Younes Mohammadi1, Mahboubeh Parsaeian2, Farshad Farzadfar3, Amir Kasaeian2, Parinaz Mehdipour2, Ali Sheidaei4, Anita Mansouri4, Sahar Saeedi Moghaddam4, Shirin Djalalinia5, Mahmood Mahmoudi6, Ardeshir Khosravi7, Kamran Yazdani6. 1. Department of Epidemiology and Biostatistics, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran, Non-Communicable Disease Research Center, Endocrinology and Metabolism Population Science Institute, Tehran University of Medical Sciences, Tehran, Iran. kyazdani@tums.ac.ir, ardeshir1344@yahoo.com. 2. Department of Epidemiology and Biostatistics, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran, Non-Communicable Disease Research Center, Endocrinology and Metabolism Population Science Institute, Tehran University of Medical Sciences, Tehran, Iran. 3. Non-Communicable Disease Research Center, Endocrinology and Metabolism Population Science Institute, Tehran University of Medical Sciences, Tehran, Iran, Endocrinology and Metabolism Research Center, Endocrinology and Metabolism Research Institute, Tehran University of Medical Sciences, Tehran, Iran. 4. Non-Communicable Disease Research Center, Endocrinology and Metabolism Population Science Institute, Tehran University of Medical Sciences, Tehran, Iran, Department of Biostatistics, Faculty of Paramedical sciences, Shahid Beheshti University of Medical Sciences, Tehran, Iran. 5. Endocrinology and Metabolism Research Center, Endocrinology and Metabolism Research Institute, Tehran University of Medical Sciences, Tehran, Iran, Non-Communicable Disease Research Center, Endocrinology and Metabolism Population Science Institute, Tehran University of Medical Sciences, Tehran, Iran. 6. Department of Epidemiology and Biostatistics, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran. 7. Non-Communicable Disease Research Center, Endocrinology and Metabolism Population Science Institute, Tehran University of Medical Sciences, Tehran, Iran, Deputy of Health, Ministry of Health and Medical Education, Tehran, Iran.
Abstract
BACKGROUND: Calculation of burden of diseases and risk factors is crucial to set priorities in the health care systems. Nevertheless, the reliable measurement of mortality rates is the main barrier to reach this goal. Unfortunately, in many developing countries the vital registration system (VRS) is either defective or does not exist at all. Consequently, alternative methods have been developed to measure mortality. This study is a subcomponent of NASBOD project, which is currently conducting in Iran. In this study, we aim to calculate incompleteness of the Death Registration System (DRS) and then to estimate levels and trends of child and adult mortality using reliable methods. METHODS: In order to estimate mortality rates, first, we identify all possible data sources. Then, we calculate incompleteness of child and adult morality separately. For incompleteness of child mortality, we analyze summary birth history data using maternal age cohort and maternal age period methods. Then, we combine these two methods using LOESS regression. However, these estimates are not plausible for some provinces. We use additional information of covariates such as wealth index and years of schooling to make predictions for these provinces using spatio-temporal model. We generate yearly estimates of mortality using Gaussian process regression that covers both sampling and non-sampling errors within uncertainty intervals. By comparing the resulted estimates with mortality rates from DRS, we calculate child mortality incompleteness. For incompleteness of adult mortality, Generalized Growth Balance, Synthetic Extinct Generation and a hybrid of two mentioned methods are used. Afterwards, we combine incompleteness of three methods using GPR, and apply it to correct and adjust the number of deaths. CONCLUSION: In this study, we develop a conceptual framework to overcome the existing challenges for accurate measuring of mortality rates. The resulting estimates can be used to inform policy-makers about past, current and future mortality rates as a major indicator of health status of a population.
BACKGROUND: Calculation of burden of diseases and risk factors is crucial to set priorities in the health care systems. Nevertheless, the reliable measurement of mortality rates is the main barrier to reach this goal. Unfortunately, in many developing countries the vital registration system (VRS) is either defective or does not exist at all. Consequently, alternative methods have been developed to measure mortality. This study is a subcomponent of NASBOD project, which is currently conducting in Iran. In this study, we aim to calculate incompleteness of the Death Registration System (DRS) and then to estimate levels and trends of child and adult mortality using reliable methods. METHODS: In order to estimate mortality rates, first, we identify all possible data sources. Then, we calculate incompleteness of child and adult morality separately. For incompleteness of child mortality, we analyze summary birth history data using maternal age cohort and maternal age period methods. Then, we combine these two methods using LOESS regression. However, these estimates are not plausible for some provinces. We use additional information of covariates such as wealth index and years of schooling to make predictions for these provinces using spatio-temporal model. We generate yearly estimates of mortality using Gaussian process regression that covers both sampling and non-sampling errors within uncertainty intervals. By comparing the resulted estimates with mortality rates from DRS, we calculate child mortality incompleteness. For incompleteness of adult mortality, Generalized Growth Balance, Synthetic Extinct Generation and a hybrid of two mentioned methods are used. Afterwards, we combine incompleteness of three methods using GPR, and apply it to correct and adjust the number of deaths. CONCLUSION: In this study, we develop a conceptual framework to overcome the existing challenges for accurate measuring of mortality rates. The resulting estimates can be used to inform policy-makers about past, current and future mortality rates as a major indicator of health status of a population.