Malihe Safari1, Hossein Mahjub2, Habib Esmaeili3, Mohammad Abbasi4, Ghodratollah Roshanaei5. 1. Department of Biostatistics, School of Public Health, Hamadan University of Medical Sciences, Hamadan, Iran. 2. Research Center for Health Sciences, Department of Biostatistics, Faculty of Public Health, Hamadan University of Medical Sciences, Hamadan, Iran. 3. PhD in Mathematics, Principal Statistician and Project Lead, Staburo GmbH, Munich, Germany. 4. Department of Internal Medicine, School of Medicine, Hamadan University of Medical Sciences, Hamadan, Iran. 5. Associate Professor in Biostatistics, Department of Biostatistics, School of Public Health, Modeling of Noncommunicable Diseases Research Center, Hamadan University of Medical Sciences, Hamadan, Iran. gh.roshanaei@umsha.ac.ir.
Abstract
PURPOSE: In survival analysis, some patients may be at risk of more than one event, for example cancer-related death and cancer-unrelated death. In this case, if the aim of study becomes to assess the impact of risk factors on different causes of death, the competing risk model should be used rather than classical survival model. The aim of the present study is to determine the risk factors for related and unrelated mortality in patients with colorectal cancer using competing risk regression models. METHODS: The present retrospective cohort study was carried out on 310 CRC patients. Death due to cancer progression was considered as the interest event, and death due to unrelated cancer was considered as a competing event. Two most popular methods, cause-specific and subdistribution hazard regression model, were used to determine the effect of covariates on incidence and cause-specific hazard. Data analysis was performed using R3.6.2 software and cmprsk and survival packages. RESULTS: The mean (SD) of patients' age was 55.84 ± 13.2 years and 53.9% of them were male. BMI, T and N stage had a significant effect on both incidence and cause specific hazard of cancer-related death. CONCLUSION: The results of this study showed that cancer-related death is strongly correlated with under-weight (BMI < 18.5) and advanced clinical stage of the disease in patients with colorectal cancer. So, in the presence of competing events, both types of regression hazard models should be applied to permit a full understanding of the impact of covariates on the incidence and the rate of occurrence of each outcome.
PURPOSE: In survival analysis, some patients may be at risk of more than one event, for example cancer-related death and cancer-unrelated death. In this case, if the aim of study becomes to assess the impact of risk factors on different causes of death, the competing risk model should be used rather than classical survival model. The aim of the present study is to determine the risk factors for related and unrelated mortality in patients with colorectal cancer using competing risk regression models. METHODS: The present retrospective cohort study was carried out on 310 CRC patients. Death due to cancer progression was considered as the interest event, and death due to unrelated cancer was considered as a competing event. Two most popular methods, cause-specific and subdistribution hazard regression model, were used to determine the effect of covariates on incidence and cause-specific hazard. Data analysis was performed using R3.6.2 software and cmprsk and survival packages. RESULTS: The mean (SD) of patients' age was 55.84 ± 13.2 years and 53.9% of them were male. BMI, T and N stage had a significant effect on both incidence and cause specific hazard of cancer-related death. CONCLUSION: The results of this study showed that cancer-related death is strongly correlated with under-weight (BMI < 18.5) and advanced clinical stage of the disease in patients with colorectal cancer. So, in the presence of competing events, both types of regression hazard models should be applied to permit a full understanding of the impact of covariates on the incidence and the rate of occurrence of each outcome.
Authors: Katrina J Sharples; Melissa J Firth; Victoria A Hinder; Andrew G Hill; Mark Jeffery; Diana Sarfati; Charis Brown; Carol Atmore; Ross A Lawrenson; Papaarangi Mj Reid; Sarah L Derrett; Jerome Macapagal; John P Keating; Adrian H Secker; Charles De Groot; Christopher Gca Jackson; Michael Pn Findlay Journal: N Z Med J Date: 2018-06-08