Category: Statistics

J Robot Surg. 2025 Dec 1;20(1):50. doi: 10.1007/s11701-025-03002-z.

ABSTRACT

Console time is one of the automated performance metrics (APM) recorded by the robot software during robotic cardiac surgery. Little is known about what this APM predicts in cardiac surgery. This study aimed to evaluate factors associated with console time during robotically assisted mitral valve repair (raMVR). A total of 150 patients underwent raMVR from 7/2021 to 12/2024. Console time and related APMs were extracted from robotic system logs. Correlation analysis, multivariable linear regression, and multivariate analysis of variance (MANOVA) were used to assess associations between console time and pre-, intra-, and post-operative outcomes. Mean console time was 123.2 ± 47.0 min. Console time correlated with body mass index (r = 0.22, p = 0.01), cardiopulmonary bypass (CPB) time (r = 0.50, p < 0.001), aortic cross-clamp (ACC) time (r = 0.60, p < 0.001), and hospital stay (r = 0.24, p = 0.003). Console time was longer with bileaflet prolapse (p = 0.003), annular calcification (p = 0.01), leaflet calcification (p = 0.04), complex repair (p < 0.001), transfusion (p = 0.01) and reoperation for bleeding (p = 0.005). Multivariable regression identified decalcification (B = + 78.6 min, p < 0.001), ACC time (p < 0.001), CPB time (p = 0.02), leaflet resection combined with neochords (p = 0.01), and annular calcification (p = 0.03) as independent predictors. MANOVA showed console time tertiles were significantly associated with postoperative outcomes (Wilks’ lambda = 0.86, p = 0.02). Patients in the lowest and middle tertile were more likely to be extubated in the operating room (p < 0.001). Console time reflects procedural complexity and operative intensity in raMVR. As an automated, objective metric, it may serve as a valuable tool for intra-operative assessment, surgical planning, and early outcome prediction in robotic cardiac programs.

PMID:41324791 | DOI:10.1007/s11701-025-03002-z

Int Endod J. 2025 Dec 1. doi: 10.1111/iej.70076. Online ahead of print.

ABSTRACT

BACKGROUND: Placing a pulp-capping material over the remaining dentine is integral to managing deep carious lesions in permanent teeth without pulp exposure. However, current guidelines do not favour any specific pulp-capping material, and there is no direct clinical evidence that pulp-capping materials maintain pulp vitality better than placing the restoration directly on dentine.

OBJECTIVES: To compare the effectiveness of various biomaterials, including pulp-capping materials and restorative materials applied directly over the remaining dentine, against one another in preserving pulp health in permanent teeth with deep carious lesions without pulp exposure.

METHODS: On June 9, 2024, MEDLINE, Embase, Scopus, and Web of Science were searched, supplemented by a screening of clinical trial registries, grey literature, and reference lists. Randomised controlled trials (RCTs) evaluating the effectiveness of indirect pulp capping in permanent teeth affected by deep carious lesions without pulp exposure were included. Risk of bias was assessed using the revised Cochrane risk-of-bias tool for randomised trials (RoB 2). Network meta-analyses and meta-regression were performed using a Bayesian approach and a random-effects model for the primary outcome (loss of pulp vitality), followed by an assessment of confidence in the evidence using the CINeMA framework.

RESULTS: Sixteen RCTs (19 reports; 1039 participants; 1093 teeth; seven biomaterials) were included. Most comparisons involving the dentine bonding agent (DBA; control) were supported by low-confidence evidence and lacked statistical significance; however, they always resulted in RRs favouring the pulp-capping materials. Notably, moderate-confidence evidence indicated that during the second follow-up year Biodentine (RR = 0.00; 95% CI: 0.00-0.53) and glass ionomer cement (GIC) (RR = 0.30; 95% CI: 0.00-0.99) outperformed the DBA. Moderate-confidence evidence also demonstrated that during the first follow-up year mineral trioxide aggregate (MTA) (RR = 0.30; 95% CI: 0.09-0.84) outperformed calcium hydroxide cement. Meta-regression found that neither study-level demographic covariates nor clinical-technique covariates were significantly associated with pulp-vitality outcome.

CONCLUSIONS: While most findings in this review were of low confidence, the evidence nevertheless supports the use of pulp-capping materials in permanent teeth with deep carious lesions. Among these materials, Biodentine, MTA, and GIC have the strongest supporting evidence for preserving pulp vitality.

TRIAL REGISTRATION: PROSPERO number: CRD42024507641.

PMID:41321278 | DOI:10.1111/iej.70076

Health Econ. 2025 Dec 1. doi: 10.1002/hec.70067. Online ahead of print.

ABSTRACT

Low birth weight and preterm birth are key indicators of neonatal health, influencing both immediate and long-term infant outcomes. While low birth weight may reflect fetal growth restrictions, preterm birth captures disruptions in gestational development. Ignoring the potential interdependence between these variables may lead to an incomplete understanding of their shared determinants and underlying dynamics. To address this, a copula distributional regression framework is adopted to jointly model both indicators as flexible functions of maternal characteristics and geographic effects. Applied to female birth data from North Carolina, the methodology identifies shared factors of low birth weight and preterm birth, and reveals how maternal health, socioeconomic conditions and geographic disparities shape neonatal risk. The joint modeling approach provides a more nuanced understanding of these birth metrics, offering insights that can inform targeted interventions, prenatal care strategies and public health planning.

PMID:41321272 | DOI:10.1002/hec.70067

Biometrics. 2025 Oct 8;81(4):ujaf156. doi: 10.1093/biomtc/ujaf156.

ABSTRACT

Matching is a commonly used causal inference study design in observational studies. Through matching on measured confounders between different treatment groups, valid randomization inferences can be conducted under the no unmeasured confounding assumption, and sensitivity analysis can be further performed to assess robustness of results to potential unmeasured confounding. However, for many common matched designs, there is still a lack of valid downstream randomization inference and sensitivity analysis methods. Specifically, in matched observational studies with treatment doses (eg, continuous or ordinal treatments), with the exception of some special cases such as pair matching, there is no existing randomization inference or sensitivity analysis method for studying analogs of the sample average treatment effect (ie, Neyman-type weak nulls), and no existing valid sensitivity analysis approach for testing the sharp null of no treatment effect for any subject (ie, Fisher’s sharp null) when the outcome is nonbinary. To fill these important gaps, we propose new methods for randomization inference and sensitivity analysis that can work for general matched designs with treatment doses, applicable to general types of outcome variables (eg, binary, ordinal, or continuous), and cover both Fisher’s sharp null and Neyman-type weak nulls. We illustrate our methods via comprehensive simulation studies and a real data application. All the proposed methods have been incorporated into $tt {R}$ package $tt {doseSens}$.

PMID:41321245 | DOI:10.1093/biomtc/ujaf156

Biometrics. 2025 Oct 8;81(4):ujaf155. doi: 10.1093/biomtc/ujaf155.

ABSTRACT

The case-cohort study design is often used in modern epidemiological studies of rare diseases, as it can achieve similar efficiency as a much larger cohort study with a fraction of the cost. Previous work focused on parameter estimation for case-cohort studies based on a particular statistical model, but few discussed the survival prediction problem under such type of design. In this article, we propose a super learner algorithm for survival prediction in case-cohort studies. We further extend our proposed algorithm to generalized case-cohort studies. The proposed super learner algorithm is shown to have asymptotic model selection consistency as well as uniform consistency. We also demonstrate our algorithm has satisfactory finite sample performances. Simulation studies suggest that the proposed super learners trained by data from case-cohort and generalized case-cohort studies have better prediction accuracy than the ones trained by data from the simple random sampling design with the same sample sizes. Finally, we apply the proposed method to analyze a generalized case-cohort study conducted as part of the Atherosclerosis Risk in Communities Study.

PMID:41321244 | DOI:10.1093/biomtc/ujaf155

Biometrics. 2025 Oct 8;81(4):ujaf157. doi: 10.1093/biomtc/ujaf157.

ABSTRACT

Effective treatment of medical conditions begins with an accurate diagnosis. However, many conditions are often underdiagnosed, either being overlooked or diagnosed after significant delays. Electronic health records (EHRs) contain extensive patient health information, offering an opportunity to probabilistically identify underdiagnosed individuals. The rationale is that both diagnosed and underdiagnosed patients may display similar health profiles in EHR data, distinguishing them from condition-free patients. Thus, EHR data can be leveraged to develop models that assess an individual’s risk of having a condition. To date, this opportunity has largely remained unexploited, partly due to the lack of suitable statistical methods. The key challenge is the positive-unlabeled EHR data structure, which consists of data for diagnosed (“positive”) patients and the remaining (“unlabeled”) that include underdiagnosed patients and many condition-free patients. Therefore, data for patients who are unambiguously condition-free, essential for developing risk assessment models, are unavailable. To overcome this challenge, we propose ascertaining condition statuses for a small subset of unlabeled patients. We develop a novel statistical method for building accurate models using this supplemented EHR data to estimate the probability that a patient has the condition of interest. We study the asymptotic properties of our method and assess its finite-sample performance through simulation studies. Finally, we apply our method to develop a preliminary model for identifying potentially underdiagnosed non-alcoholic steatohepatitis patients using data from Penn Medicine EHRs.

PMID:41321243 | DOI:10.1093/biomtc/ujaf157

Epidemiol Psychiatr Sci. 2025 Dec 1;34:e56. doi: 10.1017/S2045796025100310.

ABSTRACT

AIMS: Depression among middle-aged and older adults is a critical public health priority. Clarifying the dynamic evolution of depression is essential for establishing prevention and intervention strategies; however, relevant research is limited. The aim of this study was to elucidate the transition patterns underlying different depressive symptoms (DS) states.

METHODS: Data from the China Health and Retirement Longitudinal Study were utilised in this study, which included participants aged ≥45 years with multiple DS assessments via the Center for Epidemiological Studies Depression Scale. Multi-state Markov models were employed to estimate transition probabilities and intensities between DS states, the total length of stay and mean sojourn time in each state and the hazard ratios (HRs) of factors.

RESULTS: Among 19,991 participants (average follow-up: 7.3 years), the 10-year cumulative probabilities of transition from non-DS to depressive states increased by 19.4% in males and 31.8% in females. Mild DS was the most unstable state, with the highest transition intensities (males: 1.029; females: 0.970) and shortest sojourn time (males: 0.959 years; females: 1.022 years). Sex and age strongly influenced depressive state transitions. Compared to participants without chronic disease, those with ≥3 chronic diseases had a higher risk of developing mild DS (HR = 1.685, 95% Confidence Interval [CI]: 1.530-1.856) and transitioning to death from both the non-DS (HR = 2.905, 95% CI: 2.293-3.681) and severe-DS (HR = 3.429, 95% CI: 1.290-9.112) states, but a lower likelihood of recovery from mild DS (HR = 0.821, 95% CI: 0.749-0.900) and severe DS (HR = 0.730, 95% CI: 0.630-0.847). Compared to no participation in social activities, frequent participation was associated with a lower risk of progression to the mild-DS state (HR = 0.851, 95% CI: 0.785-0.920) and a greater likelihood of recovery from severe DS (HR = 1.169, 95% CI: 1.034-1.322). Being underweight was associated with an increased risk of mild-DS onset (HR = 1.338, 95% CI: 1.129-1.587) and transitioning to death from both the non-DS and mild-DS states, compared with individuals of normal weight.

CONCLUSIONS: Our study revealed a continuous population shift towards depressive states and identified the mild-DS state as a critical intervention state owing to its instability. In addition to sex and age, modifiable factors, including chronic disease conditions, social activity participation and weight status, significantly influenced DS-state transitions, offering actionable insights for precision prevention strategies.

PMID:41321236 | DOI:10.1017/S2045796025100310

Disaster Med Public Health Prep. 2025 Dec 1;19:e335. doi: 10.1017/dmp.2025.10264.

ABSTRACT

OBJECTIVE: Building collapses, debris removal, new construction, and increased stove use for heating have elevated air pollution in regions affected by the February 6, 2023, Kahramanmaraş earthquake. This study examines the relationship between carbon monoxide (CO) poisoning and air pollution in these areas 1 year after the disaster.

METHODS: A retrospective analysis of 151 patients from 10 hospitals in 8 cities was conducted, including data on demographics, clinical symptoms, sources of CO exposure, vital signs, laboratory findings, air pollution levels, and outcomes.

RESULTS: Indoor stove use was the primary source of CO exposure. The average Air Quality Index (AQI) was 55 (IQR 44-56), and particulate matter (PM_2.5) levels averaged 17.5 μg/m³ (IQR 10-27), exceeding EPA (Environmental Protection Agency) thresholds. AQI levels post-earthquake were significantly higher than pre-earthquake in Kahramanmaraş (AQI₁ = 48.5 [IQR 48-55], AQI₂ = 55 [IQR 55-80]; P = 0.007), Hatay (AQI₁ = 40.5 ± 13.7, AQI₂ = 56 [IQR 51-60.5]; P <0.001), and Gaziantep (AQI₁ = 44 [IQR 41-56], AQI₂ = 55 [IQR 54-55.5]; P = 0.014). Leukocytosis (P = 0.004) and myocardial injury (P <0.001) in CO poisoning cases varied significantly across provinces.

CONCLUSIONS: In conclusion, elevated AQI and PM_2.5 levels likely worsened myocardial injury in CO poisoning cases due to combined outdoor and indoor pollution effects. These findings emphasize the need for air quality monitoring and mitigation in disaster regions.

PMID:41321219 | DOI:10.1017/dmp.2025.10264

Disaster Med Public Health Prep. 2025 Dec 1;19:e336. doi: 10.1017/dmp.2025.10265.

ABSTRACT

OBJECTIVE: To develop an evaluative framework for assessing the emergency response capabilities of higher education institutions to major emerging infectious diseases, enabling institutions to identify preparedness gaps and prioritize improvements across the outbreak lifecycle.

METHODS: The Haddon Matrix was used as the foundation for the framework. A Delphi study with a Likert scale was conducted, followed by the Analytic Hierarchy Process (AHP) to determine the importance of the indicators.

RESULTS: A consensus was reached on the evaluation system, comprising 3 primary indicators: prevention and preparedness, response and handling, and recovery and rehabilitation. These indicators were further divided into 11 secondary and 34 tertiary indicators. Expert authority coefficients were 0.82 and 0.80, and Kendall’s coefficients were 0.32 and 0.161 (P < 0.001). AHP highlighted prevention and preparedness as the highest-priority domain (weight = 0.426), followed by recovery and rehabilitation (0.326). High-priority items included safety knowledge dissemination, emergency command systems, primary prevention, and timely warning and monitoring.

CONCLUSIONS: Integrating the Haddon matrix, Delphi consensus, and AHP, this study delivers a validated, prioritized framework to assess universities’ MEID response capability across phases. External validity beyond Shanghai remains to be established; cross-regional applicability should be empirically tested through multi-site validation, broader stakeholder representation, and evaluation of technology-enabled components, particularly in resource-limited settings.

PMID:41321216 | DOI:10.1017/dmp.2025.10265

Ann Pharmacother. 2025 Dec 1:10600280251384651. doi: 10.1177/10600280251384651. Online ahead of print.

ABSTRACT

BACKGROUND: Critically ill patients receiving continuous renal replacement therapy (CRRT) commonly experience hypotension necessitating fluids, ultrafiltration (UF) adjustment, or vasopressors. There is limited evidence evaluating 25% albumin on vasopressor requirements in this population.

OBJECTIVE: To evaluate the impact of 25% albumin on vasopressor requirements for critically ill patients requiring CRRT.

METHODS: This single-center, retrospective, intrapatient comparator study included adults admitted to the Cardiothoracic Surgery Intensive Care Unit (CTICU) who received 25% albumin intravenously every 6 or 8 hours for ≥2 consecutive doses while on CRRT and vasopressors. The primary endpoint was the absolute change in average vasopressor dosage from 48 hours before to 48 hours after the first albumin administration in norepinephrine equivalents (NEE, mcg/kg/min). A multivariable interrupted time series model was conducted. Notable secondary endpoints included absolute change in UF rate and 48-hour fluid balance.

RESULTS: Of 252 patients reviewed, 60 were included. The median absolute change in average vasopressor dosage from 48 hours prealbumin to 48 hours postalbumin was -0.005 mcg/kg/min (Q1: -0.035, Q3: 0.021, P = 0.24), with a median percentage change in average dosage of -9.5% (Q1: -33.2, Q3: 34.4). The multivariable regression analysis reported a 0.0038 mcg/kg/min increase in vasopressor dosage (P = 0.02) and a 0.0044 mcg/kg/min decrease in vasopressor dosage (P = 0.001) for every 4-hour increase in time in the 48 hours before and after albumin, respectively. From 48 hours prealbumin to 48 hours postalbumin, UF rate increased numerically (10.6 mL/hr [interquartile range (IQR) -24.0, 49.8]), and 48-hour fluid balance decreased numerically (-467.4 mL/48 hr [IQR -3124.5, 1306.3]).

CONCLUSION AND RELEVANCE: In CTICU patients receiving CRRT and vasopressors, 25% albumin resulted in no statistically significant difference in average vasopressor requirements in the 48 hours prealbumin compared to the 48 hours postalbumin in the unadjusted model. However, multivariable regression demonstrated a significant association between albumin administration and reduced vasopressors during the 48-hour period following albumin.

PMID:41321209 | DOI:10.1177/10600280251384651