Nevin Manimala

Psychiatry Res. 2025 Dec 11;356:116897. doi: 10.1016/j.psychres.2025.116897. Online ahead of print.

ABSTRACT

BACKGROUND: Poetry therapy whether using reading, writing, or discussing poems in a therapeutic context, is increasingly applied in mental and physical health care, yet its empirical support remains unclear. This systematic review and meta-analysis examined the effectiveness of poetry-based interventions across psychiatric and somatic outcomes.

METHODS: PubMed and Google Scholar were searched up to November 2023 for original studies evaluating poetry-based interventions on mental or physical health outcomes. Studies in English or French using individual or group poetry activities were eligible. Fifteen studies (randomized controlled, case-control, and pre-post designs) met inclusion criteria; those scoring ≥6 on the Newcastle-Ottawa Scale were included in the meta-analysis. Random-effects models were used to pool standardized mean differences for post-traumatic stress disorder (PTSD), major depressive disorder (MDD), anxiety, resilience, stress, and perceived pain. Heterogeneity, prediction intervals, and publication bias (Egger’s test) were assessed.

RESULTS: Poetry-based interventions were associated with large reductions in PTSD symptoms and significant improvements in depressive symptoms, anxiety, and stress, with effect sizes generally in the moderate-to-large range. In contrast, effects on resilience were statistically non-significant and highly imprecise, and no reliable benefit was found for perceived pain, where heterogeneity and evidence of small-study effects were substantial. Across outcomes, most trials were small, at risk of bias, and methodologically heterogeneous.

CONCLUSIONS: Poetry therapy shows promising benefits for trauma-related, depressive, anxiety, and stress outcomes, but the evidence base is limited by small samples, variable quality, and potential publication bias. High-quality, preregistered randomized controlled trials are needed before poetry-based interventions can be firmly recommended beyond an adjunctive role in routine psychiatric care.

PMID:41411711 | DOI:10.1016/j.psychres.2025.116897

JMIR Aging. 2025 Dec 18;8:e76489. doi: 10.2196/76489.

ABSTRACT

BACKGROUND: The health and economic burden of dementia has led the World Health Organization to recognize it as a public health priority. Although there currently does not exist a cure for dementia, there are multiple interventions aimed at preventing the risk of dementia and improving the quality of life of people with dementia. Voice assistants (VAs), particularly those using large language models (LLMs), have emerged as promising tools to deliver these interventions to older adults due to their accessible and natural interface.

OBJECTIVE: This pilot study aimed to evaluate the technical feasibility (ie, functional performance and usability) and user acceptance of the embodied rule-based and LLM VA GRACE, as well as the perceived strength of the collaborative relationship or working alliance, between GRACE and healthy older adults during the delivery of cognitive stimulation interventions.

METHODS: A pilot study was conducted with 21 healthy German-speaking adults aged 60 years and older. Participants interacted with GRACE in a laboratory setting for 10-15 minutes. The interaction involved a structured cognitive stimulation session using rule-based and LLM components. Data were collected using pre- and postinteraction questionnaires and semistructured interviews. Quantitative analysis included descriptive statistics and Wilcoxon signed rank tests. Qualitative data were analyzed thematically.

RESULTS: Participants rated GRACE positively, with statistically significant scores above neutral (P<.001 for perceived ease of use, usefulness, enjoyment, and working alliance; P=.009 for perceived control; and P=.009 for intention to continue interacting). Thematic analysis revealed that GRACE was perceived as easy to understand and unambiguous, friendly, and supportive, with intervention components viewed as enjoyable and appropriately challenging. Areas for improvement included personalization, response delays, and voice quality.

CONCLUSIONS: The results suggest that embodied rule-based and LLM VAs like GRACE are feasible and well-received tools for delivering cognitive interventions to older adults. Future iterations will incorporate feedback and extend testing to individuals at risk for dementia.

PMID:41411649 | DOI:10.2196/76489

JMIR Form Res. 2025 Dec 18;9:e77357. doi: 10.2196/77357.

ABSTRACT

BACKGROUND: Despite the transformative potential of artificial intelligence (AI)-based chatbots in medicine, their implementation is hindered by data privacy and security concerns. DeepSeek offers a conceivable solution through its capability for local offline operations. However, as of 2025, it remains unclear whether DeepSeek can achieve an accuracy comparable to that of conventional, cloud-based AI chatbots.

OBJECTIVE: This study aims to evaluate whether DeepSeek, an AI-based chatbot capable of offline operation, achieves answer accuracy on medical multiple-choice questions (MCQs) comparable to that of leading chatbots (ie, ChatGPT and Gemini) on German medical MCQs, thereby assessing its potential as a privacy-preserving alternative for clinical use.

METHODS: A total of 200 interdisciplinary MCQs from the German Progress Test Medicine were administered to ChatGPT (GPT-o3-mini), DeepSeek (DeepSeek-R1), and Gemini (Gemini 2.0 Flash). Accuracy was defined as the proportion of correctly solved questions. Overall differences among the 3 models were tested with the Cochran Q test, while pairwise comparisons were conducted using the McNemar test. Subgroup analyses were performed by medical domain (Fisher exact test) and question length (Wilcoxon rank-sum test). An a priori power analysis indicated a minimum sample size of 195 questions.

RESULTS: All 3 chatbots surpassed the conventional passing threshold of 60%, with accuracies of 96% (192/200) for DeepSeek, 94% (188/200) for Gemini, and 92.5% (185/200) for ChatGPT. The overall difference among models was not statistically significant (P=.10) nor were pairwise comparisons. However, incorrect responses were significantly associated with longer question length for DeepSeek (P=.049) and ChatGPT (P=.04) but not for Gemini. No significant differences in performance were observed across clinical versus preclinical domains or medical specialties (all P>.05).

CONCLUSIONS: Overall, DeepSeek demonstrates outstanding performance on German medical MCQs comparable to the widely used chatbots ChatGPT and Gemini. Similar to ChatGPT, DeepSeek’s performance declined with increasing question length, highlighting verbosity as a persistent challenge for large language models. While DeepSeek’s offline capability and lower operational costs are advantageous, its safe and reliable application in clinical contexts requires further investigation.

PMID:41411646 | DOI:10.2196/77357

JMIR Mhealth Uhealth. 2025 Dec 18;13:e75445. doi: 10.2196/75445.

ABSTRACT

BACKGROUND: Patients undergoing gastrectomy usually experience postgastrectomy syndrome and face difficulties adapting to a regular diet. Human health coaching via a mobile app has recently been applied to patients with chronic metabolic diseases, with significant improvements being observed in clinical outcomes.

OBJECTIVE: This study aimed to compare the quality of life and nutritional outcomes of human health coaching via a mobile app with those of conventional face-to-face counseling in postgastrectomy patients with gastric cancer.

METHODS: This was a prospective randomized controlled trial, and patients were enrolled between May 2020 and August 2022. The mobile coaching group received health coaching that provides personalized advice based on self-recorded health data via a mobile app from assigned coaches for 3 months after discharge, and the conventional counseling group underwent dietary consultations with a clinical dietitian 1 and 3 months postoperatively. The primary end point for sample size calculation was the eating restriction score on the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire gastric cancer module 1 month postoperatively. Secondary end points included changes in other subscales of the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 and gastric cancer module, as well as nutritional outcomes assessed preoperatively and 1, 3, 6, and 12 months postoperatively.

RESULTS: Data from 88.9% (160/180) of enrolled patients were analyzed after excluding dropouts. In the mobile coaching group (n=76), 66% (n=50) of patients who used the mobile app for ≥8 weeks were classified as active users. No significant difference in eating restriction 1 month postoperatively was found between the mobile coaching and conventional counseling groups. However, the mobile coaching group reported less dyspnea during the entire period (P=.01), less eating restriction at 6 months (P=.045), and less negative body image 3 months postoperatively (P=.04) than the conventional counseling group (n=84). Exploratory subgroup analyses based on age, sex, and operation type indicated that younger patients (<60 years), female patients, and those who underwent distal gastrectomy had better quality of life from mobile coaching. In the mobile coaching group, exploratory subgroup analyses based on mobile activity showed that active users had a better global health status than inactive users (P=.005). However, no significant differences in body composition or nutritional parameters were observed between the mobile coaching and conventional counseling groups or between active and inactive users in the mobile coaching group.

CONCLUSIONS: Although this trial did not show a significant difference in eating restriction 1 month postoperatively, human coaching via a mobile app was associated with fewer symptoms in some scales compared to conventional counseling in postgastrectomy patients with gastric cancer. The intervention might help patients manage their symptoms and adapt to their diet.

TRIAL REGISTRATION: ClinicalTrials.gov NCT04394585; https://clinicaltrials.gov/study/NCT04394585.

PMID:41411643 | DOI:10.2196/75445

JMIR Mhealth Uhealth. 2025 Dec 18;13:e69244. doi: 10.2196/69244.

ABSTRACT

BACKGROUND: Digital microinterventions have strong potential to improve the lives of adults diagnosed with cancer. However, little is known about which types of digital microinterventions are most desired and how contextual factors may influence those preferences. This potentially limits guidance for personalized and timely digital microintervention delivery.

OBJECTIVE: This study aims to identify time-varying and person-level predictors of relative digital microintervention interest among adult survivors of cancer.

METHODS: We enrolled US adults within 5 years of a cancer diagnosis in a 5-week observational study using ecological momentary assessment. Participants (N=407) were asked 3 times a day for 5 weeks which of 9 brief, mobile-delivered interventions, if any, they would have been interested in completing within the past hour. Intervention options were (1) reducing worry, (2) reducing negative thoughts, (3) problem solving, (4) increasing positive emotions, (5) connecting with values, (6) guided relaxation, (7) getting support from others, (8) setting goals, and (9) something else. Multinomial models were used to identify demographic (ie, age), cancer-related (ie, treatment status), and psychological (ie, depression symptom severity, anxiety symptom severity, history of major depressive diagnosis, history of anxiety disorder diagnosis, and psychotherapy status) predictors of individual differences in modal intervention preference. Multilevel logistic and multilevel multinomial models were used to identify momentary negative affect, positive affect, and pain predictors of relative intervention interest.

RESULTS: Participants indicated interest in completing a digital microintervention in 87% (20,429/23,472) of completed surveys. The most frequently selected intervention option was guided relaxation (8611/20,429, 42%). Neither cancer treatment status (χ²₉=6.5; P=.69) nor psychotherapy status (χ²₉=14.0; P=.12) differentiated between modal intervention preferences. Participants with greater anxiety (χ²₉=35.1; P<.001) and depression symptom severity (χ²₉=23.0; P=.006) were less likely to modally endorse guided relaxation compared to other intervention options like increasing positive emotions, reducing negative thoughts, and getting support from others. Higher momentary negative affect and pain and lower momentary positive affect each predicted a greater likelihood to endorse interest in completing an intervention (vs not completing an intervention; P<.05) and to endorse interest in completing multiple interventions (vs only one; P<.001). Finally, higher momentary negative affect generally predicted greater interest in completing an intervention other than guided relaxation, whereas higher momentary pain generally predicted greater relative interest in guided relaxation.

CONCLUSIONS: Adult survivors of cancer differ in their digital microintervention preferences between and within persons. Guided relaxation alone is less appealing to survivors of cancer when they are in greater emotional distress but may be more appealing in response to instances of increased pain.

PMID:41411642 | DOI:10.2196/69244

JCO Precis Oncol. 2025 Oct;9:e2500076. doi: 10.1200/PO-25-00076. Epub 2025 Dec 18.

ABSTRACT

PURPOSE: To investigate tumor mutation variations across different racial/ethnic groups to better understand implications for targeted cancer therapies.

METHODS: A retrospective analysis of 5,045 patients at University of New Mexico Comprehensive Cancer Center who underwent tumor genetic testing between January 2015 and April 2022 was conducted. Data were standardized from internal genetic tests, FoundationOne, and Guardant next-generation sequencing panels. Chi-square tests, one-way analysis of variance, and negative binomial regression estimated differences in mutation rates across race/ethnicity, adjusting for cancer type, age, testing year, and number of genes screened. Primary outcomes included tumor mutation rates and their variation across racial/ethnic groups. Specific focus was placed on mutation frequencies in common genes, and association between race/ethnicity and mutations detected, adjusted for covariates.

RESULTS: Among 5,045 patients-Hispanic/Latino (30%), American Indian (5.7%), Asian/Hawaiian Native (1.9%), Black (1.5%), non-Hispanic White (41%), and other/unknown (19.7%)-mutations were identified most commonly for Asian/Hawaiian Native individuals, with a rate of 0.068 mutations per gene screened (95% CI, 0.051 to 0.090), followed by White individuals (rate = 0.061, 95% CI, 0.051 to 0.072). Fewest mutations were identified for Black individuals, with a rate of 0.045 mutations per gene screened (95% CI, 0.033 to 0.061). Single-gene comparisons suggested BRAF mutations to be most prevalent in non-Hispanic Whites (5.8%, P = .015) while EGFR mutations were most common in Asian/Hawaiian Native patients (10.53%, P = .005).

CONCLUSION: This study highlights substantial heterogeneity in tumor mutations across racial/ethnic groups while emphasizing the need for wider understanding of genomics and tailored approaches in cancer treatment. Findings underscore the need for equitable genomic testing, tailored therapies, and inclusive cancer care. Further research is necessary to bridge existing disparities, ensuring comprehensive, personalized cancer treatment for all patients.

PMID:41411616 | DOI:10.1200/PO-25-00076

JCO Precis Oncol. 2025 Oct;9:e2500638. doi: 10.1200/PO-25-00638. Epub 2025 Dec 18.

ABSTRACT

PURPOSE: Metabolic dysfunction-associated steatotic liver disease (MASLD) is a major cause of hepatocellular carcinoma (HCC), yet current screening strategies overlook the genetic complexity of MASLD. We hypothesized that capturing this complexity through a multitrait polygenic approach could improve HCC prevention.

METHODS: Using genome-wide association data for 10 MASLD-related traits in individuals of European ancestry, we constructed a meta-polygenic risk score (metaPRS) in the UK Biobank. We evaluated its performance in HCC prediction and its utility in stratified screening. Risk advancement period (RAP) analysis estimated how much earlier individuals in different genetic risk groups reach comparable risk levels.

RESULTS: The metaPRS that incorporated genome-wide variants achieved a C-statistic of 0.686 for HCC prediction, outperforming existing PRSs. Individuals in the top 20% of genetic risk had a 5.33-fold higher HCC risk than those in the bottom 20%. RAP analysis showed that high-risk individuals reached the HCC risk threshold 11.91 years earlier than the intermediate group, whereas low-risk individuals reached it 5.49 years later, suggesting a shift in recommended screening age from 65 to 43 years. Genetic stratification by the metaPRS also improved the predictive performance of noninvasive fibrosis scores (eg, Forns score). Combining high genetic risk with an elevated Forns score yielded a 10-year HCC risk of 2.68%, compared with 0.01% in the lowest-risk group-reducing the number needed to screen from 7,918 to 27.

CONCLUSION: The MASLD-related metaPRS supports effective population risk stratification and enables a layered HCC screening strategy combining genetic risk profiling with targeted clinical assessment.

PMID:41411614 | DOI:10.1200/PO-25-00638

JCO Precis Oncol. 2025 Oct;9:e2500235. doi: 10.1200/PO-25-00235. Epub 2025 Dec 18.

ABSTRACT

PURPOSE: About 60% of patients with high-risk neuroblastoma relapse. Specific mRNA detection in bone marrow (BM) by reverse transcriptase quantitative PCR (RT-qPCR) is associated with survival outcomes. Peripheral blood (PB) sampling is less invasive. Therefore, we prospectively validated an RT-qPCR panel of neuroblastoma mRNA in PB of patients with high-risk neuroblastoma, treated in NB2004-HR (GPOH) and NBL2009 (DCOG).

METHODS: From 312 patients, 634 PB samples were prospectively collected (2009-2017) at diagnosis, after two cycles and end-of-induction therapy. RT-qPCR was performed using our panel of neuroblastoma mRNA markers: PHOX2B, TH, DDC, CHRNA3, and DBH. Results were compared with paired BM samples. The association between neuroblastoma detection and event-free survival (EFS) and overall survival (OS) was estimated using Kaplan-Meier’s methodology and multivariable Cox regression model.

RESULTS: Clear correlation between calculated infiltration by neuroblastoma mRNA expression in PB and BM was seen at diagnosis (r_s 0.70 [95% CI, 0.62 to 0.76]), and heterogeneity was seen after two cycles (0.37 [95% CI, 0.12 to 0.58]) and end of induction (0.61 [95% CI, 0.10 to 0.87]). mRNA expression was significantly lower in PB compared with BM. At diagnosis, PB infiltration ≥1% was a prognostic factor for survival: adjusted hazard ratio (HR) was 2.37 [95% CI, 1.56 to 3.60] for EFS and 2.60 [1.65-4.08] for OS. At the end of induction, PHOX2B positivity in PB samples (n = 11) was associated with poor outcomes: HR 3.06 [1.51-6.20] for EFS and 2.88 [1.36-6.11] for OS. In patients with ≥10% BM infiltration at diagnosis, detection of the mRNA panel in PB samples could significantly distinguish between survival groups; the adjusted HR of PB infiltration ≥1% was 2.09 [1.01-4.30] for OS.

CONCLUSION: PB infiltration is associated with EFS and OS at diagnosis; it is also significantly associated with survival outcomes of patients with ≥10% BM infiltration at diagnosis. During follow-up, neuroblastoma mRNA detection in PB can be of added value, when BM analysis is not possible.

PMID:41411610 | DOI:10.1200/PO-25-00235

Neurology. 2026 Jan 27;106(2):e214488. doi: 10.1212/WNL.0000000000214488. Epub 2025 Dec 18.

ABSTRACT

OBJECTIVES: Stroke disproportionately affects Black individuals in the United States. We aimed to assess differences between Black and White individuals in the associations between health-related behaviors, measured with the Brain Care Score (BCS; a tool encompassing 12 modifiable risk factors), and incident stroke.

METHODS: We analyzed data from REGARDS, a prospective US cohort of Black and White adults aged 45 years or older. Participants who were stroke free at baseline with complete BCS data were included. We assessed the BCS (range: 0-21; higher indicating healthier behaviors) and its associations with incident stroke in Black vs White individuals. Cox proportional hazard models were stratified by race (Black vs White) and adjusted for demographics and socioeconomic factors. Effect sizes were compared using Z-statistics.

RESULTS: Among 10,861 participants (30.6% Black, 57.4% female, mean age: 63.2 years), 696 strokes occurred over a median of 15.9 years. A five-point higher BCS was associated with lower stroke risk in both groups, with larger magnitude among Black vs White individuals (HR: 0.47 [95% CI 0.36-0.61] vs 0.75 [95% CI 0.62-0.92]; Z-statistic p value = 0.0045).

DISCUSSION: The BCS is associated with incident stroke in a biracial US cohort and shows larger effect sizes within Black compared with White individuals, suggesting that BCS improvement may yield greater stroke prevention benefits for Black populations.

PMID:41411605 | DOI:10.1212/WNL.0000000000214488

J Opt Soc Am A Opt Image Sci Vis. 2025 Dec 1;42(12):1929-1935. doi: 10.1364/JOSAA.574392.

ABSTRACT

Partially coherent vortex beams have attracted growing interest due to their enhanced robustness and unique propagation characteristics in complex media. In this work, we experimentally investigate the behavior of partially coherent fractional vortex beams as they propagate through atmospheric turbulence. The beams are generated using a phase-only spatial light modulator and a rotating ground-glass disk modeled by the Gaussian Schell framework, and their degree of partial coherence is quantitatively characterized using a Young’s double-slit interference plate. After transmission through a 1.2 m turbulence simulator, the effective beam radius exhibits a smoothed, quasi-linear growth trend between successive integer topological charges, indicating the suppression of discrete modal transitions by the combined effects of partial coherence and turbulence. The scintillation index decreases overall with increasing topological charge, while local enhancements near half-integer orders reveal the heightened turbulence sensitivity of modal interference. Moreover, partially obstructed PCFVBs show partial statistical self-reconstruction after turbulent propagation, whereas a fully coherent control under identical conditions shows no appreciable recovery, ruling out a purely diffractive fill-in. These results provide the first, to our knowledge, comprehensive experimental insight into the interplay among coherence, turbulence, and fractional vortex structure, offering new perspectives for designing turbulence-resistant structured-light systems.

PMID:41411568 | DOI:10.1364/JOSAA.574392