Multivariate methods for clustered binary data with multiple subclasses, with application to binary longitudinal data.

B Rosner

Home/Archive/PubMed

PubMed · 1420836

Multivariate methods for clustered binary data with multiple subclasses, with application to binary longitudinal data.

B Rosner

View original source ↗Export metadata ↓

Abstract

Clustered binary data occur frequently in biostatistical work. Several approaches have been proposed for the analysis of clustered binary data. In Rosner (1984, Biometrics 40, 1025-1035), a polychotomous logistic regression model was proposed that is a generalization of the beta-binomial distribution and allows for unit- and subunit-specific covariates, while controlling for clustering effects. One assumption of this model is that all pairs of subunits within a cluster are equally correlated. This is appropriate for ophthalmologic work where clusters are generally of size 2, but may be inappropriate for larger cluster sizes. A beta-binomial mixture model is introduced to allow for multiple subclasses within a cluster and to estimate odds ratios relating outcomes for pairs of subunits within a subclass as well as in different subclasses. To include covariates, an extension of the polychotomous logistic regression model is proposed, which allows one to estimate effects of unit-, class-, and subunit-specific covariates, while controlling for clustering using the beta-binomial mixture model. This model is applied to the analysis of respiratory symptom data in children collected over a 14-year period in East Boston, Massachusetts, in relation to maternal and child smoking, where the unit is the child and symptom history is divided into early-adolescent and late-adolescent symptom experience.

Explore related subjects

Adolescent Adult Child Cluster Analysis Humans Longitudinal Studies Mathematics Models, Statistical Multivariate Analysis Odds Ratio Regression Analysis Respiratory Tract Diseases Journal Article Research Support, U.S. Gov't, P.H.S.

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

B Rosner. 1992. Multivariate methods for clustered binary data with multiple subclasses, with application to binary longitudinal data.. https://pubmed.ncbi.nlm.nih.gov/1420836/

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed2026

Blinatumomab for Replacing Chemotherapy in Pediatric Acute Lymphoblastic Leukemia.

BACKGROUND: Blinatumomab, a bispecific T-cell engager targeting the CD19 antigen on B cells, may offer an option to safely replace cycles of traditional chemotherapy in pediatric patients with newly diagnosed high-risk B-cell acute lymphoblastic leukemia (ALL). METHODS: We randomly assigned, in a 1:1 ratio, children with high-risk B-cell ALL to receive two cycles of blinatumomab (blinatumomab group) or two cycles of chemotherapy (control group) after consolidation. The primary end point was event-free survival as evaluated in a time-to-event analysis; the duration of event-free survival was defined as the time from randomization to the first event among resistance to protocol treatment, relapse, second cancer, or death from any cause. Our primary objective was to evaluate whether the 4-year event-free survival would be 10 percentage points higher in the blinatumomab group than in the control group. RESULTS: Overall, 709 of 768 eligible patients (92.3%) underwent randomization; 358 were assigned to the blinatumomab group and 351 to the control group. A planned interim analysis at a median follow-up of 2.9 years showed an estimated 4-year event-free survival of 83.0% (95% confidence interval [CI], 77.4 to 87.4) in the blinatumomab group and 70.3% (95% CI, 63.8 to 75.9) in the control group (P = 0.0002 in an intention-to-treat analysis). The estimated hazard ratio for a primary end-point event (blinatumomab vs. control) was 0.51 (95% CI, 0.35 to 0.73) as assessed with a Cox model. Infection related to the trial treatment occurred in 23.9% of patients in the blinatumomab group and in 69.4% of those in the control group (P<0.001). Life-threatening adverse events occurred in 2 patients (0.5%) in the blinatumomab group, including one (in 0.3%) that was fatal, and in 16 patients (4.7%) in the control group. Neurotoxic events were reported in 12.0% and 3.2%, respectively (P<0.001). Cytokine release syndrome of grade 2 or higher occurred in 1.1% of patients in the blinatumomab group. CONCLUSIONS: In children with newly diagnosed high-risk B-cell ALL, replacement of two cycles of highly toxic conventional chemotherapy with blinatumomab resulted in a significantly greater percentage of patients with event-free survival at 4 years. (Funded by Deutsche Krebshilfe and others; AIEOP-BFM ALL 2017 EudraCT number, 2016-001935-12; EU Clinical Trials number, 2023-509856-32-00; and ClinicalTrials.gov number, NCT03643276.).

Adolescent↗

PubMed2026

Association between plasma homocysteine levels and P-wave dispersion in pediatric patients with hyperhomocysteinemia.

UNLABELLED: Hyperhomocysteinemia has been recognized as a cardiovascular risk factor associated with endothelial dysfunction, oxidative stress, and vascular inflammation. Experimental and clinical studies suggest that elevated homocysteine levels may also influence myocardial electrophysiology and contribute to arrhythmogenesis. However, data regarding the relationship between homocysteine levels and electrocardiographic markers of atrial conduction in pediatric populations remain limited. This study aimed to evaluate the association between plasma homocysteine levels and electrocardiographic parameters, particularly P-wave dispersion, in children. This multicenter retrospective case-control study included pediatric patients evaluated in four tertiary pediatric metabolism centers between January 2023 and December 2025. A total of 47 patients with hyperhomocysteinemia (plasma total homocysteine ≥ 15 µmol/L) and 43 age- and sex-matched controls with normal homocysteine levels were included. Controls were selected from the screened population among children with available homocysteine measurements, electrocardiographic and echocardiographic evaluations, and no confirmed inherited metabolic disease or cardiac disorder. Clinical, biochemical, and electrocardiographic parameters, including maximum P-wave duration and P-wave dispersion, were retrospectively analyzed. A total of 90 participants were included, comprising 47 children with hyperhomocysteinemia and 43 healthy controls. P-wave dispersion and maximum P-wave duration were significantly higher in the hyperhomocysteinemia group compared with controls (48.96 [19.48-100.0] vs. 38.57 [10.57-71.19] ms, p < 0.001). Plasma homocysteine levels showed a moderate positive correlation with P-wave dispersion (ρ = 0.441, p < 0.001). These differences were more pronounced in children with higher homocysteine levels and in younger age groups (< 2 years and 2-14 years). In contrast, PR interval (p = 0.790) and QTc interval (p = 0.183) did not differ significantly between groups. Vitamin B12 levels were significantly lower in the hyperhomocysteinemia group (p = 0.013), while folate levels were comparable (p = 0.974). Although sodium, potassium, and magnesium levels differed significantly between groups, all values remained within normal physiological ranges. CONCLUSIONS: Children with hyperhomocysteinemia showed increased P-wave dispersion compared with controls. These findings suggest an association between elevated homocysteine levels and altered atrial conduction parameters in children. Further prospective studies are needed to determine the clinical significance of these findings. WHAT IS KNOWN: • Hyperhomocysteinemia is associated with cardiovascular risk and endothelial dysfunction. • Elevated homocysteine levels have been linked to cardiac electrophysiological alterations in adult populations. WHAT IS NEW: • Elevated homocysteine levels are associated with increased P-wave dispersion in children, with more pronounced effects observed in younger age groups. • These findings support an association between hyperhomocysteinemia and altered atrial conduction parameters in children.

Adolescent↗

PubMed2026

Characteristics of post-exercise responders versus non-responders following aerobic or isometric exercise in physically inactive adults of African and South Asian descent with high-normal blood pressure or grade I hypertension.

OBJECTIVE: To investigate interindividual variability in post-exercise hypotension (PEH) and to characterise cardiovascular and autonomic differences between responders and non-responders following aerobic and isometric exercise in adults of African and South Asian descent with elevated blood pressure (BP). METHODS: Physically inactive adults of African and South Asian descent living in Suriname (18-65 years) with high-normal BP or grade I hypertension participated in a randomised controlled crossover trial. In this randomised cross-over trial, 47 adults (50.1 ± 10.8 years; 38% male) with high-normal blood pressure or grade I hypertension completed three conditions: aerobic exercise (30 min at 40-60% heart rate reserve), isometric handgrip exercise, and a non-exercise control. Ambulatory BP was assessed over 24 h. PEH was defined as the net effect: (post-exercise - pre-exercise) - (post-control - pre-control). Participants were classified as responders if daytime BP decreased ≥5 mmHg. Arterial stiffness, cardiac, and autonomic parameters were assessed. RESULTS: Following aerobic exercise, 46% of participants were classified as systolic responders compared with 28% after isometric exercise. No baseline differences were observed in demographic or clinical characteristics between responders and non-responders, suggesting that PEH variability may reflect underlying physiological rather than clinical differences. Aerobic responders demonstrated greater reductions in aortic augmentation index (-19.4% vs. -10.9%, p = 0.05), larger increases in stroke volume (+8.1 vs. -5.3 mL, p = 0.05) and cardiac output (+1.54 ± 1.89 vs. +0.58 ± 1.60 L/min, p = 0.009), and more favourable autonomic recovery. Among all variables, only the change in cardiac output was associated with PEH magnitude (r = -0.46, p = 0.006). No consistent physiological differences were observed following isometric exercise. CONCLUSION: PEH following aerobic exercise is characterised by a distinct responder phenotype associated with greater reductions in aortic augmentation index and favourable cardiac adaptations. These findings highlight substantial interindividual variability in BP responses and support the need for individualised exercise strategies in hypertension management.

Adolescent↗