دانلود رایگان مقاله پیش بینی مرگ و میر سالانه ریه و سرطان برونش با زمان بقای فردی

abstract

Accurate forecasts of the numbers of cancer deaths are critical not only for allocating government health and welfare budgets, but also for providing guidance to the related industries. We suggest a framework for predicting the annual numbers of cancer deaths by modeling individual survival times. A Weibull mixture model with individual covariates and unobserved heterogeneity is proposed for examining the effects of demographic variables on individual survival times and predicting the annual number of cancer deaths by adopting a bottom-up strategy. We apply the suggested framework to a survival analysis of lung and bronchus cancer patients in the United States and provide a comparison with the forecast results obtained from previous studies. A comparison of our results with those of various benchmarks shows that our proposed model performs better for predicting annual numbers of cancer deaths. Furthermore, by segmenting patients based on age, sex, and race, we are able to specify the differences between groups and assess the group-specific survival probabilities within a given period. Our results show that older, female, and white patients survive significantly longer than younger, male, and black patients. Also, patients diagnosed in recent years survive significantly longer than those diagnosed a long time ago. © 2015 International Institute of Forecasters. Published by Elsevier B.V. All rights reserved.

6. Concluding remarks

In this paper, we propose a Weibull mixture model with individual covariates and unobserved heterogeneity in order to examine the effects of demographic variables on individual survival times, and to predict theannual number of cancer deaths by adopting a bottom-up strategy. A comparison of our results with those from the four benchmarks – the damped trend exponential smoothing method, the joinpoint regression model, a Bayesian state space model, and PIAMOD – reveals that our proposed model performs best for predicting annual numbers of cancer deaths. Furthermore, by segmenting the patients based on age, sex, and race, we are able to specify the differences between groups, and assess the group-specific survival probabilities within a given period. Our results show that older, female, and white patients survive significantly longer than younger, male, and black patients, respectively. Also, patients diagnosed in recent years survive significantly longer than those diagnosed a long time ago.