Parametric and semi parametric survival analysis of patients with lung cancer

The lung cancer is the third cause of death and also one of the five common cancers in Iran. In this study, the efficiency of semi-parametric Cox models and Weibull parametric models in order to evaluate the effective factors of survival time of patients with lung cancer were investigated.
This study was a prospective-cohort study in which the total number of 228 patients with lung cancer followed up from 1991-2007 and necessary information such as age at the time of diagnosis, gender, place of residence, education, residence status smoking, family history of cancer, province were collected from Babol cancer registry center. Then, these added to Cox and Weibull models as demographic factors. Akaike information criterion (AIC) was used to compare the efficiency of competing models.
In this study, 75% of cases were men and only 8% of patients survived until the end of the study. The 1, 3, and 5- year survival rates were 13, 8 and 8% respectively. Among the factors studied in the Weibull model, the effect of smoking on survival time was significant (p0.05). However, AIC suggested higher efficiency for parametric Weibull model.
In spite of the importance of the Cox model is as a more common method by researchers, this study showed that the Weibull model is more efficient in survival data analysis. According to the results of this study smoking prevention is necessary to increase the longevity of patients with lung cancer.