Rana Mohammed Jaber
Robert Stempel College of Public Health and Social Work, USA
Posters-Accepted Abstracts: J Proteomics Bioinform
Dichotomous grouped-time survival analysis is a combination of the grouped cox model, the discrete time hazard model and the dichotomous approach. This method was used to test the gateway hypothesis that waterpipe smoking may predict cigarette smoking initiation among adolescents by comparing the hazards of initiating cigarette smoking between waterpipe and never smokers. In this statistical approach, survival time is represented as a set of indicators of whether or not the participant failed in each time point (until the individual experiences the eventor is censored). This approach considers the timing as well as the occurrence of the event. It also handles censoring and allows for a discrete specification of time when the data are interval-censored. Items measured at each time point were used for time-varying predictors by linking predictors pertaining to a certain time point to the risk of cigarette smoking initiation at the subsequent time points. √Ę¬?¬?ProcPhreg√Ę¬?¬? commands were used in SAS with shared frailty model considering the school as a random variable to account for the unobserved heterogeneity due to clustering. This analysis allowed for maximum data use, inclusion of time-dependent covariates and relaxing of the proportional hazards assumption as well as minimizing the standard error of the estimate.
Email: [email protected]