Department of Epidemiology and Biostatistics, University of South Carolina, Columbia, SC, USA.
Department of Statistics, North Carolina State University, Raleigh, NC, USA.
Stat Methods Med Res. 2020 Dec;29(12):3707-3720. doi: 10.1177/0962280220939123. Epub 2020 Jul 8.
To investigate the effect of fitness on cardiovascular disease and all-cause mortality using the Aerobics Center Longitudinal Study, we develop a semiparametric illness-death model account for intermittent observations of the cardiovascular disease incidence time and the right censored data of all-cause mortality. The main challenge in estimation is to handle the intermittent observations (interval censoring) of cardiovascular disease incidence time and we develop a semiparametric estimation method based on the expectation-maximization algorithm for a Markov illness-death regression model. The variance of the parameters is estimated using profile likelihood methods. The proposed method is evaluated using extensive simulation studies and illustrated with an application to the Aerobics Center Longitudinal Study data.
为了利用有氧健身中心纵向研究来调查健身对心血管疾病和全因死亡率的影响，我们开发了一种半参数疾病-死亡模型，以考虑心血管疾病发病时间的间歇性观察和全因死亡率的右删失数据。在估计中，主要的挑战是处理心血管疾病发病时间的间歇性观察（区间删失），我们基于期望最大化算法为马尔可夫疾病-死亡回归模型开发了一种半参数估计方法。使用似然比方法估计参数的方差。该方法通过广泛的模拟研究进行评估，并应用于有氧健身中心纵向研究数据进行说明。
