Tag: statistics

Optom Vis Sci. 2026 Mar;103(3):e70039. doi: 10.1002/ovs2.70039.

ABSTRACT

PURPOSE: Currently, there are neither guidelines for school-based vision program operations nor a registry of programs that exist in the United States. We aim to describe the characteristics of school-based vision programs across the United States through a survey.

METHODS: We identified 10 organizations that operated one or more school-based vision programs during the 2022-2023 school year. We developed and distributed an online survey form to collect information on program demographics, program components, and protocols for vision screenings and eye exams, as well as procedures for obtaining consent, dispensing glasses, and funding resources. Organizations were asked to complete a survey for each school district in which they operated, which we defined as an individual program. Programs that did not complete the survey were excluded. We used descriptive statistics to describe program characteristics. We compared grade levels that received vision screening by school-based vision programs and their respective screening mandate in the state where the programs operated.

RESULTS: Seven organizations participated and provided data about 184 unique school-based vision programs. These programs operated in 19 states and the District of Columbia. All 184 programs provided vision screenings, eye exams, and eyeglasses. 94.6% of programs included instrument-based screening, with 46.7% using it exclusively. School-based vision programs most frequently screened elementary school grades. For eye exams, 81% of programs utilized opt-out consent; 21.7% of programs included dilation. Vision screenings were mostly conducted by program staff (164/184, 89.1%), and eye exams were performed by optometrists (183/184, 99.5%). Nine programs (5.0%) reported providing additional care beyond eyeglasses provision, while 135 (73.4%) reported having an established relationship with community eye care providers for students referred for additional evaluation and care.

CONCLUSIONS: Many school-based vision programs operate in the United States. These programs focus on vision screening, eye exams, and eyeglasses provision. Establishing a registry of programs would help to disseminate lessons learned and track outcomes data. Future programs will benefit from guidelines to ensure they are operating according to best practices.

PMID:42020923 | DOI:10.1002/ovs2.70039

Geroscience. 2026 Apr 22. doi: 10.1007/s11357-026-02212-z. Online ahead of print.

ABSTRACT

Age is a well-known risk factor to develop severe viral respiratory infections, including severe COVID-19. This study aimed to identify the biological alterations linked to severe disease in elderly patients with COVID-19. For this purpose, we employed a derivation cohort with 450 SARS-CoV-2 infected and unvaccinated patients admitted to hospital wards and a validation cohort with 244 SARS-CoV-2 infected and unvaccinated patients admitted to hospital Intensive Care Unit (ICU). Twenty-one biomarkers were measured in plasma samples from patients upon admission, including SARS-CoV-2 RNA, IgG antibodies, and protein biomarkers. Patient cohorts were divided into two groups based on age: adult (≤ 70 years old) and elderly (> 70 years old) patients. In the derivation cohort, 90-day mortality rate observed in the adult group was 6.0% whereas in the elderly group it rises to 31.6%, same trend was noticed regarding the validation cohort, with 11.2% versus 40.3% 90-day mortality rates for adult and elderly groups, respectively. The machine-learning framework XGBoost-SHAP, fed with the plasma biomarkers information, was used to profile an age-related host response to SARS-CoV-2 infection. Based on SHAP plot, elderly patients had a strong thrombo-inflammatory response profile (significantly elevated plasma levels of: lipocalin-2, endothelin-1, D-dimer) combined with deficient adaptive and cytotoxic antiviral responses. Model performance evaluated with the validation cohort confirmed the robustness and generalizability of the model developed (AUC = 0.710). In conclusion, the machine learning approach we built allowed us to identify the presence of a deranged host response in elderly patients with COVID-19 linked to poor viral control and increased mortality.

PMID:42020919 | DOI:10.1007/s11357-026-02212-z

Nat Biotechnol. 2026 Apr 22. doi: 10.1038/s41587-026-03084-6. Online ahead of print.

ABSTRACT

Single-cell RNA sequencing (scRNA-seq) analyses typically focus on alignment and differential gene expression, being blind to other transcriptome variation. Here we present sc-SPLASH for statistics-first, reference-free discovery on barcoded scRNA-seq and spatial transcriptomics; its independent BKC submodule optimizes barcoded data preprocessing and is approximately 50-fold faster than UMI-tools. sc-SPLASH discovers secreted repeat proteins in immune-like cells of sponge (Spongilla; missing from the reference) and tunicate (Ciona).

PMID:42020848 | DOI:10.1038/s41587-026-03084-6

Support Care Cancer. 2026 Apr 23;34(5):458. doi: 10.1007/s00520-026-10645-7.

ABSTRACT

PURPOSE: Cancer diagnosed during pregnancy presents unique challenges, requiring women to navigate treatment alongside pregnancy and early parenthood. While clinical aspects are well studied, the psychosocial impact on survivorship remains underexplored. This study examined the lived experiences of women diagnosed during pregnancy, focusing on emotional, psychological and practical challenges from diagnosis through survivorship.

METHODS: A qualitative study was conducted using interview data from 20 women in the UK diagnosed with cancer during pregnancy. Participants were recruited via Mummy’s Star, a charity supporting individuals affected by cancer in pregnancy. Interviews were thematically analysed using template analysis, focusing on decision-making, psychosocial burden and support needs.

RESULTS: Six inter-related themes were identified: (1) managing cancer with uncertainty, women reported distress due to delayed investigations and treatment adjustments during pregnancy; (2) ethical decision-making, emotionally charged choices around treatment, birth and feeding were made with limited or conflicting information; (3) balancing cancer and its treatment with pregnancy and family life, early parenting was disrupted; (4) work disruption and financial strain, treatment-related costs and lost income caused hardship; (5) emotional impact of diagnosis and treatment, including lasting psychological effects; and (6) Coping and support, guilt, fear of recurrence and unmet support needs persisted post-treatment.

CONCLUSIONS: Women diagnosed with cancer in pregnancy face profound, long-term emotional and financial challenges. Fragmented care and inadequate support exacerbate these difficulties. Integrated multidisciplinary care is essential to improving survivorship.

PMID:42020845 | DOI:10.1007/s00520-026-10645-7

Support Care Cancer. 2026 Apr 23;34(5):455. doi: 10.1007/s00520-026-10688-w.

ABSTRACT

PURPOSE: Survivors of nasopharyngeal carcinoma (NPC) frequently suffer from a prolonged reduction in quality of life (QoL) following successful treatment. This study investigates the determinants of QoL among NPC survivors, with a focus on the temporal evolution post-treatment and the influence of socioeconomic and clinical factors.

METHODS: In this cross-sectional study, NPC patients who received treatment in the last 10 years and underwent follow-up at our institute from 2021 to 2023 were enrolled. They completed the EuroQol five-dimension (EQ-5D), World Health Organization Quality of Life-Brief (WHOQOL-BREF), Sinonasal Outcome Test 22 (SNOT-22), Eustachian Tube Dysfunction Questionnaire-7 (ETDQ-7), and Eating Assessment Tool-10 (EAT-10). Kernel smoothing techniques were applied to delineate the trends in QoL scores, and linear mixed models were utilized for evaluating the unadjusted and adjusted influences of household income and other predictors on QoL.

RESULTS: Out of 248 participants, 355 QoL evaluations were performed. Kernel-smoothed trajectories revealed that higher household income correlated with superior QoL scores throughout the majority of the post-treatment timeline. Although initial analyses showed that both higher household income and Karnofsky Performance Scale (KPS) scores were predictive of better generic and condition-specific QoL, the effect of income was attenuated upon adjusting for KPS in the multivariate analysis. Mediation analysis indicated that the association between income and QoL was partly mediated by the patients’ performance status.

CONCLUSION: Performance status is a pivotal mediator in the interplay between socioeconomic status and QoL outcomes in NPC survivors. These insights underscore the need for prospective studies to confirm these relationships.

PMID:42020823 | DOI:10.1007/s00520-026-10688-w

Pediatr Res. 2026 Apr 22. doi: 10.1038/s41390-026-04929-1. Online ahead of print.

ABSTRACT

BACKGROUND: Studies have reported conflicting findings regarding whether attention-deficit/hyperactivity disorder (ADHD) and neurostimulant (NS) treatment influence growth. The study aimed to evaluate the impact of ADHD and NS-treatment on adult height (AH), while accounting for individual expected height.

METHODS: Retrospective controlled cohort study using a nationwide healthcare provider database. Children born 1995-2003 were classified into: (1) untreated ADHD, (2) NS-treated ADHD, and (3) controls, followed until AH (girls ≥17 years, boys ≥19 years). Observed height index (OHI) = individual AH ÷ sex-specific mean AH of controls. Expected height index (EHI) = individual’s MPH ÷ sex-specific mean MPH of controls. Normalized height deviance (NHD) = OHI – EHI. ADHD impact = mean NHD(untreated ADHD) – mean NHD(controls). NS-treatment impact = mean NHD(NS-treated) – mean NHD(untreated ADHD). Significance was set at p ≤ 0.05. Effect sizes reported as Cohen’s d and f.

RESULTS: The cohort included 17,517 children with ADHD (5671 untreated; 11,846 NS-treated) and 47,258 controls. Girls’ ADHD impact was -0.28% (p < 0.001; Cohen’s d = 0.08), and in boys -0.11% (p = 0.12; Cohen’s d = 0.03). Girls’ NS-treatment impact was -0.26% (p = 0.005, Cohen’s d = 0.07), in boys -0.10% (p = 0.25).

CONCLUSIONS: ADHD and NS-treatment were associated with negligible effect sizes in girls and non-significant differences in boys.

IMPACT: After accounting for individual familial genetic height potential, ADHD and neurostimulant treatment are associated with statistically detectable but clinically negligible differences in adult height. Introduces normalized height deviance (NHD), a genetics-adjusted measure that distinguishes true growth effects from delayed maturation. Uses a large, nationwide cohort followed to adult height, resolving long-standing inconsistencies in prior growth studies. Provides reassurance that ADHD and neurostimulant treatment do not meaningfully impair final adult stature and supports informed clinical decision-making.

PMID:42020798 | DOI:10.1038/s41390-026-04929-1

Commun Chem. 2026 Apr 22. doi: 10.1038/s42004-026-02020-8. Online ahead of print.

ABSTRACT

Water-soluble functionalized fullerenes commonly referred to as fullerenols, fullerols or polyhydroxy fullerenes are widely used in photonics, catalysis, and biomedicine, yet their molecular structures have been assumed to consist solely of hydroxyl groups for nearly three decades. This assumption remains despite persistent mismatches between calculated and experimental vibrational and optical spectra as well as expected and observed chemical reactivity. Here we combine Fourier transform infrared (FTIR) and absorption (UV-Vis) spectroscopy, X-ray photoelectron spectroscopy (XPS), targeted chemical derivatization, and computational quantum chemistry to resolve this discrepancy. We show that only a polyoxy-fullerene architecture incorporating counterion-coordinated carbonyl and hemiketal groups alongside hydroxyls reproduces both the characteristic FTIR features and the experimental UV-Vis absorption profile. A purely hydroxylated fullerene model fails to capture the dominant FTIR band and asymmetric ultraviolet absorption. Oxime-formation experiments chemically validate the presence of carbonyl and hemiketal groups. This structural reassignment resolves long-standing inconsistencies in fullerene chemistry, corrects a pervasive misinterpretation in the literature, and establishes a framework for rationally tuning the optical and chemical properties of functionalized nanocarbons.

PMID:42020770 | DOI:10.1038/s42004-026-02020-8

Nature. 2026 Apr 22. doi: 10.1038/s41586-026-10549-w. Online ahead of print.

ABSTRACT

Large language models sometimes produce confident, plausible falsehoods (“hallucinations”), limiting their reliability^1,2. Prior work has offered numerous explanations and effective mitigations such as retrieval and tool use³, consistency-based self-verification⁴, and reinforcement learning from human feedback⁵. Nonetheless, the problem persists even in state-of-the-art language models^6,7. Here we show how next-word prediction and accuracy-based evaluations inadvertently reward unwarranted guessing. Initially, next-word pretraining creates statistical pressure toward hallucination even with idealized error-free data: using learning theory^8,9, we show that facts lacking repeated support in training data (such as one-off details) yield unavoidable errors, while recurring regularities (such as grammar) do not. Subsequent training stages aim to correct such errors. However, dominant headline metrics like accuracy systematically reward guessing over admitting uncertainty. To align incentives, we suggest two additions to the classic approach of adding error penalties to evaluations to control abstention^10,11. First, we propose “open-rubric” evaluations that explicitly state how errors are penalized (if at all), which test whether a model modulates its abstentions to stated stakes while optimizing accuracy. Second, since hallucination-specific benchmarks rarely make leaderboards¹², we suggest using open-rubric variants of existing evaluations, to reverse their guessing incentives. Reframing hallucination as an incentive problem opens a practical path toward more reliable language models.

PMID:42020757 | DOI:10.1038/s41586-026-10549-w

Nature. 2026 Apr 22. doi: 10.1038/s41586-026-10380-3. Online ahead of print.

ABSTRACT

Human cooperation is dynamic and often declines even under favourable conditions^1-4. Many prevailing theories explain the decrease of cooperation in terms of strategic behaviour or learning, framed as evidence of rational behaviour or progression towards rationality^5-9. Here we show that a key source of long-term decline derives from deviations from rational behaviour that systematically vary over time. We analyse a natural social dilemma in the field-that is, group lending in Sierra Leone-tracking cooperative dynamics over a five-year period. Borrowers enter a joint-liability contract, structured so that if the group loan is not repaid in full, all members lose access to future credit¹⁰. This produces a threshold social dilemma with incentives to free-ride^11,12. The dataset includes 47,931 group payments made by 7,108 borrowers, augmented with a two-stage cluster sample of semi-structured interviews. We find a statistically robust pattern of punctuated decline driven by behavioural mechanisms¹³. Cooperation rates start out high but gradually decline due to decreases in group members’ cooperative motivation and effort. Sharp rebounds occur when loans are restarted and clients resensitized to their cooperative responsibilities, even though the group membership and dilemma structure are largely unchanged. This pattern persists over the five-year observation window, but with each successive restart the subsequent decline is more rapid. The findings have direct implications for preventing behavioural decline in cooperative programmes and institutions.

PMID:42020732 | DOI:10.1038/s41586-026-10380-3

Mol Psychiatry. 2026 Apr 22. doi: 10.1038/s41380-026-03606-3. Online ahead of print.

ABSTRACT

Neuropsychiatric disorders exhibit complex polygenic architectures, yet the cell-type-specific mechanisms underlying most risk loci remain unclear. Here, we integrate single-cell expression quantitative trait locus (sc-eQTL) data from brain and blood tissues with genome-wide association studies (GWAS) of six neuropsychiatric disorders (schizophrenia (SCZ), Parkinson’s disease (PD), bipolar disorder (BP), major depressive disorder (MDD), attention-deficit/hyperactivity disorder (ADHD), and autism spectrum disorder (ASD)) to systematically identify putative causal genes at cellular resolution. Employing summary-data-based Mendelian randomization (SMR) across diverse neuronal and immune cell types, we discovered 345 cell-type-specific risk genes for various diseases, including both replicated candidates (such as MAPT in astrocytes for SCZ and PD and FLOT1 in excitatory neurons and inhibitory neurons for SCZ, BP and MDD) and novel associations (such as APTX in microglia for SCZ). Cross-disorder analyses revealed shared pathways in synaptic function and immune regulation. In contrast, disease-specific and tissue-specific patterns were observed across different disorders. Strikingly, we found that brain-derived risk genes exhibited significantly higher cell-type specificity than those identified in blood, underscoring the more focused cellular context of genetic effects in the central nervous system. Our findings suggest that neuropsychiatric disorders arise from a combination of neuronal dysfunction and immune system dysregulation. The study demonstrates how cell-type-specific mapping uncovers etiological mechanisms obscured in bulk-tissue analyses, proving novel information for clarifying the biological mechanism of gene expression implicated in the development of the six neuropsychiatric disorders.

PMID:42020718 | DOI:10.1038/s41380-026-03606-3