Category: Statistics

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

Stat Med. 2026 Jan;45(1-2):e70360. doi: 10.1002/sim.70360.

ABSTRACT

Unmeasured confounders pose a major challenge in accurately estimating causal effects in observational studies. To address this issue when estimating hazard ratios (HRs) using Cox proportional hazards models, several methods, including instrumental variables (IVs) approaches, have been proposed. However, these methods often face limitations, such as weak IV problems and restrictive assumptions regarding unmeasured confounder distributions. In this study, we introduce a novel nonparametric Bayesian procedure that provides accurate HR estimates while addressing these limitations. A key assumption of our approach is that unmeasured confounders exhibit a cluster structure. Under this assumption, we integrate two remarkable Bayesian techniques, the Dirichlet process mixture (DPM) and general Bayes (GB), to simultaneously (1) detect latent clusters based on the likelihood of exposure and outcome variables and (2) estimate HRs using the likelihood constructed within each cluster. Notably, leveraging DPM, our procedure eliminates the need for IVs by identifying unmeasured confounders under an alternative condition. Additionally, GB techniques remove the need for explicit modeling of the baseline hazard function, distinguishing our procedure from traditional Bayesian approaches. Simulation experiments demonstrate that the proposed Bayesian procedure outperforms existing methods in some performance metrics. Moreover, it achieves statistical efficiency comparable to the efficient estimator while accurately identifying cluster structures. These features highlight its ability to overcome challenges associated with traditional IV approaches for time-to-event data.

PMID:41569616 | DOI:10.1002/sim.70360

Prehosp Emerg Care. 2026 Jan 22:1-12. doi: 10.1080/10903127.2026.2619038. Online ahead of print.

ABSTRACT

OBJECTIVES: Emergency medical services (EMS) workforce challenges impact prehospital care provision in many United States communities. One potential strategy to address this challenge is for clinicians to actively promote the EMS profession. However, there is limited data regarding the likelihood of EMS clinicians recommending others to join the EMS profession. We aimed to describe professional promotion among EMS clinicians and factors that impact their likelihood of recommending.

METHODS: We performed a cross-sectional analysis of nationally certified civilian EMS clinicians (ages 18-85) recertifying between 10/2023 and 04/2024. Applicants completed a voluntary survey regarding EMS professional promotion measured using the Net Promoter Score®. This validated tool measures the likelihood of recommending a field to others (classified as promoters, passives, or detractors). Surveys were merged with demographic and workplace characteristics from the National EMS Certification database. We calculated descriptive statistics (n, %) and (median, interquartile range [IQR]) and performed multivariable logistic regression (odds ratio, 95% confidence interval) to identify factors associated with likelihood of promoting EMS by clinicians, including age, sex, race, certification, education, years experience, agency and service type, and self-reported burnout and job satisfaction as covariates.

RESULTS: We included 33,335 clinicians for analysis (response rate = 28.8%); respondents reflected the nationally certified EMS population (male [74.2%], non-Hispanic White [86.1%], median age 36 [IQR: 29, 49], patient care [90.8%]). Promotion score distribution balanced between promoters (33.8%), passives (33.1%), and detractors (33.1%), yielding a NPS = 0.7 (possible range: -100 to +100), indicating near-zero net promotion. Odds of promoting EMS across agency types were lower than fire agencies (p < 0.05). Odds of promotion were also lower for higher education levels (associate [0.90,0.82-0.98], bachelor’s [0.80,0.73-0.87]; [referent: ≤high school/General Educational Development]) and more years experience (3-7 [0.86,0.81-0.93], 8-15 [0.76,0.70-0.82], >15 [0.83,0.75-0.91]; [referent: 0-3]). Clinicians reporting burnout had significantly lower odds of promoting EMS (0.31,0.29-0.33), while clinicians with high levels of job satisfaction had increased odds of promoting EMS (6.27,5.08-7.74).

CONCLUSIONS: Demographic and workplace characteristics are significantly associated with the likelihood of EMS clinicians promoting the profession. The observed associations with satisfaction and burnout suggest areas that may warrant further investigation regarding their relationship to professional promotion and broader workforce dynamics.

PMID:41569613 | DOI:10.1080/10903127.2026.2619038

Stat Med. 2026 Jan;45(1-2):e70376. doi: 10.1002/sim.70376.

ABSTRACT

Synthetic patient data has the potential to advance research in the medical field by providing privacy-preserving access to data resembling sensitive personal data. Assessing the level of privacy offered is essential to ensure privacy compliance, but it is challenging in practice. Many common methods either fail to capture central aspects of privacy or result in excessive caution based on unrealistic worst-case scenarios. We present a new approach to evaluating the privacy of synthetic datasets from known probability distributions based on the maximal local privacy loss. The strategy is based on measuring individual contributions to the likelihood of generating a specific synthetic dataset, to detect possibilities of reconstructing records in the original data. To demonstrate the method, we generate synthetic time-to-event data based on pancreatic and colon cancer data from the Cancer Registry of Norway using sequential regressions including a flexible parametric survival model. This illustrates the method’s ability to measure information leakage at an individual level, which can be used to ensure acceptable privacy risks for every patient in the data.

PMID:41569604 | DOI:10.1002/sim.70376

Clin Psychol Psychother. 2026 Jan-Feb;33(1):e70226. doi: 10.1002/cpp.70226.

ABSTRACT

Research into the differences across cognitive domains has been conducted to characterize the various presentations of schizophrenia-spectrum disorders. We aimed to identify distinct combined cognitive profiles with clinical relevance in patients with early psychosis (EP) by integrating neurocognition, social cognition and cognitive biases (CBs). Seventy-five outpatients attending an EP programme were assessed on neuropsychological performance, Theory of Mind (ToM), facial emotion recognition (FER), jumping to conclusions (JTC) bias and self-reported CBs through standardized tools. A two-step cluster analysis was performed to identify latent profiles. The optimal number of clusters was determined based on the Bayesian information criterion. Symptom dimensions, depression, global functioning, antipsychotic use, duration of untreated psychosis (DUP) and sociodemographic variables were compared across the resulting clusters. Two distinct profiles were identified. The first cluster (53.3%) was characterized by significant impairments in neurocognition, ToM and FER, as well as greater JTC and self-reported CBs. The second cluster (46.7%) was defined by relatively preserved performance across cognitive domains. Furthermore, the impaired cluster showed more severe positive, negative and disorganized symptoms, poorer functioning, lower premorbid intelligence and increased use of antipsychotics. No significant differences were found for depression, DUP or sociodemographic variables. Concluding, cognitive clustering revealed a clinically impaired subgroup of patients with more severe psychotic symptoms and poorer functioning. Our results may contribute to a better understanding of the distinct cognitive profiles of patients with EP. These findings may be relevant because several interventions targeting different cognitive domains have been shown to improve clinical and functional outcomes in EP.

PMID:41569600 | DOI:10.1002/cpp.70226

JAMA Otolaryngol Head Neck Surg. 2026 Jan 22. doi: 10.1001/jamaoto.2025.5077. Online ahead of print.

ABSTRACT

IMPORTANCE: Acute invasive fungal rhinosinusitis (AIFRS) is a rapidly progressive and potentially life-threatening infection that predominantly affects immunocompromised patients. Recent advances in diagnostic imaging, antifungal therapy, and surgical techniques may have altered its incidence, morbidity, and mortality.

OBJECTIVE: To evaluate temporal trends in the pooled proportion, morbidity, and mortality of AIFRS in immunocompromised patients and assess the association of diagnostic and therapeutic advances.

DATA SOURCES: Systematic searches of Ovid MEDLINE, Ovid Embase, PubMed, Scopus, Web of Science, Cochrane, and Google Scholar from 1977 through October 20, 2025, following Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines.

STUDY SELECTION: Prospective, retrospective, and cross-sectional studies and case series reporting pooled proportion, morbidity, or mortality of AIFRS in immunocompromised patients were included. Non-English articles, reviews, editorials, and studies with fewer than 10 patients were excluded.

DATA EXTRACTION AND SYNTHESIS: Two independent reviewers extracted data using standardized templates; disagreements were resolved by consensus. Risk of bias was assessed using the Newcastle-Ottawa Scale and Murad tool for case series. Random-effects meta-analysis generated pooled proportion, morbidity, and mortality rates with 95% CIs. Heterogeneity was quantified using I2 statistics. Meta-regression and sensitivity analyses evaluated temporal trends and study-level effects.

MAIN OUTCOMES AND MEASURES: Pooled proportion, morbidity, and mortality rates of AIFRS stratified by publication period (1983-2012 vs 2013-2025).

RESULTS: A total of 205 studies comprising 48 437 immunocompromised patients (median [range] age, 49.4 [5.2-68.8] years), including 10 311 (21.3%) with AIFRS, were analyzed. The pooled proportion was 11.8% (95% CI, 7.9%-17.2%), rising to 16.6% (95% CI, 8.7%-29.2%) in studies from 2013 to 2025. Overall mortality was 31.2% (95% CI, 28.3%-34.3%), declining from 41.9% (95% CI, 35.0%-49.1%) before 2013 to 28.2% (95% CI, 25.1%-31.4%) after 2013. Morbidity was 37.0% (95% CI, 32.9%-41.4%), with similar rates across periods (39.3% before 2013 vs 36.4% after 2013). The most common complications were vision loss, exophthalmos/proptosis, and orbital exenteration.

CONCLUSIONS AND RELEVANCE: This systematic review and meta-analysis suggests that the pooled proportion of AIFRS among immunocompromised patients has increased while mortality has declined, reflecting advances in diagnostic and therapeutic approaches. Early detection and aggressive management remain critical to improving outcomes.

PMID:41569597 | DOI:10.1001/jamaoto.2025.5077

Stat Med. 2026 Jan;45(1-2):e70383. doi: 10.1002/sim.70383.

ABSTRACT

Derived variables are variables that are constructed from one or more source variables through established mathematical operations or algorithms. For example, body mass index (BMI) is a derived variable constructed from two source variables: weight and height. When using a derived variable as the outcome in a statistical model, complications arise when some of the source variables have missing values. In this paper, we propose how one can define a single fully integrated Bayesian model to simultaneously impute missing values and sample from the posterior. We compare our proposed method with alternative approaches that rely on multiple imputation (MI), with examples including an analysis to estimate the risk of microcephaly (a derived variable based on sex, gestational age, and head circumference at birth) in newborns exposed to the ZIKA virus.

PMID:41569594 | DOI:10.1002/sim.70383

Nutr Hosp. 2026 Jan 19. doi: 10.20960/nh.05908. Online ahead of print.

ABSTRACT

BACKGROUND: disordered eating behaviors (DEBs) encompass altered eating behaviors that do not meet the diagnostic criteria to be considered eating disorders, yet, like the latter, are associated with multiple medical, psychological, and social complications.

OBJECTIVE: this study aimed to analyze the influence of the internalization of beauty ideals (specifically thinness and muscularity), depressive symptoms, body mass index (BMI), and type of university on DEBs.

METHODS: a correlational, cross-sectional study was conducted with two independent samples of university students from two universities, one public and one private, in Mexico City (n = 1571; 20.8  2.07 years). Data analysis included frequency and percentage estimation, mean comparison, and linear regression analysis.

RESULTS: students enrolled at the private university scored higher for all the variables studied, with statistically significant differences, except for BMI, where public university students scored higher. When compared by BMI, overweight and obese students scored higher for DEBs and internalization. In women, DEBs were predicted by thin-ideal internalization, BMI, and type of university. In male participants, predictors included BMI, depressive symptoms, and type of university.

CONCLUSIONS: the results confirmed previous findings in the literature, with socioeconomic status being a determining factor for the presence of DEBs.

PMID:41569591 | DOI:10.20960/nh.05908