IntroductionCannabis use during pregnancy and the postpartum period has increased in recent years, raising clinical concerns regarding maternal and infant health, particularly during lactation. However, evidence regarding Δ9-THC concentrations in breast milk during the early postpartum period and their relationship with other biological matrices remains limited.ObjectiveThis study aimed to assess Δ9-THC concentrations in breast milk and saliva, and 11-nor-9-carboxy-Δ9-tetrahydrocannabinol (THC-COOH) concentrations in urine, among postpartum women with cannabis use at the time of delivery. A secondary objective was to explore correlations between these biological matrices during early lactation.MethodsA longitudinal observational study was conducted at Vall d’Hebron University Hospital (Barcelona, Spain) between April 2022 and December 2023. Thirteen postpartum women aged over 18 years with a positive urine test for cannabis at delivery and intent to breastfeed were included. Saliva, urine, and breast milk samples were collected at 24 hours, 48 hours, and one week after birth. Δ9-THC concentrations in breast milk and saliva and THC-COOH concentrations in urine were analyzed using liquid chromatography–tandem mass spectrometry (LC-MS/MS).ResultsAmong participants who remained abstinent during the first postpartum week, urinary THC-COOH concentrations progressively decreased but remained quantifiable across all study stages. In contrast, Δ9-THC concentrations in breast milk decreased over time and were below the limit of quantification (LOQ) one week postpartum. Salivary Δ9-THC concentrations were generally low and frequently below the LOQ. Breast milk Δ9-THC concentrations at the first sampling stage were significantly correlated with salivary Δ9-THC and urinary THC-COOH concentrations, whereas no significant correlations were observed at later stages.ConclusionsThis preliminary study suggests that Δ9-THC concentrations in breast milk may decline rapidly after postpartum cannabis cessation, becoming non-quantifiable within the first postpartum week among participants who discontinued use after delivery. In contrast, urinary THC-COOH remained quantifiable for a longer period. Salivary Δ9-THC showed limited concordance with breast milk Δ9-THC and should therefore be interpreted cautiously as a potential surrogate marker. Larger prospective studies are needed to confirm these findings and to support evidence-based breastfeeding counseling for women with recent cannabis use.
Stigma and quality of life in hospitalized schizophrenia patient-family caregiver dyads in Northern China: an actor-partner interdependence model analysis
BackgroundSchizophrenia is a chronic and relapsing mental disorder that is consistently associated with a severely diminished quality of life (QoL) for patients. Existing research has predominantly focused on how the stigma experienced by patients with schizophrenia relates to their own QoL. However, stigma among family caregivers has received considerably less attention, and its potential association with patients’ QoL, in particular, remains underexplored. Therefore, this study aims to systematically analyze the dyadic associations of stigma—as experienced by both patients with schizophrenia and their family caregivers—with QoL, utilizing an actor-partner interdependence model (APIM). Through this framework, this study seeks to explore the interdependence of stigma between patients and their family caregivers and its correlational links to their quality of life.MethodsTwo hundred and sixty-four pairs of schizophrenic patients and their family caregivers were included, and the subjects’ stigma was measured using the Internalized Stigma of Mental Illness Scale and the Conjunctive Stigma Scale, respectively, and the quality of life was measured using the World Health Organization Quality of Life Measurement Short Form. The actor-partner effect of stigma on quality of life was explored by constructing an actor-partner reciprocity model.ResultsThe actor effect of stigma on quality of life was significant for people with schizophrenia and their family caregivers (β=-0.472, p < 0.001, β=-0.779, p < 0.001), and the partner effect of stigma on quality of life was significant for people with schizophrenia and their family caregivers (β=-0.128, p = 0.033, β=-0.419, p < 0.001).ConclusionIn future research and interventions aimed at improving the quality of life for people with schizophrenia and their caregivers, it is important to consider not only the individual’s own stigma, but also how the other person’s stigma is associated with one’s quality of life.
Vitamins A and D Help Improve Lung Function in People with Asthma
Having higher levels of vitamins A and D in the body could help improve lung function in adults with asthma, suggests research led by Brigham and Women’s Hospital in Boston.
The study, published in the journal Thorax, also showed that higher vitamin A levels could also benefit lung health in children with asthma.
“Vitamins A and D are key regulators of gene expression involved in lung development and immune function…Both vitamins have complex roles in asthma and lung function,” write lead author Michael McGeachie, PhD, assistant professor at Brigham and Women’s Hospital, and colleagues.
“Vitamin A deficiency is more common in people with asthma and is linked to airway hyperresponsiveness. Moderate intake in childhood may improve lung function and reduce asthma risk, whereas excessive intake may increase adult-onset asthma risk. Vitamin D deficiency is associated with poorer lung function, increased exacerbations and worse disease control, although results are variable due to diet, age and sun exposure.”
In this study, the researchers evaluated data from two asthma study cohorts—the Genetic Epidemiology of Asthma in Costa Rica Study including 1,165 children with asthma and the adult Omic Determinants of Longitudinal Lung Function in Asthma cohort including 1,041 adults with the condition.
The researchers looked at both circulating vitamin A and D in both groups and also looked at links with epigenetic regulation.
In children, there were no significant links between lung function and vitamin D levels. Higher vitamin A levels were linked to better breathing capacity though. For each step up in vitamin A levels, the amount of air the children could blow out in the first second was about 2.5 percentage points higher, and the total amount of air they could breathe out was about 7.6 percentage points higher.
In the adults, one step up in vitamin A levels was linked to a 4.7 unit increase in the amount of air they could blow out in the first second and a 3.4 unit increase in the total amount of air they could breathe out. The effect of vitamin D levels in this group was smaller, but statistically significant, with lung function improvements between 0.16-0.18 units per step up in vitamin D levels.
The researchers also looked at DNA methylation and at several epigenetic clocks that estimate a person’s biological age from methylation patterns in the adult cohort. They found that adults with higher vitamin A and vitamin D levels had fewer methylation tags at key control sites in the IRF5 gene than those with lower levels and showed changes in small regulatory micro RNAs that respond to vitamin levels. These epigenetic changes were linked to better lung function and slower biological aging.
The epigenetic tests suggest that vitamins influence lung function and aging partly by producing these epigenetic changes, rather than only through a direct effect, according to the researchers.
“While future studies to replicate these findings in independent populations are needed to confirm the generalizability and robustness of these observations,” write Sze Man Tse, MD, and Geneviève Mailhot, PhD, of the CHU Sainte-Justine Research Center, Montreal, and the University of Montreal in an accompanying editorial in the same journal, “subsequent interventional studies examining the impact of vitamin supplementation on biological ageing and on IRF5 function will be particularly relevant.”
The post Vitamins A and D Help Improve Lung Function in People with Asthma appeared first on Inside Precision Medicine.
Claude Science is Anthropic’s newest flagship product
At an event for pharmaceutical executives, biotech founders, and researchers on Tuesday, Anthropic announced Claude Science, a major new product intended to support scientific research in the same way that Claude Code supports software engineering. Like Claude Code, Claude Science can autonomously carry out meaningful work when given concise, high-level instructions, and it has access to tools that make it particularly useful for research in computational biology and drug development. Along with launching and previewing Claude Science, which is now available to all paid Claude subscribers, Anthropic also announced that it will be using the product to pursue some of its own research into drugs for rare, neglected diseases.
This is not Anthropic’s first foray into AI for science. In October, the company released plug-ins that help Claude make use of scientific software and databases under the heading “Claude for Life Sciences.” But unlike this earlier release, Claude Science is a full-featured, standalone product. Anthropic’s decision to elevate Claude Science to the same rank as Claude Code and Claude Cowork indicates that the company is taking AI’s scientific applications very seriously—or at least wants to give the impression that it is.
“It represents how important this is to our mission that this is right up there with Claude Code and Claude Cowork as the next really significant product that we’re releasing,” says Eric Kauderer-Abrams, Anthropic’s head of life sciences. “Our mission is to develop AI that serves humanity’s long-term well-being, and we believe that by far the greatest opportunity to do that is in the life sciences.”
For the past decade, one company—Google DeepMind—has been at the vanguard of AI for science. CEO Demis Hassabis and researcher John Jumper won the Nobel Prize in chemistry for their work on the company’s AlphaFold model, and DeepMind has also made major contributions to meteorology, materials science, and a variety of other disciplines. But in the past several months, the fast-advancing frontier of AI progress seems to have left DeepMind in the dust. When it comes to coding, which has become the most lucrative use case for LLMs, DeepMind is stuck playing catch-up.
Anthropic is well positioned to take up DeepMind’s scientific mantle. Like Hassabis, Anthropic CEO Dario Amodei is a PhD scientist—unlike OpenAI CEO Sam Altman, who’s a businessman through and through. Many scientists are already avid users of tools such as Claude Code. These days, a lot of scientific research involves some amount of coding, but not all scientists are expert software engineers, and so tools like Claude Code can make a huge difference for their productivity. And the company has recently earned a major scientific vote of confidence: Earlier this month, Jumper announced that he is leaving DeepMind for Anthropic.
Since agents powered by LLMs, including Anthropic’s Opus model series, became capable of useful, independent work in late 2025, scientists have been seeing just how much they can do. In a blog post published on Anthropic’s website, the Harvard physicist Matthew Schwartz estimated, on the basis of his work with Claude Code and other Anthropic tools, that the company’s Opus 4.5 model is about as capable of executing scientific projects as a second-year graduate student.
According to Kauderer-Abrams, Claude Science isn’t intended to displace Claude Code and Claude Cowork in scientists’ workflows. Instead, it’s designed to build on what scientists already find useful about Anthropic’s products. For instance, it not only writes code but also helps scientists run their code on powerful computer clusters, which many many scientists need for their work but can be difficult to manage. And it prioritizes reproducibility, so that scientists can trace back the source of any figure or result and check it for accuracy and validity.
Though Claude Science could in principle assist with any area of scientific research, it seems designed and marketed as a tool for molecular and cellular biology, and for drug development in particular. It can interface with various tools used in genetics, chemistry, and protein biology, all of which could come in handy for researchers on the hunt for new drugs. During the Tuesday event, Alexander Tarashansky, who led the development of Claude Science, demonstrated how the system could autonomously identify new drug candidates for phenylketonuria, a rare genetic disease.
And Anthropic isn’t leaving all that work to the pharma companies and university labs that were represented at the event. Armed with Claude Science, it will be pursuing its own research into drug candidates for neglected diseases—both to help move science forward and to gain a clearer sense of how Claude Science works in the real world.
There are obvious humanitarian reasons to prioritize drug development when creating a general-purpose scientific research tool, and AI industry leaders often cite curing disease as a major potential upside of the technology. But it’s also notable that pharmaceutical companies have far deeper pockets than academic researchers. Anthropic says it’s set to see its first profitable quarter, and if major new contracts with pharmaceutical companies are forthcoming, they could help ensure it stays profitable as the tokenmaxxing craze dies down—something that’s ever more important as an IPO approaches later this year.
Knowledge Graphs Based on Meta-Analysis Papers Improve the Quality of Case Formulation: Mixed Methods Design
Background: Case formulation (CF) is a core skill for therapists; however, creating high-quality CFs requires considerable time. Objective: This study aims to demonstrate that providing a knowledge graph based on meta-analytic literature can enhance CF quality. Methods: Five groups were established, including 4 large language model groups and 1 human expert group, each generating 25 CFs based on 25 vignettes. The control group with Claude (Sonnet 3.7; Anthropic) produced 25 CFs. The personalization group served as the control group with additional personalization prompts. The knowledge graph group used a large language model that generated 25 CFs, which was provided with a meta-analysis knowledge graph. Further incorporation of additional personalization prompts then comprised the knowledge graph with personalization group. Finally, the expert group consisted of 25 CFs generated by a human expert. These 125 CFs in total were evaluated for general quality (ie, correctness, completeness, feasibility, and consistency) using a 7-point scale and 18 essential elements with binary scores (0 or 1) by another human expert. The CFs were also qualitatively analyzed. Results: The knowledge graph and knowledge graph with personalization groups scored significantly higher than the control group in terms of correctness, completeness, and feasibility. The expert group scored significantly higher on consistency than the machine-generated groups. Additionally, there was no significant difference in the feasibility scores among the knowledge graph, knowledge graph with personalization, and expert groups. The qualitative evaluation suggested that human CFs narrow the text to content that is easy for the client to read, whereas machine CFs are more likely to include expressions that are unnatural to the client. Conclusions: These results indicate that providing knowledge graphs to novice therapists increases the correctness, completeness, and feasibility of CF. Providing experienced therapists with knowledge graphs is suggested to improve the quality of their CF and mental health services.
<img src="https://jmir-production.s3.us-east-2.amazonaws.com/thumbs/7d68afb473b485455c0e84aa0bb111d5" />
Stakeholder Experiences With the Pneumococcal Conjugate Vaccine Chatbot as a Complementary Capacity-Building Tool for Frontline Health Workers in India: Qualitative Study
<strong>Background:</strong> Pneumonia remains the leading cause of mortality in individuals aged 5 years or younger globally, with India bearing a disproportionately high burden. The introduction of the pneumococcal conjugate vaccine (PCV) in India necessitated innovative approaches to support frontline health workers (FLHWs), particularly in remote settings. To address this, a customizable WhatsApp-based PCV chatbot was developed as a complementary tool to traditional training and reference materials. <strong>Objective:</strong> This study aimed to document the opportunities, challenges, and mitigation measures encountered during the development and rollout of the PCV chatbot, and to explore its use and user experience as a capacity-building and support tool for FLHWs during new vaccine introduction. <strong>Methods:</strong> A qualitative study was conducted across 4 Indian states—Arunachal Pradesh, Delhi, Karnataka, and West Bengal—using purposive sampling at the district and block levels. Data collection involved key informant interviews with immunization officials and chatbot developers, and focus group discussions with auxiliary nurse midwives. A Likert scale–based tool captured quantitative feedback on user satisfaction. <strong>Results:</strong> Stakeholders appreciated the chatbot’s accessibility, familiarity (through WhatsApp), and multilingual functionality. Most auxiliary nurse midwives found it easy to use and rated responses highly for completeness and usefulness. The chatbot enabled immediate access to information, saving time and bridging gaps, especially when traditional training was delayed or unavailable in hard-to-reach areas. Challenges included occasional technical issues, limited content related to dropout and left-out scenarios, and difficulties in typing regional languages. Recommendations included implementing predictive text, expanding scenario coverage, and strengthening user-centered design and field testing. <strong>Conclusions:</strong> The PCV chatbot demonstrated acceptability and perceived value as an on-demand knowledge tool among FLHWs. Continuous user-driven refinement, expanded content, and enhanced usability are essential for its scalability and sustained use in vaccine introduction and capacity-building efforts.
Expedited Transition to Digital Delivery of Recovery Support Services Due to the COVID-19 Pandemic: Mixed Methods Needs Assessment
Background: Recovery support services (RSS) are an evidence-based approach to support recovery from substance use disorders, most often composed of peer-to-peer support, referrals to housing, job training, and other forms of prosocial engagement and activities. During the COVID-19 pandemic, RSS providers quickly converted in-person services to digital delivery to avoid disruption. It is unclear if this rapid conversion impacted the delivery of services or if this delivery model could enhance RSS reach and uptake more generally by extending the reach of RSS providers and offering an alternative delivery method and access point. Objective: The goal of this study was to identify how RSS providers in Texas adapted their services for digital delivery and to what extent, if at all, technology limitations (eg, lack of digital infrastructure) were present. Methods: We conducted an electronic survey of 85 RSS providers, assessing their current capacity and methods for the digital recovery support service (D-RSS), followed by semistructured online interviews with a subset of 20 respondents. Results: Most survey respondents (74/85, 87.1%) used D-RSS, though they used many dated technologies, devices, and platforms for service delivery. Many respondents indicated that they use Zoom (Zoom Video Communications) videoconferencing to communicate with participants; however, providers also indicated that they must use several different technology platforms to accomplish their service delivery goals. Four main themes emerged from the interviews: (1) the impact of the COVID-19 pandemic on RSS, (2) barriers and facilitators to technology-delivered D-RSS, (3) awareness and expectations regarding the use of D-RSS, and (4) training needs to deliver D-RSS. Conclusions: RSS organizations have access to technology for D-RSS; however, the technology is often outdated. Because the pandemic required a rapid and unexpected shift to D-RSS to maintain and potentially expand access during a public health emergency, providers desire guidance for training staff and participants on how to best use technology. A subset of providers endorsed the potential of a unified platform for D-RSS delivery, especially for data capture. Most barriers to D-RSS identified by our respondents may be addressable through the streamlined deployment of technology resources, rigorous training and onboarding programs in best practices for providers and participants, and tailored implementation strategies for varying local contexts.
<img src="https://jmir-production.s3.us-east-2.amazonaws.com/thumbs/3a2240fcd24e97753a368fb6d403d6fb" />
Governing Ethical Tensions in Youth Digital Mental Health Research
As mental health research increasingly aims to generate societal impact, researchers operate at the intersection of innovation and ethical responsibility. Drawing on experiences from the cocreated NEON Young Norway Study on youth recovery narratives, this viewpoint identifies four ethical tensions that arise from the existing governance frameworks in youth digital mental health research: (1) balancing safeguarding against harm with youth participation, (2) protecting privacy without undermining authentic storytelling, (3) governing unpredictable outcomes of cocreated research, and (4) meeting ethical and legal standards while ensuring youth-friendly communication. These tensions highlight limitations in mental health research that adopts participatory and digital approaches, as this often struggles to accommodate iterative designs, narrative data, and cross-sector collaboration. We argue that responsible youth mental health research requires ethics to be understood as a dynamic, participatory practice that supports safe and equitable inclusion, rather than having a focus on risk prevention. Ethical governance, therefore, needs to evolve toward proportionate, context-sensitive approaches that can enable innovation while protecting young people’s rights, agency, and voices.
<img src="https://jmir-production.s3.us-east-2.amazonaws.com/thumbs/123af813428fc7e274490540157cb83a" />
Impact of an mHealth App on Digital Transformation: Randomized Clinical Trial on Strengthening Digital Skills in Older Women
Background: The rapid growth of digital technologies has transformed daily activities, health management, and social interaction. Older adults, however, continue to face challenges in adopting and using these tools due to limited previous exposure, age-related sensory or cognitive decline, and low digital confidence. In Brazil, internet access among adults aged 60 years or older has increased, yet digital exclusion persists, worsening health disparities. Mobile health (mHealth) apps offer a potential strategy to promote digital inclusion, strengthen digital competencies, and support healthy aging. Nonetheless, studies show that culturally adapted, multidisciplinary interventions for this group remain scarce and are rarely assessed through both quantitative and qualitative methods. Objective: This study aimed to evaluate the impact of a lifestyle mHealth app on improving digital skills, as well as to analyze the level of satisfaction and usability of the app. Methods: In this mixed methods study, a 14-week randomized clinical trial was conducted in Ribeirão Preto, São Paulo, Brazil. A total of 40 older adult women were randomized into an intervention group (n=21), who used the mobile app, and a control group (n=19). Digital competencies were measured before and after the intervention using a semistructured questionnaire based on the (MCDMSênior; Digital Competency Model for M-learning with a focus on older adults) framework, covering 6 domains—basic technology use, internet navigation, mobile app use, online research, digital communication, and usage of digital resources. Additionally, satisfaction with the educational content was evaluated using the suitability assessment of materials, and system usability was assessed using the System Usability Scale. Qualitative data were collected through semistructured, in-person interviews conducted immediately after the intervention with all intervention participants. Interviews explored perceptions of the app’s usability, satisfaction with its content, barriers, and facilitators to engagement, and perceived changes in digital skills. All interviews were audio-recorded, transcribed, and analyzed thematically by 2 independent researchers using an inductive coding approach. Results: Postintervention analyses revealed significant differences in specific digital competencies. The intervention group demonstrated a moderate improvement in internet navigation skills, while gains in basic technology use and digital communication were minimal. Conversely, the control group exhibited moderate improvement in basic technology skills and lower effects in online research and digital communication. Overall, satisfaction with the educational content was low, and usability was rated as average. Qualitative findings indicated that, although participants valued the clarity of navigation and cultural relevance, persistent age-related fears and insecurities in using digital technologies were reported. Participants highlighted the need for more personalized guidance, ongoing motivational support, and technical adjustments to improve usability and engagement. Conclusions: mHealth apps can effectively enhance certain digital competencies in older women, particularly internet navigation, but improvements in content suitability and usability are needed. Refinements in design and tailored support are essential to overcome age-related barriers and foster digital inclusion. Trial Registration: Brazilian Registry of Clinical Trials RBR-6wgkzs8; https://ensaiosclinicos.gov.br/rg/RBR-6wgkzs8
<img src="https://jmir-production.s3.us-east-2.amazonaws.com/thumbs/0c5daf33accff9728a4940d3c44aa93d" />
Emotion Classification in Japanese Cancer Survivor Interview Narratives Using Sentiment Polarity and Plutchik Emotion Frameworks: Model Development and Evaluation Study
Background: Cancer survivors often experience complex and coexisting emotions throughout diagnosis, treatment, and posttreatment life. Emotion classification of patient narratives may help in understanding survivorship experiences; however, evidence remains limited for multidimensional classification using cancer survivor interview narratives. Objective: This study aimed to develop and evaluate natural language processing–based emotion classification models using Japanese cancer survivor interview narratives and to examine whether polarity and multidimensional emotion labels provide complementary perspectives. Methods: We analyzed verbatim transcripts from 15 cancer survivor interviews published by the Cancer Note, Nonprofit Organization. Survivor utterances were extracted, noninformative conversational elements were removed, texts were segmented at Japanese punctuation marks, and 5 consecutive sentences were grouped into 1 chunk. Two annotators labeled 1998 text chunks with 3-class sentiment polarity labels (positive, neutral, or negative) and multilabel Plutchik 8-emotion labels (joy, trust, fear, surprise, sadness, disgust, anger, and anticipation). Japanese BERT (Bidirectional Encoder Representations from Transformers) and LUKE (Language Understanding with Knowledge-based Embeddings) were fine-tuned to build a multiclass polarity classifier and a multilabel 8-emotion classifier. Performance was evaluated using precision, recall, -score, macroaveraged metrics, Micro- for polarity, and Hamming loss for multilabel classification. For comparison, the same architectures were fine-tuned on WRIME (writers’ and readers’ intensities of emotion for their estimation), a Japanese social media emotion dataset, and evaluated on Cancer Note texts as a domain-transfer analysis. The 95% CIs were estimated using bootstrap resampling with 1000 iterations. Results: Neutral was the most frequent polarity label, trust was the most frequent 8-emotion label, and anger was the least frequent emotion label. Label distributions were imbalanced, with most-to-least frequency ratios of 3.47 for polarity and 8.10 for 8-emotion labels. In the 3-class sentiment polarity task, interview-trained models outperformed WRIME-trained transfer models. Interview Text-BERT achieved the highest micro- of 0.696 (95% CI 0.676‐0.716), whereas Interview Text-LUKE achieved the highest macro- of 0.660 (95% CI 0.639‐0.682). In the 8-emotion multilabel task, Interview Text-LUKE achieved the highest macro- of 0.427 (95% CI 0.398‐0.453) and the lowest Hamming loss of 0.078 (95% CI 0.073‐0.082). WRIME-trained transfer models showed lower performance, particularly in the 8-emotion task. Sadness and trust co-occurred most frequently, suggesting that positive and negative emotional elements may coexist in the same narratives. Conclusions: This exploratory study suggests the feasibility of domain-specific emotion classification for Japanese cancer survivor interview narratives. Models fine-tuned on target-domain narratives generally outperformed WRIME-trained transfer models, although the best architecture differed by task and metric. Polarity labels and Plutchik 8-emotion labels provided complementary perspectives on complex and coexisting emotions in survivorship narratives. However, performance for rare emotions remained limited, and the models should be regarded as preliminary research tools rather than clinically actionable systems. Larger, more diverse, prospectively or externally validated datasets, imbalance-aware methods, and user-centered evaluation are needed before clinical translation.
<img src="https://jmir-production.s3.us-east-2.amazonaws.com/thumbs/c55302eeaf28f695d0c405aa15e4c529" />