Category: Statistics

J Immunother Cancer. 2026 Aug 26;14(8):e013527. doi: 10.1136/jitc-2025-013527.

ABSTRACT

BACKGROUND: Patients with unresectable stage III non-small cell lung cancer (NSCLC) have an unmet need for new therapies that improve survival. This phase III trial investigated the safety and efficacy of concurrent ociperlimab and tislelizumab plus concurrent chemoradiotherapy (cCRT) in treatment-naïve patients with stage III NSCLC.

METHODS: In this phase III, multicenter, randomized, multiarm, open-label trial, patients with unresectable stage III NSCLC received concurrent ociperlimab plus tislelizumab and cCRT, followed by ociperlimab plus tislelizumab (arm A), tislelizumab and cCRT, followed by tislelizumab (arm B), or cCRT, followed by durvalumab (arm C) (NCT04866017). Objectives were to compare progression-free survival (PFS), overall survival (OS), objective response rate (ORR), duration of response, disease control rate, clinical benefit rate, time to death or distant metastasis (TTDM), and safety and tolerability for arms A versus C, B versus C, and A versus B.

RESULTS: 63 patients were randomized to arms A (N=22), B (N=19), and C (N=22) prior to early trial termination. In A, B, and C, respectively, 95.5% (21/22), 84.2% (16/19), and 95.5% (21/22) were current or former smokers, and 68.2% (15/22), 73.7% (14/19), and 68.2% (15/22) had PD-L1 expression in tumor cells of ≥1%. Median PFS (95% CI) was not reached (NR) (6.3-not estimable (NE)) in A, 15.0 months (7.4 to NE) in B, and 10.4 months (5.7 to NE) in C. Median OS and TTDM were NR in any arm. ORR (95% CI) was 68.2% (45.1%-86.1%) in A, 68.4% (43.4%-87.4%) in B, and 59.1% (36.4%-79.3%) in C; all responses were partial responses. Treatment-emergent adverse events (TEAEs) occurred in all patients; in arms A, B, and C, respectively, pneumonitis occurred in 18.2% (4/22), 5.6% (1/18), and 9.1% (2/22) of patients, and interstitial lung disease occurred in 13.6% (3/22), 11.1% (2/18), and 0% of patients, of which the majority of events for each were grade 1/2. Grade ≥3 treatment-related TEAEs occurred in 68.2% (15/22), 66.7% (12/18), and 68.2% (15/22) of patients in arms A, B, and C, respectively.

CONCLUSIONS: There was a trend toward improved efficacy when adding tislelizumab with or without ociperlimab to cCRT followed by tislelizumab with or without ociperlimab compared with cCRT followed by durvalumab; however, efficacy data were for descriptive purposes only. No unexpected or new safety signals were identified.

PMID:42648750 | DOI:10.1136/jitc-2025-013527

BMJ Open Qual. 2026 Aug 26;15(3):e004387. doi: 10.1136/bmjoq-2026-004387.

ABSTRACT

BACKGROUND: Transition from paediatric to adult healthcare is particularly challenging for youth and their caregivers with complex health conditions, in part due to loss of paediatric supports and fragmentation of care, contributing to adverse outcomes. To date, transition research and policy has been largely driven by paediatric providers. Adult provider insights are critical to inform gaps in care, transition planning and enable innovative and sustainable interventions.

OBJECTIVE: To describe adult subspecialist perspectives on barriers, facilitators and opportunities to improve the transition from paediatric to adult care for youth with complex health conditions.

METHODS: An adult subspecialty advisory panel was convened as part of a quality improvement initiative, with physicians from the University of Toronto’s Department of Medicine who participated in three virtual meetings through nomination and stakeholder mapping. The first meeting focused on identifying current challenges and opportunities in transition to adult care; the second generated recommendations based on experiential knowledge and the third reviewed case studies to identify gaps and potential solutions. Discussions were recorded, transcribed and analysed using thematic qualitative analysis.

RESULTS: Adult subspecialists identified five key areas for improvement: (1) preparing patients and families for the realities of adult care; (2) bridging communication gaps between paediatric and adult systems; (3) moving beyond age-based transition toward flexible, readiness-informed care; (4) strengthening collaboration among adult providers caring for young adults; and (5) addressing structural and infrastructure barriers within adult care systems.

CONCLUSION: Adult care physicians expressed and validated that young adults have unique care needs during their transition to adult care. They highlighted the need for stronger cross-system collaboration, resources and investment to ensure effective transition. These findings highlight barriers in implementing current paediatric-led transition standards and provide a foundation for system-level, disease-agnostic approaches to improve continuity of care across the paediatric-adult divide.

PMID:42648749 | DOI:10.1136/bmjoq-2026-004387

BMJ Open Qual. 2026 Aug 26;15(3):e004321. doi: 10.1136/bmjoq-2026-004321.

ABSTRACT

Placental histopathology is central to good perinatal care, offering insight into antepartum and intrapartum events, informing neonatal management and supporting counselling for future pregnancies. Despite clear national guidance, many placentas that meet criteria are not sent for evaluation. Reasons are multifactorial, including inconsistent awareness of indications and unclear post-delivery processes. Two UK maternity units, St Thomas’ Hospital in London and the Royal Berkshire Hospital in Reading, identified local variations in practice that led to placentas not being appropriately sent. At St Thomas’ Hospital, a baseline audit performed as part of a preterm birth research study revealed only 80% of eligible deliveries had placental examination. At the Royal Berkshire Hospital, a coroner’s inquest after an unexpected neonatal death, highlighted the absence of indicated placental histology, prompting a recommendation for 48-hour retention of all placentas. A quality improvement project using Plan Do Study Act cycles was implemented at both sites. Interventions included installation of dedicated labour ward fridges for 48-hour storage with temperature monitoring, targeted education on indications for placental histology and daily cross-checking of neonatal admissions to ensure eligible placentas were sent. The primary outcome was the proportion of placentas meeting criteria sent for histopathology, focusing on deliveries <32 weeks and intrauterine deaths >18 weeks. National investigation reports at St Thomas’ Hospital were also reviewed for missing histology. Labour ward storage proved feasible and effective. At the Royal Berkshire Hospital, submission rates improved from 88% to 97% for intrauterine deaths and from 71% to 100% for very preterm births. At St Thomas’ Hospital, preterm submissions increased from 80% to 96% after education and 98% after storage, with intrauterine death remaining at 100%. Missing histology in national investigations reduced from 12% to 0%. Short-term placental storage is a low-cost, scalable intervention aligned with national maternity safety priorities.

PMID:42648748 | DOI:10.1136/bmjoq-2026-004321

JMIR Form Res. 2026 Aug 26;10:e95883. doi: 10.2196/95883.

ABSTRACT

BACKGROUND: Large language models (LLMs) are increasingly used to support digital health communication, yet their reliability in patient-facing cardiovascular imaging education remains uncertain. Cardiovascular imaging involves complex terminology and procedural details that many patients struggle to understand, creating a need for accurate, clear, and reassuring explanations. While prior evaluations of conversational AI have focused primarily on diagnostic reasoning or clinician-oriented tasks, few studies have systematically compared contemporary LLMs in their ability to communicate effectively with patients.

OBJECTIVE: This study aimed to compare the accuracy, clarity, completeness, and patient-centered communication quality of responses generated by 3 state-of-the-art conversational agents (DeepSeek, GPT-o1, and GPT-4o) when addressing real-world patient questions about cardiovascular imaging.

METHODS: A prospective methodological evaluation was conducted using 84 unique patient-centered questions curated from authoritative cardiovascular information sources and online patient forums. Each question was independently submitted to DeepSeek, GPT-o1, and GPT-4o in isolated sessions to avoid contextual contamination. Two cardiovascular radiologists scored each response across 4 domains (accuracy, clarity and appropriateness, completeness, and user engagement and reassurance) using a standardized 3-point rubric (total score range 4-12). Discrepancies were resolved through predefined adjudication procedures. Because the scores were ordinal, median domain and composite scores with IQRs were summarized and compared across the 3 models using the Kruskal-Wallis test, with ε² as an effect size. Statistical significance was defined as an α value of .05.

RESULTS: Across the 84 patient questions, all 3 models produced largely accurate, clear, and complete responses, with comparably high scores across the accuracy, clarity, and completeness domains (median 3 of 3, IQR 3-3 in each). The only meaningful difference appeared in user engagement and reassurance. A “good” engagement rating was assigned to 96.4% (81/84) of DeepSeek responses and 98.8% (83/84) of GPT-o1 responses but only 53.6% (45/84) of GPT-4o responses (Kruskal-Wallis P<.001). Composite scores were correspondingly lower for GPT-4o (median 11, IQR 10-12) than for DeepSeek and GPT-o1 (both median 12, IQR 11-12; P<.001). No significant differences were observed across models for accuracy (P=.91), clarity (P=.06), or completeness (P=.65), and no unsafe statements were identified in any model.

CONCLUSIONS: DeepSeek and GPT-o1 consistently delivered accurate, clear, and patient-centered explanations of cardiovascular imaging questions, whereas GPT-4o, despite comparable technical accuracy, provided less engaging and reassuring communication. These findings suggest that affective qualities rather than factual correctness represent the main differentiator among current LLMs in patient education tasks. As conversational agents become integrated into cardiovascular imaging workflows, attention to communication tone, emotional support, and health literacy alignment will be essential to ensure safe and effective patient use.

PMID:42647860 | DOI:10.2196/95883

JMIR Serious Games. 2026 Aug 26;14:e98694. doi: 10.2196/98694.

ABSTRACT

BACKGROUND: Hemiplegic cerebral palsy (CP) commonly affects upper-limb function in children and may limit functional use of the affected arm and hand. Virtual reality (VR)-based rehabilitation can provide interactive, feedback-based, and task-oriented practice, but whether adding VR-based intervention to conventional therapy provides additional benefits remains uncertain. Previous reviews have often included broader CP populations, mixed comparators, or broad upper-limb outcome labels, limiting domain-specific interpretation.

OBJECTIVE: This systematic review and meta-analysis evaluated VR-based intervention combined with conventional therapy versus conventional therapy alone for upper-limb-related outcomes in children with hemiplegic CP.

METHODS: This review was prospectively registered with PROSPERO (CRD420261486172). PubMed, Embase, Web of Science, and the Cochrane Central Register of Controlled Trials were searched from inception to April 15, 2026. Eligible randomized controlled trials included children aged ≤12 years with hemiplegic CP and compared VR-based or interactive game-based intervention plus conventional therapy with conventional therapy alone. Outcomes were grouped into upper-limb activity, dexterity, motor control and movement quality, activities of daily living (ADLs), and grip strength. Risk of bias was assessed using the revised Cochrane risk-of-bias 2 tool. Random-effects meta-analyses with Hartung-Knapp-Sidik-Jonkman adjustment were performed; Nagashima-corrected prediction intervals (PIs) were calculated where applicable; and certainty of evidence was assessed using Grading of Recommendations Assessment, Development, and Evaluation (GRADE).

RESULTS: Thirteen randomized controlled trials involving 458 children were included. Pooled estimates generally favored VR combined with conventional therapy, but no outcome reached statistical significance; effect estimates were as follows: upper-limb activity (standardized mean difference [SMD] 0.28, 95% CI -0.07 to 0.64), dexterity (SMD 0.26, 95% CI -0.34 to 0.86), motor control and movement quality (SMD 0.29, 95% CI -0.22 to 0.80), ADLs (SMD 0.17, 95% CI -0.22 to 0.55), and grip strength (mean difference 0.89, 95% CI -1.93 to 3.72). PIs also crossed the null value for all outcomes, ranging from -0.51 to 1.04 for upper-limb activity and from -4.89 to 6.45 for grip strength, suggesting variable effects across settings. Heterogeneity was low to moderate for most outcomes but substantial for grip strength. Most trials had some risk-of-bias concerns. Certainty of evidence was low for all outcomes except grip strength, which was very low.

CONCLUSIONS: VR-based intervention combined with conventional therapy did not provide clear evidence of additional benefits over conventional therapy alone. Although effects generally favored the combined intervention, CIs, PIs, risk-of-bias concerns, and low to very low certainty of evidence support cautious interpretation. VR may be considered a potentially engaging adjunct, but current evidence does not support its use as a replacement for therapist-led rehabilitation or as a definitively effective intervention. Larger, adequately powered trials with clearer reporting of VR characteristics, intervention dose, task specificity, and longer follow-up are needed.

TRIAL REGISTRATION: PROSPERO CRD420261486172; https://www.crd.york.ac.uk/PROSPERO/view/CRD420261486172.

PMID:42647858 | DOI:10.2196/98694

J Med Internet Res. 2026 Aug 26;28:e96002. doi: 10.2196/96002.

ABSTRACT

BACKGROUND: Rehabilitation clinical practice guidelines (CPGs) have increased rapidly, but inconsistent methodological quality limits their implementation. Although Appraisal of Guidelines for Research and Evaluation II (AGREE II) and Reporting Items for Practice Guidelines in Health Care (RIGHT) provide standardized appraisal frameworks, their application is time-consuming. Large language model (LLM)-based AI agents may offer a scalable alternative with uncertain reliability.

OBJECTIVE: We evaluated rehabilitation CPGs’ methodological and reporting quality and determined whether structured guidance improves human expert-AI agent agreement.

METHODS: We systematically reviewed English- and Chinese-language rehabilitation CPGs from Embase, Scopus, PubMed, China National Knowledge Infrastructure, Wanfang Data, National Institute for Health and Care Excellence, Scottish Intercollegiate Guidelines Network, and Guidelines International Network up to June 2026. Methodological and reporting quality were assessed using AGREE II and the RIGHT checklist. Factors associated with guideline quality were examined using regression and subgroup analyses. Two AI agents were compared with human consensus with and without a structured guideline appraisal workbook, followed by external validation using 6 anterior cruciate ligament reconstruction CPGs.

RESULTS: We included 227 CPGs (163 English-language, 64 Chinese-language). After introducing a structured guideline appraisal workbook, agreement among human experts improved markedly-mean intraclass correlation coefficients (ICCs) increased from -0.09 to 0.66 to 0.84-0.92 across AGREE II domains. Overall guideline quality remained low, with 35.9% (SD 18,8%) applicability and 52% (SD 17.2%) stakeholder involvement. English-language guidelines outperformed Chinese-language guidelines in scope and purpose (mean 74.64, SD 15.4 vs mean 68.88, SD 13.9; P=.004) and applicability (mean 39.14, SD 18.6 vs mean 27.54, SD 16.7; P<.001). Backward-elimination logistic regression revealed external review as an associated process characteristic (odds ratio 20.39, 95% CI 4.66-89.27; P<.001). RIGHT assessments showed consistent reliability (ICC=0.80-0.88). Reporting was highest for basic information (70.5%) and lowest for funding, declaration, and management of interests (44.1%). Meta-analysis of RIGHT reporting rates showed lower reporting among Chinese-language than English-language guidelines (risk difference [RD] -0.07, 95% CI -0.13 to -0.02, 95% prediction interval [PI] -0.39 to 0.24) and among guidelines published before vs after RIGHT release (RD -0.19, 95% CI -0.26 to -0.13, 95% PI -0.56 to 0.17). Without additional guidance, agent-human agreement was moderate (ICC=0.608-0.629). The workbook improved agreement for both models, with DeepSeek-R1’s increasing from 0.613 to 0.709 and o1-mini’s from 0.629 to 0.687. In validation beyond rehabilitation, DeepSeek-R1 maintained stable agreement (ICC=0.711) and completed appraisals in 5.44 minutes compared to 11.18 minutes for humans.

CONCLUSIONS: Rehabilitation CPGs, particularly Chinese-language CPGs, continue showing deficiencies in applicability and stakeholder involvement. LLM-based appraisal without structured guidance provides insufficient agreement. Structured guidance improved agent-human agreement, supporting AI-assisted guideline appraisal under human oversight. Although further validation across additional clinical specialties is needed, AI agents can serve as efficient assistants in guideline appraisal instead of replacing humans. Future synthesis requires human-AI integration guided by structured, expert-defined principles.

TRIAL REGISTRATION: PROSPERO CRD420251270676; https://www.crd.york.ac.uk/PROSPERO/view/CRD420251270676.

PMID:42647856 | DOI:10.2196/96002

R I Med J (2013). 2026 Sep 1;109(9):72-77.

ABSTRACT

INTRODUCTION: Case-based learning strategies have been shown to enhance student engagement, promote active learning, and improve problem-solving ability by simulating clinical scenarios. However, few interventions focus specifically on integration within basic science domains, and even fewer have been student-led and implemented longitudinally throughout a pre-clerkship curriculum. This study evaluates the impact of a student-led initiative that incorporated interdisciplinary clinical vignettes into first-year medical coursework at a large allopathic medical school in the United States.

METHODS: Vignettes were designed to align with curricular themes across anatomy, histology, and pathology for each of the 13 curricular blocks. First-year medical students completed post-intervention surveys measuring confidence in applying course material, integration between concepts, and perceived relevance of the vignettes.

RESULTS: Across all blocks of vignettes, the majority of respondents agreed that the vignettes allowed them to think about lecture material in a new way (98%; 57% strongly agree, 41% somewhat agree), felt relevant to current material (95%; 73% strongly agree, 22% somewhat agree), and related the concepts from their anatomy, pathology, and histology courses (94%; 69% strongly agree, 25% somewhat agree). The analysis of our survey data confirms that students’ confidence in exam material increased after the use of the clinical vignettes. This was statistically significant for every discipline and across the vast majority of blocks.

CONCLUSIONS: Integrated vignettes fostered interdisciplinary learning and improved confidence in applying pre-clerkship knowledge across disciplines. This low- resource, peer-led model may serve as an adaptable tool to support early clinical reasoning and integrative learning across foundational science curricula.

PMID:42647842

R I Med J (2013). 2026 Sep 1;109(9):63-71.

ABSTRACT

BACKGROUND: In the United States, significant disparities exist in preventative health among underserved communities, including African American/Black (AA/B), Hispanic/Latino (H/L) and lesbian/gay/bisexual/transgender/ queer (LGBTQ+) individuals. We explored the influence of social determinants of health (SDOH) on preventative health outcomes in these groups.

METHODS: We conducted a comprehensive cross-sectional needs assessment at safety-net health clinics in Providence, Rhode Island, from June to September 2022.

RESULTS: Of N=1,017 individuals, 67% identified as H/L, 16% identified as AA/B, and 19% identified as LGBTQ+. The most common diagnosed health conditions were anxiety (26%), depression (23%), hypertension (20%), and diabetes (15%). People reported most commonly needing help with food (31%), housing (25%), and utilities (22%). More social needs were associated with reductions in COVID-19 booster vaccination (p<.01) and higher mental health diagnoses (p<.01). AA/B individuals with higher health literacy were also less likely to be diagnosed with a mental health condition (p=.02). LGBTQ+ individuals were more likely to report a history of mental health illness, tobacco use, hazardous alcohol use, other substance use, and more sexual partners (p<.05).

CONCLUSIONS: Addressing health literacy and other social determinants of health is important to address disparities and achieve health equity among marginalized communities.

PMID:42647841

R I Med J (2013). 2026 Sep 1;109(9):42-46.

ABSTRACT

OBJECTIVE: Partial hospitalization programs (PHP) are acute psychiatric settings that commonly treat patients with active suicidal ideation (SI), anxiety, and alcohol use. Alcohol use is common among patients with SI and anxiety, and this triple vulnerability may negatively impact treatment engagement and outcomes in PHP. This study examined whether patients presenting with SI, elevated anxiety, and alcohol use (“triple vulnerability”) demonstrated differences in PHP treatment outcomes (completing treatment, number of days attended, number of absences) compared to patients without the vulnerability.

METHODS: Participants were 493 patients admitted to an Acceptance and Commitment Therapy-based PHP program in the United States between 2014-2020. Triple vulnerability was defined as reporting (1) active SI, (2) clinically elevated anxiety (CUXOS-D ≥ 20), and (3) consuming ≥ 1 alcoholic drink on a typical drinking day within the past year. Binary and linear regressions were used to assess differences in treatment completion status, number of days attended, and number of absences.

RESULTS: The triple vulnerability was not significantly associated with treatment completion status or with the number of absences; it was, however, positively associated with the number of days attended at PHP.

CONCLUSION: People who experience SI, clinically elevated anxiety, and who consume at least one alcoholic beverage on a typical drinking day are more likely to attend PHP for a longer duration of time, but are not any more susceptible to not finishing treatment, relative to people who do not have the triple vulnerability. Future work can benefit from exploring additional treatment outcomes that may be impacted by patients with the triple vulnerability.

PMID:42647836

R I Med J (2013). 2026 Sep 1;109(9):37-41.

ABSTRACT

Rates of adolescent suicide have risen over the last 10 years, creating a public health crisis. Disclosure of suicidal thoughts and behaviors (STBs) is one help-seeking strategy available to adolescents. Estimates from prior studies on the frequency of youth STB disclosure vary considerably. It is crucial to understand patterns and rates of STB disclosure for youth at elevated risk for suicide, including those who have been psychiatrically hospitalized or who hold sexual and gender minority (SGM) status. To examine STB disclosure in this high-risk population, the current study examined self-report of STB disclosure behaviors, including to whom, how, and when disclosure occurred, as well as the outcome of the disclosure in a sample of psychiatrically hospitalized youth with prior STBs (N=163). Data showed that the most common recipients of disclosures were caregivers, health professionals, and in-person friends. The most frequently endorsed method of disclosure was doing so in person, and there were no significant differences in when adolescents disclosed STBs (before, during, or after the experience). SGM adolescents were half as likely as non-SGM adolescents to disclose to parents or caregivers, and SGM adolescents disclosed more often during the STB than non-SGM youth. These findings can inform future interventions targeted at reducing STBs in youth or increasing help-seeking behaviors for psychiatrically hospitalized adolescents.

PMID:42647835