Category: Statistics

JMIR Med Inform. 2025 Dec 17;13:e79748. doi: 10.2196/79748.

ABSTRACT

BACKGROUND: Knee cartilage injury (KCI) poses significant challenges in the early clinical diagnosis process, primarily due to its high incidence, the complexity of healing, and the limited sensitivity of initial imaging modalities.

OBJECTIVE: This study aims to employ magnetic resonance imaging and machine learning methods to enhance the classification accuracy of the classifier for KCI, improve the existing network structure, and demonstrate important clinical application value.

METHODS: The proposed methodology is a multidimensional feature cross-level fusion classification network driven by the large separable kernel attention, which enables high-precision hierarchical diagnosis of KCI through deep learning. The network first fuses shallow high-resolution features with deep semantic features via the cross-level fusion module. Then, the large separable kernel attention module is embedded in the YOLOv8 network. This network utilizes the combined optimization of depth-separable and point-by-point convolutions to enhance features at multiple scales, thereby dramatically improving the hierarchical characterization of cartilage damage. Finally, five classifications of knee cartilage injuries are performed by classifiers.

RESULTS: To overcome the limitations of network models trained with single-plane images, this study presents the first hospital-based multidimensional magnetic resonance imaging real dataset for KCI, on which the classification accuracy is 99.7%, the Kappa statistic is 99.6%, the F-measure is 99.7%, the sensitivity is 99.7%, and the specificity is 99.9%. The experimental results validate the feasibility of the proposed method.

CONCLUSIONS: The experimental outcomes confirm that the proposed methodology not only achieves exceptional performance in classifying knee cartilage injuries but also offers substantial improvements over existing techniques. This underscores its potential for clinical deployment in enhancing diagnostic precision and efficiency.

PMID:41406414 | DOI:10.2196/79748

Neurology. 2026 Jan 27;106(2):e214343. doi: 10.1212/WNL.0000000000214343. Epub 2025 Dec 17.

ABSTRACT

BACKGROUND AND OBJECTIVES: The association between dairy intake and dementia risk remains uncertain, especially for dairy products with varying fat contents. The aim of this study was to investigate the association between high-fat and low-fat dairy intake and dementia risk.

METHODS: This study used data from a prospective cohort in Sweden, the Malmö Diet and Cancer cohort, which consisted of community-based participants who underwent dietary assessment at baseline (1991-1996). Dietary intake was evaluated using a comprehensive diet history method that combined a 7-day food diary, a food frequency questionnaire, and a dietary interview. Dementia cases were identified through the Swedish National Patient Register until December 31, 2020, and cases diagnosed until 2014 were further validated. The primary outcome of the study was all-cause dementia, and the secondary outcomes were Alzheimer disease (AD) and vascular dementia (VaD). Cox proportional hazard regression models were used to estimate hazard ratio (HR) and 95% CI.

RESULTS: This study included 27,670 participants (mean baseline age 58.1 years, SD 7.6; 61% female). During a median of 25 years of follow-up, 3,208 incident dementia cases were recorded. Consumption of ≥50 g/d of high-fat cheese (>20% fat) was associated with a reduced risk of all-cause dementia (HR 0.87; 95% CI, 0.78-0.97) and VaD (HR 0.71, 95% CI 0.52-0.96) compared with lower intake (<15 g/d). An inverse association between high-fat cheese and AD was found among APOE ε4 noncarriers (HR 0.87, 95% CI 0.76-0.99, p-interaction = 0.014). Compared with no consumption, individuals consuming ≥20 g/d of high-fat cream (>30% fat) had a 16% lower risk of all-cause dementia (HR 0.84, 95% CI 0.72-0.98). High-fat cream consumption was inversely associated with the risk of AD and VaD. Consumption of low-fat cheese, low-fat cream, milk (high-fat and low-fat), fermented milk (high-fat and low-fat), and butter showed no association with all-cause dementia.

DISCUSSION: Higher intake of high-fat cheese and high-fat cream was associated with a lower risk of all-cause dementia, whereas low-fat cheese, low-fat cream, and other dairy products showed no significant association. APOE ε4 status modified the association between high-fat cheese and AD. Our study’s observational design limits causal inference.

PMID:41406402 | DOI:10.1212/WNL.0000000000214343

J Am Acad Orthop Surg. 2025 Dec 8. doi: 10.5435/JAAOS-D-25-00380. Online ahead of print.

ABSTRACT

OBJECTIVES: Increased mortality following geriatric hip fractures is well reported. However, population-level analysis of mortality trends over time are not common. This study aimed to evaluate the 3- and 12-month mortality after geriatric hip fractures from 2010 to 2019.

METHODS: The New York Statewide Planning and Research Cooperative System database from 2010 to 2020 was retrospectively queried for patients aged >65 years with a femoral neck or intertrochanteric hip fracture. Kaplan-Meier survival analysis was used to calculate mortality rates for each year. Cox proportional hazard multivariable regression controlling for sex, age, race, obesity, smoking, and Elixhauser comorbidity index was used to compare mortality hazard ratios for each year. Secondary outcomes included length of stay, discharge disposition, and 3-month readmission and emergency department visits.

RESULTS: From 2010 to 2019, 142,540 patients aged ≥65 years had a diagnosis of femoral neck fracture (62%) or intertrochanteric hip fracture (38%). The mean age was 83.29 years (SD 8.22). The mean Elixhauser comorbidity index was 7.35 (SD 7.60). Kaplan-Meier survival analysis revealed that for the complete cohort 3-month mortality rate was 9.82% (95% confidence interval 9.65% to 9.98%) and 12-month mortality rate was 16.06% (95% confidence interval 15.84% to 16.27%). The 3-month mortality rate went from 10.8% in 2010 to 8.6% in 2019 and the 12-month mortality rate went from 17.7% in 2010 to 14.8% in 2018 before rising to 16.9% in 2019. Cox multivariate proportional hazard regression demonstrated statistically significant decreased hazard ratio from 2012 to 2019 compared with reference hazard in 2010 (all P < 0.05). Reductions were also observed for length of stay (7.8 to 6.4 days, P < 0.001), 3-month readmissions rate (34% to 22%, P < 0.001), and 3-month emergency department visit rate (45% to 34%, P < 0.001).

CONCLUSION: Mortality after geriatric hip fractures has demonstrated a reduction in the past decade with 3-month mortality continuously decreasing from 2010 to 2019 and 12-month mortality decreasing from 2010 to 2018 before increasing in 2019.

PMID:41406399 | DOI:10.5435/JAAOS-D-25-00380

J Cardiovasc Pharmacol. 2025 Nov 25. doi: 10.1097/FJC.0000000000001780. Online ahead of print.

ABSTRACT

Colchicine has been studied as an anti-inflammatory treatment for cardiovascular prevention, but findings from randomized trials have been inconsistent. This meta-analysis evaluated the efficacy and safety of colchicine in reducing major adverse cardiovascular events (MACE) and its individual components, using ChatGPT as an assistant throughout the process. Randomized trials of colchicine for cardiovascular prevention were systematically identified, and data extraction, risk of bias assessment, and meta-analyses were performed with ChatGPT under human supervision. The primary outcome was MACE, while secondary outcomes included myocardial infarction (MI), stroke, revascularization, cardiovascular mortality, and all-cause mortality. Eleven trials involving 30,888 patients were included. Colchicine significantly reduced MACE (risk ratio 0.75, 95% CI 0.63-0.88), though no significant effects were observed for MI, stroke, cardiovascular mortality, or all-cause mortality. In addition to its clinical findings, this study illustrates the potential of ChatGPT to assist in systematic reviews and meta-analyses by automating screening, data extraction, bias assessment, and statistical code generation. This integration reduced researcher time by over 70% while maintaining accuracy through human validation. Overall, colchicine appears to lower the risk of MACE but the results of the CLEAR trial have lowered certainty, while the findings highlight the feasibility and efficiency gains of using large language models in evidence synthesis workflows.

PMID:41406368 | DOI:10.1097/FJC.0000000000001780

Can Urol Assoc J. 2025 Dec 15. doi: 10.5489/cuaj.9370. Online ahead of print.

ABSTRACT

INTRODUCTION: Accurate pre-intervention staging of upper tract urothelial carcinoma (UTUC) remains a significant clinical challenge, particularly in identifying muscle-invasive disease (≥pT2), where kidney-sparing surgery may not be appropriate. Current imaging and biopsy approaches are often inadequate. Radiomics, which extracts high-dimensional features from medical imaging, may improve non-invasive staging. This study assessed whether computed tomography (CT)-based radiomic features, alone or combined with clinical data, could predict ≥pT2 UTUC in a multicenter Canadian cohort.

METHODS: We retrospectively analyzed clinical, pathologic, and radiographic features of patients with UTUC who underwent extirpative surgery at five academic centers from January 2, 2001, to May 1, 2023. Radiomic features were extracted from machine-learning segmentations of the affected kidney using the excretory phase of CT. Predictive models were developed using clinical only, radiomic only, and combined data to predict stage ≥pT2. Feature selection included univariable logistic regression, correlation filtering, and LASSO. Model performance was assessed via five-fold cross-validation repeated 10 times, with area under the curve (AUC) as the primary metric.

RESULTS: Of 441 patients, 208 (47.2%) were included. Of the 208 patients, 97 (46.6%) had ≥pT2 disease. The clinical model (AUC 0.602) included age, hydronephrosis, and high-grade cytology. The radiomics model, based on two texture features, achieved an AUC of 0.653. The combined model achieved an AUC of 0.647. Radiomics and combined models significantly outperformed the clinical model (p<0.01), but did not differ from each other. For 117 patients with renal pelvis cancers, the combined model’s discrimination performance was statistically better than the clinical model (AUC 0.708 vs. AUC 0.607, p<0.001). Likewise, the radiomics’ AUC discrimination performance was statistically better than the clinical model (AUC 0.694 vs. AUC 0.607, p=0.004). In contrast, we found no significant difference in model performance in the non-renal pelvis subgroup (n=91).

CONCLUSIONS: Conventional radiomics improved the prediction of muscle-invasive UTUC compared to clinical models alone, but overall accuracy remained suboptimal for clinical use. Heterogeneity in CT protocols and challenges with tumor segmentation were the main limitations. Future work should develop more adaptable AI models trained on larger, more diverse datasets to better reflect real-world imaging conditions.

PMID:41406346 | DOI:10.5489/cuaj.9370

Can Urol Assoc J. 2025 Dec 15. doi: 10.5489/cuaj.9451. Online ahead of print.

ABSTRACT

INTRODUCTION: Type and screen testing (T&S) is routinely performed preoperatively for many endoscopic procedures, despite low transfusion rates. While important, T&S can be costly, unnecessary, and burdensome for patients to obtain in a short timeframe due to expiry. We aimed to assess and reduce unnecessary T&S in a safe and collaborative manner through a Transfusion Dashboard. We assessed the effect of reduced testing on patient safety, cost, and the environment.

METHODS: This quality improvement study used the Transfusion Dashboard, a web-based, institutional platform tracking blood transfusion trends. During the observation phase (2016-2019), procedure-specific preoperative T&S recommendations were developed. Following implementation of these recommendations in 2020, the incidence of T&S, perioperative transfusion rates, and rescue transfusion rates were assessed pre- and post-intervention using the Chi-squared test. Cost and environmental savings were also evaluated.

RESULTS: From 2016-2023, outcomes were tracked for 4375 pre-initiative and 2488 post-initiative patients who underwent endoscopic procedures. We found a statistically significant decrease in T&S following initiative implementation for transurethral resection of the prostate (TURP), percutaneous nephrolithotomy (PCNL), holmium e-nucleation of the prostate (HoLEP), and transurethral resection of bladder tumor (TURBT) by as much as 51.2%. There was no change in uncrossed or overall blood transfusions. Since the implementation of the initiative, $45 362.81 in testing materials were saved and an associated reduction of 697 kg CO₂ was observed.

CONCLUSIONS: Institutional- and procedure-specific testing guidelines decreased unnecessary tests, leading to improved resource stewardship, reduced cost, improved patient experience, and environmental savings. Initial modest cost savings and care improvements may be amplified safely in larger organizations and across more procedures.

PMID:41406342 | DOI:10.5489/cuaj.9451

J Med Microbiol. 2025 Dec;74(12). doi: 10.1099/jmm.0.002104.

ABSTRACT

Introduction. Biofilms have been implicated as a potential cause of chronic rhinosinusitis (CRS), with patients showing an increased prevalence of biofilms, likely contributing to antibiotic ineffectiveness in these individuals. In many environments, biofilms are polymicrobial, with interspecies interactions promoting bacterial survival and encouraging robust growth. Improvements in visualization techniques for biofilms have enabled species-specific identification, leading to a growing body of literature using these techniques and examining severity in different phenotypes of CRS.Gap Statement. It is unclear whether sinus biofilms are typically poly- or monomicrobial, and if they are correlated with clinical severity in CRS.Aim. We conducted a scoping review to determine how prevalent biofilms were in sinus tissue of patients with CRS. Furthermore, we correlated disease severity with the presence of biofilms.Methodology. We searched PubMed, Scopus, Medline and Web of Science databases for all studies which directly visualized biofilms on tissue from patients with CRS. After screening 1,853 search results, 39 studies were included for analysis in this review.Results. Patients with CRS had a higher prevalence of biofilms compared with controls. We found no significant difference in the proportion of biofilms detected across visualization techniques or based on CRS phenotyping. Fifteen studies reported disease severity by biofilm status; most reported greater severity in patients with biofilms, although only some were statistically significant. Nine studies used techniques capable of detecting polymicrobial biofilms, all of which found a subset of polymicrobial biofilms.Conclusion. Our findings demonstrate an increased prevalence of biofilms in patients with CRS, which may correspond to increased disease severity. The evidence for biofilms being polymicrobial is compelling, although it is based on a small number of studies.

PMID:41405936 | DOI:10.1099/jmm.0.002104

JAMA. 2025 Dec 17. doi: 10.1001/jama.2025.20541. Online ahead of print.

NO ABSTRACT

PMID:41405908 | DOI:10.1001/jama.2025.20541

JAMA Psychiatry. 2025 Dec 17. doi: 10.1001/jamapsychiatry.2025.3828. Online ahead of print.

ABSTRACT

IMPORTANCE: Veterans with serious mental illness (SMI) experience a higher prevalence of obesity than the general veteran population; weight loss programs are needed that are tailored to this population.

OBJECTIVE: To evaluate a weight loss program, CoachToFit (CTF), which includes weekly calls from a Veteran Health Administration peer specialist, a Bluetooth-enabled scale and fitness tracker, and a smartphone application that provides health education and tracks steps, goals, and weight.

DESIGN, SETTING, AND PARTICIPANTS: This randomized clinical trial was conducted within the Pittsburgh Veteran Affairs health care system and presents pre-post (6 months) analysis comparing CTF and usual care. Veterans with body mass index (BMI, calculated as weight in kilograms divided by height in meters squared) of 30 or higher and diagnosis of major depressive disorder, bipolar disorder, or schizophrenia were eligible for inclusion. Exclusion criteria included history of bariatric surgery or recent psychiatric hospitalization. The study was conducted from October 1, 2020, to September 30, 2025, and data analysis was conducted from January to October 2025.

EXPOSURE: Random assignment to CTF.

MAIN OUTCOMES AND MEASURES: The primary outcomes were weight (in kg), BMI, and cardiorespiratory fitness (meters walked in 6 minutes).

RESULTS: Among the sample (n = 256), mean (SD) age was 53.5 (13.1) years, 80 participants (31.3%) were female, and 199 (77.7%) were diagnosed with major depressive disorder. Mean (SD) weight loss at 6 months was -3.2 (6.2) kg in the CTF group (n = 128) compared to -1.6 (4.9) kg in the usual care group (P = .05). After adjustment, participants in CTF experienced greater, nonsignificant weight loss compared to usual care, with an adjusted mean difference (AMD) of -1.62 kg (95% CI, -3.38 to 0.14; P = .07). For BMI, the AMD in change between groups at 6 months was -0.56 (95% CI, -1.15 to 0.03; P = .06). Change in meters walked was not statistically significant between groups, with an AMD of 3.53 m (95% CI, -12.87 to 19.92; P = .67). At 6 months, 34 participants (36.6%) from the CTF group lost 5% or more of their body weight compared to 19 (22.4%) in usual care, representing a 1.93-fold greater likelihood in adjusted analyses (95% CI, 0.96-3.91; P = .07). More participants in CTF (n = 21 [22.6%]) lost 7% or more of their body weight compared to usual care (n = 7 [8.2%]), representing a 3.9-fold greater likelihood in adjusted analyses (95% CI, 1.45-10.36; P = .007).

CONCLUSIONS AND RELEVANCE: In this randomized clinical trial, a weight loss program tailored to veterans with SMI using remote technologies and paraprofessionals demonstrated the potential to help this population lose weight.

TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT04560335.

PMID:41405896 | DOI:10.1001/jamapsychiatry.2025.3828

JAMA Netw Open. 2025 Dec 1;8(12):e2544153. doi: 10.1001/jamanetworkopen.2025.44153.

ABSTRACT

IMPORTANCE: The Planetary Health Diet (PHD), introduced by the EAT-Lancet Commission in 2019, emphasizes a plant-based diet. Several cohorts have assessed adherence using the PHD Index (PHDI), but evidence is limited on whether maternal periconceptional and early pregnancy adherence is associated with fetal growth.

OBJECTIVE: To examine the association of maternal PHDI adherence in early pregnancy with longitudinal 2-dimensional and 3-dimensional fetal biometric measures.

DESIGN, SETTING, AND PARTICIPANTS: Secondary analysis of a prospective cohort of pregnant women with singletons (July 2009 to January 2013) at 12 US clinical sites. Women with preexisting health conditions were excluded. Participants had data on fetal body composition and organ volumes (April 2016 to September 2019). Analyses were conducted June 2024 to August 2025.

EXPOSURE: PHDI adherence categorized as low, moderate, or high, assessed by the Food Frequency Questionnaire between 8 to 13 weeks’ gestation.

MAIN OUTCOMES AND MEASURES: Estimated fetal weight (EFW), head circumference (HC), humerus and femur lengths, abdominal circumference and area, abdominal subcutaneous tissue thickness, fractional thigh and arm volumes (total, lean, and fat), midthigh areas, ratio of fractional fat thigh volume to fractional thigh volume, and ratio of midthigh fat area to midthigh area, measured up to 5 times between 15 to 42 weeks’ gestation. Trajectories were modeled using linear mixed-effects models.

RESULTS: Analyses included 1464 women. Mean (SD) maternal age was 28.1 (5.6) years, and mean (SD) gestational age at delivery was 39.2 (1.7) weeks. High vs low PHDI adherence was associated with larger EFW at 32 to 40 weeks (32 weeks: difference, 35 g; 95% CI, 22 to 49 g; 40 weeks: difference, 165 g; 95% CI, 108 to 223 g), larger HC at 37 to 39 weeks (37 weeks: difference, 1.89 mm; 95% CI, 1.10 to 2.69 mm; 39 weeks: difference, 2.44 mm; 95% CI, 1.47 to 3.42 mm), smaller fractional lean arm volume at 34 to 37 weeks; (34 weeks: difference, -0.40 cm3; 95% CI, -0.56 to -0.25 cm3), larger fractional fat arm volume at 28 to 29 weeks; (28 weeks: difference, 0.23 cm3; 95% CI, 0.13 to 0.34 cm3), and larger abdominal area at 25 to 26 weeks (25 weeks: difference, 62.2 mm2; 95% CI, 34.4 to 90.0 cm3). Moderate adherence showed similar patterns.

CONCLUSIONS AND RELEVANCE: In this cohort study of pregnant women, higher maternal PHDI adherence in early pregnancy was associated with greater fetal adiposity and reduced lean tissue, suggesting potential implications for offspring metabolic health.

PMID:41405886 | DOI:10.1001/jamanetworkopen.2025.44153