arXiv subjects

Daijiro Kabata

Publications and source records attributed to Daijiro Kabata.

3 recordsLinked to original sources

arXiv2025

A sensitivity analysis for non-inferiority studies with non-randomised data

Background: Non-inferiority studies based on non-randomised data are increasingly used in clinical research but remain prone to unmeasured confounding. The classical E-value offers a simple way to quantify such bias but has been applied almost exclusively with respect to the statistical null. We reformulated the E-value framework to make explicit its applicability to predefined clinical margins, thereby extending its utility to non-inferiority analyses. Development: Using the bias-factor formulation by Ding and VanderWeele, we defined the non-inferiority E-value as the minimum strength of association that an unmeasured confounder would need with both treatment and outcome, on the risk-ratio scale, to move the 95% confidence-limit estimate to the prespecified non-inferiority margin. Application: This approach was applied to three observational studies and one single-arm trial with external controls to illustrate interpretation and range. The resulting non-inferiority E-values for the confidence limits varied from about one to three, depending on design and findings. In the single-arm trial, a large gap between the confidence-limit and point-estimate NIEs reflected small sample size and wide confidence intervals, highlighting that both should be reported for a balanced assessment of robustness. Conclusion: This study reformulates the E-value to focus on clinically meaningful margins rather than the statistical null, enabling its application to non-inferiority analyses. Although the non-inferiority E-value inherits the limitations of the original method and cannot address all bias sources, it offers a transparent framework for interpreting non-randomised evidence and for generating insights that inform the design of future, more definitive randomised controlled trials.

stat.ME↗

arXiv2025

Generalisable prediction model of surgical case duration: multicentre development and temporal validation

Background: Accurate prediction of surgical case duration underpins operating room (OR) scheduling, yet existing models often depend on site- or surgeon-specific inputs and rarely undergo external validation, limiting generalisability. Methods: We undertook a retrospective multicentre study using routinely collected perioperative data from two general hospitals in Japan (development: 1 January 2021-31 December 2023; temporal test: 1 January-31 December 2024). Elective weekday procedures with American Society of Anesthesiologists (ASA) Physical Status 1-4 were included. Pre-specified preoperative predictors comprised surgical context (year, month, weekday, scheduled duration, general anaesthesia indicator, body position) and patient factors (sex, age, body mass index, allergy, infection, comorbidity, ASA). Missing data were addressed by multiple imputation by chained equations. Four learners (elastic-net, generalised additive models, random forest, gradient-boosted trees) were tuned within internal-external cross-validation (IECV; leave-one-cluster-out by centre-year) and combined by stacked generalisation to predict log-transformed duration. Results: We analysed 63,206 procedures (development 45,647; temporal test 17,559). Cluster-specific and pooled errors and calibrations from IECV are provided with consistent performance across centres and years. In the 2024 temporal test cohort, calibration was good (intercept 0.423, 95%CI 0.372 to 0.474; slope 0.921, 95%CI 0.911 to 0.932). Conclusions: A stacked machine-learning model using only widely available preoperative variables achieved accurate, well-calibrated predictions in temporal external validation, supporting transportability across sites and over time. Such general-purpose tools may improve OR scheduling without relying on idiosyncratic inputs.

stat.AP↗

arXiv2025

Quantifying uncertainty of individualized treatment effects in right-censored survival data: A comparison of Bayesian additive regression trees and causal survival forest

Estimation of individualized treatment effects (ITE), also known as conditional average treatment effects (CATE), is an active area of methodology development. However, much less attention has been paid to the quantification of uncertainty of ITE/CATE estimates in right-censored survival data. Here we undertake an extensive simulation study to examine the coverage of interval estimates from two popular estimation algorithms, Bayesian additive regression trees (BART) and causal survival forest (CSF). We conducted simulation designs from 3 different settings: first, in a setting where BART was developed for an accelerated failure time model; second, where CSF was developed; and finally, a ``neutral'' simulation taken from a setting where neither BART nor CSF was developed. BART outperformed CSF in all three simulation settings. Both the BART and CSF algorithms involve multiple hyperparameters, and BART credible intervals had better coverage than the CSF confidence intervals under the default values, as well as under optimized values, of these hyperparameters.

stat.ME↗