著者
Masahiko Gosho Tomohiro Ohigashi Kengo Nagashima Yuri Ito Kazushi Maruo
出版者
Japan Epidemiological Association
雑誌
Journal of Epidemiology (ISSN:09175040)
巻号頁・発行日
pp.JE20210089, (Released:2021-09-25)
参考文献数
46
被引用文献数
8

Background: Logistic regression models are widely used to evaluate the association between a binary outcome and a set of covariates. However, when there are few study participants at the outcome and covariate levels, the models lead to bias of the odds ratio (OR) estimated using the maximum likelihood (ML) method. This bias is known as sparse data bias, and the estimated OR can yield impossibly large values because of data sparsity. However, this bias has been ignored in most epidemiological studies.Methods: We review several methods for reducing sparse data bias in logistic regression. The primary aim is to evaluate the Bayesian methods in comparison with the classical methods, such as the ML, Firth’s, and exact methods using a simulation study. We also apply these methods to a real data set.Results: Our simulation results indicate that the bias of the OR from the ML, Firth’s, and exact methods is considerable. Furthermore, the Bayesian methods with hyper-g prior modeling of the prior covariance matrix for regression coefficients reduced the bias under the null hypothesis, whereas the Bayesian methods with log F-type priors reduced the bias under the alternative hypothesis.Conclusion: The Bayesian methods using log F-type priors and hyper-g prior are superior to the ML, Firth’s, and exact methods when fitting logistic models to sparse data sets. The choice of a preferable method depends on the null and alternative hypothesis. Sensitivity analysis is important to understand the robustness of the results in sparse data analysis.
著者
Akihiko Nogami Kyoko Soejima Itsuro Morishima Kenichi Hiroshima Ritsushi Kato Satoru Sakagami Fumiharu Miura Keisuke Okawa Tetsuya Kimura Takashi Inoue Atsushi Takita Kikuya Uno Koichiro Kumagai Takashi Kurita Masahiko Gosho Kazutaka Aonuma for the RYOUMA Investigators
出版者
The Japanese Circulation Society
雑誌
Circulation Journal (ISSN:13469843)
巻号頁・発行日
pp.CJ-22-0290, (Released:2022-08-20)
参考文献数
41
被引用文献数
7

Background: Optimal periprocedural oral anticoagulant (OAC) therapy before catheter ablation (CA) for atrial fibrillation (AF) and the safety profile of OAC discontinuation during the remote period (from 31 days and up to 1 year after CA) have not been well defined.Methods and Results: The RYOUMA registry is a prospective multicenter observational study of Japanese patients who underwent CA for AF in 2017–2018. Of the 3,072 patients, 82.3% received minimally interrupted direct-acting OACs (DOACs) and 10.2% received uninterrupted DOACs. Both uninterrupted and minimally interrupted DOACs were associated with an extremely low thromboembolic event rate. Female, long-standing persistent AF, low creatinine clearance, hepatic disorder, and high intraprocedural heparin dose were independent factors associated with periprocedural major bleeding. At 1 year after CA, DOAC was continued in 55.9% of patients and warfarin in 56.4%. The incidence of thromboembolic and major bleeding events for 1 year was 0.3% and 1.2%, respectively. Age ≥73 years, dementia, and AF recurrence were independently associated with major bleeding events. Univariate analyses revealed that warfarin continuation and off-label overdose of DOACs were risk factors for major bleeding after CA.Conclusions: High intraprocedural dose of heparin was associated with periprocedural major bleeding events. At 1 year after CA, over half of the patients had continued OAC therapy. Thromboembolic events were extremely low; however, major bleeding occurred in 1.2%. Age ≥73 years, dementia, and AF recurrence were independently associated with major bleeding after CA.