Tag: statistics

JAMA Netw Open. 2026 Jun 1;9(6):e2618199. doi: 10.1001/jamanetworkopen.2026.18199.

ABSTRACT

IMPORTANCE: Violent injuries are the leading cause of death among children and young adults in the United States. Violent injury may increase the risk of subsequent injury (ie, violent reinjury), but it is unclear how this risk varies across different injury types (eg, blunt assault, stabbing, firearm).

OBJECTIVE: To describe patterns of violent reinjury across different injury types to inform tailored prevention strategies.

DESIGN, SETTING, AND PARTICIPANTS: This multicenter cohort study included patients who presented with violent injuries to 2 adult and 2 pediatric level I trauma centers in St Louis, Missouri, from January 2016 to December 2023.

EXPOSURES: Violent injuries were classified into 3 injury types: blunt assault, stabbing, and firearm.

MAIN OUTCOMES AND MEASURES: The main outcome was time to recurrent violent injury in years. We used Kaplan-Meier time-to-event analyses to estimate the probability of violent reinjury across different initial (index) and recurrent injury types at 1, 5, and 7 years from index injury.

RESULTS: There were 29 810 individuals with violent injuries during the study period. The age distribution was as follows: 17 228 (57.8%) older than 26 years; 7690 (25.8%), 18 to 26 years; and 4727 (15.8%) younger than 18 years. Most were male (19 933 [66.9%]), and there were 21 126 (70.9%) Black and 7031 (23.6%) White individuals, with 792 (2.7%) belonging to an additional racial group (American Indian or Alaska Native, Asian, Native Hawaiian or Pacific Islander, other, or multiracial), having unknown race, or declining to disclose race. Among 29 038 patients (97.4%) who survived their index injury, 3899 (13.4%) experienced reinjury. The 7-year probability of any reinjury was highest among individuals with index stabbing (19.5% [95% CI, 17.5%-21.8%]) and blunt assault (17.4% [95% CI, 16.6%-18.2%]) injuries. Patients with index blunt assault injuries were more likely to experience blunt assault reinjury (7-year probability: 13.2% [95% CI, 12.6%-13.8%]) compared with other reinjury types. The 7-year probability of firearm reinjury was highest among those with index firearm injuries (12.6% [95% CI, 11.8%-13.4%]).

CONCLUSIONS AND RELEVANCE: In this cohort study of individuals with violent injuries, reinjury was common, but this risk varied across injury types, which may be due to differing contextual factors. Preventing violent reinjury requires attention to these issues to better address the root causes of repeated violence exposure.

PMID:42307949 | DOI:10.1001/jamanetworkopen.2026.18199

JAMA Netw Open. 2026 Jun 1;9(6):e2618581. doi: 10.1001/jamanetworkopen.2026.18581.

ABSTRACT

IMPORTANCE: Antigenically drifted influenza A(H3N2) J.2.4.1 (subclade K) viruses predominated during the 2025-2026 Northern Hemisphere influenza season.

OBJECTIVE: To describe influenza activity and burden, characterize subclade K, evaluate susceptibility to influenza antivirals and postinfluenza vaccination antibodies, and estimate vaccine effectiveness.

DESIGN, SETTING, AND PARTICIPANTS: This surveillance study used multiple data sources, including (1) national surveillance of influenza-positive respiratory specimens collected by approximately 300 clinical laboratories and 100 public health laboratories from October 1, 2025, through March 14, 2026, a subset of which were further characterized; (2) serologic data of people who received 2025-2026 influenza vaccines; (3) influenza admissions data from the Influenza Hospitalization Surveillance Network (ie, 10% of US population) and the associated estimates of US burden; and (4) test-negative, case-control vaccine effectiveness estimates from the Virtual SARS-CoV-2, Influenza, and Other Respiratory Viruses Network.

EXPOSURES: Influenza infection, hospitalization, and vaccination.

MAIN OUTCOMES AND MEASURES: Outcomes included influenza virus type, subtype, and clade; antiviral susceptibility; immunogenicity; influenza-associated outpatient and emergency department visits, hospitalizations, and mortality; estimated influenza illnesses, hospitalizations, and death; and estimated vaccine effectiveness.

RESULTS: As of March 14, 2026, of the 55 318 influenza-positive respiratory specimens tested by public health laboratories, most (50 291 specimens [90.9%]) were influenza A, of which 40 779 (81.1%) were subtyped and 35 801 (87.8%) were A(H3N2). Of the 1754 characterized A(H3N2) viruses, most (1626 specimens [92.7%]) were subclade K. Postinfluenza vaccination neutralizing geometric mean antibody titers against subclade K were reduced 1.62 (95% CI, 1.29-2.02)-fold compared with the vaccine virus. All 1729 tested A(H3N2) viruses were sensitive to antivirals. Of the 27 881 recorded influenza hospitalizations, 15 426 (54.7%) were among female patients, and 15 051 (54.0%) were among patients aged 65 years or older. The estimated cumulative influenza-associated hospitalization rate was 80.0 per 100 000 which would correlate with estimates of between 28 000 000 to 49 000 000 illnesses, 360 000 to 740 000 hospitalizations, and 22 000 to 74 000 deaths in the US during the 2025-2026 season. Adjusted interim vaccine effectiveness estimates against influenza-associated emergency department or urgent care encounters and hospitalizations were 35% (95% CI, 33%-38%) and 27% (95% CI, 21%-34%), respectively.

CONCLUSIONS AND RELEVANCE: This surveillance study found that while antigenically drifted viruses predominated and caused substantial morbidity and mortality, influenza vaccines were associated with a reduced risk of influenza among those who were vaccinated, and recommended antivirals remained effective.

PMID:42307948 | DOI:10.1001/jamanetworkopen.2026.18581

JAMA Netw Open. 2026 Jun 1;9(6):e2618883. doi: 10.1001/jamanetworkopen.2026.18883.

ABSTRACT

IMPORTANCE: Gestational environmental chemical exposures are widespread. Some chemicals are known to adversely affect birth outcomes, but many remain understudied.

OBJECTIVE: To evaluate associations of gestational exposure to a priori identified chemicals in 10 classes with birth outcomes in a large, diverse US cohort.

DESIGN, SETTING, AND PARTICIPANTS: In the prospective Environmental influences on Child Health Outcomes Cohort study, 5318 mother-child pairs were enrolled from January 1, 2000, to December 31, 2021, with data on gestational urinary chemical concentrations, gestational age at birth, and birth weight. Statistical analysis was performed from January 2024 to February 2026.

EXPOSURES: In single, midgestation (median, 25 weeks [IQR, 21-30 weeks]) urine samples, concentrations of 113 analytes (chemicals or their metabolites) from 10 chemical classes were simultaneously measured: fungicides and herbicides (n = 11), insecticides (n = 20), halogenated phenols (n = 5), organophosphate esters (n = 10), benzophenones (n = 6), bisphenols (n = 14), parabens (n = 6), antimicrobials (n = 2), phthalates or alternative plasticizers (n = 32), and polycyclic aromatic hydrocarbons (PAHs) (n = 7).

MAIN OUTCOMES AND MEASURES: Linear mixed-effects regression models with a random effect for site were used to estimate covariate-adjusted differences in gestational age at birth (days) and birth weight-for-gestational age (BW-GA) z scores per IQR increase in urinary analyte concentrations. In secondary analyses, odds ratios (ORs) for preterm birth and small for gestational age (SGA) were estimated.

RESULTS: In the sample of 5318 mother-child pairs, most infants (2667 female [50%]; median gestational age at birth, 39.0 weeks [IQR, 38.0-40.0 weeks]) were born to college-educated (67% [3218 of 4785]), parous (56% [2815 of 5007]) mothers (median age at delivery, 30.7 years [IQR, 26.1-34.3 years]). A total of 43 of 113 analytes (38%) were detected in 50% or more of samples. Multiple phthalates or alternative plasticizers were associated with younger gestational age at birth or lower BW-GA z scores; for example, summed diisononyl phthalate metabolites were associated with a 0.6-day (95% CI, -1.0 to -0.1 days) younger gestational age (preterm birth OR, 1.16 [95% CI, 1.01-1.34]), and summed phthalate or alternative plasticizers were associated with a 0.06 (95% CI, -0.11 to -0.02) lower BW-GA z score (SGA OR, 1.09 [95% CI, 0.93-1.27]). Two halogenated phenols, benzophenone 8, bisphenol F, and several PAHs were associated with lower BW-GA z scores; for example, 1- and 9-hydroxphenanthrene were associated with a 0.04 (95% CI, -0.08 to -0.01) lower BW-GA z score (SGA OR, 1.13 [95% CI, 1.01-1.27]).

CONCLUSIONS AND RELEVANCE: This large cohort study of diverse US pregnancies found widespread exposure to 10 classes of environmental chemicals, many of which were associated with differences in gestational age at birth or lower BW-GA z scores. These findings indicate that reducing gestational exposure to chemicals, particularly phthalates or alternative plasticizers and PAHs, could promote healthy deliveries and better child outcomes.

PMID:42307947 | DOI:10.1001/jamanetworkopen.2026.18883

JAMA Netw Open. 2026 Jun 1;9(6):e2619054. doi: 10.1001/jamanetworkopen.2026.19054.

ABSTRACT

IMPORTANCE: With increasing abortion restrictions, more people self-manage their abortions with medications. Little is known about geographic and demographic characteristics of those self-managing abortions in the US, particularly after the Dobbs v Jackson Women’s Health Organization decision.

OBJECTIVE: To describe requests for self-managed medication abortion support in the US through Mexico-based accompaniment collectives.

DESIGN, SETTING, AND PARTICIPANTS: This is a cross-sectional study of records from 2023 from 2 Mexico-based accompaniment collectives. Participants were people in the US requesting support from Mexico-based accompaniment collectives to self-manage an abortion.

EXPOSURES: State abortion policy context (complete abortion bans, 6-week and/or embryonic cardiac activity bans, 12- to 15-week bans, and abortion available past 15 weeks), sociodemographic characteristics (poverty, insurance status, and Spanish language), and driving distance from the nearest abortion facility.

MAIN OUTCOMES AND MEASURES: The primary outcomes were the sociodemographic characteristics of those requesting support from accompaniment collectives and the rate of requests to these collectives by zip code tabulation areas (ZCTAs). Demographic characteristics, pregnancy characteristics, and state policy context were described. Differences in request characteristics by state-level abortion policy category were tested using χ2 tests. Also, rates of requests per 100 000 female individuals aged 15 to 49 years within categories of ZCTA-level characteristics were estimated.

RESULTS: A total of 2850 requests were analyzed, with 700 requests (24.6%) from people aged 20 to 24 years, 995 (34.9%) from people aged 25 to 34 years, and 192 (6.7%) from people younger than 20 years. Pregnancy duration was 6 weeks or less for 1441 requests (50.6%), 7 to 9 weeks for 677 requests (23.8%), and 12 weeks or more for 160 requests (5.6%). In total, 2363 requests (82.9%) were from states with complete abortion bans. Requests with pregnancies at 12 weeks or more ranged from 98 (4.2%) in complete-ban states to 36 requests (14.8%) in states where abortion was available past 15 weeks (P < .001). Spanish was spoken with accompaniers by 65 requesters (38.0%) in states where abortion was available past 15 weeks and by 282 requesters (13.2%) in complete-ban states (P < .001). The mean ZCTA-level request rate was 2.44 (95% CI, 1.81-3.08) per 100 000 female individuals aged 15 to 49 years and increased with ZCTA-level poverty (from 0.35 [95% CI, 0.23-0.53] to 1.63 [95% CI, 1.14-2.34] requests per 100 000 female individuals aged 15 to 49 years) and Spanish-speaking population share (from 0.25 [95% CI, 0.12-0.54] to 1.56 [95% CI, 1.03-2.39] requests per 100 000 female individuals aged 15 to 49 years). ZCTAs more than 8 hours from an abortion facility in complete ban states had the highest request rate (7.85 [95% CI, 4.53-13.61] requests per 100 000 female individuals aged 15 to 49 years).

CONCLUSIONS AND RELEVANCE: In this cross-sectional study of accompaniment network records, requests for Mexico-based accompaniment support were highest in states with abortion bans and in communities with elevated poverty, uninsurance, and Spanish-speaking populations, underscoring the role of accompaniment collectives in serving communities facing the greatest structural barriers to facility-based abortion care.

PMID:42307945 | DOI:10.1001/jamanetworkopen.2026.19054

JAMA Netw Open. 2026 Jun 1;9(6):e2619285. doi: 10.1001/jamanetworkopen.2026.19285.

ABSTRACT

IMPORTANCE: Validating polygenic risk scores (PRSs) as a breast cancer (BC) risk stratification tool in high-risk clinical settings is essential for potential clinical implementation.

OBJECTIVE: To evaluate the performance of African-ancestry PRSs for overall BC and triple-negative BC (TNBC) in an independent African American cohort at elevated familial risk.

DESIGN, SETTING, AND PARTICIPANTS: This case-control study was conducted among women with self-reported African ancestry in a clinical hereditary cancer genetic testing setting who had negative results for pathogenic or likely pathogenic variants in known BC-associated genes between 2016 to 2024. Analyses were performed from September 2025 to April 2026.

EXPOSURE: Six African-ancestry PRSs and a European-ancestry PRS (PRS-313).

MAIN OUTCOMES AND MEASURES: Performance of the PRSs was measured with covariate-adjusted area under the receiver operating characteristic curve (AUC) and adjusted odds ratio (OR) per 1 SD from logistic regression adjusting for age, top 10 genetic principal components, and family history.

RESULTS: This study included 31 522 women (mean [SD] age, 47.1 [13.3] years): 12 067 women with BC (2311 with TNBC) and 19 455 women unaffected by BC. Women unaffected by BC were younger at testing than women with BC at diagnosis (mean [SD] age, 42.1 [11.7] years vs 55.2 [11.8] years) and were more likely to report a first- and second-degree family history of BC than women with BC (14 524 women [74.7%] vs 6483 women [53.7%]). For overall BC, PRS-313 showed an AUC of 0.567 (95% CI, 0.560-0.574) and an OR of 1.28 (95% CI, 1.25-1.31) per 1 SD of individuals unaffected by BC. In comparison, the African overall BC models 1 (2 324 063-variant model) and 2 (175 173-variant model) performed better, with AUCs of 0.588 (95% CI, 0.580-0.595) and 0.584 (95% CI, 0.576-0.591) and ORs of 1.39 (95% CI, 1.35-1.43) and 1.37 (95% CI, 1.34-1.41) per 1 SD of individuals unaffected by BC, respectively. Importantly, for TNBC, one 162-variant PRS consistently outperformed all other models, with an AUC of 0.609 (95% CI, 0.596-0.622) and an OR of 1.47 (95% CI, 1.40-1.55) per 1 SD of individuals unaffected by BC.

CONCLUSIONS AND RELEVANCE: In this clinically ascertained case-control study of self-reported Black or African American women, the PRSs demonstrated good performance among women with a strong family history of BC, reflecting populations in whom early PRS testing is most relevant. The high accuracy of the 162-variant TNBC PRS supported its potential as a cost-effective risk assessment tool to promote equitable care.

PMID:42307944 | DOI:10.1001/jamanetworkopen.2026.19285

JAMA. 2026 Jun 17. doi: 10.1001/jama.2026.7745. Online ahead of print.

ABSTRACT

IMPORTANCE: Patients with epidermal growth factor receptor (EGFR) gene variant nonsquamous non-small cell lung cancer (NSCLC) who have disease progression after prior EGFR tyrosine kinase inhibitor (TKI) therapy have limited treatment options, creating a need for more effective subsequent therapies.

OBJECTIVE: To provide final overall results of a trial assessing whether adding ivonescimab (a bispecific antibody targeting programmed cell death protein 1 and vascular endothelial growth factor) to chemotherapy improves overall survival in this population.

DESIGN, SETTING, AND PARTICIPANTS: Randomized, double-blind, placebo-controlled phase 3 trial conducted at 55 sites in China. From January 25 to November 2, 2022, a total of 322 adult patients with locally advanced or metastatic EGFR-variant nonsquamous NSCLC who had received prior EGFR-TKI therapy were enrolled. The data cutoff date was April 12, 2025.

INTERVENTIONS: Patients were randomized 1:1 to receive ivonescimab (20 mg/kg; n = 161) or placebo (n = 161) plus chemotherapy with pemetrexed and carboplatin once every 3 weeks for 4 cycles, followed by maintenance therapy.

MAIN OUTCOMES AND MEASURES: This final results report focuses on overall survival, the key secondary end point, tested in a hierarchical manner (the primary end point was progression-free survival assessed by an independent radiology review committee).

RESULTS: The 322 enrolled patients had a median age of 59.4 years, and 51.6% were female. During a median follow-up of 32.5 months, ivonescimab plus chemotherapy improved overall survival compared with chemotherapy alone (median survival, 16.8 months vs 14.1 months; stratified hazard ratio, 0.74; 95% CI, 0.58-0.95; P = .02). The absolute difference in median overall survival was 2.7 months. Estimated 30-month survival rates were 29.1% (95% CI, 22.1%-36.4%) with ivonescimab and 18.4% (95% CI, 12.8%-24.8%) with placebo. Grade 3 or higher treatment-emergent adverse events occurred in 67.1% and 54.7% of patients receiving ivonescimab and placebo, respectively.

CONCLUSIONS AND RELEVANCE: Ivonescimab plus chemotherapy provided a statistically significant and clinically meaningful improvement in overall survival with an acceptable safety profile in patients with EGFR-variant NSCLC after EGFR-TKI therapy.

TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT05184712.

PMID:42307937 | DOI:10.1001/jama.2026.7745

Health Promot Chronic Dis Prev Can. 2026 Jun;46(6):237-244. doi: 10.24095/hpcdp.46.6.03.

ABSTRACT

INTRODUCTION: Chronic conditions are highly prevalent in Canada and are commonly examined as a single, aggregated exposure in population research on social isolation. Such approaches emphasize overall disease burden but make it difficult to distinguish the independent contributions of diagnostic category, chronic pain and disability. In this study, we examine these dimensions separately to assess how each is associated with social isolation among Canadian adults.

METHODS: Using the 2022 Mental Health and Access to Care Survey (n = 9861), the association between chronic conditions, chronic pain, and disability in relation to social support was assessed, using the Social Provisions Scale (SPS-10), applying multivariable linear regression.

RESULTS: More severe disability was negatively associated with social support (B = -0.09, 95% CI = -0.11, -0.08). Those with more functional impairments experienced lower social support which typically indicates greater social isolation.

CONCLUSION: When examined jointly, functional disability, but not chronic disease category or chronic pain, was independently associated with lower social support. These findings indicate that social isolation among Canadian adults is more closely related to functional limitation than to diagnostic labels, underscoring the importance of function-focused approaches in research and intervention.

PMID:42307906 | DOI:10.24095/hpcdp.46.6.03

Health Promot Chronic Dis Prev Can. 2026 Jun;46(6):215-225. doi: 10.24095/hpcdp.46.6.01.

ABSTRACT

INTRODUCTION: The objective of this study is to evaluate adherence to seven Canadian Lower-Risk Cannabis Use Guideline (LRCUG) recommendations among Canadian university students and identify subgroups of high-risk users.

METHODS: We analyzed survey data collected across four Canadian universities under the World Mental Health-International College Student (WMH-ICS) initiative. Seven of the ten 2017 LRCUG recommendations were evaluated. Zero-inflated Poisson models were employed to examine the sociodemographic correlates of (1) any lifetime cannabis use; and (2) the number of unmet LRCUG recommendations, conditional on lifetime use. Additionally, multivariable binary logistic regression models examined the sociodemographic correlates of adherence to individual recommendations.

RESULTS: Among the 27 236 respondents, the prevalence of lifetime cannabis use was 33.8%. Of the seven recommendations evaluated, “choosing lower-strength cannabis products” had the lowest adherence rate (29.0%), followed by “not smoking cannabis” (36.7%). “Not using synthetic cannabis” had the highest adherence rate (96.1%), followed by “delaying cannabis use until age 16” (91.2%). Men, non-heterosexual students, students living in shared housing, and domestic students were more likely to use cannabis and, among users, reported risky use. While White students were more likely to use cannabis, among users, many non-White student groups reported riskier use.

CONCLUSION: Although most students did not use cannabis and many of the LRCUG recommendations had high rates of adherence, there were low rates of choosing lower-strength cannabis products and avoiding smoking cannabis among users. Study findings highlight specific recommendations and subpopulations to inform tailoring of future interventions targeting university students.

PMID:42307904 | DOI:10.24095/hpcdp.46.6.01

Spine Deform. 2026 Jun 17. doi: 10.1007/s43390-026-01481-0. Online ahead of print.

ABSTRACT

BACKGROUND: Delayed presentation of Adolescent Idiopathic Scoliosis (AIS) is common in the UK and is characterised by larger Cobb angles at diagnosis. The size of the Cobb angle influences the mode and effectiveness of treatment strategies and worsens the overall outcome and prognosis. Severe scoliosis typically necessitates surgical correction. We hypothesised that social deprivation may be one factor that contributes to delayed presentation.

AIMS: To assess whether social deprivation is associated with greater spinal curvature at first consultation.

METHODS: A retrospective study of children presenting with AIS to a major UK children’s hospital between 2013 and 2022. Cobb angle and the English Index of Multiple Deprivation (IMD) 2019 rank were initially analysed as continuous variables using scatter plot visualisation and linear regression. Subsequently, a case-control analysis was performed with revised groupings: cases were defined as patients with large, neglected curves (Cobb angle > 70°) and controls as patients with nonoperative scoliosis (Cobb angle < 50°). Patients with Cobb angles between 50° and 70° (n = 31) were excluded from the case-control analysis to reduce misclassification. Statistical analysis included descriptive statistics, linear regression, and logistic regression.

RESULTS: A total of 364 patients were identified. Linear regression of IMD rank against Cobb angle across the full cohort demonstrated a non-significant regression coefficient (β = 19.90, p = 0.335), with a Pearson correlation of r = 0.051. After excluding 31 patients with Cobb angles 50-70°, the case-control analysis included 113 cases and 220 controls. Logistic regression yielded an odds ratio for IMD decile of 0.992 (95% CI 0.923-1.066, p = 0.828), confirming no significant association between social deprivation and curve severity.

CONCLUSION: This study did not find a significant association between social deprivation and severity of scoliosis at presentation. Findings were consistent across both continuous and categorical analyses. However, the lack of variability in the studied population and potential overmatching may have limited the ability to detect meaningful differences. Large-scale, multicentre studies are needed.

PMID:42307890 | DOI:10.1007/s43390-026-01481-0

Behav Res Methods. 2026 Jun 17;58(7):201. doi: 10.3758/s13428-026-03070-5.

ABSTRACT

In this study we examine how a mixed-effects model with crossed random effects for individuals and variables estimates within- and between-variability in longitudinal multivariate trajectories from cohort-sequential designs. These designs are characterized by large proportions of planned missing data, and they usually require continuous-time metrics. Via simulations, we evaluated different model outcomes under various conditions regarding the size of clusters (individuals and variables) and the complexity of the trajectories. Results show that (a) this model can estimate the general trajectories (common to all individuals and variables) and their variability, plus the variable-specific trajectories through the predictions of the levels of the random factors; (b) the standard errors of the random effects are wide, yet they are important for making substantive decisions for specific variables; and (c) the model predictions can adequately forecast individual and variable-specific complete trajectories from just a few observations per individual. These results are supported in an empirical illustration using cognitive developmental data. These findings show that researchers can obtain complete individual trajectories for multiple variables throughout a target age range. The relative simplicity of this model in comparison with other alternatives makes it a promising and accessible tool for multivariate longitudinal data analysis.

PMID:42307878 | DOI:10.3758/s13428-026-03070-5