Meta-analysis Of A Joint Additive Frailty-copula Model Based On Semi-competing Risks

Recently there has been a great interest in studying the meta-analysis method in medical and statistical literature.For the meta-analysis synthesizing existing several studies,a joint additive frailty-copula model between recurrence and death has been studied under semi-competing risks,where the frailty reflects the specific study dependence and is assumed following gamma distribution in the paper.In this paper it maximize the penalized likelihood to get estimators of these parameters where smoothing parameters can be confirmed by the likelihood cross-validation.The article uses additive model to present the hazard functions of recurrence and death,and its baseline hazard functions are approximated by splines where there use the cubic M-spline.What's more,the article consider the Clayton copula to describe the dependence between recurrence and death time.The simulation conducts for four groups with different true values and the censored rate.The method is based on real data which are ovarian cancer patients data to verify validity.