Tag: statistics

Nutr J. 2026 Jan 22. doi: 10.1186/s12937-025-01279-2. Online ahead of print.

NO ABSTRACT

PMID:41572290 | DOI:10.1186/s12937-025-01279-2

Cardiovasc Diabetol. 2026 Jan 22. doi: 10.1186/s12933-025-03073-0. Online ahead of print.

ABSTRACT

BACKGROUND: The COLCOT trial showed that patients with diabetes may benefit from low-dose colchicine, suggesting a potential interplay between insulin resistance (IR) and inflammation. Whether their combined assessment improves mortality risk stratification in the general population remains unclear.

METHODS: We analyzed 50,654 adults from NHANES 1999-2018 linked to the National Death Index. IR and inflammation were assessed using estimated glucose disposal rate (eGDR) and the log₂-transformed aggregate index of systemic inflammation (AISI), respectively. Survey-weighted Cox proportional hazards models were used for all-cause mortality. For cardiovascular (CVD) mortality, cumulative incidence functions (CIFs) were estimated with Gray’s test for between-group comparisons, and Fine-Gray subdistribution hazard models were fitted treating non-CVD death as a competing event. Discrimination was assessed using time-dependent ROC curves at 5 and 10 years. Robustness was evaluated through sensitivity analyses excluding immune-modifying conditions/treatments, applying a 24-month lag, and excluding extreme absolute lymphocyte counts.

RESULTS: Over a median follow-up of 120 months, 6,936 all-cause deaths and 2,170 CVD deaths occurred. Higher eGDR was inversely associated with mortality (all-cause HR per 1-unit increase 0.90, 95% CI 0.88-0.92; CVD sHR 0.88, 95% CI 0.85-0.91), whereas higher log₂(AISI) was positively associated (all-cause HR per doubling 1.10, 95% CI 1.06-1.15; CVD sHR 1.13, 95% CI 1.06-1.20). In joint analyses, participants with low eGDR (≤ 8.40) and high log₂(AISI) (> 7.98) had the highest risks of all-cause mortality (HR 1.58, 95% CI 1.38-1.81) and CVD mortality (cause-specific HR 2.09, 95% CI 1.58-2.77; Fine-Gray sHR 2.13, 95% CI 1.66-2.74), with graded separation of CIFs (Gray’s test P < 0.001). The combined model showed improved discrimination (AUCs at 5/10 years: all-cause 0.705/0.723; CVD 0.754/0.769). Results were consistent across sensitivity analyses.

CONCLUSION: In a nationally representative U.S. cohort, eGDR and log₂(AISI) were independently and jointly associated with all-cause and CVD mortality. Their combined assessment improves risk stratification and may help identify individuals most likely to benefit from targeted preventive and anti-inflammatory strategies.

PMID:41572288 | DOI:10.1186/s12933-025-03073-0

J Neuroeng Rehabil. 2026 Jan 22;23(1):28. doi: 10.1186/s12984-025-01829-z.

ABSTRACT

BACKGROUND: Digital health technologies (DHTs) can quantify movements in daily routines but rely heavily on participant adherence over prolonged wear times.

METHODS: We analyzed accelerometry data from wrist-worn devices during short at-home episodes of prescribed exercises performed by 329 individuals living with amyotrophic lateral sclerosis (ALS) in a longitudinal study. We developed an automated and interpretable signal processing method to estimate four metrics describing exercise repetitions, i.e., their count, duration, intensity, and similarity. We examined their associations with time elapsed from enrollment and ALS Functional Rating Scale-Revised (ALSFRS-R) using linear mixed effect models. We also compared them with previously validated free-living metrics that require substantial sensor wear-time. Finally, we studied how many repetitions are sufficient to determine participants’ upper limb functioning.

RESULTS: Three out of four exercise metrics (all but count) demonstrated significant association with ALSFRS-R outcomes. The duration of exercise repetitions increased, while intensity and similarity of movement decreased over time (all p-value < 0.001), indicating longer but less vigorous and less consistent upper limb movements over time. Exercise intensity was determined as the most robust exercise-based predictor of changes in upper limb function, and it was comparable to free-living metrics, which required at 21 h of sensor wear time (R-squared 0.899 vs. 0.860, respectively). Sensitivity analysis indicated that as few as five exercise repetitions were sufficient to yield statistically significant associations with ALSFRS-R.

CONCLUSIONS: These results suggest that prescribed exercise can effectively quantify upper limb function and track longitudinal decline comparably to free-living observation. The proposed method may serve as an alternative that decreases participation burden, increases study adherence, and extends diagnostic accessibility.

PMID:41572285 | DOI:10.1186/s12984-025-01829-z

BMC Vet Res. 2026 Jan 23. doi: 10.1186/s12917-026-05303-3. Online ahead of print.

ABSTRACT

Feline immunodeficiency virus (FIV) induces immunosuppression in affected cats, increasing susceptibility to chronic and secondary infections. Rapid and accurate detection of FIV-specific antibodies is essential for effective clinical management and epidemiological monitoring. This study conducted a comparative evaluation of nine commercially available lateral flow assays (LFAs) for detecting FIV antibodies in whole blood, serum, or plasma, using a newly developed in-house enzyme-linked immunosorbent assay (ELISA) as a reference method. All tested LFAs demonstrated 100% specificity. While Vet Expert new and VetFor achieved 100% across all metrics indicating the best performance, formal statistical comparison did not reveal significant differences between the evaluated kits. Overall, the results confirm that all tested LFAs offer comparable reliability. Importantly, our in-house ELISA exhibited 100% concordance for positive samples with the commercial ELISA treated as the reference standard, confirming its reliability as a comparator. These findings emphasize the importance of selecting high-performing diagnostic tools to ensure reliable FIV detection and effective disease control strategies.

PMID:41572271 | DOI:10.1186/s12917-026-05303-3

BMC Med. 2026 Jan 22. doi: 10.1186/s12916-026-04649-7. Online ahead of print.

ABSTRACT

BACKGROUND: The anatomical and pathophysiological characteristics of coronary artery disease vary between the sexes. This study investigated the impact of sex on outcomes in patients with de novo coronary artery lesions treated with drug-coated balloons (DCB) or drug-eluting stents (DES).

METHODS: REC-CAGEFREE I was an investigator-initiated, non-inferiority trial conducted at 43 sites in China from Feb 5, 2021, to May 1, 2022, which randomized 2,272 patients for treating de novo coronary lesions, regardless of vessel diameter. After successful lesion pre-dilatation, eligible patients were randomized (1:1) to either DCB angioplasty with the option of rescue stenting or intended DES deployment. In this prespecified subgroup analysis, patients were analyzed by sex based on their medical records. The primary endpoint was device-oriented composite endpoint (DoCE), including cardiovascular death, target-vessel myocardial infarction, and clinically and physiologically indicated target lesion revascularization at 2 years. Between-group differences were compared by Cox proportional-hazards models, and imbalances in baseline characteristics were adjusted with inverse probability of treatment weighting (IPTW). The analyses were conducted in the intention-to-treat population.

RESULTS: A total of 2,272 participants underwent randomization, of which 698 (30.7%) were female and 1,574 (69.3%) were male. At 2 years, no statistically significant differences in the incidence of DoCE were observed between sexes (36 [5.2%] for females and 74 [4.7%] for males, HR_IPTW:1.04, 95%CI:0.67 to 1.61, P = 0.877). Compared with DES, DCB was associated with a numerically higher risk of DoCE in females (6.3% versus 3.9%, HR_IPTW:1.55, 95%CI:0.78 to 3.11, P = 0.210) and a statistically significant higher risk in males (6.4% versus 3.1%, HR_IPTW:2.28, 95%CI:1.40 to 3.70, P = 0.001), respectively, with no significant sex-by-treatment (DCB/DES) interaction observed (P_interaction = 0.575). The prognosis of DCB and DES differed significantly between small vessel disease (SVD) and non-SVD among females (P_interaction = 0.007), but not among males (P_interaction = 0.408).

CONCLUSIONS: For patients with de novo, non-complex coronary artery disease, DCB was associated with a significantly higher risk of 2-year DoCE compared with DES in males, whereas a consistent but non-significant trend was observed in females.

TRIAL REGISTRATION: ClinicalTrials.gov identifier: NCT04561739.

PMID:41572254 | DOI:10.1186/s12916-026-04649-7

Stat Med. 2026 Jan;45(1-2):e70392. doi: 10.1002/sim.70392.

ABSTRACT

Clustering longitudinal biomarkers in clinical trials uncovers associations between clinical outcomes, disease progression, and treatment effects. Finite mixtures of multivariate $t$ linear mixed-effects (FM-MtLME) models have proven effective for modeling and clustering multiple longitudinal trajectories that exhibit grouped patterns with strong within-group similarity. Motivated by an AIDS study with plasma viral loads measured under assay-specific detection limits, this article extends the FM-MtLME model to account for censored outcomes. The proposed model is called the FM-MtLME with censoring (FM-MtLMEC). To allow covariate-dependent mixing proportions, we further extend it with a logistic link, resulting in the EFM-MtLMEC model. Two efficient EM-based algorithms are developed for parameter estimation of both FM-MtLMEC and EFM-MtLMEC models. The utility of our methods is demonstrated through comprehensive analyses of the AIDS data and simulation studies.

PMID:41569638 | DOI:10.1002/sim.70392

JMIR Public Health Surveill. 2026 Jan 22;12:e83488. doi: 10.2196/83488.

ABSTRACT

BACKGROUND: Cardiovascular disease (CVD) remains the leading cause of death. Primary prevention relies heavily on health risk assessments and lifestyle changes, which can reduce long-term risk and mortality. Digital health offers an accessible and cost-effective approach to support prevention, enabling data sharing and visualization of key indicators such as blood pressure and glucose fluctuations. These visual insights may help patients better understand the effects of lifestyle changes and enhance communication with health care providers.

OBJECTIVE: This research aims to evaluate whether the use of CVD risk visualization (Petal-X) and continuous glucose monitoring (CGM), alone or in combination, is associated with lifestyle changes and the perception of person-centered care (PCC) among patients at increased risk of CVD.

METHODS: We conducted a 4-arm, single-blind, 2×2 factorial randomized controlled feasibility trial in primary care. A total of 119 participants were enrolled, of whom 101 completed the 6-month follow-up. Participants were randomized to 1 of 4 arms: (1) Petal-X CVD risk visualization+CGM; (2) CGM only; (3) Petal-X only; or (4) standard care with routine lifestyle counseling and no digital tools. CGM was used for 10 days in the CGM arms. Since this was a feasibility trial, no formal sample size calculation was performed. Primary outcomes are healthy lifestyle (Health Lifestyle and Personal Control Questionnaire [HLPCQ]) and perception of PCC (Person-Centered Practice Inventory-Service User [PCPI-SU]), and secondary outcomes (Systematic Coronary Risk Evaluation 2 [SCORE2], anthropometrics, and biological age) were assessed at baseline and 6 months. Descriptive statistics and Kruskal-Wallis tests (K independent samples) were used for analyses.

RESULTS: At baseline, mean SCORE2 values ranged from 3.84 (SD 2.08) in intervention group 3 to 4.87 (SD 2.61) in intervention group 1, with the control group having a mean value of 4.53 (SD 3.63). Regarding the assessment of a healthy lifestyle, the domain of daily routine had the highest baseline scores across all groups (eg, mean 19.24, SD 5.87 in intervention group 1), and these scores improved by the final evaluation, although there were no statistically significant differences (P=.42) in changes between the groups. The perception of PCC was rated highest across all groups in the domain of shared decision-making, with no statistically significant differences (P=.26) between the groups. Results indicated improvements in healthy lifestyle habits, but the impact of interventions on perceived changes remained insignificant.

CONCLUSIONS: Healthy lifestyle and perceived PCC scores improved, although no statistically significant between-group differences were found. Risk visualization appears to be a key tool for increasing CVD awareness and strengthening patient involvement in care planning. Longer interventions with larger samples are needed to clarify these effects and optimize digital tools for lifestyle change.

PMID:41569629 | DOI:10.2196/83488

Stat Med. 2026 Jan;45(1-2):e70350. doi: 10.1002/sim.70350.

ABSTRACT

We present a novel framework for concomitant dimension reduction and clustering. This framework is based on a novel class of Bayesian clustering factor models. These models assume a factor model structure where the vectors of common factors follow a mixture of Gaussian distributions. We develop a Gibbs sampler to explore the posterior distribution and propose an information criterion to select the number of clusters and the number of factors. Simulation studies show that our inferential approach appropriately quantifies uncertainty. In addition, when compared to two previously published competitor methods, our information criterion has favorable performance in terms of correct selection of number of clusters and number of factors. Finally, we illustrate the capabilities of our framework with an application to data on recovery from opioid use disorder where clustering of individuals may facilitate personalized health care.

PMID:41569628 | DOI:10.1002/sim.70350

Stat Med. 2026 Jan;45(1-2):e70333. doi: 10.1002/sim.70333.

ABSTRACT

Combining predictions from multiple models into an ensemble is a widely used practice across many fields with demonstrated performance benefits. Popularized through domains such as weather forecasting and climate modeling, multi-model ensembles are becoming increasingly common in public health and biological applications. For example, multi-model outbreak forecasting provides more accurate and reliable information about the timing and burden of infectious disease outbreaks to public health officials and medical practitioners. Yet, understanding and interpreting multi-model ensemble results can be difficult, as there are a diversity of methods proposed in the literature with no clear consensus on which is best. Moreover, a lack of standard, easy-to-use software implementations impedes the generation of multi-model ensembles in practice. To address these challenges, we provide an introduction to the statistical foundations of applied probabilistic forecasting, including the role of multi-model ensembles. We introduce the hubEnsembles package, a flexible framework for ensembling various types of predictions using a range of methods. Finally, we present a tutorial and case-study of ensemble methods using the hubEnsembles package on a subset of real, publicly available data from the FluSight Forecast Hub.

PMID:41569627 | DOI:10.1002/sim.70333

Stat Med. 2026 Jan;45(1-2):e70397. doi: 10.1002/sim.70397.

ABSTRACT

Deep learning has excelled in the field of statistical learning. In the field of survival analysis, some studies have combined deep learning methods with partially linear structures to propose deep partially linear structures. We extend it to the field of competing risks and propose the deep partially linear subdistribution hazard model (DPLSHM). To evaluate the predictive performance of the model, we further develop a time-dependent AUC method specifically tailored for competing risks data and provide an estimator for AUC. Theoretical results for the proposed model demonstrate the asymptotic normality of the parameter component at a rate of $\sqrt{n}$ and provide the convergence rate of the nonparametric component, which achieves the minimal limit convergence rate (multiplicative logarithmic factors). The theory of consistency and rate of convergence of AUC-related estimates is also developed, while we prove that the regression component of DPLSHM maximizes theoretical AUC asymptotically. Subsequently, the paper validates the excellent performance of DPLSHM in estimation and prediction through numerical simulations and real-world datasets.

PMID:41569618 | DOI:10.1002/sim.70397