### Data collection

The annual age- and sex-specific demographic data was provided and permitted to use by local bureau of Statistics affiliated Government, major responsibility of which was for the statistics of population. In China, the population census was carried out every ten years, and the sample survey of population was carried out every five years. The annual number of population was obtained based on the census and sample survey.

The data on death was provided and permitted to use by the Yunnan centers for disease control and prevention. Data in three periods, 1973–1975, 1990–1992 and 2004–2005, was collected as a part of the national retrospective sampling surveys of death cause. All people lived in Xuanwei was included in subjects of the surveys. The sample size was 2528237 in 1973–1975, 3566456 in1990–1992, and 2729497 in 2004–2005.

For a variety of reasons, the coverage of vital registration system in China was low. To understand the death cause of Chinese peoples, the national retrospective sampling surveys of death cause were conducted for three times, 1973–1975, 1990–1992 and 2004–2005, under the leadership of the Ministry of Health of the People Republic of China. Data was collected in a standard way by the trained investigators. First, the name lists of the dead were obtained from police departments, funeral parlour, medical institutions, and elder peoples in villages. Second, the investigators entered the home of the dead to obtain the information with a standard questionnaire concerning name, sex, birth date, death date, cause of death, and medical treatment. If the information of death was unknown for a little dead person, Verbal Autopsy was used. The Chinese Code for Death Causes and Population was used to categorize and code cause of death in 1973–1975 and 1990–1992, and the categories of the 10th revision of the International Classification of Diseases was used in 2004–2005. The definition of lung cancer has kept consistently in the three surveys.

We had complied with the Declaration of Helsinki Ethical Principles for Medical Research Involving Human Subjects (World Medical Association 1989). Informed consent was obtained from participants’ relatives, such as parents or son and daughters.

### Statistical analysis

Data was stratified by gender and 5-year age group. Mortality of lung cancer was expressed as deaths per 100,000 populations. The method of decomposing the differences of mortality rate was used to evaluate the contribution of AF and NAF to the lung cancer. The calculations can be performed according to the formulas expressed in the following steps [12].

Step 1 To estimate the contribution value

$$ \begin{array}{l} diff\\ {}=CD{R}_{(y1)}-CD{R}_{(y2)}\kern1.25em \\ {} = \Sigma {\mathrm{M}}_{\mathrm{i}\left(\mathrm{y}1\right)}*{\mathrm{C}}_{\mathrm{i}\left(\mathrm{y}1\right)}-\Sigma {\mathrm{M}}_{\mathrm{i}\left(\mathrm{y}2\right)}*{\mathrm{C}}_{\mathrm{i}\left(\mathrm{y}2\right)}\\ {}=\Sigma \left[{\mathrm{C}}_{\mathrm{i}\left(\mathrm{y}1\right)}-{\mathrm{C}}_{\mathrm{i}\left(\mathrm{y}2\right)}\right]\ast \frac{\left[{\mathrm{M}}_{\mathrm{i}\left(\mathrm{y}1\right)}+{\mathrm{M}}_{\mathrm{i}\left(\mathrm{y}2\right)}\right]}{2}+\Sigma \left[{\mathrm{M}}_{\mathrm{i}\left(\mathrm{y}1\right)}-{\mathrm{M}}_{\mathrm{i}\left(\mathrm{y}2\right)}\right]*\frac{\left[{\mathrm{C}}_{\mathrm{i}\left(\mathrm{y}1\right)}+{\mathrm{C}}_{\mathrm{i}\left(\mathrm{y}2\right)}\right]}{2}\end{array} $$

Diff = difference of total mortality

CDR(y1) = total mortality in a population in a period

CDR(y2) = total mortality in a population in another period

Mi = the age specific mortality rates

Ci = the age specific proportion of population.

The contribution value of AF=

$$ \Sigma \left[{\mathrm{C}}_{\mathrm{i}\left(\mathrm{y}1\right)}-{\mathrm{C}}_{\mathrm{i}\left(\mathrm{y}2\right)}\right]\ast \frac{\left[{\mathrm{M}}_{\mathrm{i}\left(\mathrm{y}1\right)}+{\mathrm{M}}_{\mathrm{i}\left(\mathrm{y}2\right)}\right]}{2} $$

The value of NAF=

$$ \Sigma \left[{\mathrm{M}}_{\mathrm{i}\left(\mathrm{y}1\right)}-{\mathrm{M}}_{\mathrm{i}\left(\mathrm{y}2\right)}\right]*\frac{\left[{\mathrm{C}}_{\mathrm{i}\left(\mathrm{y}1\right)}+{\mathrm{C}}_{\mathrm{i}\left(\mathrm{y}2\right)}\right]}{2} $$

Step 2 To estimate the contribution proportion

The contribution proportion of AF=

$$ \Sigma \left[{\mathrm{C}}_{\mathrm{i}\left(\mathrm{y}1\right)}-{\mathrm{C}}_{\mathrm{i}\left(\mathrm{y}2\right)}\right]\ast \frac{\left[{\mathrm{M}}_{\mathrm{i}\left(\mathrm{y}1\right)}+{\mathrm{M}}_{\mathrm{i}\left(\mathrm{y}2\right)}\right]}{2}/ diff*100 $$

The contribution proportion of NAF=

$$ \Sigma \left[{\mathrm{M}}_{\mathrm{i}\left(\mathrm{y}1\right)}-{\mathrm{M}}_{\mathrm{i}\left(\mathrm{y}2\right)}\right]*\frac{\left[{\mathrm{C}}_{\mathrm{i}\left(\mathrm{y}1\right)}+{\mathrm{C}}_{\mathrm{i}\left(\mathrm{y}2\right)}\right]}{2}/ diff*100 $$

Data analysis was performed in Excel 2010.