Related Experiment Video
Updated: Dec 24, 2025

Inverse Probability of Treatment Weighting Propensity Score using the Military Health System Data Repository and National Death Index
Published on: January 8, 2020
The methodology for assessing smoking-attributed mortality based on All Causes of Death Surveillance in Tianjin,
Wei Li1, Dezheng Wang1, Hui Zhang1
1Department of Non-communicable Disease Control and Prevention, Tianjin Centers for Disease Control and Prevention, Tianjin, China.
Introduction:
The All Causes of Death Surveillance (ACDS) system was used to measure smoking-attributed mortality by inserting questions on smoking on death certificates. Smoking status information of the deceased has been routinely collected in death certificates since 2010. We describe a death registry-based case-control study using smoking and cause-of-death data for the period 2010-15.
Methods:
From 2010, three questions about the smoking status of the deceased were inserted in a revised death certificate: 1) Smoking status (current smoker, quit smoking, never smoker); 2) Number of cigarettes per day smoked; and 3) Number of years of smoking. A data-accuracy survey of 1788 telephone interviews of the family of the deceased was also conducted. Smoking habits (current/ex-smoker vs non-smoker) were compared in study cases (persons who died of lung cancer and other diseases known to be caused by smoking) and the controls (never smokers). Multivariate logistic regression analysis was conducted to estimate relative risks, RR (odds ratios) for smoking-attributed mortality, for lung cancer and all causes of death related to smoking, adjusted for 5-year interval age groups, education, marital status, and year of death.
Results:
During the study period (2010-15), the annual crude death reporting rates ranged from 6.5‰ to 7.0‰. The reporting rates of smoking status, smoking history and the number of cigarettes smoked daily were 95.5%, 98.6% and 98.6%, respectively. Compared to never smokers, the RR of ever smoking in males was 1.38 (95% CI: 1.33-1.43) for all causes of smoking-related deaths and 3.07 (95% CI: 2.91-3.24) for lung cancer, while in females the values were 1.46 (95% CI: 1.39-1.54) for all causes of smoking-related deaths and 4.07 (95% CI: 3.81-4.35) for lung cancer. The results in Tianjin are in accord with published results from previous studies.
Conclusions:
Levels and trends in smoking attributed mortality can be measured at low cost by using the stable, complete and effective ACDS system in Tianjin.
Related Concept Videos
Statistical Methods for Analyzing Epidemiological Data
Observational Studies
There are three types of observational studies – Prospective, retrospective, and cross-sectional.
Prospective Study
Prospective studies, also known as longitudinal or cohort studies, are carried out by collecting future data from groups sharing similar characteristics. One...
Causality in Epidemiology
Actuarial Approach
Consider the example of a high-risk surgical procedure with significant early-stage mortality. A two-year clinical study is conducted,...
Criteria for Causality: Bradford Hill Criteria - II
Cancer Survival Analysis
