Category: Statistics

Eur J Public Health. 2025 Nov 29:ckaf217. doi: 10.1093/eurpub/ckaf217. Online ahead of print.

ABSTRACT

There is a need to efficiently identify groups at risk of unmet health service needs. In response, we developed and evaluated the performance of a regression model to assess unmet health service needs in the Finnish population. The study population consisted of population-based Healthy Finland 2022-cohort participants (N = 18 442), aged 20-104. The primary outcome was self-reported unmet need for physician’s or nurse’s services. A total of 38 potential risk factors were evaluated. Statistical models were developed using bootstrap-enhanced LASSO regression (bolasso). Of the participants, 5875 (32%) were classified as experiencing unmet health care need. The C-index from the final model including 15 predictors from the best bolasso models varied between 0.73 and 0.76 and pooled C-index over the imputed data sets was 0.75 (95% CI 0.70-0.79). Fifteen factors-including health-related, socioeconomic variables, heavy alcohol use, experiences with health services, caregiving for others, and language group-were found to be strongly associated with an increased risk of unmet health care needs and may be a useful targets for preventing unmet health care need.

PMID:41351301 | DOI:10.1093/eurpub/ckaf217

J Clin Lab Anal. 2025 Dec 5:e70136. doi: 10.1002/jcla.70136. Online ahead of print.

ABSTRACT

BACKGROUND: This study investigates the potential interplay between gut microbiota and circulating cytokines in Sjögren’s syndrome (SS) through a bidirectional and mediation Mendelian randomization (MR) approach.

METHODS: Summary-level statistics of 473 gut microbiota (n = 5959), 41 circulating cytokines (n = 8293), and SS (ncase = 2735, ncontrol = 399,355) were obtained from genome-wide association studies (GWAS) in European populations. A two-sample MR analysis was employed to investigate the bidirectional causal effects of gut microbiota and circulating cytokines on SS, and mediation analyses were applied to discover potential mediating gut microbiota and circulating cytokines. A series of sensitivity analyses were conducted to address heterogeneity and pleiotropy concerns.

RESULTS: Fifteen taxa were found to be causally associated with SS, and SS had a causal effect on 26 taxa. A bidirectional causal relationship was identified between CAG-269 sp001916065 and SS, and between UBA7703 and SS. Genetically predicted levels of five circulating cytokines-MIG, IL-5, IL-1RA, IL-2RA, and SCGF-β-were found to potentially affect SS, and genetically predicted SS was associated with increased levels of two circulating cytokines, IL-1β and IL-5. A bidirectional causal relationship was identified between circulating IL-5 and SS. Mediation analyses further revealed that circulating cytokines do not mediate the gut microbiome’s influence on SS, and conversely, the gut microbiome does not influence circulating cytokines to affect SS.

CONCLUSION: This study provides compelling evidence for causal effects of gut microbiome composition and circulating cytokines on SS risk. Further mediation analysis suggests that these biological factors may operate independently to influence SS development.

PMID:41351294 | DOI:10.1002/jcla.70136

Transl Behav Med. 2025 Jan 16;15(1):ibaf081. doi: 10.1093/tbm/ibaf081.

ABSTRACT

BACKGROUND: Rising use of electronic nicotine delivery systems (ENDS) and tobacco products among rural youth challenges tobacco control efforts, with higher use rates compared to urban youth. Exposure to tobacco retail outlets (TROs) and marketing influences tobacco use, but the longitudinal pathways of these factors in rural youth are underexplored.

PURPOSE: This study examines the feasibility of app-based ecological momentary assessment (EMA) in capturing real-time exposure to TROs and tobacco marketing and its link to tobacco use and susceptibility among rural youth.

METHODS: Participants (n = 25) residing in Southwest Virginia were recruited for a 14-day EMA sub-study via the Effortless Assessment Research System (EARS) app. We assessed exposure to TROs and tobacco marketing as predictors and tobacco use susceptibility and past 24-hour tobacco use as outcomes. GPS data identified exposure to TROs within 100 m, and associations with tobacco outcomes were examined using generalized linear mixed-effect regression models.

RESULTS: Most participants were 9th graders (36%; range 9th grade-12th grade) and white (56%), with slightly more female (56%) than male (44%). One-third were tobacco-susceptible, and 13% had used tobacco. TRO exposure was higher in activity space outside of the home and school (M = 7.2 exposures) than near home (4.1) or school (2.1). Over 14 days, 328 EMA responses were collected from 25 participants (72.9% response rate), demonstrating EMA’s feasibility. TRO exposure was positively associated with recent tobacco use (P = .04) and negatively associated with recalling combustible tobacco ads (P = .03).

CONCLUSIONS: This study examined the feasibility of app-based EMAs to track rural youth’s exposure to TROs, tobacco marketing, and tobacco outcomes. Findings demonstrate the feasibility of EMA for capturing tobacco-related exposures among adolescents.

PMID:41351280 | DOI:10.1093/tbm/ibaf081

Equine Vet J. 2025 Dec 5. doi: 10.1111/evj.70126. Online ahead of print.

ABSTRACT

BACKGROUND: The widespread assumption that there is minimal potential for pathogen transmission between British racehorse and sport horse populations remains unverified by empirical evidence.

OBJECTIVES: To characterise spatiotemporal patterns of horse attendance at racing and other sport events in Great Britain in 2018.

STUDY DESIGN: Spatiotemporal analysis.

METHODS: Publicly available data from British Horseracing Authority, British Dressage, British Eventing, Endurance GB, and British Showjumping events in Great Britain during 2018 were analysed. Horse attendance was summarised by discipline, month, and season. Venue density was mapped with kernel density estimation. Sport venues within 5 km of racecourses with horse attendance within 24 h of racing were identified and Kulldorff’s spatial scan statistic was used to detect significant time-space clustering of venue use.

RESULTS: Excluding showjumpers, 49,012 horses competed in 8314 events across 598 venues during 2018, generating over 400,000 horse-venue attendances. Most horses (97.2%; n = 47,635/49,012) competed in a single discipline. Venue attendances peaked in summer and were concentrated in southeast England. There were five significant space-time clusters of venue-events within 5 km and 24 h of each other involving 5 racecourses and 8 sport venues. The most likely cluster was in the southeast of England, between January and July, with a relative risk of 62.54.

MAIN LIMITATIONS: Inconsistent horse identification precluded horse-level analysis of showjumping data.

CONCLUSIONS: Racehorse and sport horse populations competing in Great Britain are largely separate, but limited opportunities for local or indirect pathogen spread do exist during peak seasons in areas with high venue density.

PMID:41351275 | DOI:10.1111/evj.70126

HGG Adv. 2025 Dec 4:100552. doi: 10.1016/j.xhgg.2025.100552. Online ahead of print.

ABSTRACT

Despite considerable advances in identifying risk factors for obesity, gaps remain in our understanding about its etiology. Genetic variants explain only a small portion of variation in obesity-related traits such as body mass index (BMI). Epigenetic regulation, which controls gene expression and is influenced by environmental and genetic factors, may account for additional variability in BMI. Epigenetic studies of BMI have largely been conducted in European ancestry populations, despite the disproportionate burden of obesity in African Americans (AAs). We conducted a sex-stratified BMI epigenome-wide association study (EWAS) meta-analysis in AA participants from the Jackson Heart Study (JHS, n=1604) and the Multi-Ethnic Study of Atherosclerosis (MESA, n=179) with Illumina EPIC (850k) array data. Linear regression models with methylation as the outcome and continuous BMI as the predictor were stratified by study and sex and meta-analyzed. We identified 208 methylation sites (CpGs, p< 8.72×10^-8) significantly associated with BMI; 151 not been previously reported in literature. Replication was performed in a separate sample of AA participants with 450k array data, which lacks many CpGs present in the 850k array. Replication testing was possible for only 29 of the 151 CpGs; 19 were statistically significant (p<1.72×10^-3). Sex-specific results showed 4 female-only and 3 male-only BMI-CpGs not identified in the sex-combined results. Differentially methylated region (DMR) analysis resulted in 66 DMRs, including several regions near genes previously implicated for obesity (e.g., SOCS3 and TGFB1). Further analyses showed enrichment of genes and traits related to the immune system and inflammation-related pathways (e.g., the IL-6/JAK/STAT pathway).

PMID:41351263 | DOI:10.1016/j.xhgg.2025.100552

Stat Med. 2025 Dec;44(28-30):e70334. doi: 10.1002/sim.70334.

ABSTRACT

With the rapid advancements in wearable device technologies, there is a growing interest in learning useful digital biomarkers from wearable device data as objective, low-cost, real-time alternatives to use in healthcare settings. They have the potential to facilitate disease progression monitoring, medication tailoring, and supplementing clinical trial endpoints. For example, triaxial accelerometer sensor data is promising for monitoring symptoms of movement-related diseases, such as tremors in Parkinson’s disease (PD). However, existing methods for accelerometer studies based on hidden Markov models (HMM) often analyze each individual’s activity data separately, leading to inefficiency and limited generalizability. This paper proposes a joint nonparametric Bayesian method that extends the hierarchical Dirichlet process autoregressive HMM (HDP-AR-HMM) to incorporate subject-specific transition parameters. This approach allows for simultaneous estimation across multiple subjects and repeated measurements, accounts for between-subject variability, and provides consistent hidden state estimation without pre-specifying the number of states. We validate our method on simulated data and show that it can achieve higher accuracy in detecting the true hidden states compared to alternative methods. We apply the method to a free-living study, the Biomarker & Endpoint Assessment to Track Parkinson’s disease (BEAT-PD) DREAM Challenge CIS-PD study, to demonstrate its utility in monitoring disease symptoms in PD patients.

PMID:41351259 | DOI:10.1002/sim.70334

J Obstet Gynaecol Res. 2025 Dec;51(12):e70164. doi: 10.1111/jog.70164.

ABSTRACT

AIM: Previous studies reported inconsistent results on the effects of maternal smoking and drinking on pre-eclampsia (PE). Furthermore, some studies indicated that the development of PE may also be attributable to unhealthy behaviors from men. Hence, this study aimed to evaluate the association between smoking and drinking by couples and PE in the Asian population.

METHODS: A prospective cohort study including 34 104 couples was conducted. Multivariate logistic regression models were used to estimate the odds ratios (ORs) of couples’ smoking and drinking behaviors and their interaction. The assignment score method was used to explore the cumulative effect of adverse behavioral exposures to alcohol and tobacco on PE.

RESULTS: Seven hundred eighty-eight-pregnant women (2.31%) were diagnosed with PE. Maternal active smoking 3 months before pregnancy (OR: 2.010, 95% CI: 1.136-3.555), periconceptional active smoking (OR: 1.560, 95% CI: 1.027-2.368), drinking 3 months before pregnancy (OR: 2.101, 95% CI: 1.397-3.158), and periconceptional drinking (OR: 1.829, 95% CI: 1.318-2.539) were positively associated with PE. Spousal smoking was also a risk factor (OR: 1.174, 95% CI: 1.009-1.366). Additionally, there was an antagonistic effect between maternal active smoking and drinking during the periconceptional period. Moreover, with the increase of bad behaviors of couples, the risk of PE also increased to a certain extent.

CONCLUSIONS: Maternal active smoking and drinking as well as spousal smoking were risk enhancers of PE in the Asian population. We encourage both couples to actively quit smoking and drinking from the beginning of pregnancy preparation.

PMID:41351256 | DOI:10.1111/jog.70164

Clin Pharmacol Ther. 2025 Dec 5. doi: 10.1002/cpt.70158. Online ahead of print.

ABSTRACT

The European Medicines Regulatory Network plays a critical role in safeguarding public health through the assessment and supervision of human medicines. However, rapid scientific and technological advancements have exposed capability deficits across the network, threatening timely and effective regulatory decision-making. To identify the most critical learning needs within the European Medicines Regulatory Network-specifically those related to capability gaps-the EU4Health Joint Action, IncreaseNET, conducted a learning needs analysis. This study presents a learning needs analysis conducted via two complementary approaches: stakeholder collaboration with European Medicines Agency expert committees and a network-wide survey of National Competent Authorities. The analysis identified six priority subject areas requiring urgent capability development: Advanced Therapy Medicinal Products, Biologically Active Substances, Clinical Trials, Modeling and Simulation, Pharmacokinetics, and Statistics. Additional training needs-such as Environmental Risk Assessment and Radiopharmaceuticals-were highlighted through the National Competent Authority survey. Key barriers to capability development include uneven distribution of expertise, limited time for training due to high workloads, and recruitment challenges. The study also revealed methodological variation in how learning needs analyses are conducted across the European Medicines Regulatory Network, underscoring the need for a standardized yet context-sensitive framework. IncreaseNET will pilot innovative training development strategies, including the use of pedagogical specialists, formalized training processes, and targeted expert recruitment initiatives. These efforts aim to build a more agile, capable, and resilient regulatory workforce across Europe.

PMID:41351255 | DOI:10.1002/cpt.70158

Clin Spine Surg. 2025 Oct 31. doi: 10.1097/BSD.0000000000001960. Online ahead of print.

ABSTRACT

STUDY DESIGN: Retrospective cohort study.

OBJECTIVE: To compare the treatment outcomes of lateral lumbar interbody fusion (LLIF) combined with lateral plate fixation (LLIF+LP) and LLIF stand-alone (LLIF-SA).

BACKGROUND: Biomechanical studies suggest that LLIF+LP provides greater biomechanical strength than LLIF-SA. However, limited clinical research has shown no significant differences in fusion and cage subsidence rates between the 2 techniques at follow-up periods longer than 1 year. This study observed the occurrence of a unique fusion pattern, EVB, during the early postoperative period.

MATERIALS AND METHODS: The study included 57 patients in the LLIF-SA group (88 levels) and 68 patients in the LLIF+LP group (110 levels). The outcomes assessed included patient-reported outcomes, cage subsidence rates, fusion rates, lateral cage migration (LCM), and the incidence of EVB at various postoperative time points.

RESULTS: At 6 months postoperatively, the incidence of EVB significantly differed between the 2 groups: 44.9% in the LLIF+LP group and 27.7% in the LLIF-SA group (P<0.05). However, no significant differences were observed at 3 months and 12 months postoperatively (P>0.05). At the end of the follow-up period, the overall fusion rate was comparable between the groups: 98.1% for LLIF+LP and 97.6% for LLIF-SA (P>0.05). No statistically significant difference was found in overall cage subsidence rates, but a significant difference was observed in severe subsidence rates: 27.3% in the LLIF-SA group versus 15.5% in the LLIF+LP group (P<0.05). The LCM rates were 6% in the LLIF-SA group and 0% in the LLIF+LP group (P<0.05).

CONCLUSIONS: Both surgeries yielded satisfactory treatment outcomes. The addition of a lateral plate reduced severe subsidence and lateral migration rates of the cage. While lateral plate fixation did not significantly affect long-term fusion rates, it showed superior fusion advantages in the early postoperative period (6 mo).

PMID:41351240 | DOI:10.1097/BSD.0000000000001960

J Cachexia Sarcopenia Muscle. 2025 Dec;16(6):e70145. doi: 10.1002/jcsm.70145.

ABSTRACT

BACKGROUND: Being underweight is an underrecognized risk factor for mortality among individuals with type 2 diabetes (T2D). This study aimed to evaluate the associations of underweight status with mortality among individuals with T2D.

METHODS: This nationwide, retrospective, population-based cohort study used data from the Korean National Health Insurance Service. We included 1 788 996 adults with T2D who underwent baseline screening between 1 January 2015 and 31 December 2016, with follow-up through December 31, 2022. Underweight was defined as body mass index (BMI) < 18.5 kg/m² and further categorized as mild (17.0-18.4 kg/m²), moderate (16.0-16.9 kg/m²), and severe (< 160 kg/m²). The primary outcome was all-cause mortality analysed using multivariable Cox proportional hazards regression.

RESULTS: During a median follow-up of 6.96 years, 176 056 deaths occurred. Compared with the non-underweight group (BMI ≥ 18.5 kg/m²), all-cause mortality increased stepwise across underweight categories, with adjusted hazard ratios (aHRs) of 2.037 (95% CI, 1.982-2.094) for mild underweight, 2.719 (2.587-2.857) for moderate underweight, and 3.876 (3.646-4.119) for severe underweight. Cause-specific mortality showed similar gradients. For diabetes-related mortality, the aHRs were 2.265 (2.073-2.474), 3.306 (2.845-3.843) and 5.136 (4.300-6.134) across mild, moderate and severe underweight, respectively; for cardiovascular mortality, 1.881 (1.721-2.055), 2.087 (1.751-2.487) and 2.825 (2.267-3.519); and for cerebrovascular mortality, 2.100 (1.881-2.344), 2.300 (1.846-2.866) and 3.501 (2.702-4.537). Notably, the severe underweight group (BMI < 16.0 kg/m²) showed significantly higher all-cause mortality than the severe obesity group (BMI ≥ 35.0 kg/m²) when compared to the BMI 25.0-29.9 kg/m² reference group (aHR [95% CI]: 5.168 [4.857-5.499] vs. 1.504 [1.418-1.595]).

CONCLUSIONS: Among individuals with T2D, underweight status was associated with substantially elevated mortality risks, with severe underweight exhibiting greater risk than severe obesity.

PMID:41351230 | DOI:10.1002/jcsm.70145