Category: Statistics

Optom Vis Sci. 2025 Feb 14. doi: 10.1097/OPX.0000000000002237. Online ahead of print.

ABSTRACT

SIGNIFICANCE: Intervention strategies for post-mild traumatic brain injury (mTBI) ocular motor disorders vary across disciplines and include watchful waiting, vestibular rehabilitation, vision rehabilitation/vision therapy, and optical intervention. However, evidence supporting their effectiveness is limited, highlighting the need for high-quality randomized controlled trials with standardized testing, diagnostic criteria, and reassessment of ocular motor function after intervention.

BACKGROUND: Ocular motor disorders occur frequently after mTBI.

OBJECTIVES: This study aimed to conduct a scoping review of interventions for mTBI-related ocular motor disorders in children and adults.

DATA SOURCES: The following electronic bibliographic databases were searched: PubMed, Embase, PEDro, OVID, Clinical Key, Google Scholar, and REHABDATA.

STUDY SELECTION: Intervention studies published in English between 2003 and 2024 involving mTBI participants who had an ocular motor assessment prior to intervention were included in this study.

DATA EXTRACTION AND SYNTHESIS: Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews guidelines were followed for reporting. Study population, intervention, and outcomes were extracted and synthesized in tabular and graphical formats.

RESULTS: Sixty-seven eligible studies were included, with only three (4%) judged as low risk of bias. Intervention strategies included watchful waiting (n = 31, 46%), vestibular rehabilitation (n = 13, 19%), vision rehabilitation/vision therapy (n = 10, 15%), optical intervention (n = 4, 6%), and alternative interventions or multifaceted interventions (n = 9, 14%). Among the studies providing statistically supported results, improvements in one or more ocular motor outcome domains were reported in nearly 80% of the studies on watchful waiting (19/24) and 100% of the studies on vestibular rehabilitation (4/4), vision rehabilitation/vision therapy (7/7), or optical intervention (1/1).

CONCLUSIONS AND IMPLICATIONS: Although post-mTBI ocular motor deficits improved with watchful waiting strategy, vestibular rehabilitation, vision rehabilitation/vision therapy, and optical interventions, most studies had significant risk of bias. This review emphasizes the necessity for high-quality randomized controlled trials with standardized testing protocols and diagnostic criteria and reassessment of ocular motor functions after intervention to evaluate the effectiveness of these interventions in different age groups and recovery stages.

PMID:39951336 | DOI:10.1097/OPX.0000000000002237

J Chem Inf Model. 2025 Feb 14. doi: 10.1021/acs.jcim.4c02163. Online ahead of print.

ABSTRACT

Metal-embedded complexes (MECs), including transition metal complexes (TMCs) and metal-organic frameworks (MOFs), are important in catalysis, materials science, and molecular devices due to their unique metal atom centrality and complex coordination environments. However, modeling and predicting their properties accurately is challenging. A new metal attention (MA) framework for graph neural networks (GNNs) was proposed to address the limitations of traditional methods, which fail to differentiate core coordination structures from ordinary covalent bonds. This MA framework converts heterogeneous graphs of complexes into homogeneous ones with distinct metal features by highlighting key metal-feature coordination through hierarchical pooling and a metal cross-attention. To assess its performance, 11 widely used GNN algorithms, three of which are heterogeneous, were compared. Experimental results indicate significant improvements in accuracy: an average of 32.07% for predicting TMC properties and up to 23.01% for MOF CO₂ absorption. Moreover, tests on the framework’s robustness regarding data set size variation and comparison with a larger non-MA model show that the enhanced performance stems from the architecture, not merely increasing model capacity. The MA framework’s potential in predicting metal complex properties offers a potent statistical tool for optimizing and designing new materials like catalysts and gas storage systems.

PMID:39951325 | DOI:10.1021/acs.jcim.4c02163

Environ Toxicol Chem. 2025 Feb 14:vgaf047. doi: 10.1093/etojnl/vgaf047. Online ahead of print.

ABSTRACT

Excess metal accumulation in organisms can result in adverse impacts at the levels of the individual, population, and community. A detectable increase of metal concentrations in organisms does not necessarily imply that there are such impacts, but to our knowledge no field study has directly tested this hypothesis. To test this hypothesis, we investigated the accumulation of six elements (Cu, Zn, Cd, Pb, As, and Se) in masu salmon (Oncorhynchus masou, Salmonidae) at a total of nine study sites in a metal-contaminated river receiving mine discharge and in a nearby reference river. Multiple fish community surveys in 2018 and 2019 consistently indicated that the abundances and condition factors of the four dominant fish species, including masu salmon, were comparable in the two rivers. In contrast, despite sample sizes of only five individuals per site, statistically significant increases in the concentrations of Cu, Cd, Pb, As, and Se in the muscle of masu salmon were observed at multiple sites in the metal-contaminated river, where no detectable impacts on the abundance or condition factor of the fish were observed. The muscle concentrations of Cu, Zn, Cd, Pb, As, and Se at these sites were 1.4-2.5, 1.5-1.9, 188-520, 4.6-68.0, 1.1-3.9, and 2.8-3.5 times, respectively, the mean concentrations at the reference sites, although the increase of the Zn concentration in muscle at these metal-contaminated river sites was not significant. These results provide clear empirical evidence that a detectable increase in metal accumulation does not necessarily imply that population and community-level impacts are simultaneously occurring in the environment.

PMID:39951317 | DOI:10.1093/etojnl/vgaf047

JAMA Health Forum. 2025 Feb 7;6(2):e245220. doi: 10.1001/jamahealthforum.2024.5220.

ABSTRACT

IMPORTANCE: Understanding the drivers of health care spending across US counties is important for developing policies and assessing the allocation of health care services.

OBJECTIVE: To estimate the amount of cross-county health care spending variation explained by (1) population age, (2) health condition prevalence, (3) service utilization, and (4) service price and intensity.

DESIGN, SETTING, AND PARTICIPANTS: In this cross-sectional study, data for 4 key drivers of per capita spending were extracted for 3110 US counties, 148 health conditions, 38 age-sex groups, 4 payers, and 7 types of care for 2019. Service utilization was measured as service volume per prevalent case, while price and intensity was measured as spending per visit, admission, or prescription. Das Gupta and Shapley decomposition methods and linear regression were used to estimate the contribution of each factor. The data analysis was conducted between March 2024 and July 2024.

EXPOSURES: Age, disease prevalence, service utilization, or service price and intensity.

MAIN OUTCOMES AND MEASURES: Variation in health care spending across US counties.

RESULTS: In 2019, 76.6% of personal health care spending was included in this study. Overall, 64.8% of cross-county health care spending variation among 3110 US counties was explained by service utilization, while population age, disease prevalence, and price and intensity of services explained 4.1%, 7.0%, and 24.1%, respectively. The rate at which these factors contributed to variation in spending differed by payer, type of care, and health condition. Service utilization was associated with insurance coverage, median income, and education. An increase in each of these from the median to the 75th percentile was associated with a 7.8%, 4.4%, and 3.8% increase in ambulatory care utilization, respectively. The fraction of Medicare beneficiaries with Medicare Advantage was associated with less utilization. An increase in Medicare Advantage coverage from the median to the 75th percentile was associated with a 1.9% decrease in ambulatory care utilization. Differences in cross-state spending levels were also attributed to different factors. For Utah, the state with the least health care spending per capita, spending rates were lower for all types of care due principally to the young age profile. For New York, the state with the highest spending, spending rates were relatively high for hospital inpatient and prescribed pharmaceutical spending. For both types of care, high service price and intensity contributed to the above-average spending.

CONCLUSIONS AND RELEVANCE: In this cross-sectional study, variation in health care spending among US counties was largely related to variation in service utilization. Understanding the drivers of spending variation in the US may help policymakers assess the allocation of health care resources.

PMID:39951314 | DOI:10.1001/jamahealthforum.2024.5220

JAMA Health Forum. 2025 Feb 7;6(2):e245376. doi: 10.1001/jamahealthforum.2024.5376.

ABSTRACT

IMPORTANCE: Private equity (PE) acquisition of physician practices is increasing, with owners targeting sales, or exits, in 3 to 7 years. Little is known about the association of exit with physician retention and subsequent employment.

OBJECTIVE: To examine whether PE exit of physician practices is associated with changes in physician retention and subsequent choice of practice size.

DESIGN, SETTING, AND PARTICIPANTS: Using data from the Centers for Medicare & Medicaid Services Doctors and Clinicians National Downloadable File from December 31, 2014, to December 31, 2020, this case-control study compared employment changes for physicians at PE-exiting practices sold between January 1, 2016, and December 31, 2018 (treatment group), with employment changes for matched control physicians in practices not sold by PE owners but with the same specialty, hospital referral region, practice size, and time period. Physicians billing fee-for-service Medicare during the study period were eligible for inclusion. A difference-in-differences design was used to compare retention between the treatment and control groups in the 2 years before and after exit using a multinomial logit model that adjusted for physician decade of graduation. Data were analyzed from August 1, 2023, to November 9, 2024.

EXPOSURE: Exit of a PE-owned physician practice.

MAIN OUTCOMES AND MEASURES: Physician employment outcomes included staying (continuing to bill through the initial practice), working elsewhere (only billing through other practices), and retirement (no longer billing). Whether a physician left to a join large (>120-physician) practice was also evaluated.

RESULTS: Of the 1215 physicians included in the analysis (405 at 70 PE-exiting practices and 810 matched controls; 814 [67.0%] male and 401 (33.0%) were female. Physicians in all PE-exiting practices were typically in practices of more than 20 physicians (471 [65.2%]) and often in the South (373 [51.7%]). Dermatology was the leading specialty (216 [29.9%]), followed by family medicine (94 [13.0%]). Physicians employed in PE-exiting practices were 16.5 (95% CI, 10.6-22.3) percentage points less likely to continue working in that practice 2 years after exit compared with matched controls. There was no significant change in the probability of retirement (0 percentage points; 95% CI, -4.1 to 4.0). Physicians in PE-exiting practices were 10.1 (95% CI, 6.5 to 13.7) percentage points likelier than matched controls to join a large practice of more than 120 physicians.

CONCLUSIONS AND RELEVANCE: In this case-control study, PE exit was followed by an increase in physician turnover and subsequent employment at a large (>120-physician) practice relative to matched controls, notwithstanding similar turnover rates between these physicians and matched controls prior to exit. The increase in physician turnover and consolidation following PE exits has important implications for patients, physicians, investors, and physician markets.

PMID:39951313 | DOI:10.1001/jamahealthforum.2024.5376

JAMA Health Forum. 2025 Feb 7;6(2):e245385. doi: 10.1001/jamahealthforum.2024.5385.

ABSTRACT

IMPORTANCE: For patients with heart failure with reduced ejection fraction (HFrEF), treatment with sacubitril-valsartan, an angiotensin receptor-neprilysin inhibitor, has become increasingly preferred over angiotensin-converting enzyme inhibitors (ACE-Is) and angiotensin II receptor blockers (ARBs). However, sacubitril-valsartan is much more expensive than generic ACE-I/ARBs. It is unknown whether the high cost of sacubitril-valsartan is offset by lower spending on hospitalizations and other treatments.

OBJECTIVE: To compare total and out-of-pocket health care spending among Medicare beneficiaries initiating sacubitril-valsartan vs ACE-I/ARBs for HFrEF.

DESIGN, SETTING, AND PARTICIPANTS: This was a cohort study using data from Medicare fee-for-service claims with propensity score matching of Medicare beneficiaries with HFrEF. Data analysis was performed from November 2022 to December 2023.

EXPOSURE: Initiation of sacubitril-valsartan or an ACE-I/ARB. Patients were matched by propensity score based on 104 covariates, including demographic characteristics, comorbidities, baseline annual spending, and baseline use of health care services.

MAIN OUTCOMES AND MEASURES: Mean total and out-of-pocket health care expenditures during the 365 days after initiating sacubitril-valsartan or an ACE-I/ARB. Censoring for incomplete follow-up was addressed using Kaplan-Meier probability weighting. Cost differences, cost ratios, and 95% CIs were calculated using a nonparametric bootstrapping method with 500 samples drawn with replacement.

RESULTS: Among 13 755 matched pairs of Medicare patients with HFrEF (mean [SD] age, 77.5 [7.5] years; 5138 [39%] 80 years or older; 9949 females [36%] and 17 561 males [64%]), mean annual total health care spending per person was similar for sacubitril-valsartan initiators and ACE-I/ARB initiators (difference, $701; 95% CI, -$132 to $1593). Sacubitril-valsartan initiators had higher prescription drug costs (difference, $1911; 95% CI, $1704 to $2113), lower inpatient costs (difference, -$790; 95% CI, -$1468 to -$72), lower outpatient costs (difference, -$330; 95% CI, -$664 to -$11), and higher annual out-of-pocket spending (difference, $109; 95% CI, $13 to $208).

CONCLUSIONS AND RELEVANCE: This cohort study found that Medicare beneficiaries initiating sacubitril-valsartan to treat HFrEF had similar total health care spending as those initiating ACE-I/ARBs; higher prescription drug spending was offset by lower inpatient and outpatient spending. However, sacubitril-valsartan use was associated with higher patient out-of-pocket costs, which may exacerbate health disparities and limit access and affordability.

PMID:39951312 | DOI:10.1001/jamahealthforum.2024.5385

Rheumatology (Oxford). 2025 Feb 14:keaf103. doi: 10.1093/rheumatology/keaf103. Online ahead of print.

ABSTRACT

OBJECTIVE: Rheumatoid arthritis (RA) is a chronic inflammatory disease associated with several comorbidities, including an increased risk of certain cancers. This study aimed to investigate the potential associations between RA and increased risk of urological cancers-specifically kidney, bladder, prostate and testicular cancers-and the influence of RA serological status on this risk.

METHODS: This retrospective cohort study used data from the Korean National Health Insurance System database (2010-2020), including patients with RA and a 1:5 matched non-RA population. RA patients were grouped according to serological status. The primary outcome was newly diagnosed urological cancer, and its association with RA was analyzed by Cox proportional hazards regression analyses adjusting for potential confounders.

RESULTS: RA patients had an increased risk of kidney cancer compared with the non-RA population (adjusted hazard ratio [aHR], 1.34 [95% confidence interval (CI), 1.04-1.78]). The risk of kidney cancer was even higher in women with RA (aHR 1.57 [95% CI: 1.10, 2.24]). However, the risk of bladder, prostate and testicular cancers was not associated with RA (bladder cancer, aHR 1.24, 95% CI 0.95-1.62; prostate cancer, aHR 1.13, 95% CI 0.94-1.35; testicular cancer, aHR 2.31, 95% CI 0.44-12.20). No significant difference in urological cancer risk was found according to serological status.

CONCLUSIONS: RA patients have a higher risk of kidney cancer than the general population. Further research is needed to understand the mechanisms underlying the association between RA and kidney cancer to optimize cancer prevention and screening strategies.

PMID:39951303 | DOI:10.1093/rheumatology/keaf103

Bioinformatics. 2025 Feb 14:btaf071. doi: 10.1093/bioinformatics/btaf071. Online ahead of print.

ABSTRACT

SUMMARY: Polygenic risk scores (PRS) hold promise for early disease diagnosis and personalized treatment, but their overall discriminative power remains limited for many diseases in the general population. As a result, numerous novel PRS modeling techniques have been developed to improve predictive performance, but determining the most effective method for a specific application remains uncertain until tested. Hence, we introduce a novel, versatile tool for building an optimized PRS model by integrating candidate models from multiple existing PRS building methods that use target population data and/or incorporating information from other populations through a trans-ethnic approach. Our tool, PNL is based on PairNet algorithm, a Convolutional Neural Network with low computation complexity through simple paring operation. In the case studies for asthma, type 2 diabetes, and vertigo, the optimal PRS model generated with PNL using only TWB data achieved AUCs that matched or improved the best results using other methods individually. Incorporating UKBB data further improved performance of PNL for asthma and type 2 diabetes. For vertigo, unlike the other diseases, individual method analysis showed that UKBB data alone generally produced lower AUCs compared to TWB data alone. As a result, incorporating UKBB data did not improve AUC with PNL, suggesting that increasing the number of candidate models does not necessarily result in higher AUC values, alleviating concerns about overfitting.

AVAILABILITY AND IMPLEMENTATION: The python code for PairNet algorithm incorporated in PNL is freely available on: https://github.com/FannLab/pairnet. An archived, citable version is stored on: https://doi.org/10.5281/zenodo.14838227.

CONTACT: Correspondence should be addressed to corresponding authors.

SUPPLEMENTARY INFORMATION: Detailed implementation procedures can be found in the Supplementary Materials.

PMID:39951285 | DOI:10.1093/bioinformatics/btaf071

JAMA Netw Open. 2025 Feb 3;8(2):e2458531. doi: 10.1001/jamanetworkopen.2024.58531.

ABSTRACT

IMPORTANCE: Whether population-based racial and ethnic survival disparities for children with high-risk neuroblastoma persist in the clinical trial setting is unknown.

OBJECTIVE: To investigate racial and ethnic survival disparities among children with high-risk neuroblastoma treated on frontline clinical trials.

DESIGN, SETTING, AND PARTICIPANTS: This retrospective cohort study used data from Children’s Oncology Group (COG) high-risk neuroblastoma trials from January 1, 2007, to December 31, 2016, with a data freeze on June 30, 2021. Children with high-risk neuroblastoma were analyzed in 2 cohorts: induction/consolidation trial participants and post-consolidation trial participants. Statistical analyses were performed from September 2, 2021, to December 30, 2024.

EXPOSURES: Race and ethnicity were the primary exposures, categorized as Hispanic, non-Hispanic Black, non-Hispanic other (American Indian or Alaska Native, Asian, and Native Hawaiian or Other Pacific Islander), or non-Hispanic White.

MAIN OUTCOMES AND MEASURES: Primary outcomes included overall survival (OS) and event-free survival (EFS) from time of trial enrollment, estimated by Kaplan-Meier methods. Associations with race and ethnicity were evaluated by log-rank tests and Cox proportional hazards regression models. Secondary outcomes included induction delays, early trial withdrawal, relapse as first event, death as first event, postrelapse OS, and early phase trial enrollment.

RESULTS: The induction/consolidation cohort (median follow-up, 8.3 years [IQR, 6.1-9.8 years]) included 696 patients (404 males [58.1%]; 79 Hispanic patients [11.4%], 109 non-Hispanic Black patients [15.7%], 27 patients of non-Hispanic other race [3.9%], and 481 non-Hispanic White patients [69.1%]). The post-consolidation cohort (median follow-up, 7.5 years [IQR, 5.8-9.4 years]) included 935 patients (567 males [60.6%]; 87 Hispanic patients [9.3%], 145 non-Hispanic Black patients [15.5%], 41 patients of non-Hispanic other race [4.4%], and 662 non-Hispanic White patients [70.8%]). In multivariable Cox proportional hazards regression models, Hispanic children experienced significantly inferior OS (hazard ratio [HR], 1.78; 95% CI, 1.25-2.53; P = .01) on induction/consolidation studies compared with non-Hispanic White children; EFS did not differ. Non-Hispanic Black (HR, 1.54; 95% CI, 1.13-2.11) and Hispanic children (HR, 1.63; 95% CI, 1.09-2.43) experienced inferior OS on post-consolidation studies compared with non-Hispanic White children (P = .009); Hispanic children in post-consolidation studies experienced inferior EFS (HR, 1.68; 95% CI, 1.14-2.47; P = .02). Death as first event and postrelapse OS also differed by race and ethnicity.

CONCLUSIONS AND RELEVANCE: This study suggests that Black and Hispanic children with high-risk neuroblastoma experienced inferior OS despite uniform planned treatment on frontline COG clinical trials. Investigated mechanisms did not completely explain survival disparities. Future evaluation of disparate treatment-related toxicities and postrelapse care as explanatory mechanisms are key next steps to promote equity.

PMID:39951269 | DOI:10.1001/jamanetworkopen.2024.58531

JAMA Netw Open. 2025 Feb 3;8(2):e2459567. doi: 10.1001/jamanetworkopen.2024.59567.

ABSTRACT

IMPORTANCE: Dizziness symptoms account for nearly 2 million annual emergency department (ED) visits and present a diagnostic challenge for clinicians. Most dizziness research has focused on improving guideline-concordant care among clinicians, with little focus on developing patient-centered interventions to improve dizziness-related disability.

OBJECTIVE: To examine the feasibility of ED vestibular rehabilitation therapy (ED-VeRT) using a protocolized diagnostic classification algorithm and collection of longitudinal patient-reported outcomes.

DESIGN, SETTING, AND PARTICIPANTS: A pilot nonrandomized clinical trial of ED-VeRT vs usual care for patients presenting to the ED with dizziness at a single urban US ED was conducted from November 16, 2021, to February 6, 2023, with collection of 3-month outcomes through May 1, 2023. Patients were allocated to ED-VeRT or usual care at the discretion of the treating physician.

INTERVENTIONS: Use of ED-VeRT was delivered by an ED physical therapist via a protocolized diagnostic classification and treatment algorithm based on a diagnosis of benign paroxysmal positional vertigo, triggered undifferentiated dizziness, spontaneous undifferentiated dizziness, or unilateral peripheral hypofunction.

MAIN OUTCOMES AND MEASURES: Feasibility outcomes included participant screening, enrollment, and retention rates to inform the design of a future randomized clinical trial; retention was defined as completing any of 4 follow-up surveys over 3 months. The primary efficacy outcome was change in the Dizziness Handicap Inventory score; the secondary efficacy outcome was change in the Vestibular Activities Avoidance Inventory-9 score.

RESULTS: Of 366 patients screened, 125 participants were enrolled (median age, 52 [IQR, 40-66] years, 73 [58%] female, 61 [49%] White), and 105 retained (84.0%) in longitudinal data collection. Sixty-three participants (50.4%) received ED vestibular therapy and were assigned to primary diagnostic classifications of benign paroxysmal positional vertigo (23 [37.1%]), triggered undifferentiated dizziness (14 [22.6%]), spontaneous undifferentiated dizziness (14 [22.6%]), or unilateral peripheral hypofunction (9 [14.5%]). Despite having higher Dizziness Handicap Inventory and Vestibular Activities Avoidance Inventory scores at baseline, ED-VeRT participants reported lower dizziness handicap (difference: -1.68; 95% CI, -11.30 to 7.90) and vestibular activities avoidance (difference: -2.27; 95% CI, -8.40 to 3.86) at 3 months, although these differences were not statistically significant.

CONCLUSIONS AND RELEVANCE: In this nonrandomized clinical trial, ED vestibular therapy was feasibly delivered to patients presenting to the ED with undifferentiated dizziness symptoms. For participants receiving vestibular therapy the findings for dizziness-related disability over 3 months were not statistically significant, pointing to the need for a fully powered randomized clinical trial.

TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT05122663.

PMID:39951266 | DOI:10.1001/jamanetworkopen.2024.59567