mid-60s with treatment-resistant bipolar depression and severe symptom burden (MADRS 45; HAMD-17 35; HAMD-21 39) in whom brain MRI incidentally revealed a 6×4 mm cavernous hemangioma in the left precentral gyrus without signs of prior hemorrhage. Following interdisciplinary consultation with neurosurgery and anesthesiology, an index course of ECT was initiated with careful hemodynamic monitoring and pharmacological blood pressure control. The patient received 16 sessions; initial bifrontal stimulation yielded insufficient seizure quality despite maximal stimulus intensity, prompting a switch to right unilateral electrode placement, which produced adequate seizures. Hypertensive peaks up to 210/150 mmHg were successfully managed with metoprolol and intravenous urapidil. Apart from transient anisocoria after the first session with unremarkable cranial CT findings, no neurological complications occurred. Depressive symptoms markedly improved by the end of treatment (MADRS 12; HAMD-17 15; HAMD-21 17). This case suggests that ECT can be performed safely and effectively in patients with cerebral cavernous hemangioma when multidisciplinary assessment and targeted hemodynamic management are implemented, and that the combination of metoprolol and urapidil may represent a useful strategy to mitigate blood pressure surges during ECT.
The COVID-19 pandemic caused significant disruption in health care services. Many essential services were delayed by health care facilities, including voluntary counseling and testing (VCT) services for people living with HIV. There were many reports of interruption in HIV testing, antiretroviral therapy (ART) initiation, and also ART access for people living with HIV during the pandemic. Patients were unable to attend follow-ups and acute care visits due to fear and anxiety. This situation also caused stress for people living with HIV. This study aimed to determine the level of COVID-19 knowledge, anxiety, and access to VCT services for people living with HIV. This cross-sectional, correlational study was conducted with 200 participants at 1 public hospital in Samarinda (n=140, 70%) and 1 public hospital in South Jakarta (n=60, 30%), Indonesia, from August 2022 to April 2023. Sampling was done using convenience methods and predefined inclusion criteria. Data collection included a demographic information form, COVID-19 knowledge questionnaire, the Coronavirus Anxiety Scale, and a questionnaire assessing access to health services. Both COVID-19 knowledge (odds ratio 11.246, 95% CI 11.246; P=.001) and anxiety (odds ratio 2.258, 95% CI 2.216; P=.03) had a positive and significant relationship with access to health services. A multivariate analysis showed that the most influential factor affecting access to VCT services was knowledge (P=.001; B=2.289). This study highlights the need for enhanced support and education for people living with HIV or AIDS regarding their knowledge of and anxiety related to COVID-19, particularly considering their vulnerabilities. To ensure compliance with health protocols in future pandemics, it is crucial to improve access to health care services. One key recommendation is to enhance the VCT service system, especially for people living with HIV or AIDS, during public health emergencies such as the COVID-19 pandemic. Additionally, services such as telemedicine and telehealth should be further developed to allow people living with HIV or AIDS to receive ART without the need for in-person hospital visits.
In Brazil, diagnosing and treating non-small cell lung cancer (NSCLC) with actionable molecular alterations pose substantial challenges because of health care disparities. Anaplastic lymphoma kinase (ALK) rearrangements represent a clinically relevant subset with highly effective targeted therapies. However, real-world access to ALK diagnostics and treatments across different Brazilian health care sectors remains inadequately characterized. We conducted a cross-sectional survey of Brazilian oncologists between October 2024 and March 2025 to assess the availability of ALK testing and targeted therapies, alongside perceived implementation barriers. Of 197 responses collected, 156 were included in the final analytic cohort. Data were analyzed using descriptive statistics, and categorical variables were reported as proportions with 95% CIs. Within the final analytic cohort (N = 156), 93.9% of the respondents practicing in the private sector (n = 147) reported access to ALK testing, whereas only 43.9% of those practicing in the public health care system (n = 107) had access. Access to ALK-targeted therapies was limited for the public health care population: 7.1% received crizotinib and <2% received newer-generation ALK-targeted therapies available in the first-line setting. By contrast, in the private sector, 75.6% and 60.9% reported access to alectinib and lorlatinib, respectively. Chemotherapy remained predominant in the public health care system. Main barriers included lack of reimbursement (58.3%), insufficient tissue (40.4%), and urgency to initiate treatment (36.5%). Despite robust evidence supporting ALK-targeted therapies, this study highlights substantial disparities in access to diagnostics and treatment for ALK-rearranged NSCLC in Brazil, particularly among patients reliant on the public health care system. Findings underscore the need for policies to strengthen testing infrastructure, ensure equitable access to guideline-recommended therapies, and enhance provider education. Addressing these gaps is essential for equitable precision oncology and improved outcomes.
Underdiagnosis and undercoding are common across mental health conditions, particularly suicide and self-harm. This leaves health care datasets lacking reliable negative examples needed for predictive modeling, phenotype prevalence estimation, and identification of individuals at elevated risk. We use positive and unlabeled (PU) learning to address this challenge. This study aims to identify US Veterans whose self-harm events were not explicitly captured through diagnostic codes in electronic health records (EHRs) and estimate the underlying prevalence using a novel PU learning algorithm. We performed a retrospective cohort study using Veterans Health Administration EHRs (from October 1, 1999, to August 31, 2019), selecting a random 25% sample of 1,329,120 Veterans out of 5,316,480 (1,193,563 males and 135,557 females) with at least 2 years of observation. The study cohort comprised 24,625 Veterans with coded self-harm and 1,304,495 uncoded, with the mean ages of 38.39 (SD 12.17) and 48.76 (SD 15.04) years, respectively. We applied our PULSNAR (positive unlabeled learning selected not at random) algorithm to estimate the proportion of individuals with uncoded self-harm. Covariates included age, medical conditions, procedures, and clinical observations. Four experts (raters) independently reviewed charts of 97 uncoded Veterans, each selected from 1% intervals of calibrated PULSNAR probabilities from 0.01 to 0.97. Agreement was assessed among raters, PULSNAR classifications, and consensus review decisions. Post hoc calibration was used to refine prevalence estimates. Of the 159,049 covariates in the dataset, PULSNAR's Extreme Gradient Boosting (XGBoost) model identified 1302 (0.82%) as informative for classification. Only 1.85% (24,625/1,329,120) of Veterans had diagnostic codes indicating self-harm events, while PULSNAR estimated an overall prevalence of 10.46% (139,026/1,329,120) by identifying an additional α=8.77% (114,404/1,304,495) of self-harm cases among the uncoded population. Of the 97 chart-reviewed patients, 39 had documented but uncoded self-harm. PULSNAR probabilities were post hoc calibrated such that their sum over the 97 cases equaled 39, which adjusted the combined coded and imputed prevalence downward from 10.46% to 7.91% (105,133/1,329,120). By applying this calibration to shift the probabilities of all uncoded Veterans, with bootstrapping for confidence intervals, PULSNAR estimates that coded self-harm represents only 23.4% (95% CI 17.76% to 31.51%) of all documented (coded+notes) self-harm. Under the "selected not at random" assumption, PULSNAR provides an innovative and scalable framework for estimating the clinically documented prevalence of mental health conditions and identifying the uncoded individuals with calibrated prediction, without requiring confirmed negative labels. This method offers an alternative to time-consuming chart reviews for detecting likely cases missing structured coding capture. By addressing diagnostic undercoding of mental health conditions in EHRs, this approach has the potential to enhance the estimation of mental health prevalence and support screening, activation of automated clinical decision support, targeted intervention, better resource allocation, and research to improve outcomes in real-world settings.
Most studies assessing digital interventions for people with heart failure (HF) focus on clinical outcomes, and few include patient perspectives. Understanding patient experiences of the use of a digital HF platform along with community health worker (CHW) care as part of a digitally enabled CHW intervention can inform management of HF at home and improve the postdischarge phase of care. This study aimed to identify patient perceptions related to the use of a digitally enabled CHW intervention. This qualitative study included interviews with adults (aged ≥18 years) with HF who were assigned to the intervention arm of a pilot randomized controlled trial from September 2022 to June 2023. For 30 days after hospital discharge, intervention participants were paired with a CHW and instructed to use a digital platform that tracked biometrics (eg, heart rate, oxygenation, blood pressure, body weight, steps taken, and symptoms) and offered educational videos. In-depth interviews were conducted after the 30-day intervention was complete (between 31 and 45 days after hospital discharge). Key interview domains included barriers and facilitators to the intervention, use of remote monitoring in HF, and the role of CHWs in HF home care. Interviews with participants (N=19; mean age 62.1, SD 15.1 years) yielded five key themes: (1) the combined intervention was well received, and CHWs made the use of the digital platform more approachable; (2) the digital platform enhanced HF knowledge and confidence in self-care; (3) digital platform use was easy to integrate into daily routines; (4) in addition to assisting with navigation of unmet social needs (eg, transportation, insurance benefits, and food access), CHWs provided emotional support and increased motivation for clinical care plan adherence and platform use; and (5) connectivity issues and other technical challenges occurred with digital platform use. The digital platform was easily integrated into patients' daily routines. CHWs played a key role in making the platform more approachable for participant use. Further research is needed to better understand the impact of this intervention in larger HF populations over more extended time intervals.
In South Asia, climate change caused rising urban heat, and rapid urbanisation has become a major public health concern, particularly for cardiovascular health. However, socioeconomic factors (including income, education, and housing quality) have not been comprehensively included in any of the previous studies. This study aimed to examine the association between prolonged heat extended exposure and the prevalence of cardiovascular disease. A cross-sectional study was conducted with 1300 adults aged ≥ 19 years, recruited from different urban centers. Data was collected from June to August 2025 using a structured questionnaire, assessing cardiovascular health indicators, heat exposure levels, socioeconomic status, and perceptions of public health policies. Multinomial logistic regression and Wilcoxon rank sum test were performed by using the SPSS-26 and R-Studio. The sample comprised 55.61% males and 44.38% females. Approximately 44.7% of participants reported cardiovascular condition. Prolonged heat exposure was significantly associated with increased cardiovascular risk (OR = 1.29, 95% CI: 1.09-1.79, p = 0.010). Heat related stress in urban areas was strongly associated with adverse outcomes (OR = 3.67, 95% CI: 2.23-6.80, p < 0.001). Awareness of heat related campaigns was associated with lower cardiovascular risk (OR = 0.54, 95% CI: 0.31-0.81, p = 0.01). Cardiovascular health is associated with urban heat exposure. Socioeconomic factors may modify this relationship. Improved education, healthcare system and targeted health policies are needed to reduce the impact.