Category: Statistics

Stat Comput. 2026;36(3):131. doi: 10.1007/s11222-026-10866-0. Epub 2026 May 5.

ABSTRACT

In an effort to develop topic modeling methods that can be quickly applied to large data sets, we revisit the problem of maximum-likelihood estimation in topic models. It is known, at least informally, that maximum-likelihood estimation in topic models is closely related to non-negative matrix factorization (NMF). Yet, to our knowledge, this relationship has not been exploited previously to fit topic models. We show that recent advances in NMF optimization methods can be leveraged to fit topic models very efficiently, often resulting in much better fits and in less time than existing algorithms for topic models. We also formally make the connection between the NMF optimization problem and maximum-likelihood estimation for the topic model, and using this result we show that the expectation maximization (EM) algorithm for the topic model is essentially the same as the classic multiplicative updates for NMF. Our methods are implemented in the R package “fastTopics”.

PMID:42100650 | PMC:PMC13144203 | DOI:10.1007/s11222-026-10866-0

Cureus. 2026 May 6;18(5):e108357. doi: 10.7759/cureus.108357. eCollection 2026 May.

ABSTRACT

BACKGROUND: The fragility index (FI) is intended to quantify how many outcome changes would be required to convert a statistically significant two-arm trial result into a nonsignificant one. A reliable statistical metric should produce a result for every valid case it evaluates. This study examined whether a fragility value is always attainable for every statistically significant trial result.

METHODS: FI was analyzed as follows: baseline significance was required (p < 0.05), one-way movement only, and outcome changes were restricted to converting a nonevent to an event in the arm with fewer events, while keeping the arm size fixed. Nonattainability was assessed by determining whether valid 2×2 tables exist for which no finite FI can be obtained under these rules. Evidence is provided through formal counterexamples, complete enumeration of all valid nondegenerate 2 × 2 tables up to total sample size N = 60, and empirical evaluation of published two-arm trials with binary outcomes.

RESULTS: Valid baseline-significant 2 × 2 tables exist for which FI is not attainable. A simple counterexample is {3,0,4,11}: baseline two-sided Fisher’s exact p = 0.0429, the arm with fewer events is uniquely identified, but that arm has no nonevents available for the required toggle; thus, no legal FI path exists. Enumeration revealed that unattainable cases first appeared at N = 18 and then recurred at every larger sample size through N = 60; by N = 60, a total of 2,390 of 20,774 evaluable baseline-significant tables were unattainable (11.5%). In an empirical dataset of published trials, 2 of 82 baseline-significant evaluable trials (2.4%) were not attainable.

CONCLUSIONS: The FI is not universally attainable. This is a structural property of the FI algorithm, confirmed by mathematical proof, a complete table enumeration, and published trial data.

PMID:42100648 | PMC:PMC13148188 | DOI:10.7759/cureus.108357

Cureus. 2026 May 6;18(5):e108363. doi: 10.7759/cureus.108363. eCollection 2026 May.

ABSTRACT

Complex wound reconstruction may progress through stage-dependent clinical and economic trajectories in which wound failure leads to infection, reoperation, prolonged hospitalization, post-acute care, outpatient wound management, and possible readmission. Procedural-cost evaluation may not fully capture these downstream consequences in high-risk reconstructive settings. This paper presents the CASCADE (Cost Analysis of Surgical Complications and Downstream Expenditure) model, a conceptual, literature-derived decision framework rather than a primary data analysis or statistically validated predictive model. CASCADE applies an expected-value framework to evaluate the reconstructive strategy at the index operation, using three bounded inputs: failure probability (P), failure trajectory cost (C), and incremental reconstruction cost (ΔC). These inputs are structured into the decision rule ΔC < ΔP × C. A Clinical Risk Score (CRS) is used to standardize the assignment of cases into risk tiers. Across illustrative high-risk wound environments, failure trajectories may increase total episode expenditure from approximately $80,000-$150,000 after successful reconstruction to $400,000-$1,000,000+ after failure-driven care. These values are literature-informed illustrative estimates used to parameterize the framework, not observed case-level measurements. Sensitivity analysis demonstrates that at CRS ≥6, the CASCADE decision threshold often exceeds typical incremental reconstructive procedure costs across plausible input combinations. CASCADE provides a reproducible, bidirectional framework for trajectory-based cost evaluation in complex wound reconstruction. It may support surgical decision-making, institutional planning, and reimbursement analysis when total episode cost, rather than index procedural cost alone, is the appropriate unit of economic evaluation. The framework evaluates cost relationships but does not define reimbursement levels and requires future validation against institutional or claims-based datasets.

PMID:42100647 | PMC:PMC13148454 | DOI:10.7759/cureus.108363

Ecol Evol. 2026 May 5;16:e73605. doi: 10.1002/ece3.73605. eCollection 2026 May.

ABSTRACT

Forest structural complexity is critical for ecosystem functions, yet standardized metrics for its quantification remain elusive. This study compares two LiDAR-derived three-dimensional indices, the box dimension ( $D_b$ ) as a fractal-based measure, and canopy entropy ( $\mathrm{CE}$ ), an entropy-based metric, to evaluate their methodological, computational, and conceptual differences. Using mobile LiDAR scans from $15m\times 15m$ forest plots in Maine, USA, and Nova Scotia and New Brunswick, Canada, we analyzed 170 point clouds to assess correlation, computation time, and theoretical underpinnings. Statistical analysis revealed a strong linear relationship between $D_b$ and $\mathrm{CE}$ (Pearson’s $r=0.823,p<0.001$ ), with Deming regression indicating $\widehat{\mathrm{CE}}=4.75\times D_b–1.07$ . Also, $\mathrm{CE}$ computation averaged 40 times slower than $D_b$ , scaling roughly linearly with point cloud size. Conceptually, $D_b$ reflects fractal dimensionality linked to physiological process optimization, while $\mathrm{CE}$ quantifies biomass distribution homogeneity. ${\mathrm{CE}}^{‘}$ s unit dependence on plot size limits cross-study comparability, whereas ${D_b}^{‘}$ s dimensionless fractal interpretation offers broader intuitiveness. Both indices address sampling density bias but differ in parameterization and data efficiency. Despite ${\mathrm{CE}}^{‘}$ s theoretical novelty, it does not surpass $D_b$ in interpretability, precision, or speed, and its proposed advantage in capturing higher complexity remains unsubstantiated. Despite their conceptual distinctions, their strong correlation suggests competitive rather than complementary roles. Future research should explore biome-specific variability and physiological links to ecosystem functions to refine their utility in forest management under climate change.

PMID:42100628 | PMC:PMC13143574 | DOI:10.1002/ece3.73605

Contracept Reprod Med. 2026 May 7. doi: 10.1186/s40834-026-00455-x. Online ahead of print.

ABSTRACT

BACKGROUND: India is the most populous country in the world, with over 1.4 billion people, facing significant challenges in maternal and reproductive health services. Family planning is a key strategy to prevent unintended pregnancies and optimize spacing between births. Post-placental intrauterine contraceptive device (PPIUCD) insertion is a safe, long-acting, reversible, and cost-effective method of immediate postpartum contraception. This study aimed to compare continuation rates, complications, and expulsion rates of PPIUCD insertion between preterm and term deliveries, and to explore factors influencing outcomes.

SETTINGS AND DESIGN: Hospital-based observational study conducted at VMMC & Safdarjung Hospital, New Delhi.

MATERIALS AND METHODS: Among 2000 antenatal women counseled for immediate postpartum IUCD insertion, 240 (12%) consented and were enrolled. Participants were divided equally into preterm (n = 120) and term (n = 120) groups. CuT 380 A devices were inserted immediately post-placentally. Follow-up was conducted at 3, 6, and 12 months to assess continuation, satisfaction, expulsion, and complications. Counseling was provided antenatally and postnatally using standardized IEC materials, models, and checklists.

STATISTICAL ANALYSIS: Categorical variables were presented as numbers and percentages. Chi-square tests and t-tests were applied where appropriate using SPSS 24.0, with p < 0.05 considered statistically significant.

RESULTS: The mean age of participants was 26.9 ± 3.2 years. Continuation rates at 6 and 12 months were significantly higher in term deliveries (96.6% and 85.0%) than in preterm deliveries (88.3% and 63.3%; p < 0.05). Spontaneous IUCD expulsion was higher in preterm women at 6 months (7.5% vs. 1.6%) and 12 months (15.3% vs. 6.4%; p < 0.05), potentially due to smaller uterine size, underdeveloped uterine cavity, and variations in uterine involution. Satisfaction rates were high and comparable between groups at 12 months (81% preterm vs. 85% term; p > 0.05). Acceptance rates were low overall (12%), highlighting the need for structured counseling, provider training, and peer education to improve uptake.

CONCLUSION: Post-placental IUCD insertion is safe, effective, and cost-efficient in both preterm and term deliveries. Higher expulsion rates in preterm deliveries warrant focused counseling and follow-up. Structured antenatal and postnatal counseling, combined with provider confidence and patient education, is essential to increase acceptance and continuation of PPIUCD, particularly in populations at higher risk of early postpartum fertility.

PMID:42098888 | DOI:10.1186/s40834-026-00455-x

Knee Surg Relat Res. 2026 May 7;38(1):19. doi: 10.1186/s43019-026-00320-w.

ABSTRACT

BACKGROUND: The quadriceps tendon (QT) has emerged as a reliable autograft for anterior cruciate ligament reconstruction (ACLR), but uncertainty remains regarding several key comparative aspects-particularly donor-site morbidity, long-term graft survival, knee stability, and complication rates-when evaluated against hamstring tendon (HT) and bone-patellar tendon-bone (BPTB) autografts. High-level evidence restricted to randomized controlled trials directly comparing QT with HT or BPTB remains limited. To compare clinical outcomes, graft failure, donor-site morbidity, and knee stability among QT, HT, and BPTB autografts for primary ACLR using level-I and level-II randomized controlled trials (RCTs).

METHODS: The MEDLINE (PubMed), Embase (Elsevier), and Cochrane Library databases were searched on 1 September 2025, and repeated 2 weeks later. Only level-I or -II RCTs comparing QT to HT or BPTB in primary ACLR were included. Random-effects meta-analyses were performed for International Knee Documentation Committee (IKDC) and Lysholm scores, instrumented laxity, graft failure, donor-site morbidity, and reoperation. Risk of bias was assessed with RoB 2.0, and small-study effects with funnel and doi plots.

RESULTS: Eleven RCTs (mean follow-up, 2-10 years) were included. Pooled IKDC scores averaged 84.8 (95% CI 81.9-87.9) and Lysholm scores averaged 93.1 (95% CI 91.6-94.6), with no significant differences between QT and either comparator (P > 0.05). Side-to-side anterior tibial translation averaged 1.2 mm (95% CI 0.99-1.54 mm) across all grafts, also without significant differences (P > 0.05). Pooled graft failure and ipsilateral reoperation rates were 0.7% (95% CI 0.0-1.9%) and 2.3% (95% CI 0.6-4.7%), respectively, again with no between-graft differences (P > 0.05). Donor-site morbidity did not differ significantly between QT and HT (mean 13.83 [95% CI 9.6-19.83]; P > 0.05).

CONCLUSION: This meta-analysis of level-I/II randomized controlled trials found no statistically significant differences among quadriceps tendon, hamstring tendon, and bone-patellar tendon-bone autografts in patient-reported outcomes, knee stability, graft re-rupture, or additional knee surgery. Donor-site morbidity comparisons were limited by incomplete reporting, particularly for BPTB. These findings suggest that contemporary surgical techniques and rehabilitation protocols may minimize graft-specific differences in mid-term outcomes, although interpretation should consider the limited number of direct comparative trials across all three graft types. Level of evidence Systematic review and meta-analysis; level of evidence, 1 and 2.

PMID:42098886 | DOI:10.1186/s43019-026-00320-w

BMC Psychol. 2026 May 7. doi: 10.1186/s40359-026-04642-y. Online ahead of print.

ABSTRACT

BACKGROUND: Disability poses significant challenges to subjective well-being (SWB), yet many individuals with physical disabilities report unexpectedly high levels of life satisfaction-a phenomenon known as the disability paradox. While quantitative research has identified correlates such as social support and psychological capital, these studies often rely on cross-sectional surveys that capture group trends but fail to illuminate the dynamic, narrative processes through which SWB emerges over the life course, particularly in collectivistic cultures like China where relational and cultural factors may uniquely shape adaptation.

METHODS: This study utilized a mixed-methods approach, employing life tone and life theme analyses within a life story interview framework, supplemented by grounded theory coding procedures. We conducted a qualitative analysis of narratives from 35 Chinese adults with physical disabilities, who also completed psychometric measures of subjective well-being (SWB), social support, and psychological capital. To enhance findings’ robustness, we quantitized narrative indicators and performed triangulation by conducting statistical correlation analyses between these narrative metrics and questionnaire scores within the primary sample (N = 35).

RESULTS: The results indicate: (1) The factors influencing SWB among individuals with physical disabilities primarily consist of four dimensions: family establishment, life experiences, social support, and self-personality; (2) Although the overall emotional tone among Chinese individuals with physical disabilities leans toward negativity, the overall personality tone exhibits a positive orientation, with the frequency of positive emotional words surpassing negative ones starting from mid-adulthood; (3) Life theme analysis demonstrates that relational themes predominate in the narratives, with interpersonal relationships serving as a central pathway for meaning reconstruction and the generation of subjective well-being; and (4) Significant associations were observed between social support, psychological capital, and the emotional and personality features in the narratives, indicating the joint involvement of internal and external resources in shaping SWB among individuals with physical disabilities.

CONCLUSIONS: In summary, this study, grounded in a localized narrative perspective across the full life course, elucidates the factors influencing SWB among Chinese individuals with physical disabilities and their cultural specificity, particularly emphasizing the key role of relational factors in fostering resilience and meaning construction within a collectivistic value framework.

PMID:42098881 | DOI:10.1186/s40359-026-04642-y

BMC Psychol. 2026 May 7. doi: 10.1186/s40359-026-04691-3. Online ahead of print.

ABSTRACT

This study aimed to investigate the associations linking nature connection and pro-environmental behavior (PEB) in adolescents by constructing a chain mediation model, with a focus on the mediating effects of empathy with nature and moral identity. A cross-sectional questionnaire survey was administered to 379 adolescents aged 12-16 from three cities in Yunnan Province, China. The findings revealed that: (1) significant positive correlations were observed among adolescents’ nature connection, empathy with nature, moral identity, and pro-environmental behaviors; (2) moral identity and empathy with nature are consistent with a chain mediating role between adolescents’ nature connections and pro-environmental behaviors. This study suggests cognitive-emotional pathways that are statistically consistent with a link from nature connection to pro-environmental behavior among adolescents, providing a potential theoretical basis and practical guidance for fostering such behavior.

PMID:42098880 | DOI:10.1186/s40359-026-04691-3

J Cardiothorac Surg. 2026 May 7. doi: 10.1186/s13019-026-04251-1. Online ahead of print.

ABSTRACT

BACKGROUND: Large language models (LLMs) are increasingly being explored in surgical training and clinical knowledge assessment. Although these models have demonstrated promising performance in standardized examinations, their performance in highly specialized fields such as cardiovascular surgery remains insufficiently investigated. This study aimed to evaluate the performance of current large language models in answering and generating board-level cardiovascular surgery questions reflecting guideline-based clinical reasoning.

METHODS: In this cross-sectional evaluation study, three large language models (ChatGPT-5.1, Gemini 3, and DeepSeek v3.2) were evaluated in two stages. In the first stage, the models answered 150 multiple-choice questions developed and validated by five cardiovascular surgery specialists using a Delphi process, designed to reflect the content scope and difficulty level of the American Board of Thoracic Surgery certification examination. Accuracy rates and pairwise comparisons were analyzed using the McNemar test. In the second stage, model-generated questions were evaluated by expert cardiovascular surgeons in terms of medical accuracy, clinical relevance, exam-level appropriateness, error type, and difficulty level. Statistical analyses included Spearman correlation, Wilcoxon signed-rank test, and chi-square analysis.

RESULTS: The models demonstrated comparable accuracy rates (ChatGPT 80.7%; Gemini 78.7%; DeepSeek 82.0%), with no statistically significant differences between them. Question difficulty level was not associated with model accuracy. Error distribution differed significantly between models (χ² = 8.1; p = 0.02), with Gemini demonstrating the highest rate of valid question generation and DeepSeek showing a higher rate of major errors. A significant positive correlation was observed between model- and expert-assigned difficulty levels.

CONCLUSIONS: Current large language models demonstrate strong performance in board-level cardiovascular surgery knowledge assessment. However, the presence of major errors and variability in difficulty calibration, together with known limitations in clinical reasoning, indicate that these systems should be used cautiously as supportive tools in surgical training and knowledge assessment rather than as substitutes for clinical decision-making.

PMID:42098878 | DOI:10.1186/s13019-026-04251-1

Genome Med. 2026 May 7. doi: 10.1186/s13073-026-01665-3. Online ahead of print.

ABSTRACT

BACKGROUND: Compared to European American women, African American women are more likely to be diagnosed with triple-negative breast cancer (TNBC). This difference may be partially due to genetic factors. This study aims to investigate associations of African ancestry and risk variants with TNBC among African American women.

METHODS: We used data from 2,335 TNBC cases, 8,159 estrogen receptor (ER)-positive cases, and 9,814 controls included in the African-ancestry Breast Cancer Genetics (AABCG) Consortium. The proportion of African ancestry (%AFR) and local ancestry were estimated using samples from the 1000 Genomes Project as reference. Logistic regressions were performed for case-control (TNBC vs. control) and case-case (TNBC vs. ER-positive) comparisons, adjusted for age, study, genotype principal components 2-5, body mass index, and reproductive factors. Local ancestry-aware association analyses were conducted in 12 TNBC risk loci to identify ancestry-specific risk variants.

RESULTS: In case-control analyses, no statistically significant association was found between %AFR and TNBC risk after adjustment for potential confounders. However, TNBC cases had a significantly higher mean % AFR (mean = 0.811, standard deviation, SD = 0.104) compared to ER-positive cases (mean = 0.798, SD = 0.110, P < 0.001). Females with %AFR of ≥ 95% had 1.62 times higher odds (95% confidence interval, CI: 1.16-2.25) of having TNBC rather than ER-positive breast cancer, compared to those with %AFR of 55.0-64.9%. Local ancestry-aware association analyses identified seven subtype-informative variants in or near MDM4, RP11-19E11.1, TERT, MRPL36, TCF7L2, C11orf65, and ANKLE1. All of them were significantly associated with TNBC as compared with ER-positive cases, and six of them were also associated with TNBC risk in case-control analyses. Large allelic odds ratios of 1.25 or higher were found in association with TNBC risk or subtype classification. The risk allele frequency for five of them is substantially higher in haplotypes of African ancestry than those of European ancestry.

CONCLUSIONS: These findings support a significant role of African-ancestry specific genetic factors in determining breast cancer subtypes and highlight the need for future research to uncover possible pathways driving TNBC susceptibility.

PMID:42098873 | DOI:10.1186/s13073-026-01665-3