Tag: pubmed

Annu Rev Nutr. 2025 May 12. doi: 10.1146/annurev-nutr-111124-092627. Online ahead of print.

ABSTRACT

The US Department of Agriculture’s Gus Schumacher Nutrition Incentive Program (GusNIP) funds produce prescription (PPR) programs that allow healthcare to support patients in accessing fruits and vegetables. This hybrid systematic narrative review identified 16 studies of PPR programs associated with GusNIP funding in some way that examined health outcomes, including clinical measures and healthcare utilization. Program designs were heterogeneous, sample sizes were generally small, and methodological rigor was often low, with most studies using a prepost design and none using a randomized control group. Fewer than half of the studies examining clinical values showed an association between PPR participation and improved health outcomes (for example, three of eight studies measuring weight or body mass index showed a statistically significant reduction, as well as two of the six studies measuring glycosylated hemoglobin). Only three studies examined healthcare utilization, two of which showed improvements in hospitalization and/or emergency department utilization. Overall, evidence for the health impact of PPRs is nascent but growing. PPRs with capacity should engage in rigorous study designs and examine a variety of downstream health and utilization outcomes.

PMID:40354556 | DOI:10.1146/annurev-nutr-111124-092627

Proc Natl Acad Sci U S A. 2025 May 20;122(20):e2409426122. doi: 10.1073/pnas.2409426122. Epub 2025 May 12.

ABSTRACT

Intermolecular interactions underlie all cellular functions, yet visualizing these interactions at the single-molecule level remains challenging. Single-molecule localization microscopy (SMLM) offers a potential solution. Given a nanoscale map of two putative interaction partners, it should be possible to assign molecules either to the class of coupled pairs or to the class of noncoupled bystanders. Here, we developed a probabilistic algorithm that allows accurate determination of both the absolute number and the proportion of molecules that form coupled pairs. The algorithm calculates interaction probabilities for all possible pairs of localized molecules, selects the most likely interaction set, and corrects for any spurious colocalizations. Benchmarking this approach across a set of simulated molecular localization maps with varying densities (up to ∼55 molecules μm^-2) and localization precisions (1 to 50 nm) showed typical errors in the identification of correct pairs of only a few percent. At molecular densities of ∼5 to 10 molecules μm^-2 and localization precisions of 20 to 30 nm, which are typical parameters for SMLM imaging, the recall was ∼90%. The algorithm was effective at differentiating between noninteracting and coupled molecules both in simulations and experiments. Finally, it correctly inferred the number of coupled pairs over time in a simulated reaction-diffusion system, enabling determination of the underlying rate constants. The proposed approach promises to enable direct visualization and quantification of intermolecular interactions using SMLM.

PMID:40354526 | DOI:10.1073/pnas.2409426122

PLoS One. 2025 May 12;20(5):e0319784. doi: 10.1371/journal.pone.0319784. eCollection 2025.

ABSTRACT

BACKGROUND: Many datasets in medicine and other branches of science are incomplete. In this article we compare various imputation algorithms for missing data.

OBJECTIVES: We take the point of view that it has already been decided that the imputation should be carried out using multiple imputation by chained equation and the only decision left is that of a subroutine for the one-dimensional imputations. The subroutines to be compared are predictive mean matching, weighted predictive mean matching, sampling, classification or regression trees and random forests.

METHODS: We compare these subroutines on real data and on simulated data. We consider the estimation of expected values, variances and coefficients of linear regression models, logistic regression models and Cox regression models. As real data we use data of the survival times after the diagnosis of an obstructive coronary artery disease with systolic blood pressure, LDL, diabetes, smoking behavior and family history of premature heart diseases as variables for which values have to be imputed. While we are mainly interested in statistical properties like biases, mean squared errors or coverage probabilities of confidence intervals, we also have an eye on the computation time.

RESULTS: Weighted predictive mean matching had to be excluded from the statistical comparison due to its enormous computation time. Among the remaining algorithms, in most situations we tested, predictive mean matching performed best.

NOVELTY: This is by far the largest comparison study for subroutines of multiple imputation by chained equations that has been performed up to now.

PMID:40354495 | DOI:10.1371/journal.pone.0319784

PLoS Comput Biol. 2025 May 12;21(5):e1013098. doi: 10.1371/journal.pcbi.1013098. Online ahead of print.

ABSTRACT

Uncertainty quantification (UQ) is the process of systematically determining and characterizing the degree of confidence in computational model predictions. In systems biology, and particularly with dynamic models, UQ is critical due to the nonlinearities and parameter sensitivities that influence the behavior of complex biological systems. Addressing these issues through robust UQ enables a deeper understanding of system dynamics and more reliable extrapolation beyond observed conditions. Many state-of-the-art UQ approaches in this field are grounded in Bayesian statistical methods. While these frameworks naturally incorporate uncertainty quantification, they often require the specification of parameter distributions as priors and may impose parametric assumptions that do not always reflect biological reality. Additionally, Bayesian methods can be computationally expensive, posing significant challenges when dealing with large-scale models and seeking rapid, reliable uncertainty calibration. As an alternative, we propose using conformal predictions methods and introduce two novel algorithms designed for dynamic biological systems. These approaches can provide non-asymptotic guarantees, improving robustness and scalability across various applications, even when the predictive models are misspecified. Through several illustrative scenarios, we demonstrate that these conformal algorithms can serve as powerful complements-or even alternatives-to conventional Bayesian methods, delivering effective uncertainty quantification for predictive tasks in systems biology.

PMID:40354480 | DOI:10.1371/journal.pcbi.1013098

PLoS One. 2025 May 12;20(5):e0323618. doi: 10.1371/journal.pone.0323618. eCollection 2025.

ABSTRACT

BACKGROUND: Kidney disease is prevalent among people living with HIV (PLHIV), especially in Sub-Saharan Africa (SSA), due to complications of HIV infection, co-morbidities, and antiretroviral therapy (ART). Despite SSA shouldering a disproportionate burden of HIV, there is limited data on the effect of clinical and demographic factors on the kidney with the introduction of the Test and Treat policy. This study aimed to determine the incidence and factors associated with kidney impairment among PLHIV on ART in the Southern Province of Zambia.

METHODS: We conducted a retrospective cohort study among 1216 adult individuals living with HIV who initiated ART between January 1, 2014, and July 31, 2016 [before test-and-treat cohort (BTT), n = 814] and August 1, 2016, and October 1, 2020 [after test-and-treat cohort (ATT), n = 402] without kidney function impairment at baseline, followed for 6 months in 12 districts of the Southern Province. The primary outcome was kidney function impairment, defined by an estimated glomerular filtration rate (eGFR) of < 60 ml/min/1.73m² estimated using the Modification of Diet in Renal Disease (MDRD) equation. We used multivariable logistic regression (xtlogit model) to identify factors associated with kidney function impairment. Statistical significance was set at p < 0.05.

RESULTS: The median age was 36.4 years (interquartile range (IQR): 29.9, 43.3), and the majority of participants were women (57.2%, n = 695). Tenofovir Disoproxil Fumarate (TDF) and XTC exposure was noted among 1,173/1216 (96.5%) enrolled participants and 92.9% (26/28)of those with renal impairment. The overall cumulative incidence of kidney impairment was 2.3% (n = 28/1216: 95% confidence interval (CI) 3%, 5%), and it was higher BTT compared to the ATT (2.8% vs. 1.2%). Every unit increase in age was associated with an increased odds of having kidney function impairment (adjusted odds ratio (AOR):1.05, 95% CI: 1.01-1.09, p = 0.008).. Participants from urban facilities also had a higher risk (AOR: 5.14, 95% CI: 1.95-13.55, p < 0.001). In contrast, being enrolled after the implementation of the “test-and-treat” policy was associated with lower odds of having kidney function impairment (AOR: 0.45, 95% CI: 0.12-0.97, p = 0.042).

CONCLUSIONS: This study found a 2.3% incidence of kidney function impairment among PLHIV within 6 months of initiating ART. An increase in age and receiving care at an urban facility were positively associated with kidney function impairment, whereas ART enrollment following the implementation of the “test-and-treat” policy was negatively associated. This study highlights the benefits of early ART initiation on kidney function, reinforcing the need to maintain the universal test-and-treat policy.

PMID:40354477 | DOI:10.1371/journal.pone.0323618

Elife. 2025 May 12;13:RP100840. doi: 10.7554/eLife.100840.

ABSTRACT

Understanding brain function relies on the collective work of many labs generating reproducible results. However, reproducibility has not been systematically assessed within the context of electrophysiological recordings during cognitive behaviors. To address this, we formed a multi-lab collaboration using a shared, open-source behavioral task and experimental apparatus. Experimenters in 10 laboratories repeatedly targeted Neuropixels probes to the same location (spanning secondary visual areas, hippocampus, and thalamus) in mice making decisions; this generated a total of 121 experimental replicates, a unique dataset for evaluating reproducibility of electrophysiology experiments. Despite standardizing both behavioral and electrophysiological procedures, some experimental outcomes were highly variable. A closer analysis uncovered that variability in electrode targeting hindered reproducibility, as did the limited statistical power of some routinely used electrophysiological analyses, such as single-neuron tests of modulation by individual task parameters. Reproducibility was enhanced by histological and electrophysiological quality-control criteria. Our observations suggest that data from systems neuroscience is vulnerable to a lack of reproducibility, but that across-lab standardization, including metrics we propose, can serve to mitigate this.

PMID:40354112 | DOI:10.7554/eLife.100840

J Med Internet Res. 2025 May 12;27:e65317. doi: 10.2196/65317.

ABSTRACT

BACKGROUND: Synthetic electronic health records (EHRs) generated by large language models (LLMs) offer potential for clinical education and model training while addressing privacy concerns. However, performance variations and demographic biases in these models remain underexplored, posing risks to equitable health care.

OBJECTIVE: This study aimed to systematically assess the performance of various LLMs in generating synthetic EHRs and to critically evaluate the presence of gender and racial biases in the generated outputs. We focused on assessing the completeness and representativeness of these EHRs across 20 diseases with varying demographic prevalence.

METHODS: A framework was developed to generate 140,000 synthetic EHRs using 10 standardized prompts across 7 LLMs. The electronic health record performance score (EPS) was introduced to quantify completeness, while the statistical parity difference (SPD) was proposed to assess the degree and direction of demographic bias. Chi-square tests were used to evaluate the presence of bias across demographic groups.

RESULTS: Larger models exhibited superior performance but heightened biases. The Yi-34B achieved the highest EPS (96.8), while smaller models (Qwen-1.8B: EPS=63.35) underperformed. Sex polarization emerged: female-dominated diseases (eg, multiple sclerosis) saw amplified female representation in outputs (Qwen-14B: 973/1000, 97.3% female vs 564,424/744,778, 75.78% real; SPD=+21.50%), while balanced diseases and male-dominated diseases skewed the male group (eg, hypertension Llama 2-13 B: 957/1000, 95.7% male vs 79,540,040/152,466,669, 52.17% real; SPD=+43.50%). Racial bias patterns revealed that some models overestimated the representation of White (eg, Yi-6B: mean SPD +14.40%, SD 16.22%) or Black groups (eg, Yi-34B: mean SPD +14.90%, SD 27.16%), while most models systematically underestimated the representation of Hispanic (average SPD across 7 models is -11.93%, SD 8.36%) and Asian groups (average SPD across 7 models is -0.77%, SD 11.99%).

CONCLUSIONS: Larger models, such as Yi-34B, Qwen-14B, and Llama 2 to 13 B, showed improved performance in generating more comprehensive EHRs, as reflected in higher EPS values. However, this increased performance was accompanied by a notable escalation in both gender and racial biases, highlighting a performance-bias trade-off. The study identified 4 key findings as follows: (1) as model size increased, EHR generation improved, but demographic biases also became more pronounced; (2) biases were observed across all models, not just the larger ones; (3) gender bias closely aligned with real-world disease prevalence, while racial bias was evident in only a subset of diseases; and (4) racial biases varied, with some diseases showing overrepresentation of White or Black populations and underrepresentation of Hispanic and Asian groups. These findings underline the need for effective bias mitigation strategies and the development of benchmarks to ensure fairness in artificial intelligence applications for health care.

PMID:40354109 | DOI:10.2196/65317

Int J Geriatr Psychiatry. 2025 May;40(5):e70094. doi: 10.1002/gps.70094.

ABSTRACT

BACKGROUND: Delirium is an acute cognitive change characterized by behavioral and psychological features, such as visual and auditory hallucinations, sleep disturbances, and emotional confusion. It can lead to extended hospital stays, increased mortality risk, and higher nursing costs. In postoperative hip fracture patients, delirium results in a higher complication rate, poorer functional recovery, increased readmission rates, repeat surgeries, and elevated mortality. Despite these serious consequences, the literature provides limited information on the incidence of postoperative delirium following hip fracture surgeries in Asians. Additionally, there is a lack of long-term, comprehensive nationwide population-based studies, highlighting an important area for future research. This study aims to understand the incidence and risk factors of postoperative delirium in hip fracture patients using representative population data.

METHODS: We conducted a retrospective cohort study using the Taiwan National Health Insurance Research Database (NHIRD) from 2009 to 2020. The cohort consisted of 118,682 patients aged 65 years or older who were diagnosed with hip fractures. The delirium incidence was observed per 1000 person-years. The Cox proportional hazards model was used to investigate the incidence of delirium among hip fracture patients.

RESULTS: The incidence of the first episode of delirium after hip surgery in the elderly was 1.87 events per 1000 PYs. Factors associated with delirium included being female (adjusted hazard ratio [aHR]: 0.59; 95% confidence interval [CI]: 0.53-0.64), age ≥ 95 years (aHR: 3.52; 95% CI: 2.74-4.51), comorbid dementia (aHR: 2.63; 95% CI: 2.38-2.92), and ICU stay 2-3 days (aHR: 2.85; 95% CI: 1.28-6.37). The occurrence of delirium was significantly associated with an ICU stay of ≥ 4 days, dementia, as well as 30-day, 90-day, and 1-year mortality (p < 0.001).

CONCLUSIONS: This study highlights the relatively low incidence of postoperative delirium in elderly hip fracture patients in Taiwan. Key risk factors identified include advanced age, female gender, comorbid dementia, and prolonged ICU stays. These findings underscore the need for targeted prevention and early intervention strategies to improve patient outcomes.

PMID:40354105 | DOI:10.1002/gps.70094

JAMA Pediatr. 2025 May 12. doi: 10.1001/jamapediatrics.2025.0807. Online ahead of print.

ABSTRACT

IMPORTANCE: Previous studies suggest that administration of erythropoiesis-stimulating agents darbepoetin or erythropoietin to preterm infants results in fewer transfusions, fewer donor exposures, and improved neurodevelopmental outcome.

OBJECTIVE: To determine if, compared with placebo, preterm infants randomized to weekly darbepoetin would have greater red cell mass during hospitalization and better neurocognitive outcome at 22 to 26 months’ corrected age.

DESIGN, SETTING, AND PARTICIPANTS: This randomized clinical trial was conducted between September 2017 and November 2019 for infants 23 0/7 to 28 6/7 weeks’ gestation in 19 US Neonatal Research Network centers comprising 33 neonatal intensive care units. Follow-up occurred through January 2023. Infants were randomized by 36 hours after birth to weekly placebo or darbepoetin (10 μg/kg) through 35 weeks’ postmenstrual age. Iron administration and transfusions were administered by protocol. Study data were analyzed from June to October 2023.

MAIN OUTCOMES AND MEASURES: The primary outcome was the mean cognitive composite score on the Bayley Scales of Infant Development, third edition (Bayley-III) at 22 to 26 months’ corrected age. The lowest possible score (54) was assigned to infants who died.

RESULTS: A total of 650 infants (322 darbepoetin; 328 placebo; mean [SD] gestational age, 26.2 [1.7] weeks; 328 female [50.5%]) were enrolled. Five hundred eighty-three infants (291 darbepoetin; 292 placebo) had the primary outcome determined (90% of those enrolled). Mean (SD) cognitive scores were similar between groups: 80.7 (19.5) darbepoetin vs 80.1 (18.7) placebo, adjusted mean difference, -0.23 (95% CI, -3.09 to 2.64). Compared with infants receiving placebo, more infants in the darbepoetin group were transfusion free (40% [127 of 319] vs 21% [70 of 327]; adjusted relative risk [RR], 1.3; 95% CI, 1.2-1.5), received fewer transfusions (mean [SD], 2.3 [3.1] vs 3.3 [3.5]), were exposed to fewer donors (mean [SD], 1.6 [2.3] vs 2.2 [2.3]), had higher red cell mass by week 2 of age (adjusted mean difference, 3.2; 95% CI, 1.7-4.7), and higher mean hematocrit by week 2 of age (adjusted mean difference, 2.8; 95% CI, 2.1-3.6), and were less likely to have bronchopulmonary dysplasia greater than grade 1 (35% [91 of 261] vs 46% [128 of 277]; RR, 0.78; 95% CI, 0.64-0.96). The incidence of retinopathy of prematurity stage greater than 2 was similar between groups, 13% (35 of 273) in the darbepoetin group vs 16% (45 of 279) in the placebo group. There were no differences in adverse effects between groups.

CONCLUSIONS AND RELEVANCE: Results of this randomized clinical trial reveal that this dose and dosing schedule of darbepoetin did not improve cognitive scores of preterm infants at 22 to 26 months’ corrected age. Darbepoetin significantly increased red cell mass resulting in higher hematocrit values, fewer transfusions, and fewer donor exposures.

TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT03169881.

PMID:40354084 | DOI:10.1001/jamapediatrics.2025.0807

JAMA Pediatr. 2025 May 12. doi: 10.1001/jamapediatrics.2025.0828. Online ahead of print.

NO ABSTRACT

PMID:40354076 | DOI:10.1001/jamapediatrics.2025.0828