BMJ Open Is Committed to Open Peer Review. As Part of This Commitment We Make the Peer Review History of Every Article We Publish Publicly Available

BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

BMJ Open is committed to open peer review. As part of this commitment we make the peer review history of every article we publish publicly available.

When an article is published we post the peer reviewers’ comments and the authors’ responses online. We also post the versions of the paper that were used during peer review. These are the versions that the peer review comments apply to.

The versions of the paper that follow are the versions that were submitted during the peer review process. They are not the versions of record or the final published versions. They should not be cited or distributed as the published version of this manuscript.

BMJ Open is an open access journal and the full, final, typeset and author-corrected version of record of the manuscript is available on our site with no access controls, subscription charges or pay-per-view fees (http://bmjopen.bmj.com).

If you have any questions on BMJ Open’s open peer review process please email [email protected] http://bmjopen.bmj.com/ on September 27, 2021 by guest. Protected copyright. BMJ Open BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Clustering analysis of modifiable risk factors for upper digestive tract cancer among residents aged 40 to 69 years in Yangzhong City, China: a cross-sectional study ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2020-042006

Article Type: Original research

Date Submitted by the 23-Jun-2020 Author:

Complete List of Authors: xiang Feng, xiang Feng; Yangzhong People's Hospital, Institute of Cancer Prevention and Treatment, Hua, Zhao-lai; Yangzhong People's Hospital, Institute of Cancer Prevention and Treatment, Zhou, Qin; Yangzhong People's Hospital, Institute of Cancer Prevention and Treatment, Shi, Ai-wu; Yangzhong People's Hospital, Institute of Cancer Prevention and Treatment, Song, Tong-qiu; Yangzhong People's Hospital, Institute of Cancer Prevention and Treatment, Qian, Dongfu; Nanjing Medical University, School of Health Policy &

Management http://bmjopen.bmj.com/ Chen, Ru; National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Wang, Gui-qi; National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College Wei, Wen-Qiang; National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences

and Peking Union Medical College, on September 27, 2021 by guest. Protected copyright. Zhou, Jin-Yi; Jiangsu Provincial Center for Disease Control and Prevention, Institute of Chronic Non-communicable Diseases Prevention and Control Wang, Jie-jun Shao, Gang Wang, Xi

EPIDEMIOLOGY, Endoscopy < GASTROENTEROLOGY, Risk management Keywords: < HEALTH SERVICES ADMINISTRATION & MANAGEMENT

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 37 BMJ Open

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 7 8 9 I, the Submitting Author has the right to grant and does grant on behalf of all authors of the Work (as defined 10 in the below author licence), an exclusive licence and/or a non-exclusive licence for contributions from authors 11 who are: i) UK Crown employees; ii) where BMJ has agreed a CC-BY licence shall apply, and/or iii) in accordance 12 with the terms applicable for US Federal Government officers or employees acting as part of their official 13 duties; on a worldwide, perpetual, irrevocable, royalty-free basis to BMJ Publishing Group Ltd (“BMJ”) its 14 licensees and where the relevant Journal is co-owned by BMJ to the co-owners of the Journal, to publish the 15 Work in this journal and any other BMJ products and to exploit all rights, as set out in our licence. 16 17 The Submitting Author accepts and understands that any supply made under these terms is made by BMJ to 18 the Submitting Author Forunless you peer are acting as review an employee on behalf only of your employer or a postgraduate 19 student of an affiliated institution which is paying any applicable article publishing charge (“APC”) for Open 20 Access articles. Where the Submitting Author wishes to make the Work available on an Open Access basis (and 21 intends to pay the relevant APC), the terms of reuse of such Open Access shall be governed by a Creative 22 Commons licence – details of these licences and which Creative Commons licence will apply to this Work are set 23 out in our licence referred to above. 24 25 Other than as permitted in any relevant BMJ Author’s Self Archiving Policies, I confirm this Work has not been 26 accepted for publication elsewhere, is not being considered for publication elsewhere and does not duplicate 27 material already published. I confirm all authors consent to publication of this Work and authorise the granting 28 of this licence. 29 30 31 32 33 34 35

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 37

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 Clustering analysis of modifiable risk 6 7 factors for upper digestive tract cancer 8 9 10 among residents aged 40 to 69 years in 11 12 Yangzhong City, China: a cross-sectional 13 14 15 study 16 17 Xiang Feng,1,2 Zhaolai Hua,1 Qin Zhou1, Aiwu Shi,1 Tongqiu Song,1 Dongfu Qian,2 Ru Chen,3 18 For peer review only 19 GuiQi Wang,3 Wenqiang Wei,3 Jinyi Zhou,4 Jiejun Wang,5 Gang Shao,6 Xi Wang6 20 21 22 23 1Institute of tumor prevention and control, People’s Hospital of Yangzhong City, Yangzhong, 24 25 China 26 27 2School of Medicine and Politics, Nanjing Medical University, Nanjing, China 28 29 3 30 National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese 31 Academy of Medical Science and Peking Union Medical College, Beijing, China 32 33 4Depatment of Non-communicable Disease Prevention, Jiangsu Center for Disease Control and 34 35 Prevention, Nanjing, China

36 http://bmjopen.bmj.com/ 37 5Department of Oncology, Changzheng Hospital, Naval Military Medical University, Shanghai 38 39 200070, China. [email protected] 40 41 6Department of Oncology, the 903rd Hospital of PLA, Hangzhou, China. 42 43 [email protected]

44 on September 27, 2021 by guest. Protected copyright. 45 6Department of Oncology, the 903rd Hospital of PLA, Hangzhou, China. [email protected] 46 47 48 49 50 Correspondence to Zhaolai Hua; [email protected]; Dr Wenqiang Wei; 51 [email protected] 52 53 54 55 56 ABSTRACT 57 58 Objectives To describe the prevalence of modifiable risk factors for upper digestive tract cancer 59 (UDTC) and its clustering, and investigate relevant influencing factors of modifiable UDTC risk 60 1

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 37 BMJ Open

1 2 3 factors clustering among residents aged 40-69 years in Yangzhong City, China.

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 Design Cross-sectional study 7 8 Participants A total of 21 175 participants aged 40-69 years enrolled in the study. 1 962 subjects 9 10 were excluded due to missing age, marital status or some other key information. Eventually, 19 11 213 participants were available for the present analysis. 12 13 Main outcomes measures Prevalence and clustering of eight modifiable UDTC risk factors 14 15 (overweight or obesity, current smoking, excessive alcohol consumption, insufficient vegetables 16 intake, insufficient fruit intake, and the consumption of pickled, fried and hot food) were analyzed. 17 18 For peer review only 19 Results The prevalence of overweight/obesity, current smoking, excessive alcohol consumption, 20 21 insufficient vegetable intake, insufficient fruit intake and the consumption of pickled, fried and hot 22 23 food in this study was 45.3%, 24.1%, 16.2%, 66.1%, 94.5%, 68.1%, 36.0% and 88.4%, 24 25 respectively. Nearly all(99.9%) participants showed one or more UDTC risk factors, 98.7% of the 26 27 participants showed at least two risk factors and 92.2% of the participants had at least three risk 28 29 factors. Multivariate logistic regression analysis revealed that men, younger age, single, higher 30 31 education, higher annual family income and smaller household size were more likely to present 32 33 modifiable UDTC risk factors clustering. 34 35 Conclusions A large proportion of the participants had modifiable UDTC risk factors with a

36 http://bmjopen.bmj.com/ 37 tendency of clustering in Yangzhong City. Extra attention is required to pay on these groups who 38 39 are susceptible to risk factors clustering during the progress of screening. Relative departments 40 41 also need to make effective public health programs that aim to decrease modifiable UDTC risk 42 43 factors clustering among residents aged 40-69 years from high-risk areas of UDTC.

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 Strengths and limitations of this study 50 51 This is the first study examining the prevalence and clustering of modifiable upper digestive tract 52 53 cancer (UDTC) risk factors, and investigating relevant influencing factors in Yangzhong City, 54 with large and representative residents aged 40-69 years from the Upper Digestive Tract Cancer 55 56 Early Diagnosis and Treatment (UDTCEDAT). 57 58 59 60 2

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 37

1 2 3 Participants who volunteered to take part in our study are considered to be the high-risk group for

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 UDTC, which is significantly important for the prevention and control of UDTC in China where 6 the morbidity and mortality of UDTC are increasing. 7 8 9 10 11 A cross-section study cannot exam the causality or temporal relationship. 12 13 14 15 16 The modifiable UDTC risk factors included in our study were self 17 18 -reported by participants,For which peer may contribute review to recall and reporting only bias, except BMI. 19 20 21 22 23 We only focus on the eight modifiable UDTC risk factors, but there are far more than eight risk 24 factors for UDTC, which may underestimate the clustering of modifiable UDTC risk factors. 25 26 27 28 29 INTRODUCTION 30 31 According to the WHO, more than 70% of the total deaths worldwide were related to 32 33 non-communicable diseases (NCDs) in 2016.1 Cancer is the second cause of NCDs, accounting 34 35 for 22% of total global deaths related to NCDs. Globally, in 2018, an estimated 572 034

36 http://bmjopen.bmj.com/ 37 individuals diagnosed with oesophageal cancer (OC) and 1 033 701 individuals diagnosed with 38 39 stomach cancer (SC), with approximately 50% of new cases occurring in China. There are an 40 41 estimated 508 585 cancer deaths of OC and 782 685 cancer deaths of SC in 2018, which accounts 42 43 for 5.3% and 8.2% of cancer-cause deaths, respectively.2 Apparently, upper digestive tract cancer

44 on September 27, 2021 by guest. Protected copyright. 45 (UDTC) (oesophagus, stomach) has become a major source of morbidity and mortality related to 46 47 cancer. According to the National Cancer Center,3 OC has been the sixth most common cancer 48 49 and the fourth most common cancer cause of cancer-related death. SC has been the second most 50 51 common cancer and the third most common cancer cause of cancer-related death. The incidence 52 53 rate of OC and SC were 17.87/100 000 and 29.31/100 0000, the mortality rate of them were 54 55 13.68/100 0000 and 21.16/100 0000, respectively, in China in 2015.3 Hence, UDTC (oesophagus, 56 57 stomach) has become a major public health challenge in China and the disease burden of it is also 58 4 59 considerable. Due to the health-seeking behavior of the population and the character of diseases, 60 3

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 37 BMJ Open

1 2 3 UDTC is mostly diagnosed at a late stage, which is leading to a poor survival rate. It was 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 estimated that its five-year survival rate was less than 20% if diagnosed at an advanced stage but

7 5 6 7 8 is as high as 95% if detected at an earlier stage. 9 10 Although the cause of UDTC is not clear, it is believed by researchers that the epidemic of 11 12 UDTC in China is attributed to the multiplicity of demographic factors, diet, lifestyle, family 13 14 health, environment, gastrointestinal history and genetic factors.8 It is well known that tobacco, 15 16 alcohol consumption, overweight or obesity, thermal irritation (drinking very hot liquids) and 17 18 insufficient intakeFor of vegetable peer and fruit, review consumption of pickled only and fried food are eight risk 19 20 factors that can be altered by a tangible action for UDTC control.9 10 11 12 13 21 22 23 Although parts of these risk factors have dramatically decreased because of a set of 24 25 interventions implemented by Chinese government,14 15 the others have increased and will 26 27 continue to increase in the next decades because of the rapid transition of urbanization, 28 16 17 29 industrialization and aging. Further more, several studies have indicated that these risk factors 30 18 19 31 clustering was common in population which would further increase the risk of UDTC. A 32 33 comprehensive assessment of the distribution and the status of UDTC risk factors clustering is 34 35 significant for cancer prevention and control. Once we have such data, interventions can be

36 http://bmjopen.bmj.com/ planned and implemented efficiently to minimize these modifiable risk factors, thereby 37 38 minimizing the health risks of increasing UDTC-related mortality and morbidity. 39 40 41 Yangzhong City of Jiangsu Province is one of the high-risk areas of UDTC, especially in the 42 43 rural area.8 In 2015, the incidence rate of OC was 69.2/100 000, the mortality rate of OC and SC

44 on September 27, 2021 by guest. Protected copyright. 45 were 70.24/100 0000 and 81.89/100 0000, respectively, in Yangzhong City, which is higher 46 47 compared with the average of the nation.20 21 Hence, Yangzhong City had been one of the project 48 49 sites of the Upper Digestive Tract Cancer Early Diagnosis and Treatment (UDTCEDAT) since the 50 51 2006.8 A number of studies have estimated the risk factors for UDTC in different areas 52 53 worldwide.9 10 11 12 13 The results reveal that risk factors for UDTC are widespread, and the 54 55 modifiable risk factors are significantly important for the prevention and control of UDTC, 56 57 because these factors can be changed by some healthy education or other interventions 58 59 implemented by doctors and government. Or improve with the increase of personal health 60 4

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 37

1 2 3 awareness. However, the evidence on the clustering of these modifiable risk factors in high-risk 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 areas is still limited. Thus, we aimed to report the prevalence and clustering of modifiable UDTC 7 8 risk factors, and analyze the relevant factors influencing modifiable UDTC risk factor clustering 9 among residents aged 40-69 years in Yangzhong City which is a high-risk area of UDTC. 10 11 12 METHODS 13 14 15 Study population 16 17 18 For the present study,For we used peer secondary datareview collected from theonly screening of UDTC, focusing on 19 20 the early diagnosis and treatment of UDTC among high-risk populations in Yangzhong City, 21 22 China, from 2006 to 2017. In brief, villages from three towns (Baqiao, Youfang, and Xinglong) 23 24 were selected from Yangzhong City by cluster random sampling with all men and women aged 25 26 40-69 years invited in each village for participation from 2006 to 2017, unless they met the 27 28 following exclusion criteria: (1) history of UDTC or mental disorder; (2) contraindications for 29 30 endoscopic examinations and (3) inability to complete the whole interview or informed consent. 31 Inclusion criteria for participants were as following: (1) aged 40-69 years; (2) permanent residents 32 33 in Yangzhong City and (3) willing to accept endoscopic examination. Before the screening, we 34 35 obtained written informed consent from all participants after informing them about the

36 http://bmjopen.bmj.com/ 37 backgrounds, objectives, procedures, benefits, confidentiality agreement of personal information 38 39 and possible consequences of the whole program and then questionnaire-based interview, physical 40 41 examinations, laboratory tests performed by professional investigators. At last, the endoscopic 42 43 examinations, pathological diagnosis and necessary therapy for participants are conducted by

44 on September 27, 2021 by guest. Protected copyright. 45 well-trained doctors in People’s Hospital of Yangzhong City. The screening procedure follows 46 47 China's cancer screening and early diagnosis and treatment technology program strictly.22 The 48 49 data used in this study derived mainly from the questionnaire and physical examinations. Finally, 50 51 in total of 21 175 individuals were surveyed, 1 962 residents were excluded due to missing age, 52 53 marital status or some other factors, leaving 19 213 participants available for the present survey. 54 55 We provided health education about UDTC and the potential role of modifiable risk factors related 56 57 to UDTC to all eligible participants. Besides, we combined active and passive follow-ups to 58 59 collect outcome information for participants diagnosed with UDTC or precancerous lesions and 60 5

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 37 BMJ Open

1 2 3 performed a regular re-examination for patients according to the diagnosis. 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 7 8 9 Questionnaire data collection 10 11 12 Before implementing data collection, training sessions organized by the expert group on 13 14 UDTCEDAT were provided for all staff. The aim of this study, the standard measurement 15 16 methods, how to perform questionnaires properly, and the concrete study procedure were included 17 18 in the training contents.For At thepeer end of the trainingreview sessions, all only staff participated in the assessment 19 20 and proved to be qualified. 21 22 23 We use uniformly structured questionnaires to collect information through face-to-face 24 25 interviews. Each questionnaire took approximately 25 mins to complete. The questionnaire 26 information included demographic factors (gender, birthday, address, ID, marital status, and 27 28 household size), socioeconomic characteristic (education and annual family income), behavioral 29 30 factors (excessive alcohol consumption, current smoking), dietary habit (insufficient fresh 31 32 vegetables intake and insufficient fruit intake, consumption of pickled, fried and hot food), and 33 34 body mass index (BMI). 35

36 http://bmjopen.bmj.com/ 37 Physical examination 38 39 40 Physical examination included height and weight. Height and weight measurements were taken by 41 42 height scale and digital weight, respectively, with the help of trained examiners based on a 43 standardized program. All subjects were asked to remove any footwear, hats and heavy clothing 44 on September 27, 2021 by guest. Protected copyright. 45 46 before height and weight were measured. Height was measured to the nearest 0.1 cm, while 47 48 weight was measured to the nearest 0.1 kg. BMI was calculated by dividing body weight (in 49 50 kilograms) by the square of height (in meters). 51 52 Assessment criteria 53 54 55 Definitions of UDTC modifiable risk factors and the clustering of these risk factors 56 57 58 Eight modifiable UDTC risk factors were defined based on current national guidelines or 59 60 related references. Overweight/obesity was defined as BMI ≥24.0 kg/m2.23 24 25 Excessive alcohol 6

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 37

1 2 3 consumption, insufficient fresh vegetables and fruit intake were defined according to the Dietary 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 26 6 Guidelines for Chinese residents (2016). Accordingly, excessive alcohol consumption was 7 8 defined as self-reported consumption of more than 25 gram (for males) or 15 gram (for females) 9 alcohol drinks per day, insufficient vegetable intake as self-reported consumption of vegetables 10 11 less than 7 times per week and insufficient fruit intake as self-reported consumption of fruit less 12 13 than 7 times per week.26 Current smoking was defined as self-reported having used of any tobacco 14 15 products, including cigarettes, cigars or pipes daily continuously.27 Ex-smokers (those who 16 17 smoked previously but quit subsequently) and those who never used any tobacco products were 18 For peer review only 19 classified as non-smokers in our study.23 In addition, self-reported consumption of pickled, fried 20 21 or hot food at least once a week were classified as “Yes” in dietary habits, respectively. 22 23 24 Clustering of modifiable UDTC risk factors was defined as presenting at least three related 25 26 risk factors in one individual.24 28 27 28 29 Covariates 30 31 Covariates included in this study were demographic and socioeconomic information ascertained 32 33 by questionnaire, including age (40-44 years, 45-49 years, 50-54 years, 55-59 years, 60-64 years, 34 35 65-69 years), gender (male and female), marital status (single, currently married,

36 http://bmjopen.bmj.com/ 37 divorced/widowed/separated), educational status (no institutional education, primary school, 38 39 junior high school, senior high school and higher), household size (0-3, 4-5, ≥6) and annual family 40 41 income (tertiles: lower, middle and higher).29 42 43

44 Statistical analysis on September 27, 2021 by guest. Protected copyright. 45 46 47 Descriptive statistics were applied to describe the socio-demographic characteristics of the 48 49 sampled population. The difference in continuous variables was analyzed by student’s t-test, and 50 51 by χ² test to assess the differentials in the prevalence values among categorical variables. 52 53 Differences in men and women, the prevalence of each modifiable UDTC risk factor and the 54 55 distribution of modifiable UDTC risk factors clustering (0, 1 and 2, ≥3) in socio-demographic and 56 57 other characteristics were described in the overall population, respectively. Multiple logistic 58 regression models were adopted to explore the association between relevant characteristics and 59 60 7

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 37 BMJ Open

1 2 3 UDTC risk factors clustering. Only the variables that we found statistically significant at P <0.05 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 in the univariate analysis were included in the multiple logistic regression models. The result of 7 8 multiple logistic regression analyses was presented in terms of adjusted odds ratios (AOR) and 9 their respective 95% confidence intervals (CIs). All statistical analyses were performed by SPSS 10 11 software V.17.0. A two-sided P <0.05 was considered to be statistically significant. 12 13 14 Participant involvement statement 15 16 17 No participants or public were included in the design phase of this study. No participants were 18 For peer review only 19 asked to advise on interpretation or writing up of results. Dissemination of the result of the 20 21 research to participants and relevant participants community was prohibited. All the participants 22 23 had the right to receive the result of health check if they wanted. 24 25 26 27 28 29 30 31 32 33 34 35

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 8

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

BMJ Open Page 10 of 37

1 2 3 4 5 RESULTS 6 7 8 Socio-demographic and other characteristics of participants 9 10 11 The description of socio-demographic and other characteristics of 19 213 participants are presented in Table 1. Of the participants (40-69 years, mean age 53.2±7.8 12 For peer± review only 13 years) surveyed, 57.0% were women, with a mean age of 52.9 7.8 years. More than 95.0% (man 95.6%, woman 94.8%) of participants were married, nearly half 14 15 (44.3%) (man 41.6%, woman 46.4%) had an education level of primary school, 44.7% (man 44.6%, woman 44.7%) had a lower level of annual family income, and 16 http://bmjopen.bmj.com/ ± ± ± ± 2 ± ± 17 the mean of household size and BMI were 4.0 1.4 (man 4.0 1.4, woman 4.0 1.4) and 23.8 3.0 kg/m (man 23.9 3.0, woman 23.7 3.0), respectively. The 18 differences between men and women in age, marriage, education, household size and BMI were significant (all P 0.01) (Table 1). 19 < 20 21 22 23 24 on September 27, 2021 by guest. Protected copyright. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 9 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Page 11 of 37 BMJ Open

1 2 3 4 5 Prevalence of modifiable UDTC risk factors 6 7 8 The prevalence of overweight or obesity, current smoking and excessive alcohol consumption in this study was 45.3%, 24.1% and 16.2%, respectively. The 9 10 prevalence of current smoking and excessive alcohol consumption in men was significantly higher than that in women (all P<0.05). In addition,insufficient vegetable 11 12 intake,insufficient fruit intake and the consumptionFor of pickled,friedpeer and hot review food in participants accounted only for 66.1%,94.5%,68.1%,36.0%, and 88.4%, respectively. 13 14 The prevalence of consumption of pickled and hot food was higher in women than in men (all P<0.01) (Table 2). As shown in table 3, there were significant 15 16 differences with age and the level of education and annual family income in the eight UDTC risk factors (all P<0.05). Thehttp://bmjopen.bmj.com/ prevalence of these eight modifiable 17 18 UDTC risk factors tends to be higher in participants who were single, except for overweight or obesity (all P<0.001). Moreover, the prevalence of excessive alcohol 19 20 consumption, insufficient vegetables intake, insufficient fruit intake and the consumption of pickled, fried and hot food varied significantly with the household size 21 22 (Table 3). 23 24 on September 27, 2021 by guest. Protected copyright. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 10 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

BMJ Open Page 12 of 37

1 2 3 4 5 Clustering of modifiable UDTC risk factors 6 7 8 The prevalence of zero to eight modifiable UDTC risk factors participants had simultaneously in the study (Overweight/obesity, current smoking, excessive alcohol 9 10 consumption, insufficient vegetable intake, insufficient fruit intake, the consumption of pickled, fried and hot food) was 0.1%, 1.3%, 6.4%, 22.6%, 23.2%, 22.4%, 11 12 17.5%, 5.5% and 1.0%, respectively (dataFor shown partly peer in Table 4). Among review them, nearly all (99.9%) participants only showed one or more modifiable UDTC risk factors, 13 14 98.7% of the participants showed two or more modifiable UDTC risk factors and 92.2% of the participants had at least three modifiable UDTC risk factors. The 15 16 prevalence of clustering of modifiable UDTC risk factors was higher in men than in women (P<0.001). The prevalence of clusteringhttp://bmjopen.bmj.com/ of modifiable UDTC risk factors 17 18 was the highest in the age group 50-54 and among participants who were single (all P<0.001). There was an increasing trend towards modifiable UDTC risk factors 19 20 clustering with an increasing level of education and annual family income (all P<0.001). Moreover, modifiable UDTC risk factors clustering was the highest among 21 22 those participants who had 4-5 family members (P<0.001) (Table 4). 23 on September 27, 2021 by guest. Protected copyright. 24 The multivariable logistic regression analysis revealed that men were more likely to have 3 or more modifiable UDTC risk factors compared with women (OR 25 26 2.030, 95% CI 1.791 to 2.300). The prevalence of modifiable UDTC risk factors clustering increased with age and then decreased, peaking at 50-54 (OR 2.250, 95% 27 28 CI 1.801 to 2.810). Participants who were divorced/widowed/separated (OR 0.392, 95% CI 0.192 to 0.802) were less likely to have 3 or more modifiable UDTC risk 29 30 factors than those who were single. In addition, we divided the education into four groups, which showed clearly that increasing the level of educations was a risk 31 32 factor for modifiable UDTC risk factors clustering. Compared with participants who had a higher level of annual family income, those who had middle (OR 0.449, 33 34 95% CI 0.379 to 0.532) and lower (OR 0.311, 95% CI 0.263 to 0.368) level of annual family income were less likely to have 3 or more modifiable UDTC risk 35 36 factors. Modifiable UDTC risk factors clustering were less common among participants who had more than 6 family members than those who had less than 3 ones 37 38 (OR 0.698, 95% CI 0.577 to 0.844) (Table 5). 39 40 41 11 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Page 13 of 37 BMJ Open

1 2 3 4 5 6 7 8 9 10 11 12 For peer review only 13 14 15 16 http://bmjopen.bmj.com/ 17 18 19 20 21 22 23 24 on September 27, 2021 by guest. Protected copyright. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 12 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 14 of 37

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 DISCUSSION 7 8 9 10 11 As far as we know, this is the first large population-based survey not only investigated the 12 13 prevalence and clustering of eight modifiable UDTC risk factors but also described the 14 15 socio-demographic and socio-economic factors associated with these among Yangzhong City 16 residents aged 40-69 from southeast China. The present study revealed that the prevalence of 17 18 overweight or obesity,For current peer smoking, excessivereview alcohol consumption,only insufficient vegetables 19 20 intake, insufficient fruit intake and consumption of pickled, fried and hot food is significantly high 21 22 with a tendency of clustering, which implies the health risk of UDTC residents have in Yangzhong 23 24 City. We found that the prevalence of insufficient fruit intake and consumption of hot and fried 25 26 food were the top three modifiable UDTC risk factors in the population surveyed. Besides, 92.2% 27 28 of the participants presented at least three UDTC risk factors. 29 30 31 The prevalence of overweight or obesity (45.3%) in our study was higher than that observed 32 33 in Nanjing (35.6%),23 and nationwide population (42.0%),30 but our findings were closed to that 34 35 observed in some other regional.24 31 The prevalence of current smoking in this population

36 http://bmjopen.bmj.com/ 37 (24.1%) was much higher than that in a cross-sectional study in Shenzhen (10.5%)24 and Barbados 38 39 (9.2%),32 which was consistent with a study in Nanjing (24.5%).23 However, the rate was not as 40 30 41 high as reported (28.1%) in the China national nutrition and chronic disease survey (2015). Our 42 43 findings showed a high prevalence of excessive alcohol consumption in the Yangzhong population

44 on September 27, 2021 by guest. Protected copyright. 30 45 (16.2%) relative to the national average of 11.1% in men and 2.0% in women. The rate of 46 47 excessive alcohol consumption we found was similar to the Barbados, Nanbu and the prospective 48 study of China Kadoorie Biobank, where excessive alcohol consumption rates were around 49 50 14.5%,32 16.7%19and 14.9%,33 respectively. We found higher levels of insufficient intake of 51 52 vegetable(66.1%) in this population than those observed in the Tanzania34 and Hubei Province35 53 54 where insufficient intake of vegetable or fruit is 55.8% and 29.7%, respectively, while the levels 55 56 of insufficient intake of fruit (94.5%) in our study were also much higher than that observed in the 57 58 region mentioned above.34 35 Moreover, the proportion of the Yangzhong population had dietary 59 60 habits of consumption of pickled, fried and hot food were greater than the levels in Huaian 13

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 37 BMJ Open

1 2 3 (22.7%,7.1%, and 10.9%),36 as well as in the Nanbu (28.63%, 1.95%, and 6.11%),19 both of these 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 region mentioned above are high-risk areas of UDTC in China. 7 8 The clustering of risk factors for chronic diseases is widespread.19 Several previous studies 9 10 reported the clustering of chronic diseases in the Chinese population. For example, among 49 247 11 12 Chinese aged 15 to 69 years from the 2007 China Chronic Disease and Risk Factor Surveillance, 13 14 the prevalence of having zero, one, two and at least three chronic disease risk factors were 9.1%, 15 16 33.9%, 32.4% and 24.6%, respectively.37 Also other regional studies have examined the clustering 17 18 of some specific chronicFor diseases peer in local reviewresidents. Hong et alonly reported that 30.1% and 35.2% of 19 20 the Nanjing population presented one and at least two cardiovascular diseases (CVD) risk 21 22 factors.23 Conversely, a much higher rate of CVD risk factors clustering was noticed by Ni et al in 23 24 Shenzhen City.24 25 26 27 In our present study, it was observed that 0.1%, 1.3%, 6.4% and 92.2% of participants had 28 29 zero, one, two and three or more modified UDTC risk factors, respectively, among residents aged 30 31 40-69 years. The tendency of modified UDTC risk factors clustering was prevalent considerably 32 33 in Yangzhong City. Different estimates of the risk factors clustering for UDTC were found in the 34 35 literature. He et al. 2019 showed that among residents aged 40-69 years in UDTC high-risk areas,

36 http://bmjopen.bmj.com/ 33.08%, 35.99%, 16.76% and 11.93% of participants had one, two, three and at least four 37 38 esophageal cancer risk factors, respectively.19 In another case-control study of 2 266 Chinese 39 40 adults, 32.5% and 41.1% of the participants presented three and four or more risk factors, 41 42 respectively, for esophageal or stomach cancer.18 Compared with these two studies mentioned 43

44 above,18 19 a much higher clustering of risk factors was noticed in our present study. The variations on September 27, 2021 by guest. Protected copyright. 45 46 could likely due to the difference of diagnostic criteria, the number and kind of risk factors 47 48 included in the study and the age group of the participants. Overall, there are some other 49 50 nationwide and worldwide studies on the clustering of some common or specific chronic diseases 51 52 risk factors, but the study on modifiable UDTC risk factors is obviously limited. 53 54 55 The factors found to be associated with modifiable UDTC risk factors clustering included 56 57 gender, age, marriage status, education, annual family income and household size. We found the 58 59 prevalence of modifiable UDTC risk factors clustering was lower in women compared with men, 60 14

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 37

1 2 3 which was consistent with findings from other settings.19 23 37 The possible reason could be 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 Chinese men are less aware of self-protect for chronic diseases and have worse health-seeking 7 8 behavior, and may also attend more social occasions, tend to consume more tobacco/cigarette 9 high-salt, high-fat and high-calorie food compared with women.23 31 10 11 12 In addition, this study revealed that the prevalence of modifiable UDTC risk factors 13 14 clustering was decreased with age, which was inconsistent with previous studies.19 23 24 37 Studies 15 16 showed that in Nanbu, China, the clustering of esophageal cancer risk factors increased with age 17 18 may attribute to For the lower peer level of awareness, review practice and only willingness for health among the 19 20 elder.19 A possible reason for this difference is as follows: with the aging of the body and the 21 22 deterioration of organ function, an elder individual possesses a higher risk of health disorder and 23 24 has a greater demand for medical care. As a result, this creates more opportunities to get 25 26 diagnosed with some health screening, including UDTC. Correspondingly, the elderly have more 27 28 chance to get healthy education from physicians than the younger.38 Our study also showed that 29 30 single participants had more prevalence of modifiable UDTC risk factors clustering compared 31 32 with participants who were divorced/widowed/separated, in accordance with a previous study.19 A 33 34 possible explanation is being a single older resident comes with its own economic and emotional 35 challenges, which may contribute to the unhealthy habit of lifestyle and diet.39 40 36 http://bmjopen.bmj.com/ 37 38 Our study demonstrated that the level of socioeconomic status (SES, education, annual 39 40 family income) was positively associated with modifiable UDTC risk factors clustering which was 41 42 inconsistent with other reports.19 23 37 Residents with the higher level of SES are more aware of 43

44 control and prevention of chronic disease and have better health-seeking behavior compared with on September 27, 2021 by guest. Protected copyright. 45 46 those with lower level of SES.19 23 41 Moreover, the poor or lower education participants may have 47 48 relatively more inaccessibility and unaffordability to medical services.38 This paradox may be due 49 50 to most of the participants enrolled in our study were from rural areas and their SES was generally 51 52 low. However, it may also imply that the higher income may contribute to unhealthy lifestyle42 53 54 and knowledge alone may not be sufficient to change unhealthy lifestyles. Therefore, the level of 55 56 education and income are two of the most important SES factors for modifiable UDTC risk factors 57 58 clustering. 59 60 15

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 37 BMJ Open

1 2 3 It is, however, important to note that participants who have more than 6 family members had 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 a lower prevalence of modifiable UDTC risk factors clustering compared with those who have less 7 8 than 3 ones. Changes in household size are bound to affect the adjustment of the family diet. As 9 the household size increase, it is more likely to increase dietary diversity (e.g. fruits, vegetables 10 11 and milk) every day.43 Besides, the affection, information and economic support among family 12 13 members will also increase significantly, which can adjust and correct the unhealthy lifestyle of 14 15 individuals.44 16 17 18 Our study exploredFor that peer the prevalence review and influencing only factors of modifiable UDTC risk 19 20 factors clustering in the UDTC high-risk area, Yangzhong City, which was based on the 21 22 community-based project for UDTC screening with a large sample size. Additionally, the physical 23 24 measurement and the data collection implemented by trained interviewers strictly according to 25 26 standard protocol and instrument, which increase the validity of our results. The findings from our 27 28 study may provide the reference for departments in charge of the prevention and control of UDTC 29 30 in Yangzhong City, Jiangsu province and relevant departments in other UDTC high-risk areas 31 32 (e.g. Linzhou, Feicheng, Yanting). 33 34 35 There were also several possible limitations in our study. Firstly, cross-section study cannot

36 http://bmjopen.bmj.com/ exam the causality or temporal relationship between the clustering of modifiable UDTC risk 37 38 factors and its influencing factors. Secondly, the modifiable UDTC risk factors included in our 39 40 study were self-reported by participants, which may contribute to recall and reporting bias, except 41 42 BMI. The results of our study were from Yangzhong City only, and cannot be generalized to the 43

44 other high-risk areas and the whole of southeast China. Additionally, the study response rate was on September 27, 2021 by guest. Protected copyright. 45 46 relatively low, particularly among males, which may affect the representativeness of the results. 47 48 Finally, our study only focused on the 8 modifiable UDTC risk factors, but there are far more than 49 50 eight risk factors for UDTC. Hence, further studies are needed. 51 52 53 Conclusion 54 55 56 In summary, this cross-sectional study shows that a large proportion of the participants had 57 58 modifiable UDTC risk factors with a tendency of clustering in Yangzhong City. Our analyses 59 60 indicate that men, younger adults, single adults, participants with higher levels of SES or smaller 16

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 37

1 2 3 household size are susceptible to modifiable UDTC risk factors clustering. Policies to prevent 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 UDTC have already been developed in the strategic plan and operational plan, however, the 7 8 accuracy and validity of implementing the undertaken policies are still insufficient. Consequently, 9 extra attention is required to pay on these high-risk groups during the progress of screening. 10 11 Relative departments also need to make effective public health programs targeting modifiable 12 13 UDTC risk factors that aim to decrease UDTC risk factors clustering in high-risk groups from 14 15 high-risk areas of UDTC. 16 17 18 AcknowledgmentsFor Wepeer would like review to thank all participants only of the present study. We are 19 20 also grateful to the Upper Digestive Tract Cancer Early Diagnosis and Treatment Program. 21 22 23 24 25 26 Contribution FX and CR are joint first authors. QDF, WWQ, WGQ, ZJY, WJJ, SG and 27 28 WX contributed to the study design. ZQ, SAW, STQ performed the survey and collected study 29 30 data. FX and CR wrote the manuscript and QDF, WWQ, ZJY, WJJ, SG,WX were responsible for 31 manuscript revision. 32 33 34 35

36 http://bmjopen.bmj.com/ 37 Funding This study was supported by the National key research and development program 38 39 (Grant NO. 2016YFC0901400; 2016YFC1302800); National Natural Science Foundation of 40 41 China (Grant NO. 81974493); and National Key Technology R&D Program (2006BAI02A05). 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 Competing interests None declared 48 49 50 51 52 Patient consent Obtained 53 54 55 56 57 58 Ethics approval This study was approved by the academic and ethical committee of the 59 60 cancer hospital of Chinese academy of medical sciences. 17

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 37 BMJ Open

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 Provenance and peer review Not commissioned; externally reviewed. 7 8 9 10 11 12 Data sharing statement No additional data are available. 13 14 15 16 17 18 Open accessFor This is peer an Open Access review article distributed only in accordance with the Creative 19 Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to 20 21 distribute, remix, adapt, build upon this work non-commercially and license their derivative works 22 23 on different terms, provided the original work is properly cited and the use is non-commercial. 24 25 See:http://creativecommons.org/licenses/by-nc/4.0/ 26 27 28 29 30 31 32 33 34 35

36 http://bmjopen.bmj.com/ 37 38 39 REFERENCES 40 41 [1] World Health Organization. World health statistics 2018: monitoring health for the SDGs, 42 sustainable development goals. Geneva: World Health Organization, 43

44 2018.https://www.who.int/gho/publications/world_health_statistics/2018/en/. on September 27, 2021 by guest. Protected copyright. 45 46 [2] Bray F, Ferlay J, Soerjomataram I, et al. Global cancer statistics 2018: GLOBOCAN 47 48 estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J 49 Clin,2018,68(6):394-424. 50 51 52 [3] Zheng RS, Sun KX, Zhang SW, et al. Report of cancer epidemiology in China, 2015. Chin J 53 Oncol,2019,41(1):19-28.In Chinese. 54 55 [4] Chen WQ, Zheng RS, Baade PD, et al. Cancer statistics in China, 2015. CA Cancer J 56 57 Clin,2016,66(2):115-32. 58 59 [5] Ajani JA, Bentrem DJ, Besh S, et al. Gastric cancer, version 2.2013: featured updates to the 60 18

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 37

1 2 3 NCCN Guidelines. J Natl Compr Canc Netw,2013,11(5):531-46.

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 [6] Ciocirlan M, Lapalus MG, Hervieu V, et al. Endoscopic mucosal resection for squamous 7 premalignant and early malignant lesions of the esophagus. Endoscopy,2007,39(1):24-9. 8 9 10 [7] Ma D, Yang F, Liao Z, et al. Expert opinion on early screening and endoscopic diagnosis and 11 treatment esophageal cancer in China (2014, Beijing). China Prac Med,2015,4:320-337.In 12 13 Chinese. 14 15 [8] Chen R, Ma S, Guan C, et al. The National Cohort of Esophageal Cancer-Prospective Cohort 16 Study of Esophageal Cancer and Precancerous Lesions based on High-Risk Population in 17 18 China (NCEC-HRP):For study peer protocol. BMJreview Open,2019,9(4):e027360. only 19 20 [9] Chen MJ, Chiou YY, Wu DC, et al. Lifestyle habits and gastric cancer in a hospital-based 21 22 case-control study in Taiwan. Am J Gastroenterol,2000,95(11):3242-9. 23 24 [10] Somi MH, Mousavi SM, Naghashi S, et al. Is there any relationship between food habits in 25 26 the last two decades and gastric cancer in North-Western Iran?.. Asian Pac J Cancer Prev. 27 2015,16(1):283-90. 28 29 [11] Navarro-Silvera SA, Mayne ST, Risch HA, et al. Principal component analysis of dietary and 30 31 lifestyle patterns in relation to risk of subtypes of esophageal and gastric cancer. Ann 32 Epidemiol,2011,21(7):543-50. 33 34 35 [12] Andrici J, Eslick GD. Hot Food and Beverage Consumption and the Risk of Esophageal

36 Cancer: A Meta-Analysis. Am J Prev Med.2015,49(6):952–960. http://bmjopen.bmj.com/ 37 38 [13] Gupta B, Kumar N, Johnson NW. Relationship of Lifetime Exposure to Tobacco, Alcohol 39 40 and Second Hand Tobacco Smoke with Upper aero-digestive tract cancers in India: a 41 Case-Control Study with a Life-Course Perspective. Asian Pac J Cancer 42 43 Prev,2017,18(2):347-56.

44 on September 27, 2021 by guest. Protected copyright. 45 [14] Chen WQ, ZUO TT. Initial effect achievement of battles on upper digestive tract cancer in 46 47 China. Chin J Prev Med,2017, 51(5):378-80.In Chinese. 48 49 [15] Luan DC, Li SJ, Li H, et al. Change trends in health behaviors among residents in Liaoning 50 51 province,1991-2006. Chin J Public Health, 2013, 29(10):1509-1511.In Chinese. 52 53 [16] McCormack VA, Boffetta P. Today's lifestyles, tomorrow's cancers: trends in lifestyle risk 54 factors for cancer in low- and middle-income countries. Ann Oncol. 2011,22(11):2349–2357. 55 56 57 [17] Li FX, Robson PJ, Chen Y, et al. Prevalence, trend, and sociodemographic association of five 58 modifiable lifestyle risk factors for cancer in Alberta and Canada. Cancer Causes 59 60 Control,2009,20(3):395-407. 19

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 37 BMJ Open

1 2 3 [18] Gu XP, Wang YC, Zhi HK, et al. Risk factors of esophageal and stomach cancer and their

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 clustering in Dafeng municipality ： a case-control study. Chin J Public Health,2016, 6 32(10):1406-1409.In Chinese. 7 8 9 [19] He Q, Jing YH, Huang HR, et al. Prevalence and clustering of esophageal cancer-related risk 10 factors among rural residents in Nanbu County, Sichuan Province. Chin H Cancer Prev 11 12 Treat,2019,26(22):1675-1680.In Chinese. 13 14 [20] Tong HY, Zhang MM, Zhang HY. Analysis on the epidemiology trend and disease burden of 15 16 esophagus cancer in Yangzhong from 2004 to 2015. Modern Prev 17 Med,2016,43(20):3665-3668,3687.In Chinese. 18 For peer review only 19 [21] Tong HY, Zhang MM, Sun LP, et al. Quantitative study on death caused by main chronic 20 21 diseases in Yangzhong city. Jiangsu J Prev Med,2017,28(5):502-504,508. In Chinese. 22 23 [22] DONG ZW. China's cancer screening and early diagnosis and treatment technology 24 25 program. Beijing: People's medical publishing house, 2009.In Chinese. 26 27 [23] Hong X, Ye Q, He J, et al. Prevalence and clustering of cardiovascular risk factors: a 28 cross-sectional survey among Nanjing adults in China. BMJ Open,2018,8(6):e020530. 29 30 31 [24] Ni W, Weng RX, Yuan X, et al. Clustering of cardiovascular disease biological risk factors 32 among older adults in Shenzhen City, China: a cross-sectional study. BMJ 33 34 Open,2019,9(3):e024336. 35

36 [25] Department of disease control, ministry of health, PRC. Guidelines for the prevention and http://bmjopen.bmj.com/ 37 38 control of overweight and obesity in Chinese adults. Beijing: people's medical publishing 39 house, 2006:2-4.In Chinese. 40 41 [26] Chinese nutrition society. Dietary guidelines for Chinese residents (2016). Beijing: people's 42 43 medical publishing house, 2016In Chinese.

44 on September 27, 2021 by guest. Protected copyright. 45 [27] Howitt C, Hambleton IR, Rose AM, et al. Social distribution of diabetes, hypertension and 46 47 related risk factors in Barbados: a cross-sectional study. BMJ Open,2015,5(12):e008869. 48 49 [28] Zaman MM, Bhuiyan MR, Karim MN, et al. Clustering of noncommunicable diseases risk 50 51 factors in Bangladeshi adults: An analysis of STEPS survey 2013. BMC Public 52 Health.2015,15:659. 53 54 [29] Xu F, Yin XM, Zhang M, et al. Family average income and body mass index above the 55 56 healthy weight range among urban and rural residents in regional Mainland China. Public 57 Health Nutr,2005,8(1):47-51. 58 59 60 [30] National Commission of Health Bureau of disease control and Prevention. Report on China 20

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 22 of 37

1 2 3 national nutrition and chronic disease survey (2015). Beijing: People’s Medical Publishing

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 House, 2015,11:33-50.In Chinese. 6 7 [31] Wang R, Zhang P, Gao C, et al. Prevalence of overweight and obesity and some associated 8 9 factors among adult residents of northeast China: a cross-sectional study. BMJ Open 10 2016,6:e010828. 11 12 13 [32] Howitt C, Hambleton IR, Rose AMC, et al. Social distribution of diabetes, hypertension and 14 related risk factors in Barbados: a cross-sectional study. BMJ Open.2015,5:e008869. 15 16 [33] Millwood IY, Walters RG, Mei XW, et al. Conventional and genetic evidence on alcohol and 17 18 vascular diseaseFor aetiology: peer a prospective review study of 500 000 only men and women in China. Lancet. 19 2019,393(10183):1831-42. 20 21 22 [34] Msambichaka B, Eze IC, Abdul R, et al. Insufficient Fruit and Vegetable Intake in a Low- 23 and Middle-Income Setting: A Population-Based Survey in Semi-Urban Tanzania.Nutrients. 24 25 2018,10(2):222. 26 27 [35] Guo YL, Tan XD, Liu XZ, et al. Fruit and vegetable intake of adults and its influencing 28 factors in some cities of Hubei province .J of Pub health and Prev Med,2016,27(5):82-85.In 29 30 Chinese. 31 32 [36] Wen JB, Sun ZM, Miao DD, et al. Influencing factors about early cancer of upper-digestive 33 34 tract among high-risk population in Huai’an City of Jiangsu Province. China 35 cancer,2019,28(10):749-756.In Chinese.

36 http://bmjopen.bmj.com/ 37 38 [37] Li Y, Zhang M, Jiang Y, Wu F. Co-variations and clustering of chronic disease behavioral 39 risk factors in China: China Chronic Disease and Risk Factor Surveillance, 2007. PLoS One. 40 41 2012,7(3):e33881. 42 43 [38] Ahmed S, Tariqujjaman M, Rahman MA, et al. Inequalities in the prevalence of undiagnosed

44 on September 27, 2021 by guest. Protected copyright. hypertension among Bangladeshi adults: evidence from a nationwide survey. Int J Equity 45 46 Health. 2019,18(1):33. 47 48 [39] Floud S, Balkwill A, Canoy D, et al. Marital status and ischemic heart disease incidence and 49 50 mortality in women: a large prospective study. BMC Med. 2014,12:42. 51 52 [40] Cao Z, Wang R, Cheng Y, et al. Adherence to a healthy lifestyle counteracts the negative 53 54 effects of risk factors on all-cause mortality in the oldest-old. Aging (Albany NY). 55 2019,11(18):7605-7619. 56 57 [41] Prom-Wormley EC, Clifford JS, Bourdon JL, et al. Developing community-based health 58 59 education strategies with family history: Assessing the association between community 60 21

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 23 of 37 BMJ Open

1 2 3 resident family history and interest in health education. Soc Sci Med. 2019. [Online ahead of

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 print]Doi:10.1016/j.socscimed.2019.02.011. 6 7 [42] Yang F, Qian D, Liu X, et al. Socioeconomic disparities in prevalence, awareness, treatment, 8 9 and control of hypertension over the life course in China. Int J Equity Health, 2017, 10 16(1):100. 11 12 13 [43] Workicho A, Belachew T, Feyissa GT, et al. Household dietary diversity and Animal Source 14 Food consumption in Ethiopia: evidence from the 2011 Welfare Monitoring Survey. BMC 15 16 Public Health. 2016,16(1):1192. 17 18 [44] Bot SD, MackenbachFor peer JD, Nijpels review G, et al. Association only between Social Network 19 Characteristics and Lifestyle Behaviours in Adults at Risk of Diabetes and Cardiovascular 20 21 Disease. PLoS One. 2016,11(10):e0165041. 22 23 [45] Verheijden MW, Bakx JC, van Weel C, et al. Role of social support in lifestyle-focused 24 25 weight management interventions. Eur J Clin Nutr. 2005:59 Suppl 1:S179-S186. 26 27 28 29 30 31 32 33 34 35

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 Table 1 Socio-demographic and socio-economic characteristics of participants in Yangzhong 50 51 City, China 52 53 Total(n=192 Men(n=826 Women(n=109 54 Category t/ χ² value P value 55 13) 8) 45) 56 57 Number,n(%) 19213(100.0) 8268(43.0) 10945(57.0) 58 59 60 22

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 24 of 37

1 2 3 Age,years,mean(SD) 53.2±7.8 53.5±7.8 52.9±7.8 5.61 <0.001 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 Age group,n(%) 33.29 <0.001 7 8 40-44 3142(16.4) 1270(15.4) 1872(17.1) 9 10 11 45-49 3937(20.5) 1642(19.9) 2295(21.0) 12 13 50-54 3743(19.5) 1560(18.9) 2183.0(19.9) 14 15 55-59 3571(18.6) 1582(19.1) 1989(18.2) 16 17 18 60-64 For peer3088(16.1) review1416(17.1) 1672(15.3) only 19 20 65-69 1732(9.0) 798(9.7) 934(8.5) 21 22 23 Marrige,n(%) 93.41 <0.001 24 25 Single 203(1.1) 137(1.7) 66(0.6) 26 27 Currently married 18285(95.2) 7907(95.6) 10378(94.8) 28 29 30 Divorced, widowed or 725(3.8) 224(2.7) 501(4.6) 31 separated 32 33 34 Education,n(%) 886.879 <0.001 35

36 No institutional education 1594(8.3) 228(2.8) 1366(12.5) http://bmjopen.bmj.com/ 37 38 39 Primary school 8510(44.3) 3436(41.6) 5074(46.4) 40 41 Junior high school 7591(39.5) 3647(44.1) 3944(36.0) 42 43 Senior high school and

44 1518(7.9) 957(11.6) 561(5.1) on September 27, 2021 by guest. Protected copyright. 45 higher 46 47 Annual family income, 48 0.04 0.981 49 n(%) 50 51 lower 8585(44.7) 3689(44.6) 4896(44.7) 52 53 Middle 5420(28.2) 2338(28.3) 3082(28.2) 54 55 56 Higher 5208(27.1) 2241(27.1) 2967(27.1) 57 58 4.0±1.4 4.0±1.4 4.0±1.4 -3.17 0.002 59 Household size,n,mean 60 23

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 25 of 37 BMJ Open

1 2 3 (SD) 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 BMI,kg/m2,mean(SD) 23.8±3.0 23.9±3.0 23.7±3.0 3.47 0.001 7 8 9 10 11 12 13 14 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 Table 2 Prevalence of modifiable UDTC risk factors in Yangzhong City adults aged 40-69 29 30 Total Women P 31 Factor Men (n=8268) χ² value 32 (n=19213) (n=10945) value 33 34 BMI,n(%) 2.62 0.106 35

36 http://bmjopen.bmj.com/ 37 Overweight or obesity 8695(45.3) 3797(45.9) 4898(44.8) 38 39 Normal weight or underweight 10518(54.7) 4471(54.1) 6047(55.2) 40 41 0.001 42 Current smoking,n(%) 6093.25 < 43

44 No 14589(75.9) 3988(48.2) 10601(96.9) on September 27, 2021 by guest. Protected copyright. 45 46 Yes 4624(24.1) 4280(51.8) 344(3.1) 47 48 49 Excessive alcohol consumption,n(%) 3197.31 <0.001 50 51 No 16109(83.8) 5504(66.6) 10605(96.9) 52 53 Yes 3104(16.2) 2764(33.4) 340(3.1) 54 55 56 insufficient vegetable intake,n(%) 0.32 0.572 57 58 No 6519(33.9) 2787(33.7) 3732(34.1) 59 60 24

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 26 of 37

1 2 3 Yes 12694(66.1) 5481(66.3) 7213(65.9) 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 insufficient fruit intake,n(%) 0.10 0.749 7 8 No 1055(5.5) 459(5.6) 596(5.4) 9 10 11 Yes 18158(94.5) 7809(94.4) 10349(94.6) 12 13 Pickled food consumption,n(%) 8.38 <0.05 14 15 No 6138(31.9) 2734(33.1) 3404(31.1) 16 17 18 Yes For peer13075(68.1) review5534(66.9) only7541(68.9) 19 20 Fried food consumption,n(%) 1.00 0.318 21 22 23 No 12293(64.0) 5323(64.4) 6970(63.7) 24 25 Yes 6920(36.0) 2945(35.6) 3975(36.3) 26 27 Hot food consumption,n(%) 27.10 <0.001 28 29 30 No 2221(11.6) 1070(12.9) 1151(10.5) 31 32 Yes 16992(88.4) 7198(87.1) 9794(89.5) 33 34 35

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 25

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Page 27 of 37 BMJ Open

1 2 3 4 5 Table 3 Prevalence of modifiable UDTC risk factors by relevant characters in Yangzhong City adults aged 40-69 6 7 Category Overweight or obesity Smoking Drinking Vegetables Fruit Pickled food Fried food Hot food 8 9 10 Age group,years,n(%) 11 12 40-44 For1492(47.5) peer533(17.0) review348(11.1) 1888(60.1) only2902(92.4) 2033(64.7) 782(24.9) 2971(94.6) 13 14 15 45-49 1953(49.6) 827(21.0) 540(13.7) 2648(67.3) 3735(94.9) 2594(65.9) 1451(36.9) 3894(98.9) 16 http://bmjopen.bmj.com/ 17 50-54 1729(46.2) 1068(28.5) 723(19.3) 2506(67.0) 3538(94.5) 2669(71.3) 1325(35.4) 3723(99.5) 18 19 55-59 1607(45.0) 987(27.6) 695(19.5) 2314(64.8) 3380(94.7) 2366(66.3) 1312(36.7) 3464(97.0) 20 21 22 60-64 1271(41.2) 796(25.8) 537(17.4) 2131(69.0) 2949(95.5) 2158(69.9) 1285(41.6) 2343(75.9) 23 on September 27, 2021 by guest. Protected copyright. 24 65-69 643(37.1) 413(23.8) 261(15.1) 1207(69.7) 1654(95.5) 1255(72.5) 765(44.2) 597(34.5) 25 26 27 P value <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 28 29 Marrige,n(%) 30 31 Single 77(37.9) 73(36.0) 43(21.2) 177(87.2) 200(98.5) 169(83.3) 143(70.4) 186(91.6) 32 33 34 Currently married 8334(45.6) 4439(24.3) 2979(16.3) 12087(66.1) 17253(94.4) 12424(67.9) 6518(35.6) 16304(89.2) 35 36 Divorced, widowed or separated 284(39.2) 112(15.4) 82(11.3) 430(59.3) 705(97.2) 482(66.5) 259(35.7) 502(69.2) 37 38 39 40 41 26 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

BMJ Open Page 28 of 37

1 2 3 4 5 P value <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 6 7 Education,n(%) 8 9 10 No institutional education 692(43.4) 75(4.7) 76(4.8) 781(49.0) 1479(92.8) 1286(80.7) 675(42.3) 1198(75.2) 11 12 Primary school For3699(43.5) peer2030(23.9) review1328(15.6) 5898(69.3) only8149(95.8) 4785(56.2) 2205(25.9) 7194(84.5) 13 14 15 Junior high school 3521(46.4) 2045(26.9) 1373(18.1) 4878(64.3) 7087(93.4) 5819(76.7) 3146(41.4) 7187(94.7) 16 http://bmjopen.bmj.com/ 17 Senior high school and higher 783(51.6) 474(31.2) 327(21.5) 1137(74.9) 1443(95.1) 1185(78.1) 894(58.9) 1413(93.1) 18 19 P value <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 20 21 22 Annual family income,n(%) 23 on September 27, 2021 by guest. Protected copyright. 24 lower 3711(43.2) 2128(24.8) 1551(18.1) 4844(56.4) 8234(95.9) 5535(64.5) 1300(15.1) 7168(83.5) 25 26 27 Middle 2458(45.4) 1303(24.0) 857(15.8) 3566(65.8) 4953(91.4) 2923(53.9) 1521(28.1) 4876(90.0) 28 29 Higher 2526(48.5) 1193(22.9) 696(13.4) 4284(82.3) 4971(95.4) 4617(88.7) 4099(78.7) 4948(95.0) 30 31 P value <0.001 <0.05 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 32 33 34 Household size,n(%) 35 36 0-3 3583(45.6) 1916(24.4) 1381(17.6) 5137(65.4) 7384(94.1) 5512(70.2) 2989(38.1) 7051(89.8) 37 38 39 40 41 27 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Page 29 of 37 BMJ Open

1 2 3 4 5 4-5 4285(44.6) 2285(23.8) 1428(14.8) 6574(68.4) 9159(95.2) 6216(64.6) 3153(32.8) 8467(88.0) 6 7 ≥6 827(47.4) 423(24.2) 295(16.9) 983(56.3) 1615(92.5) 1347(77.1) 778(44.6) 1474(84.4) 8 9 10 P value 0.064 0.601 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 11 12 Notes: Smoking, current smoking; Drinking,For excessive peer alcohol consumption; review Vegetables, insufficient vegetableonly intake, Fruit, insufficient fruit intake; Pickled food, 13 14 the consumption of pickled food, Fried food, the consumption of fried food; Hot food, the consumption of hot food 15 16 http://bmjopen.bmj.com/ 17 18 19 20 21 22 23 24 on September 27, 2021 by guest. Protected copyright. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 28 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

BMJ Open Page 30 of 37

1 2 3 4 5 Table 4 The different number and clustering of modifiable UDTC risk factors in residents by relevant characters in Yangzhong City adults aged 40-69 6 7 Category None(0) Single(1) Two(2) Non-clustering(≤2) Clustering(≥3) χ² value P value 8 9 10 Total 14(0.1) 244(1.3) 1235(6.4) 1493(7.8) 17720(92.2) 11 12 Gender,n(%) For peer review only 144.02 <0.001 13 14 15 men 2(0.0) 64(0.8) 356(4.3) 422(5.1) 7846(94.9) 16 http://bmjopen.bmj.com/ 17 women 12(0.1) 180(1.6) 879(8.0) 1071(9.8) 9874(90.2) 18 19 Age group,years,n(%) 688.53 <0.001 20 21 22 40-44 1(0.0) 66(2.1) 193(6.1) 260(8.3) 2882(91.7) 23 on September 27, 2021 by guest. Protected copyright. 24 45-49 0(0.0) 26(0.7) 165(4.2) 191(4.9) 3746(95.1) 25 26 27 50-54 0(0.0) 15(.4) 112(3.0) 127(3.4) 3616(96.6) 28 29 55-59 0(0.0) 48(1.3) 175(4.9) 223(6.2) 3348(93.8) 30 31 60-64 0(0.0) 33(1.1) 273(8.8) 306(9.9) 2782(90.1) 32 33 34 65-69 13(0.8) 56(3.2) 317(18.3) 386(22.3) 1346(77.7) 35 36 Marrige,n(%) 145.73 <0.001 37 38 39 40 41 29 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Page 31 of 37 BMJ Open

1 2 3 4 5 Single 0(0.0) 0(0.0) 9(4.4) 9(4.4) 194(95.6) 6 7 Currently married 12(0.1) 201(1.1) 1130(6.2) 1343(7.3) 16942(92.7) 8 9 10 Divorced, widowed or separated 2(0.3) 43(5.9) 96(13.2) 141(19.4) 584(80.6) 11 12 Education,n(%) For peer review only 308.68 <0.001 13 14 15 No institutional education 4(0.3) 56(3.5) 241(15.1) 301(18.9) 1293(81.1) 16 http://bmjopen.bmj.com/ 17 Primary school 8(0.1) 71(0.8) 503(5.9) 582(6.8) 7928(93.2) 18 19 Junior high school 2(0.0) 107(1.4) 428(5.6) 537(7.1) 7054(92.9) 20 21 22 Senior high school and higher 0(0.0) 10(0.7) 63(4.2) 73(4.9) 1445(95.2) 23 on September 27, 2021 by guest. Protected copyright. 24 Annual family income,n(%) 160.69 <0.001 25 26 27 lower 6(0.1) 102(1.2) 588(6.8) 696(8.1) 7889(91.9) 28 29 Middle 5(0.1) 105(1.9) 470(8.7) 580(10.7) 4840(89.3) 30 31 Higher 3(0.1) 37(0.7) 177(3.4) 217(4.2) 4991(95.8) 32 33 34 Household size,n(%) 20.88 <0.001 35 36 0-3 5(0.1) 165(2.1) 438(5.6) 608(7.7) 7242(92.3) 37 38 39 40 41 30 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

BMJ Open Page 32 of 37

1 2 3 4 5 4-5 7(0.1) 68(0.7) 627(6.5) 702(7.3) 8915(92.7) 6 7 ≥6 2(0.1) 11(0.6) 170(9.7) 183(10.5) 1563(89.5) 8 9 10 11 12 For peer review only 13 14 15 16 http://bmjopen.bmj.com/ 17 18 19 20 21 22 23 24 on September 27, 2021 by guest. Protected copyright. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 31 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Page 33 of 37 BMJ Open

1 2 3 4 5 6 7 8 9 10 11 12 For peer review only 13 Table 5 The multivariable logistic regression analysis of modifiable UDTC risk factors clustering 14 15 16 Category Wald P value β SE http://bmjopen.bmj.com/ OR 95% CI 17 18 Gender,n(%) 19 20 21 women - - - - 1.000 - 22 23 men 123.254 <0.001 0.708 0.064 2.030 1.791 to 2.300 24 on September 27, 2021 by guest. Protected copyright. 25 Age group,years,n(%) 26 27 28 40-44 - - - - 1.000 - 29 30 45-49 21.177 <0.001 0.459 0.100 1.582 1.301 to 1.924 31 32 33 50-54 51.048 <0.001 0.811 0.113 2.250 1.801 to 2.810 34 35 55-59 4.776 0.029 0.223 0.102 1.250 1.023 to 1.526 36 37 38 39 40 41 32 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

BMJ Open Page 34 of 37

1 2 3 4 5 60-64 11.690 0.001 -0.342 0.100 0.710 0.584 to 0.864 6 7 65-69 163.181 <0.001 -1.291 0.101 0.275 0.226 to 0.335 8 9 10 Marrige,n(%) 11 12 Single For peer- review- -only- 1.000 - 13 14 15 Currently married 0.588 0.443 -0.270 0.352 0.764 0.383 to 1.521 16 http://bmjopen.bmj.com/ 17 Divorced, widowed or separated 6.579 0.010 -0.935 0.365 0.392 0.192 to 0.802 18 19 Education,n(%) 20 21 22 No institutional education - - - - 1.000 - 23 on September 27, 2021 by guest. Protected copyright. 24 Primary school 87.294 <0.001 0.786 0.084 2.195 1.861 to 2.589 25 26 27 Junior high school 5.811 0.016 0.233 0.097 1.263 1.045 to 1.527 28 29 Senior high school and higher 9.342 0.002 0.458 0.150 1.581 1.179 to 2.121 30 31 Annual family income,n(%) 32 33 34 Higher - - - - 1.000 - 35 36 Middle 85.857 <0.001 -0.800 0.086 0.449 0.379 to 0.532 37 38 39 40 41 33 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Page 35 of 37 BMJ Open

1 2 3 4 5 lower 184.469 <0.001 -1.168 0.086 0.311 0.263 to 0.368 6 7 Household size,n(%) 8 9 10 0-3 - - - - 1.000 - 11 12 4-5 For peer0.626 review0.429 -0.049 only0.062 0.952 0.842 to 1.076 13 14 15 ≥6 13.766 <0.001 -0.360 0.097 0.698 0.577 to 0.844 16 http://bmjopen.bmj.com/ 17 Constant 62.514 <0.001 2.918 0.369 - - 18 19 20 21 22 23 24 on September 27, 2021 by guest. Protected copyright. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 34 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 36 of 37

1 2 STROBE Statement—Checklist of items that should be included in reports of cross-sectional studies 3

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from Item Page 5 No Recommendation No 6 7 Title and abstract 1 (a) Indicate the study’s design with a commonly used term in the title 1 8 9 or the abstract 10 11 (b) Provide in the abstract an informative and balanced summary of 1-2 12 what was done and what was found 13 14 Introduction 15 16 17 Background/rationale 2 Explain the scientific background and rationale for the investigation 2-4 18 Forbeing peer reported review only 19 20 Objectives 3 State specific objectives, including any prespecified hypotheses 4 21 22 Methods 23 24 Study design 4 Present key elements of study design early in the paper 4 25 26 Setting 5 Describe the setting, locations, and relevant dates, including periods of 4-5 27 28 recruitment, exposure, follow-up, and data collection 29 30 Participants 6 (a) Give the eligibility criteria, and the sources and methods of 4 31 selection of participants 32 33 Variables 7 Clearly define all outcomes, exposures, predictors, potential 5-6 34 confounders, and effect modifiers. Give diagnostic criteria, if 35

36 applicable http://bmjopen.bmj.com/ 37 38 Data sources/ 8* For each variable of interest, give sources of data and details of 5-6 39 measurement methods of assessment (measurement). Describe comparability of 40 assessment methods if there is more than one group 41 42 Bias 9 Describe any efforts to address potential sources of bias 5 43

44 on September 27, 2021 by guest. Protected copyright. 45 Study size 10 Explain how the study size was arrived at 4 46 47 Quantitative variables 11 Explain how quantitative variables were handled in the analyses. If 5-6 48 applicable, describe which groupings were chosen and why 49 50 Statistical methods 12 (a) Describe all statistical methods, including those used to control for 6-7 51 confounding 52 53 (b) Describe any methods used to examine subgroups and interactions 6-7 54 55 56 (c) Explain how missing data were addressed 4-5 57 58 (d) If applicable, describe analytical methods taking account of Not 59 sampling strategy included 60

1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 37 of 37 BMJ Open

1 2 (e) Describe any sensitivity analyses Not 3 included

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 Results 7 8 Participants 13* (a) Report numbers of individuals at each stage of study—eg numbers 4-5 9 potentially eligible, examined for eligibility, confirmed eligible, 10 included in the study, completing follow-up, and analysed 11 12 (b) Give reasons for non-participation at each stage Not 13 14 included 15 16 (c) Consider use of a flow diagram Not 17 included 18 For peer review only 19 Descriptive data 14* (a) Give characteristics of study participants (eg demographic, clinical, 8 20 social) and information on exposures and potential confounders 21 22 23 (b) Indicate number of participants with missing data for each variable Not 24 of interest included 25 26 Outcome data 15* Report numbers of outcome events or summary measures 8-10 27 28 Main results 16 (a) Give unadjusted estimates and, if applicable, confounder-adjusted 10 29 estimates and their precision (eg, 95% confidence interval). Make clear 30 31 which confounders were adjusted for and why they were included 32 33 (b) Report category boundaries when continuous variables were Not 34 categorized included 35

36 (c) If relevant, consider translating estimates of relative risk into Not http://bmjopen.bmj.com/ 37 absolute risk for a meaningful time period included 38 39 Other analyses 17 Report other analyses done—eg analyses of subgroups and Not 40 41 interactions, and sensitivity analyses included 42 43 Discussion

44 on September 27, 2021 by guest. Protected copyright. 45 Key results 18 Summarise key results with reference to study objectives 11-14 46 47 Limitations 19 Discuss limitations of the study, taking into account sources of 14 48 49 potential bias or imprecision. Discuss both direction and magnitude of 50 any potential bias 51 52 Interpretation 20 Give a cautious overall interpretation of results considering objectives, 14 53 limitations, multiplicity of analyses, results from similar studies, and 54 other relevant evidence 55 56 57 Generalisability 21 Discuss the generalisability (external validity) of the study results 14 58 59 Other information 60

2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 38 of 37

1 2 Funding 22 Give the source of funding and the role of the funders for the present 15 3 study and, if applicable, for the original study on which the present

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 article is based 6 7 8 9 *Give information separately for exposed and unexposed groups. 10 11 12 13 Note: An Explanation and Elaboration article discusses each checklist item and gives methodological background and 14 published examples of transparent reporting. The STROBE checklist is best used in conjunction with this article (freely 15 available on the Web sites of PLoS Medicine at http://www.plosmedicine.org/, Annals of Internal Medicine at 16 17 http://www.annals.org/, and Epidemiology at http://www.epidem.com/). Information on the STROBE Initiative is 18 available at www.strobe-statement.org.For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

3 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Prevalence and clustering of modifiable risk factors for upper digestive tract cancer among residents aged 40 to 69 years in Yangzhong City, China: a cross-sectional study ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2020-042006.R1

Article Type: Original research

Date Submitted by the 17-Dec-2020 Author:

Complete List of Authors: Feng, xiang; Yangzhong People's Hospital, Institute of Cancer Prevention and Treatment, Hua, Zhao-lai; Yangzhong People's Hospital, Institute of Cancer Prevention and Treatment, Zhou, Qin; Yangzhong People's Hospital, Institute of Cancer Prevention and Treatment, Shi, Ai-wu; Yangzhong People's Hospital, Institute of Cancer Prevention and Treatment, Song, Tong-qiu; Yangzhong People's Hospital, Institute of Cancer Prevention and Treatment, Qian, Dongfu; Nanjing Medical University, School of Health Policy &

Management http://bmjopen.bmj.com/ Chen, Ru; National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Cancer Registry Office Wang, Gui-qi; National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Cancer Registry Office Wei, Wen-Qiang; National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences

and Peking Union Medical College, Cancer Registry Office on September 27, 2021 by guest. Protected copyright. Zhou, Jin-Yi; Institute of Chronic Non-communicable Diseases Prevention and Control, Depatment of Non-communicable Disease Prevention Wang, Jie-jun; Changzheng Hospital, Department of Oncology; Naval Military Medical University Shao, Gang; the 903rd Hospital of PLA, Department of Oncology Wang, Xi; the 903rd Hospital of PLA, Department of Oncology

Primary Subject Epidemiology Heading:

Secondary Subject Heading: Public health, Epidemiology, Oncology

EPIDEMIOLOGY, Endoscopy < GASTROENTEROLOGY, Risk management Keywords: < HEALTH SERVICES ADMINISTRATION & MANAGEMENT, Epidemiology < ONCOLOGY

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39

40 on September 27, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 35

1 2 3

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 35 BMJ Open

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 Prevalence and clustering of modifiable risk 6 7 factors for upper digestive tract cancer 8 9 10 among residents aged 40 to 69 years in 11 12 Yangzhong City, China: a cross-sectional 13 14 15 study 16 17 Xiang Feng,1,2 Zhaolai Hua,1 Qin Zhou1, Aiwu Shi,1 Tongqiu Song,1 Dongfu Qian,2 Ru Chen,3 18 For peer review only 19 GuiQi Wang,3 Wenqiang Wei,3 Jinyi Zhou,4 Jiejun Wang,5 Gang Shao,6 Xi Wang6 20 21 22 23 1 Institute of tumour prevention and control, People’s Hospital of Yangzhong City, Yangzhong, 24 25 China 26 27 2 School of Medicine and Politics, Nanjing Medical University, Nanjing, China 28 29 30 3 Cancer Registry Office, National Cancer Center/National Clinical Research Center for 31 Cancer/Cancer Hospital, Chinese Academy of Medical Science and Peking Union Medical 32 33 College, Beijing, China 34 35 4 Depatment of Non-communicable Disease Prevention, Jiangsu Center for Disease Control and

36 http://bmjopen.bmj.com/ Prevention, Nanjing, China 37 38 39 5 Department of Oncology, Changzheng Hospital, Naval Military Medical University, Shanghai, 40 China. 41 42 43 6 Department of Oncology, the 903rd Hospital of PLA, Hangzhou, China.

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 Correspondence to Zhaolai Hua; [email protected]; Dr. Wenqiang Wei; 48 49 [email protected] 50 51 52 53 ABSTRACT 54 55 56 Objectives To describe the prevalence of modifiable risk factors for upper digestive tract cancer 57 (UDTC) and its clustering, and investigate relevant influencing factors of modifiable UDTC risk 58 59 factors clustering among residents aged 40-69 years in Yangzhong City, China. 60 1

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 35

1 2 3 Design Cross-sectional study

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 Participants A total of 21 175 participants aged 40-69 years enrolled in the study. 1 962 subjects 7 were excluded due to missing age, marital status, or some other essential information. Eventually, 8 9 19 213 participants were available for the present analysis. 10 11 Main outcomes measures Prevalence and clustering of eight modifiable UDTC risk factors 12 13 (overweight or obesity, current smoking, excessive alcohol consumption, insufficient vegetables 14 intake, insufficient fruit intake, and the consumption of pickled, fried, and hot food) were 15 16 analyzed. 17 18 Results The prevalenceFor of overweight/obesity,peer review current smoking, only excessive alcohol consumption, 19 20 insufficient vegetables intake, insufficient fruit intake, and the consumption of pickled, fried, and 21 22 hot food in this study was 45.3%, 24.1%, 16.2%, 66.1%, 94.5%, 68.1%, 36.0%, and 88.4%, 23 24 respectively. Nearly all (99.9%) participants showed one or more UDTC risk factors, 98.6% of the 25 26 participants showed at least two risk factors,92.2% of the participants had at least three risk 27 28 factors, and 69.7% of the participants had four or more risk factors. Multivariate logistic 29 30 regression analysis revealed that men, younger age, single, higher education, higher annual family 31 32 income and, smaller household size were more likely to present modifiable UDTC risk factors 33 34 clustering. 35

36 http://bmjopen.bmj.com/ Conclusions A large proportion of the participants had modifiable UDTC risk factors with a 37 38 clustering tendency in Yangzhong City. Extra attention is required to pay to these groups who are 39 40 susceptible to risk factors clustering during screening progress. Relative departments also need to 41 42 make significant public health programs that aim to decrease modifiable UDTC risk factors 43

44 clustering among residents aged 40-69 years from high-risk areas of UDTC. on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 50 Strengths and limitations of this study 51 52 This is the first study examining the prevalence and clustering of modifiable upper digestive tract 53 54 cancer (UDTC) risk factors and investigating relevant influencing factors in Yangzhong City, with 55 large and representative residents aged 40-69 years from the Upper Digestive Tract Cancer Early 56 57 Diagnosis and Treatment (UDTCEDAT). 58 59 60 2

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 35 BMJ Open

1 2 3 Participants who volunteered to participate in our study are considered to be the high-risk group

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 for UDTC, which is significantly important for the prevention and control of UDTC in China, 6 where the morbidity and mortality of UDTC are increasing. 7 8 9 10 11 A cross-section study cannot exam the causality or temporal relationship. 12 13 14 15 16 The modifiable UDTC risk factors included in our study were self-reported by participants, which 17 may contribute to recall and reporting bias, except BMI. 18 For peer review only 19 20 21 22 We only focus on the eight modifiable UDTC risk factors. Still, there are far more than eight risk 23 factors for UDTC, which may underestimate the clustering of modifiable UDTC risk factors. 24 25 26 27 28 INTRODUCTION 29 30 31 According to the WHO, more than 70% of the total deaths worldwide were related to

32 1 33 non-communicable diseases (NCDs) in 2016. Cancer is the second cause of NCDs, accounting 34 for 22% of total global deaths related to NCDs. Globally, in 2018, an estimated 572 034 35

36 http://bmjopen.bmj.com/ individuals were diagnosed with oesophageal cancer (OC), and 1 033 701 individuals were 37 38 diagnosed with stomach cancer (SC), with approximately 50% of new cases occurring in China. 39 40 There are an estimated 508 585 cancer deaths of OC and 782 685 cancer deaths of SC in 2018, 41 42 which accounts for 5.3% and 8.2% of cancer-cause deaths.2 Apparently, upper digestive tract 43

44 cancer (UDTC) (oesophagus, stomach) has become a significant morbidity and mortality source on September 27, 2021 by guest. Protected copyright. 45 46 related to cancer. According to the National Cancer Center,3 OC has been the sixth most common 47 48 cancer and the fourth most common cancer cause of cancer-related death. SC has been the second 49 50 most common cancer and the third most common cancer cause of cancer-related death. The 51 52 incidence rate of OC and SC was 17.87/100 000, and 29.31/100 0000, the mortality rate of them 53 54 were 13.68/100 0000 and 21.16/100 0000, respectively, in China in 2015.3 Hence, UDTC has 55 56 become a major public health challenge in China, and the disease burden of it is also 57 58 considerable.4 Due to the population health-seeking behaviour and the diseases’ character, UDTC 59 60 is mostly diagnosed at a late stage, which is leading to a low survival rate. It was estimated that its 3

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 35

1 2 3 five-year survival rate was less than 20% if diagnosed at an advanced stage but is as high as 95% 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 5 6 7 6 if detected at an earlier stage. 7 8 Although the cause of UDTC is not exact, it is believed by researchers that the epidemic of 9 10 UDTC in China is attributed to the multiplicity of demographic factors, diet, lifestyle, family 11 12 health, environment, gastrointestinal history and genetic factors.8 It is well known that tobacco, 13 14 alcohol consumption, overweight or obesity, thermal irritation (drinking scalding liquids) and 15 16 insufficient intake of vegetables and fruit, consumption of pickled and fried food are eight risk 17 18 factors that can beFor altered by peer a tangible action review for UDTC control. only9 10 11 12 13 19 20 21 Although parts of these risk factors have dramatically decreased because of a set of 22 23 interventions implemented by Chinese government,14 15 the others have increased and will 24 25 continue to grow in the next decades because of the rapid transition of urbanization, 26 27 industrialization and ageing.16 17 Furthermore, several studies have indicated that these risk factors 28 18 19 29 clustering was common in the population which would further increase the risk of UDTC. A 30 31 comprehensive assessment of the distribution and the status of UDTC risk factors clustering is 32 33 significant for cancer prevention and control. Once we have such data, interventions can be 34 35 planned and implemented efficiently to minimize these modifiable risk factors, thereby

36 http://bmjopen.bmj.com/ minimizing the health risks of increasing UDTC-related mortality and morbidity. 37 38 39 The Yangzhong City of Jiangsu Province is one of the high-risk areas of UDTC, especially in 40 41 the rural area.8 In 2015, the incidence rate of OC was 69.2/100 000, the mortality rate of OC and 42 43 SC was 70.24/100 0000 and 81.89/100 0000, respectively, in Yangzhong City, which is higher

44 on September 27, 2021 by guest. Protected copyright. 45 compared with the average of the nation.20 21 Hence, Yangzhong City had been one of the project 46 47 sites of the Upper Digestive Tract Cancer Early Diagnosis and Treatment (UDTCEDAT) since the 48 49 2006.8 Many studies have estimated the risk factors for UDTC in different areas worldwide.9 10 11 50 51 12 13 The results reveal that risk factors for UDTC are widespread. The modifiable risk factors are 52 53 significantly crucial for the prevention and control of UDTC because these factors can be changed 54 55 by some healthy education or other interventions implemented by doctors and government, and 56 57 improved with the increase of personal health awareness. However, the evidence on the clustering 58 59 of these modifiable risk factors in high-risk areas is still limited. Thus, we aimed to report the 60 4

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 35 BMJ Open

1 2 3 prevalence and clustering of modifiable UDTC risk factors, and analyze the relevant factors 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 influencing modifiable UDTC risk factor clustering among residents aged 40-69 years in 7 8 Yangzhong City which is a high-risk area of UDTC. 9 10 METHODS 11 12 13 Study population 14 15 16 For the present study, we used secondary data collected from the screening of UDTC, focusing on 17 18 the early diagnosisFor and treatment peer of UDTC review among high-risk only populations (aged 40-69 years) in 19 20 Yangzhong City, China, from 2006 to 2017.9 22 We use the method of multistage stratified 21 22 cluster sampling to select the study sample. In the first stage, we stratified Yangzhong City 23 24 into six regions (Sanmao, Baqiao, Youfang, Xinglong, Xilai, and Xinba ) covering the whole of 25 26 Yangzhong. In the second stage, we randomly selected clusters of three regions (Baqiao, Youfang, 27 28 and Xinglong) by region distribution and economic level based on Yangzhong Yearbook data. In 29 30 the third stage, administration villages or neighbourhood communities in each chosen regions 31 were randomly selected with probability proportional to size. In the fourth stage, each resident 32 33 group or village group was selected from chosen administration villages or neighbourhood 34 35 communities. In the fifth stage, all man or woman eligible from each household in the sites

36 http://bmjopen.bmj.com/ 37 mentioned above were invited for cancer screening, unless they met the following exclusion 38 39 criteria: (1) history of UDTC or mental disorder; (2) contraindications for endoscopic 40 41 examinations and (3) inability to complete the whole interview or informed consent. Inclusion 42 43 criteria for participants were as following: (1) aged 40-69 years; (2) permanent residents in

44 on September 27, 2021 by guest. Protected copyright. 45 Yangzhong City and (3) willing to accept endoscopic examination. 46 47 48 Before the screening, we obtained written informed consent from all participants after informing 49 50 them about the backgrounds, objectives, procedures, benefits, confidentiality agreement of 51 52 personal information and possible consequences of the whole program. Then questionnaire-based 53 54 interview, physical examinations, laboratory tests were performed by professional investigators. 55 56 At last, the endoscopic examinations, pathological diagnosis and necessary therapy for 57 58 participants were conducted by well-trained doctors in People’s Hospital of Yangzhong City. The 59 60 screening procedure follows China's cancer screening and early diagnosis and treatment 5

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 35

1 2 3 technology program strictly.22 The data used in this study derived mainly from the questionnaire 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 and physical examinations. Finally, a total of 21 175 individuals were surveyed, with a response 7 8 rate of 55.3% (19213 / 34743), 1 962 residents were excluded due to missing age, marital status, 9 or some other factors, leaving 19 213 participants available for the present survey. The sample 10 11 size accounted for about 17.3% of the total target population of Yangzhong City. We provided 12 13 health education about UDTC and the potential role of modifiable risk factors related to UDTC to 14 15 all eligible participants after the screening. Besides, we combined active and passive follow-ups to 16 17 collect outcome information for participants diagnosed with UDTC or precancerous lesions. We 18 For peer review only 19 also performed a regular re-examination for patients according to the diagnosis. 20 21 22 23 24 25 Questionnaire data collection 26 27 28 Before implementing data collection, training sessions organized by the expert group on 29 30 UDTCEDAT were provided for all staff. The aim of this study, the standard measurement 31 methods, how to perform questionnaires properly, and the concrete study procedure were included 32 33 in the training contents. At the end of the training sessions, all staff participated in the assessment 34 35 and proved to be qualified.

36 http://bmjopen.bmj.com/ 37 38 We use uniformly structured questionnaires to collect information through face-to-face 39 40 interviews. Each questionnaire took approximately 25 mins to complete. The questionnaire 41 42 information included demographic factors (gender, birthday, address, ID, marital status, and 43

44 household size), socioeconomic characteristic (education and annual family income), behavioural on September 27, 2021 by guest. Protected copyright. 45 46 factors (excessive alcohol consumption, current smoking), dietary habits (insufficient fresh 47 48 vegetables intake and insufficient fruit intake, consumption of pickled, fried and hot food), and 49 50 body mass index (BMI). 51 52 53 Physical examination 54 55 Physical examination included height and weight. Height and weight measurements were taken by 56 57 height scale and digital weight, respectively, with the help of trained examiners based on a 58 59 standardized program. All subjects were asked to remove any footwear, hats, and heavy clothing 60 6

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 35 BMJ Open

1 2 3 before height and weight were measured. Height was measured to the nearest 0.1 cm, while 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 weight was measured to the nearest 0.1 kg. BMI was calculated by dividing body weight (in 7 8 kilograms) by the square of height (in meters). 9 10 Assessment criteria 11 12 13 Definitions of UDTC modifiable risk factors and the clustering of these risk factors 14 15 16 Eight modifiable UDTC risk factors were defined based on current national guidelines or 17 18 related references.For Overweight/obesity peer was review defined as BMI ≥ 24.0only kg/m2.23 24 25 Excessive alcohol 19 20 consumption, insufficient fresh vegetables and fruit intake were defined according to the Dietary 21 22 Guidelines for Chinese residents (2016).26 Accordingly, excessive alcohol consumption was 23 24 defined as consumption of more than 25 gram (for males) or 15 gram (for females) alcohol drinks 25 26 per day after calculating pure alcohol based on the type of alcohol they chose, insufficient 27 28 vegetables intake as self-reported consumption of vegetables less than 7 times per week and

29 26 30 insufficient fruit intake as self-reported consumption of fruit less than 7 times per week. Current 31 smoking was defined as self-reported having used any tobacco products, including cigarettes, 32 33 cigars or pipes daily continuously.27 Besides, self-reported consumption of pickled, fried or hot 34 35 food at least once a week were classified as “Yes” in dietary habits, respectively.

36 http://bmjopen.bmj.com/ 37 38 Based on considering the literature and the average number of risk factors in the research 39 40 population, clustering of modifiable UDTC risk factors was defined as presenting at least four 41 42 related risk factors in one individual.24 28 43

44 on September 27, 2021 by guest. Protected copyright. 45 Covariates 46 47 48 Covariates included in this study were demographic and socioeconomic information ascertained 49 50 by questionnaire, including age (40-44 years, 45-49 years, 50-54 years, 55-59 years, 60-64 years, 51 52 65-69 years), gender (male and female), marital status (single, currently married, 53 divorced/widowed/separated), educational status (no institutional education, primary school, 54 55 junior high school, senior high school and higher), household size (0-3, 4-5, ≥6) and annual family 56 57 income (tertiles: lower, middle and higher).29 58 59 60 Statistical analysis 7

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 35

1 2 3 Descriptive statistics were applied to describe the socio-demographic characteristics of the 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 sampled population. The difference in continuous variables was analyzed by student’s t-test, and 7 8 by χ² test to assess the differentials in the prevalence values among categorical variables. 9 Differences in men and women, the prevalence of each modifiable UDTC risk factor and the 10 11 distribution of modifiable UDTC risk factors clustering (0, 1, 2, 3 and, ≥4) in socio-demographic 12 13 and other characteristics were described in the overall population, respectively. Multiple logistic 14 15 regression models were adopted to explore the association between relevant characteristics and 16 17 UDTC risk factors clustering. Only the variables that we found statistically significant at P <0.05 18 For peer review only 19 in the univariate analysis were included in the multiple logistic regression models. The result of 20 21 multiple logistic regression analyses was presented in terms of adjusted odds ratios (AOR) and 22 23 their respective 95% confidence intervals (CIs). All statistical analyses were performed by SPSS 24 25 software V.17.0. A two-sided P <0.05 was considered to be statistically significant. 26 27 28 Participant involvement statement 29 30 31 No participants or public were included in the design phase of this study. No participants were 32 33 asked to advise on interpretation or writing up of results. Dissemination of the result of the 34 35 research to participants and relevant participants community was prohibited. All the participants

36 http://bmjopen.bmj.com/ had the right to receive the result of health check if they wanted. 37 38 39 40 41 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 8

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 35 BMJ Open

1 2 3 RESULTS 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 Socio-demographic and other characteristics of participants 7 8 9 The description of socio-demographic and other characteristics of 19 213 participants are 10 11 presented in Table 1. Of the participants (40-69 years, mean age 53.2±7.8 years) surveyed, 57.0% 12 13 were women, with a mean age of 52.9±7.8 years. More than 95.0% (man 95.6%, woman 94.8%) 14 15 of participants were married, nearly half (44.3%) (man 41.6%, woman 46.4%) had an education 16 17 level of primary school, 44.7% (man 44.6%, woman 44.7%) had a lower level of annual family 18 For peer review only 19 income, and the mean of household size and BMI were 4.0±1.4 (man 4.0±1.4, woman 4.0±1.4) 20 21 and 23.8±3.0 kg/m2 (man 23.9±3.0, woman 23.7±3.0), respectively. The differences between 22 23 men and women in age, marriage, education, household size and BMI were significant (all 24 25 P<0.01) (Table 1). 26 27 28 29 30 31 32 33 34 35

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 9

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 35

1 2 3 Prevalence of modifiable UDTC risk factors 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 The prevalence of overweight or obesity, current smoking and excessive alcohol consumption in 7 8 this study was 45.3%, 24.1% and 16.2%, respectively. The prevalence of current smoking and 9 10 excessive alcohol consumption in men was significantly higher than that in women (all P<0.05). 11 12 In addition, insufficient vegetables intake, insufficient fruit intake and the consumption of pickled, 13 14 fried and hot food in participants accounted for 66.1%, 94.5%, 68.1%, 36.0%, and 88.4%, 15 16 respectively. The prevalence of consumption of pickled and hot food was higher in women than in 17 18 men (all P<0.01) For (Table 2). peer As shown inreview table 3, there were only significant differences with age, 19 20 marriage status, the level of education and annual family income in the eight UDTC risk factors 21 22 (all P<0.05). The prevalence of these eight modifiable UDTC risk factors tends to be higher in 23 24 single participants, except for overweight or obesity (all P<0.001). Moreover, the prevalence of 25 26 excessive alcohol consumption, insufficient vegetables intake, insufficient fruit intake and the 27 28 consumption of pickled, fried and hot food varied significantly with the household size (Table 3). 29 30 31 32 33 34 35

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 10

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 35 BMJ Open

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 Clustering of modifiable UDTC risk factors 7 8 9 The prevalence of zero to eight modifiable UDTC risk factors participants had simultaneously in 10 11 the study (Overweight/obesity, current smoking, excessive alcohol consumption, insufficient 12 13 vegetables intake, insufficient fruit intake, the consumption of pickled, fried and hot food) was 14 15 0.1%, 1.3%, 6.4%, 22.6%, 23.2%, 22.4%, 17.5%, 5.5% and 1.0%, respectively (data shown partly 16 17 in Table 4). Among them, nearly all (99.9%) participants showed one or more modifiable UDTC 18 For peer review only risk factors, 98.6% showed two or more modifiable UDTC risk factors, 92.2% had three or more 19 20 modifiable UDTC risk factors and 69.7% presented at least four modifiable UDTC risk factors. 21 22 The average count of modifiable UDTC risk factors per participant in this study is 4.39. The 23 24 prevalence of clustering of modifiable UDTC risk factors was higher in men than in women 25 26 (P<0.001). The prevalence of clustering of modifiable UDTC risk factors was the highest in the 27 28 age group 50-54 and among single participants (all P<0.001). There was an increasing trend 29 30 towards modifiable UDTC risk factors clustering with an increasing level of education and annual 31 32 family income (all P<0.001). Moreover, modifiable UDTC risk factors clustering was the highest 33 34 among those participants who had 0-3 family members (P<0.001) (Table 4). 35

36 http://bmjopen.bmj.com/ 37 The multivariable logistic regression analysis revealed that men were more likely to have 4 or 38 39 more modifiable UDTC risk factors compared with women (OR 2.302, 95% CI 2.145 to 2.471). 40 41 The prevalence of modifiable UDTC risk factors clustering increased with age and then decreased, 42 43 peaking at 50-54 (OR 2.044, 95% CI 1.825 to 2.288). Participants who were married (OR 0.511,

44 on September 27, 2021 by guest. Protected copyright. 45 95% CI 0.330 to 0.792) were less likely to have 4 or more modifiable UDTC risk factors than 46 47 those who were single. In addition, we divided the education into four groups, which showed 48 49 clearly that increasing the level of educations was a risk factor for modifiable UDTC risk factors 50 51 clustering. Compared with participants who had a higher level of annual family income, those who 52 had middle (OR 0.218, 95% CI 0.197 to 0.241) and lower (OR 0.223, 95% CI 0.201 to 0.247) 53 54 level of annual family income were less likely to have 4 or more modifiable UDTC risk factors. 55 56 Modifiable UDTC risk factors clustering were less common among participants who had more 57 58 than 6 family members than those who had less than 3 ones (OR 0.598, 95% CI 0.527 to 0.678) 59 60 (Table 5). 11

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 35

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 7 8 DISCUSSION 9 10 11 12 13 As far as we know, this is the first large population-based survey investigated the prevalence and 14 15 clustering of eight modifiable UDTC risk factors and described the socio-demographic and 16 17 socio-economic factors associated with these among Yangzhong City residents aged 40-69 from 18 For peer review only 19 southeast China. The present study revealed that the prevalence of overweight or obesity, current 20 21 smoking, excessive alcohol consumption, insufficient vegetables intake, insufficient fruit intake 22 23 and consumption of pickled, fried and hot food is significantly high with a tendency of clustering, 24 25 which implies the health risk of UDTC residents have in Yangzhong City. We found that the 26 27 prevalence of insufficient fruit intake and consumption of hot and pickled food were the top three 28 29 modifiable UDTC risk factors in the population surveyed. Besides, 69.7% of the participants 30 presented at least four UDTC risk factors. 31 32 33 The prevalence of overweight or obesity (45.3%) in our study was higher than that observed 34 35 in Nanjing (35.6%),23 and nationwide population (42.0%),30 but our findings were closed to that

36 http://bmjopen.bmj.com/ 37 observed in some other regional.24 31 The prevalence of current smoking in this population 38 39 (24.1%) was much higher than that in a cross-sectional study in Shenzhen (10.5%)24 and Barbados 40 41 (9.2%),32 which was consistent with a survey in Nanjing (24.5%).23 However, the rate was not as 42 43 high as reported (28.1%) in the China national nutrition and chronic disease survey (2015).30 Our

44 on September 27, 2021 by guest. Protected copyright. 45 findings showed a high prevalence of excessive alcohol consumption in the Yangzhong population 46 47 (16.2%) relative to the national average of 11.1% in men and 2.0% in women.30The rate of 48 49 excessive alcohol consumption we found was similar to Barbados, Nanbu and the prospective 50 51 study of China Kadoorie Biobank, where excessive alcohol consumption rates were around 52 53 14.5%,32 16.7%19and 14.9%,33 respectively. We found higher levels of insufficient intake of 54 55 vegetables (66.1%) in this population than those observed in the Tanzania34 and Hubei Province35 56 57 where the insufficient intake of vegetables or fruit is 55.8% and 29.7%, respectively, while the 58 59 levels of insufficient intake of fruit (94.5%) in our study were also much higher than that observed 60 12

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 35 BMJ Open

1 2 3 in the region mentioned above.34 35 Moreover, the proportion of the Yangzhong population had 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 dietary habits of consumption of pickled, fried and hot food were greater than the levels in Huaian

7 36 19 8 (22.7%, 7.1%, and 10.9%), as well as in Nanbu (28.63%, 1.95%, and 6.11%), both of these 9 region mentioned above are high-risk areas of UDTC in China. 10 11 12 The clustering of risk factors for chronic diseases is widespread.19 Several previous studies 13 14 reported the clustering of chronic diseases in the Chinese population. For example, among 49 247 15 16 Chinese aged 15 to 69 years from the 2007 China Chronic Disease and Risk Factor Surveillance, 17 18 the prevalence of Forhaving zero, peer one, two and review at least three chronic only disease risk factors were 9.1%, 19 20 33.9%, 32.4% and 24.6%, respectively.37 Also, other regional studies have examined the 21 22 clustering of some specific chronic diseases in residents. Hong et al. reported that 30.1% and 23 24 35.2% of the Nanjing population presented one and at least two cardiovascular diseases (CVD) 25 26 risk factors.23 Conversely, a much higher rate of CVD risk factors clustering was noticed by Ni et 27 28 al. in Shenzhen City.24 29 30 31 In our present study, it was observed that 0.1%, 1.3%, 6.4%, 22.6% and 69.7% of participants 32 33 had zero, one, two, three and at least four modified UDTC risk factors, respectively, among 34 35 residents aged 40-69 years. The tendency of modified UDTC risk factors clustering was prevalent

36 http://bmjopen.bmj.com/ considerably in Yangzhong City. Different estimates of the risk factors clustering for UDTC were 37 38 found in the literature. He et al. 2019 showed that among residents aged 40-69 years in UDTC 39 40 high-risk areas, 33.08%, 35.99%, 16.76% and 11.93% of participants had one, two, three and at 41 42 least four oesophageal cancer risk factors, respectively.19 In another case-control study of 2 266 43

44 Chinese adults, 32.5% and 41.1% of the participants presented three and four or more risk factors, on September 27, 2021 by guest. Protected copyright. 45 46 respectively, for oesophageal or stomach cancer.18 Compared with these two studies mentioned 47 48 above,18 19 a much higher clustering of risk factors was noticed in our present study. The variations 49 50 could change likely due to the difference of diagnostic criteria, the number and kind of risk factors 51 52 included in the research and the age group of the participants. Overall, there are some other 53 54 nations, and worldwide studies on the clustering of some common or specific chronic diseases risk 55 56 factors. Still, the study on modifiable UDTC risk factors is limited. 57 58 59 The factors found to be associated with modifiable UDTC risk factors clustering included 60 13

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 35

1 2 3 gender, age, marriage status, education, annual family income and household size. We found the 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 prevalence of modifiable UDTC risk factors clustering was lower in women compared with men,

7 19 23 37 8 which was consistent with findings from other settings. The possible reason could be 9 Chinese men are less aware of self-protect for chronic diseases and have worse health-seeking 10 11 behaviour, and may also attend more social occasions, tend to consume more tobacco/cigarette 12 13 high-salt, high-fat and high-calorie food compared with women.23 31 14 15 16 In addition, this study revealed that the prevalence of modifiable UDTC risk factors 17 18 clustering was increasedFor with peer age, which review was consistent with only previous studies.19 23 24 37 Studies 19 20 showed that in Nanbu, China, the clustering of oesophageal cancer risk factors increased with age 21 22 may attribute to the lower level of awareness, practice and willingness for health among the 23 24 elder.19 Meanwhile, it is also showed in table 5 that being over 65 protects against having more 25 26 than four risk factors. A possible reason for this difference is as follows: with the ageing of the 27 28 body and the deterioration of organ function, an elder individual possesses a higher risk of health 29 30 disorder and has a greater demand for medical care. As a result, this creates more opportunities to 31 32 get diagnosed with some health screening, including UDTC. Correspondingly, the elderly have 33 34 more chance to get a healthy education from physicians than the younger.38 Our study also showed 35 that single participants had more prevalence of modifiable UDTC risk factors clustering compared 36 http://bmjopen.bmj.com/ 37 19 38 with participants who were currently married, following a previous study. A possible 39 40 explanation is that being a single older resident comes with its own economic and emotional

41 39 40 42 challenges, which may contribute to the unhealthy habit of lifestyle and diet. 43

44 Our study demonstrated that the level of socioeconomic status (SES, education, annual on September 27, 2021 by guest. Protected copyright. 45 46 family income) was positively associated with modifiable UDTC risk factors clustering, which 47 48 was inconsistent with other reports.19 23 37 Residents with a higher level of SES are more aware of 49 50 control and prevention of chronic disease and have better health-seeking behaviour compared with 51 52 those with a lower level of SES.19 23 41 Moreover, the poor or lower education participants may 53 54 have relatively more inaccessibility and unaffordability to medical services.38 This paradox may 55 56 be due to most of the participants enrolled in our study were from rural areas, and their SES was 57 58 generally low. However, it may also imply that the higher income may contribute to unhealthy 59 60 lifestyles,42 and knowledge alone may not be sufficient to change unhealthy lifestyles. Therefore, 14

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 35 BMJ Open

1 2 3 the level of education and income are two of the essential SES factors for modifiable UDTC risk 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 factors clustering. 7 8 It is, however, important to note that participants who have more than 6 family members had 9 10 a lower prevalence of modifiable UDTC risk factors clustering compared with those who have less 11 12 than 3 ones. Changes in household size are bound to affect the adjustment of the family diet. As 13 14 the household size increase, it is more likely to increase dietary diversity (e.g. fruits, vegetables 15 16 and milk) every day.43 Besides, the affection, information and economic support among family 17 18 members will alsoFor increase peersignificantly, reviewwhich can adjust and only correct the unhealthy lifestyle of 19 20 individuals.44 45 21 22 23 Our study explored that the prevalence and influencing factors of modifiable UDTC risk 24 25 factors clustering in the UDTC high-risk area, Yangzhong City, which was based on the 26 27 community-based project for UDTC screening with large sample size. Additionally, the physical 28 29 measurement and the data collection implemented by trained interviewers strictly according to 30 31 standard protocol and instrument, which increase the validity of our results. In order to reduce the 32 33 prevalence and clustering rate of UTDC risk factors, the screening teams should focus on 34 35 individuals with clustering of risk factors in screening and improve their unhealthy lifestyles

36 http://bmjopen.bmj.com/ continually through a range of methods such as post-screening health education, personalized 37 38 interventions, and disease follow-up. The social impact of screening should be expanded to 39 40 improve the compliance of high-risk groups, thereby increasing the output of screening health 41 42 benefits. Meanwhile, the government should also help high-risk groups (especially the older and 43

44 male groups) to improve their health literacy and awareness of UTDC prevention through on September 27, 2021 by guest. Protected copyright. 45 46 diversified education, motivation, and publicity methods, such as health education, health talks, 47 48 and mass media campaigns. By guiding the culture of smoking, drinking, and other food culture, 49 50 promote high-cultural groups to transform their cognitive and economic advantages into 51 52 advantages in UTDC prevention and health care, and effectively change unhealthy habits. Besides, 53 54 the government should focus on single and residents with small household size in the high-risk 55 56 groups in the process of health education and the development of prevention strategies. The 57 58 findings may also provide the reference for departments in charge of the prevention and control of 59 60 UDTC in Yangzhong City, Jiangsu province, and relevant departments in other UDTC high-risk 15

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 35

1 2 3 areas (e.g. Linzhou, Feicheng, Yanting). 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 There were also several possible limitations to our study. Firstly, a cross-section study cannot 7 8 exam the causality or temporal relationship between the clustering of modifiable UDTC risk 9 10 factors and its influencing factors. Secondly, the modifiable UDTC risk factors included in our 11 12 study were self-reported by participants, which may contribute to recall and reporting bias, except 13 14 BMI. The results of our study were from Yangzhong City only, and cannot be generalized to the 15 16 other high-risk areas and the whole of southeast China. Additionally, the study response rate was 17 18 relatively low, particularlyFor amongpeer males, reviewwhich may affect theonly representativeness of the results. 19 20 Finally, our study only focused on the 8 modifiable UDTC risk factors, but there are far more than 21 22 eight risk factors for UDTC. Hence, further studies are needed. 23 24 25 Conclusion 26 27 28 In summary, this cross-sectional study shows that a large proportion of the participants had 29 30 modifiable UDTC risk factors with a tendency of clustering in Yangzhong City. Our analyses 31 indicate that men, younger adults, single adults, participants with higher levels of SES or smaller 32 33 household size are susceptible to modifiable UDTC risk factors clustering. Policies to prevent 34 35 UDTC have already been developed in the strategic plan and operational plan, however, the

36 http://bmjopen.bmj.com/ 37 accuracy and validity of implementing the undertaken policies are still insufficient. Consequently, 38 39 extra attention is required to pay on these high-risk groups during the progress of screening. 40 41 Relative departments also need to make effective public health programs targeting modifiable 42 43 UDTC risk factors that aim to decrease UDTC risk factors clustering in high-risk groups from

44 on September 27, 2021 by guest. Protected copyright. 45 high-risk areas of UDTC. 46 47 48 Acknowledgements We would like to thank all participants of the present study. We are 49 50 also grateful to the Upper Digestive Tract Cancer Early Diagnosis and Treatment Program. 51 52 53 54 55 Contribution XF and RC are joint first authors. ZLH, DFQ, WQW, GQW, JYZ, JJW, GS 56 57 and XW contributed to the study design. ZLH, QZ, AWS, TQS performed the survey and 58 59 collected study data. XF and RC wrote the manuscript and ZLH, DFQ, WQW, JYZ, JJW, GS, 60 16

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 35 BMJ Open

1 2 3 XW were responsible for manuscript revision. 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 7 8 9 Funding This study was supported by the National key research and development program 10 11 (Grant NO. 2016YFC0901400; 2016YFC1302800); National Natural Science Foundation of 12 13 China (Grant NO. 81974493); and National Key Technology R&D Program (Grant NO. 14 15 2006BAI02A05). 16 17 18 For peer review only 19 20 21 Competing interests None declared. 22 23 24 25 26 Patient consent Obtained. 27 28 29 30 31 32 Ethics approval This study was approved by the academic and ethical committee of the 33 34 cancer hospital of Chinese academy of medical sciences. 35

36 http://bmjopen.bmj.com/ 37 38 39 Provenance and peer review Not commissioned; externally reviewed. 40 41 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 Data sharing statement No additional data are available. 46 47 48 49 50 Open access This is an Open Access article distributed in accordance with the Creative 51 52 Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to 53 54 distribute, remix, adapt, build upon this work non-commercially and license their derivative works 55 56 on different terms, provided the original work is properly cited and the use is non-commercial. 57 58 See:http://creativecommons.org/licenses/by-nc/4.0/ 59 60 17

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 35

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 7 8 9 10 11 12 13 14 REFERENCES 15 16 [1] World Health Organization. World health statistics 2018: monitoring health for the SDGs, 17 18 sustainable For development peer goals. review Geneva: Worldonly Health Organization, 19 2018.https://www.who.int/gho/publications/world_health_statistics/2018/en/. 20 21 22 [2] Bray F, Ferlay J, Soerjomataram I, et al. Global cancer statistics 2018: GLOBOCAN 23 estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J 24 25 Clin,2018,68(6):394-424. 26 27 [3] Zheng RS, Sun KX, Zhang SW, et al. Report of cancer epidemiology in China, 2015. Chin J 28 29 Oncol,2019,41(1):19-28.In Chinese. 30 31 [4] Chen WQ, Zheng RS, Baade PD, et al. Cancer statistics in China, 2015. CA Cancer J 32 Clin,2016,66(2):115-32. 33 34 35 [5] Ajani JA, Bentrem DJ, Besh S, et al. Gastric cancer, version 2.2013: featured updates to the

36 NCCN Guidelines. J Natl Compr Canc Netw,2013,11(5):531-46. http://bmjopen.bmj.com/ 37 38 39 [6] Ciocirlan M, Lapalus MG, Hervieu V, et al. Endoscopic mucosal resection for squamous 40 premalignant and early malignant lesions of the esophagus. Endoscopy,2007,39(1):24-9. 41 42 [7] Ma D, Yang F, Liao Z, et al. Expert opinion on early screening and endoscopic diagnosis and 43

44 treatment esophageal cancer in China (2014, Beijing). China Prac Med,2015,4:320-337.In on September 27, 2021 by guest. Protected copyright. 45 Chinese. 46 47 48 [8] Chen R, Ma S, Guan C, et al. The National Cohort of Esophageal Cancer-Prospective Cohort 49 Study of Esophageal Cancer and Precancerous Lesions based on High-Risk Population in 50 51 China (NCEC-HRP): study protocol. BMJ Open,2019,9(4):e027360. 52 53 [9] Chen MJ, Chiou YY, Wu DC, et al. Lifestyle habits and gastric cancer in a hospital-based 54 case-control study in Taiwan. Am J Gastroenterol,2000,95(11):3242-9. 55 56 57 [10] Somi MH, Mousavi SM, Naghashi S, et al. Is there any relationship between food habits in 58 the last two decades and gastric cancer in North-Western Iran?. Asian Pac J Cancer Prev. 59 60 2015,16(1):283-90. 18

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 35 BMJ Open

1 2 3 [11] Navarro-Silvera SA, Mayne ST, Risch HA, et al. Principal component analysis of dietary and

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 lifestyle patterns in relation to risk of subtypes of esophageal and gastric cancer. Ann 6 Epidemiol,2011,21(7):543-50. 7 8 9 [12] Andrici J, Eslick GD. Hot Food and Beverage Consumption and the Risk of Esophageal 10 Cancer: A Meta-Analysis. Am J Prev Med.2015,49(6):952–960. 11 12 13 [13] Gupta B, Kumar N, Johnson NW. Relationship of Lifetime Exposure to Tobacco, Alcohol 14 and Second Hand Tobacco Smoke with Upper aero-digestive tract cancers in India: a 15 16 Case-Control Study with a Life-Course Perspective. Asian Pac J Cancer 17 Prev,2017,18(2):347-56. 18 For peer review only 19 [14] Chen WQ, ZUO TT. Initial effect achievement of battles on upper digestive tract cancer in 20 21 China. Chin J Prev Med,2017, 51(5):378-80.In Chinese. 22 23 [15] Luan DC, Li SJ, Li H, et al. Change trends in health behaviors among residents in Liaoning 24 25 province,1991-2006. Chin J Public Health, 2013, 29(10):1509-1511.In Chinese. 26 27 [16] McCormack VA, Boffetta P. Today's lifestyles, tomorrow's cancers: trends in lifestyle risk 28 factors for cancer in low- and middle-income countries. Ann Oncol. 2011,22(11):2349–2357. 29 30 31 [17] Li FX, Robson PJ, Chen Y, et al. Prevalence, trend, and sociodemographic association of five 32 modifiable lifestyle risk factors for cancer in Alberta and Canada. Cancer Causes 33 34 Control,2009,20(3):395-407. 35

36 [18] Gu XP, Wang YC, Zhi HK, et al. Risk factors of esophageal and stomach cancer and their http://bmjopen.bmj.com/ 37 38 clustering in Dafeng municipality ： a case-control study. Chin J Public Health,2016, 39 32(10):1406-1409.In Chinese. 40 41 [19] He Q, Jing YH, Huang HR, et al. Prevalence and clustering of esophageal cancer-related risk 42 43 factors among rural residents in Nanbu County, Sichuan Province. Chin H Cancer Prev

44 on September 27, 2021 by guest. Protected copyright. Treat,2019,26(22):1675-1680.In Chinese. 45 46 47 [20] Tong HY, Zhang MM, Zhang HY. Analysis on the epidemiology trend and disease burden of 48 esophagus cancer in Yangzhong from 2004 to 2015. Modern Prev 49 50 Med,2016,43(20):3665-3668,3687.In Chinese. 51 52 [21] Tong HY, Zhang MM, Sun LP, et al. Quantitative study on death caused by main chronic 53 54 diseases in Yangzhong city. Jiangsu J Prev Med,2017,28(5):502-504,508. In Chinese. 55 56 [22] DONG ZW. China's cancer screening and early diagnosis and treatment technology 57 program. Beijing: People's medical publishing house, 2009.In Chinese. 58 59 60 [23] Hong X, Ye Q, He J, et al. Prevalence and clustering of cardiovascular risk factors: a 19

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 22 of 35

1 2 3 cross-sectional survey among Nanjing adults in China. BMJ Open,2018,8(6):e020530.

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 [24] Ni W, Weng RX, Yuan X, et al. Clustering of cardiovascular disease biological risk factors 7 among older adults in Shenzhen City, China: a cross-sectional study. BMJ 8 9 Open,2019,9(3):e024336. 10 11 [25] Department of disease control, ministry of health, PRC. Guidelines for the prevention and 12 13 control of overweight and obesity in Chinese adults. Beijing: people's medical publishing 14 house, 2006:2-4.In Chinese. 15 16 [26] Chinese nutrition society. Dietary guidelines for Chinese residents (2016). Beijing: people's 17 18 medical publishingFor house, peer 2016In Chinese. review only 19 20 [27] Howitt C, Hambleton IR, Rose AM, et al. Social distribution of diabetes, hypertension and 21 22 related risk factors in Barbados: a cross-sectional study. BMJ Open,2015,5(12):e008869. 23 24 [28] Zaman MM, Bhuiyan MR, Karim MN, et al. Clustering of noncommunicable diseases risk 25 26 factors in Bangladeshi adults: An analysis of STEPS survey 2013. BMC Public 27 Health.2015,15:659. 28 29 [29] Xu F, Yin XM, Zhang M, et al. Family average income and body mass index above the 30 31 healthy weight range among urban and rural residents in regional Mainland China. Public 32 Health Nutr,2005,8(1):47-51. 33 34 35 [30] National Commission of Health Bureau of disease control and Prevention. Report on China

36 national nutrition and chronic disease survey (2015). Beijing: People’s Medical Publishing http://bmjopen.bmj.com/ 37 38 House, 2015,11:33-50.In Chinese. 39 40 [31] Wang R, Zhang P, Gao C, et al. Prevalence of overweight and obesity and some associated 41 factors among adult residents of northeast China: a cross-sectional study. BMJ Open 42 43 2016,6:e010828.

44 on September 27, 2021 by guest. Protected copyright. 45 [32] Howitt C, Hambleton IR, Rose AMC, et al. Social distribution of diabetes, hypertension and 46 47 related risk factors in Barbados: a cross-sectional study. BMJ Open.2015,5:e008869. 48 49 [33] Millwood IY, Walters RG, Mei XW, et al. Conventional and genetic evidence on alcohol and 50 51 vascular disease aetiology: a prospective study of 500 000 men and women in China. Lancet. 52 2019,393(10183):1831-42. 53 54 [34] Msambichaka B, Eze IC, Abdul R, et al. Insufficient Fruit and Vegetable Intake in a Low- 55 56 and Middle-Income Setting: A Population-Based Survey in Semi-Urban Tanzania.Nutrients. 57 2018,10(2):222. 58 59 60 [35] Guo YL, Tan XD, Liu XZ, et al. Fruit and vegetable intake of adults and its influencing 20

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 23 of 35 BMJ Open

1 2 3 factors in some cities of Hubei province . J of Pub health and Prev Med,2016,27(5):82-85.In

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 Chinese. 6 7 [36] Wen JB, Sun ZM, Miao DD, et al. Influencing factors about early cancer of upper-digestive 8 9 tract among high-risk population in Huai’an City of Jiangsu Province. China 10 cancer,2019,28(10):749-756.In Chinese. 11 12 13 [37] Li Y, Zhang M, Jiang Y, Wu F. Co-variations and clustering of chronic disease behavioral 14 risk factors in China: China Chronic Disease and Risk Factor Surveillance, 2007. PLoS One. 15 16 2012,7(3):e33881. 17 18 [38] Ahmed S, TariqujjamanFor peer M, Rahman MA,review et al. Inequalities only in the prevalence of undiagnosed 19 hypertension among Bangladeshi adults: evidence from a nationwide survey. Int J Equity 20 21 Health. 2019,18(1):33. 22 23 [39] Floud S, Balkwill A, Canoy D, et al. Marital status and ischemic heart disease incidence and 24 25 mortality in women: a large prospective study. BMC Med. 2014,12:42. 26 27 [40] Cao Z, Wang R, Cheng Y, et al. Adherence to a healthy lifestyle counteracts the negative 28 effects of risk factors on all-cause mortality in the oldest-old. Aging (Albany NY). 29 30 2019,11(18):7605-7619. 31 32 [41] Prom-Wormley EC, Clifford JS, Bourdon JL, et al. Developing community-based health 33 34 education strategies with family history: Assessing the association between community 35 resident family history and interest in health education. Soc Sci Med. 2019. [Online ahead of

36 http://bmjopen.bmj.com/ 37 print]Doi:10.1016/j.socscimed.2019.02.011. 38 39 [42] Yang F, Qian D, Liu X, et al. Socioeconomic disparities in prevalence, awareness, treatment, 40 41 and control of hypertension over the life course in China. Int J Equity Health, 2017, 42 16(1):100. 43

44 on September 27, 2021 by guest. Protected copyright. [43] Workicho A, Belachew T, Feyissa GT, et al. Household dietary diversity and Animal Source 45 46 Food consumption in Ethiopia: evidence from the 2011 Welfare Monitoring Survey. BMC 47 Public Health. 2016,16(1):1192. 48 49 50 [44] Bot SD, Mackenbach JD, Nijpels G, et al. Association between Social Network 51 Characteristics and Lifestyle Behaviours in Adults at Risk of Diabetes and Cardiovascular 52 53 Disease. PLoS One. 2016,11(10):e0165041. 54 55 [45] Verheijden MW, Bakx JC, van Weel C, et al. Role of social support in lifestyle-focused 56 57 weight management interventions. Eur J Clin Nutr. 2005:59 Suppl 1:S179-S186. 58 59 60 21

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 24 of 35

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 22

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Page 25 of 35 BMJ Open

1 2 3 4 5 Table legends 6 7 8 9 10 Table 1 Socio-demographic and socio-economic characteristics of participants in Yangzhong City, China 11 12 Category ForTotal(n=19213) peer Men(n=8268)reviewWomen(n=10945) only t/ χ² value P value 13 14 15 Number, n(%) 19213(100.0) 8268(43.0) 10945(57.0) 16 http://bmjopen.bmj.com/ 17 Age, years, mean(SD) 53.2±7.8 53.5±7.8 52.9±7.8 5.61 <0.001 18 19 Age group, n(%) 33.29 <0.001 20 21 22 40-44 3142(16.4) 1270(15.4) 1872(17.1) 23 on September 27, 2021 by guest. Protected copyright. 24 45-49 3937(20.5) 1642(19.9) 2295(21.0) 25 26 27 50-54 3743(19.5) 1560(18.9) 2183.0(19.9) 28 29 55-59 3571(18.6) 1582(19.1) 1989(18.2) 30 31 60-64 3088(16.1) 1416(17.1) 1672(15.3) 32 33 34 65-69 1732(9.0) 798(9.7) 934(8.5) 35 36 Marrige, n(%) 93.41 <0.001 37 38 39 40 41 23 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

BMJ Open Page 26 of 35

1 2 3 4 5 Single 203(1.1) 137(1.7) 66(0.6) 6 7 Currently married 18285(95.2) 7907(95.6) 10378(94.8) 8 9 10 Divorced, widowed or separated 725(3.8) 224(2.7) 501(4.6) 11 12 Education, n(%) For peer review only 886.879 <0.001 13 14 15 No institutional education 1594(8.3) 228(2.8) 1366(12.5) 16 http://bmjopen.bmj.com/ 17 Primary school 8510(44.3) 3436(41.6) 5074(46.4) 18 19 Junior high school 7591(39.5) 3647(44.1) 3944(36.0) 20 21 22 Senior high school and higher 1518(7.9) 957(11.6) 561(5.1) 23 on September 27, 2021 by guest. Protected copyright. 24 Annual family income, n(%) 0.04 0.981 25 26 27 lower 8585(44.7) 3689(44.6) 4896(44.7) 28 29 Middle 5420(28.2) 2338(28.3) 3082(28.2) 30 31 Higher 5208(27.1) 2241(27.1) 2967(27.1) 32 33 34 Household size, n, mean(SD) 4.0±1.4 4.0±1.4 4.0±1.4 -3.17 0.002 35 36 BMI,kg/m2, mean(SD) 23.8±3.0 23.9±3.0 23.7±3.0 3.47 0.001 37 38 39 40 41 24 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Page 27 of 35 BMJ Open

1 2 3 4 5 6 7 8 9 10 11 12 Table 2 Prevalence of modifiable UDTCFor risk factors inpeer Yangzhong City adultsreview aged 40-69 only 13 14 15 Factor Total (n=19213) Men (n=8268) Women (n=10945) χ² value P value 16 http://bmjopen.bmj.com/ 17 18 BMI, n(%) 2.62 0.106 19 20 Overweight or obesity 8695(45.3) 3797(45.9) 4898(44.8) 21 22 23 Normal weight or underweight 10518(54.7) 4471(54.1) 6047(55.2) 24 on September 27, 2021 by guest. Protected copyright. 25 Current smoking, n(%) 6093.25 <0.001 26 27 No 14589(75.9) 3988(48.2) 10601(96.9) 28 29 30 Yes 4624(24.1) 4280(51.8) 344(3.1) 31 32 Excessive alcohol consumption, n(%) 3197.31 <0.001 33 34 35 No 16109(83.8) 5504(66.6) 10605(96.9) 36 37 Yes 3104(16.2) 2764(33.4) 340(3.1) 38 39 40 41 25 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

BMJ Open Page 28 of 35

1 2 3 4 5 insufficient vegetables intake, n(%) 0.32 0.572 6 7 No 6519(33.9) 2787(33.7) 3732(34.1) 8 9 10 Yes 12694(66.1) 5481(66.3) 7213(65.9) 11 12 insufficient fruit intake, n(%) For peer review only 0.10 0.749 13 14 15 No 1055(5.5) 459(5.6) 596(5.4) 16 http://bmjopen.bmj.com/ 17 Yes 18158(94.5) 7809(94.4) 10349(94.6) 18 19 Pickled food consumption, n(%) 8.38 <0.05 20 21 22 No 6138(31.9) 2734(33.1) 3404(31.1) 23 on September 27, 2021 by guest. Protected copyright. 24 Yes 13075(68.1) 5534(66.9) 7541(68.9) 25 26 27 Fried food consumption, n(%) 1.00 0.318 28 29 No 12293(64.0) 5323(64.4) 6970(63.7) 30 31 Yes 6920(36.0) 2945(35.6) 3975(36.3) 32 33 34 Hot food consumption, n(%) 27.10 <0.001 35 36 No 2221(11.6) 1070(12.9) 1151(10.5) 37 38 39 40 41 26 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Page 29 of 35 BMJ Open

1 2 3 4 5 Yes 16992(88.4) 7198(87.1) 9794(89.5) 6 7 8 9 10 11 12 For peer review only 13 14 15 Table 3 Prevalence of modifiable UDTC risk factors by relevant characters in Yangzhong City adults aged 40-69 16 http://bmjopen.bmj.com/ 17 Category Overweight or obesity Smoking Drinking Vegetables Fruit Pickled food Fried food Hot food 18 19 Age group,years, n(%) 20 21 22 40-44 1492(47.5) 533(17.0) 348(11.1) 1888(60.1) 2902(92.4) 2033(64.7) 782(24.9) 2971(94.6) 23 on September 27, 2021 by guest. Protected copyright. 24 45-49 1953(49.6) 827(21.0) 540(13.7) 2648(67.3) 3735(94.9) 2594(65.9) 1451(36.9) 3894(98.9) 25 26 27 50-54 1729(46.2) 1068(28.5) 723(19.3) 2506(67.0) 3538(94.5) 2669(71.3) 1325(35.4) 3723(99.5) 28 29 55-59 1607(45.0) 987(27.6) 695(19.5) 2314(64.8) 3380(94.7) 2366(66.3) 1312(36.7) 3464(97.0) 30 31 60-64 1271(41.2) 796(25.8) 537(17.4) 2131(69.0) 2949(95.5) 2158(69.9) 1285(41.6) 2343(75.9) 32 33 34 65-69 643(37.1) 413(23.8) 261(15.1) 1207(69.7) 1654(95.5) 1255(72.5) 765(44.2) 597(34.5) 35 36 P value <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 37 38 39 40 41 27 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

BMJ Open Page 30 of 35

1 2 3 4 5 Marrige, n(%) 6 7 Single 77(37.9) 73(36.0) 43(21.2) 177(87.2) 200(98.5) 169(83.3) 143(70.4) 186(91.6) 8 9 10 Currently married 8334(45.6) 4439(24.3) 2979(16.3) 12087(66.1) 17253(94.4) 12424(67.9) 6518(35.6) 16304(89.2) 11 12 Divorced, widowed or separated For284(39.2) peer112(15.4) review82(11.3) 430(59.3) only705(97.2) 482(66.5) 259(35.7) 502(69.2) 13 14 15 P value <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 16 http://bmjopen.bmj.com/ 17 Education, n(%) 18 19 No institutional education 692(43.4) 75(4.7) 76(4.8) 781(49.0) 1479(92.8) 1286(80.7) 675(42.3) 1198(75.2) 20 21 22 Primary school 3699(43.5) 2030(23.9) 1328(15.6) 5898(69.3) 8149(95.8) 4785(56.2) 2205(25.9) 7194(84.5) 23 on September 27, 2021 by guest. Protected copyright. 24 Junior high school 3521(46.4) 2045(26.9) 1373(18.1) 4878(64.3) 7087(93.4) 5819(76.7) 3146(41.4) 7187(94.7) 25 26 27 Senior high school and higher 783(51.6) 474(31.2) 327(21.5) 1137(74.9) 1443(95.1) 1185(78.1) 894(58.9) 1413(93.1) 28 29 P value <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 30 31 Annual family income, n(%) 32 33 34 lower 3711(43.2) 2128(24.8) 1551(18.1) 4844(56.4) 8234(95.9) 5535(64.5) 1300(15.1) 7168(83.5) 35 36 Middle 2458(45.4) 1303(24.0) 857(15.8) 3566(65.8) 4953(91.4) 2923(53.9) 1521(28.1) 4876(90.0) 37 38 39 40 41 28 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Page 31 of 35 BMJ Open

1 2 3 4 5 Higher 2526(48.5) 1193(22.9) 696(13.4) 4284(82.3) 4971(95.4) 4617(88.7) 4099(78.7) 4948(95.0) 6 7 P value <0.001 <0.05 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 8 9 10 Household size, n(%) 11 12 0-3 For3583(45.6) peer1916(24.4) review1381(17.6) 5137(65.4) only7384(94.1) 5512(70.2) 2989(38.1) 7051(89.8) 13 14 15 4-5 4285(44.6) 2285(23.8) 1428(14.8) 6574(68.4) 9159(95.2) 6216(64.6) 3153(32.8) 8467(88.0) 16 http://bmjopen.bmj.com/ 17 ≥6 827(47.4) 423(24.2) 295(16.9) 983(56.3) 1615(92.5) 1347(77.1) 778(44.6) 1474(84.4) 18 19 P value 0.064 0.601 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 20 21 22 Notes: Smoking, current smoking; Drinking, excessive alcohol consumption; Vegetables, insufficient vegetables intake, Fruit, insufficient fruit intake; Pickled food, 23

the consumption of pickled food, Fried food, the consumption of fried food; Hot food, the consumption of hot food on September 27, 2021 by guest. Protected copyright. 24 25 26 27 28 29 30 31 32 33 34 35 36 37 Table 4 The different number and clustering of modifiable UDTC risk factors in residents by relevant characters in Yangzhong City adults aged 40-69 38 39 40 41 29 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

BMJ Open Page 32 of 35

1 2 3 4 5 Category None(0) single(1) Two(2) Three(3) Non-clustering(≤3) Clustering(≥4) χ² value P value 6 7 8 Total 14(0.1) 244(1.3) 1235(6.4) 4336(22.6) 5829(30.3) 13384(69.7) 9 10 Gender, n(%) 632.397 <0.001 11 12 men 2(0.0)For64(0.8) peer356(4.3) review1293(15.6) 1715(20.7) only 6553(79.3) 13 14 15 women 12(0.1) 180(1.6) 879(8.0) 3043(27.8) 4114(37.6) 6831(62.4) 16 http://bmjopen.bmj.com/ 17 Age group, years, n(%) 373.748 <0.001 18 19 40-44 1(0.0) 66(2.1) 193(6.1) 891(28.4) 1151(36.6) 1991(63.4) 20 21 22 45-49 0(0.0) 26(0.7) 165(4.2) 857(21.8) 1048(26.6) 2889(73.4) 23 on September 27, 2021 by guest. Protected copyright. 24 50-54 0(0.0) 15(.4) 112(3.0) 759(20.3) 886(23.7) 2857(76.3) 25 26 27 55-59 0(0.0) 48(1.3) 175(4.9) 724(20.3) 947(26.5) 2624(73.5) 28 29 60-64 0(0.0) 33(1.1) 273(8.8) 712(23.1) 1018(33.0) 2070(67.0) 30 31 65-69 13(0.8) 56(3.2) 317(18.3) 393(22.7) 779(45.0) 953(55.0) 32 33 34 Marrige, n(%) 67.346 <0.001 35 36 Single 0(0.0) 0(0.0) 9(4.4) 16(7.9) 25(12.3) 178(87.7) 37 38 39 40 41 30 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Page 33 of 35 BMJ Open

1 2 3 4 5 Currently married 12(0.1) 201(1.1) 1130(6.2) 4167(22.8) 5510(30.1) 12775(69.9) 6 7 Divorced, widowed or separated 2(0.3) 43(5.9) 96(13.2) 153(21.1) 294(40.6) 431(59.4) 8 9 10 Education, n(%) 417.766 <0.001 11 12 No institutional education 4(0.3)For56(3.5) peer241(15.1) review379(23.8) 680(42.7) only 914(57.3) 13 14 15 Primary school 8(0.1) 71(0.8) 503(5.9) 2396(28.2) 2978(35.0) 5532(65.0) 16 http://bmjopen.bmj.com/ 17 Junior high school 2(0.0) 107(1.4) 428(5.6) 1367(18.0) 1904(25.1) 5687(74.9) 18 19 Senior high school and higher 0(0.0) 10(0.7) 63(4.2) 194(12.8) 267(17.6) 1251(82.4) 20 21 22 Annual family income, n(%) 1078.75 <0.001 23 on September 27, 2021 by guest. Protected copyright. 24 lower 6(0.1) 102(1.2) 588(6.8) 2500(29.1) 3196(37.2) 5389(62.8) 25 26 27 Middle 5(0.1) 105(1.9) 470(8.7) 1403(25.9) 1983(36.6) 3437(63.4) 28 29 Higher 3(0.1) 37(0.7) 177(3.4) 433(8.3) 650(12.5) 4558(87.5) 30 31 Family member, n(%) 17.913 <0.001 32 33 34 0-3 5(0.1) 165(2.1) 438(5.6) 1642(20.9) 2250(28.7) 5600(71.3) 35 36 4-5 7(0.1) 68(0.7) 627(6.5) 2318(24.1) 3020(31.4) 6597(68.6) 37 38 39 40 41 31 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

BMJ Open Page 34 of 35

1 2 3 4 5 ≥6 2(0.1) 11(0.6) 170(9.7) 376(21.5) 559(32.0) 1187(68.0) 6 7 8 9 10 11 12 For peer review only 13 14 15 Table 5 The multivariable logistic regression analysis of modifiable UDTC risk factors clustering 16 http://bmjopen.bmj.com/ 17 Category Wald P value β SE OR 95% CI 18 19 Gender, n(%) 20 21 22 women - - - - 1.000 - 23 on September 27, 2021 by guest. Protected copyright. 24 men 533.13 <0.001 0.834 0.036 2.302 2.145 to 2.471 25 26 27 Age group, years, n(%) 28 29 40-44 - - - - 1.000 - 30 31 45-49 48.302 <0.001 0.381 0.055 1.464 1.315 to 1.631 32 33 34 50-54 153.543 <0.001 0.715 0.058 2.044 1.825 to 2.288 35 36 55-59 143.780 <0.001 0.710 0.059 2.033 1.811 to 2.283 37 38 39 40 41 32 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Page 35 of 35 BMJ Open

1 2 3 4 5 60-64 32.963 <0.001 0.351 0.061 1.421 1.260 to 1.602 6 7 65-69 10.685 <0.001 -0.229 0.070 0.796 0.694 to 0.913 8 9 10 Marrige, n(%) 11 12 Single For peer- review- -only- 1.000 - 13 14 15 Currently married 9.054 0.003 -0.671 0.223 0.511 0.330 to 0.792 16 http://bmjopen.bmj.com/ 17 Divorced, widowed or separated 10.813 0.001 -0.777 0.236 0.460 0.289 to 0.731 18 19 Education, n(%) 20 21 22 No institutional education - - - - 1.000 - 23 on September 27, 2021 by guest. Protected copyright. 24 Primary school 17.448 <0.001 0.256 0.061 1.291 1.145 to 1.456 25 26 27 Junior high school 51.436 <0.001 0.485 0.068 1.624 1.423 to 1.855 28 29 Senior high school and higher 40.562 <0.001 0.604 0.095 1.829 1.519 to 2.202 30 31 Annual family income, n(%) 32 33 34 Higher - - - - 1.000 - 35 36 Middle 874.464 <0.001 -1.524 0.052 0.218 0.197 to 0.241 37 38 39 40 41 33 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

BMJ Open Page 36 of 35

1 2 3 4 5 lower 799.154 <0.001 -1.502 0.053 0.223 0.201 to 0.247 6 7 Household size, n(%) 8 9 10 0-3 - - - - 1.000 - 11 12 4-5 For peer82.359 review0.429 -0.334 only 0.037 0.716 0.666 to 0.770 13 14 15 ≥6 64.364 <0.001 -0.514 0.064 0.598 0.527 to 0.678 16 http://bmjopen.bmj.com/ 17 Constant 64.447 <0.001 1.878 0.234 - - 18 19 20 21 22 23 24 on September 27, 2021 by guest. Protected copyright. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 34 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Page 37 of 35 BMJ Open

1 2 3 4 5 6 7 8 9 10 11 12 For peer review only 13 14 15 16 http://bmjopen.bmj.com/ 17 18 19 20 21 22 23 24 on September 27, 2021 by guest. Protected copyright. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 35 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Prevalence and co-prevalence of modifiable risk factors for upper digestive tract cancer among residents aged 40 to 69 years in Yangzhong City, China: a cross-sectional study ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2020-042006.R2

Article Type: Original research

Date Submitted by the 18-Feb-2021 Author:

Complete List of Authors: Feng, xiang; People’s Hospital of Yangzhong City, Institute of tumour prevention and control Hua, Zhao-lai; People’s Hospital of Yangzhong City, Institute of tumour prevention and control Zhou, Qin; People’s Hospital of Yangzhong City, Institute of tumour prevention and control Shi, Ai-wu; People’s Hospital of Yangzhong City, Institute of tumour prevention and control Song, Tong-qiu; People’s Hospital of Yangzhong City, Institute of tumour prevention and control Qian, Dongfu; Nanjing Medical University, School of Health Policy &

Management http://bmjopen.bmj.com/ Chen, Ru; Chinese Academy of Medical Sciences and Peking Union Medical College, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Cancer Registry Office Wang, Gui-qi; Chinese Academy of Medical Sciences and Peking Union Medical College, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Cancer Registry Office Wei, Wen-Qiang; Chinese Academy of Medical Sciences and Peking Union Medical College, National Cancer Center/National Clinical Research

Center for Cancer/Cancer Hospital, Cancer Registry Office on September 27, 2021 by guest. Protected copyright. Zhou, Jin-Yi; Jiangsu Province Center for Disease Control and Prevention, Depatment of Non-communicable Disease Prevention Wang, Jie-jun; Second Military Medical University, Changzheng Hospital, Department of Oncology Shao, Gang; 923rd Hospital of PLA, Department of Oncology Wang, Xi; 923rd Hospital of PLA, Department of Oncology

Primary Subject Epidemiology Heading:

Secondary Subject Heading: Public health, Epidemiology, Oncology

EPIDEMIOLOGY, Endoscopy < GASTROENTEROLOGY, Risk management Keywords: < HEALTH SERVICES ADMINISTRATION & MANAGEMENT, Epidemiology < ONCOLOGY

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39

1 2 3

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 35 BMJ Open

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 Prevalence and co-prevalence of modifiable 6 7 risk factors for upper digestive tract cancer 8 9 10 among residents aged 40 to 69 years in 11 12 Yangzhong City, China: a cross-sectional 13 14 15 study 16 17 Xiang Feng,1,2 Zhaolai Hua,1 Qin Zhou1, Aiwu Shi,1 Tongqiu Song,1 Dongfu Qian,2 Ru Chen,3 18 For peer review only 19 GuiQi Wang,3 Wenqiang Wei,3 Jinyi Zhou,4 Jiejun Wang,5 Gang Shao,6 Xi Wang6 20 21 22 23 1 Institute of tumour prevention and control, People’s Hospital of Yangzhong City, Yangzhong, 24 25 China 26 27 2 School of Health Policy & Management, Nanjing Medical University, Nanjing, China 28 29 30 3 Cancer Registry Office, National Cancer Center/National Clinical Research Center for 31 Cancer/Cancer Hospital, Chinese Academy of Medical Science and Peking Union Medical 32 33 College, Beijing, China 34 35 4 Department of Non-communicable Disease Prevention, Jiangsu Center for Disease Control and

36 http://bmjopen.bmj.com/ Prevention, Nanjing, China 37 38 39 5 Department of Oncology, Changzheng Hospital, Second Military Medical University, Shanghai, 40 China. 41 42 43 6 Department of Oncology, 903rd Hospital of PLA, Hangzhou, China.

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 Correspondence to Zhaolai Hua; [email protected]; Dr Wenqiang Wei; 48 49 [email protected] 50 51 52 53 ABSTRACT 54 55 56 Objectives To describe the prevalence of modifiable risk factors for upper digestive tract cancer 57 (UDTC) and its co-prevalence, and investigate relevant influencing factors of modifiable UDTC 58 59 risk factors co-prevalence among residents aged 40-69 years in Yangzhong City, China. 60 1

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 35

1 2 3 Design Cross-sectional study

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 Participants A total of 21 175 participants aged 40-69 years were enrolled in the study. 1 962 7 subjects were excluded due to missing age, marital status, or some other selected information. 8 9 Eventually, 19 213 participants were available for the present analysis. 10 11 Main outcomes measures Prevalence and co-prevalence of eight modifiable UDTC risk factors 12 13 (overweight or obesity, current smoking, excessive alcohol consumption, insufficient vegetables 14 intake, insufficient fruit intake, and the consumption of pickled, fried, and hot food) were 15 16 analyzed. 17 18 Results The prevalenceFor of overweight/obesity,peer review current smoking, only excessive alcohol consumption, 19 20 insufficient vegetables intake, insufficient fruit intake, and the consumption of pickled, fried, and 21 22 hot food in this study was 45.3%, 24.1%, 16.2%, 66.1%, 94.5%, 68.1%, 36.0%, and 88.4%, 23 24 respectively. Nearly all (99.9%) participants showed one or more UDTC risk factors, 98.6% of the 25 26 participants showed at least two risk factors, 92.2% of the participants had at least three risk 27 28 factors, and 69.7% of the participants had four or more risk factors. Multivariate logistic 29 30 regression analysis revealed that men, younger age, single, higher education, higher annual family 31 32 income and, smaller household size were more likely to present modifiable UDTC risk factors 33 34 co-prevalence. 35

36 http://bmjopen.bmj.com/ Conclusions The prevalence and co-prevalence of modifiable UDTC risk factors are high among 37 38 participants in Yangzhong City. Extra attention must be paid to these groups who are susceptible 39 40 to risk factors co-prevalence during screening progress. Relative departments also need to make 41 42 significant public health programs that aim to decrease modifiable UDTC risk factors 43

44 co-prevalence among residents aged 40-69 years from high-risk areas of UDTC. on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 50 Strengths and limitations of this study 51 52 This is the first study examining the prevalence and co-prevalence of modifiable upper digestive 53 54 tract cancer (UDTC) risk factors and investigating relevant influencing factors in Yangzhong City, 55 with large and representative residents aged 40-69 years from the Upper Digestive Tract Cancer 56 57 Early Diagnosis and Treatment (UDTCEDAT). 58 59 60 2

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 35 BMJ Open

1 2 3 Participants who volunteered to participate in our study are considered to be the high-risk group

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 for UDTC, which is significantly important for the prevention and control of UDTC in China, 6 where the morbidity and mortality of UDTC are increasing. 7 8 9 10 11 A cross-section study cannot exam the causality or temporal relationship. 12 13 14 15 16 The modifiable UDTC risk factors included in our study were self-reported by participants, which 17 may contribute to recall and reporting bias, except BMI. 18 For peer review only 19 20 21 22 We only focus on the eight modifiable UDTC risk factors. Still, there are far more than eight risk 23 factors for UDTC, which may underestimate the average number of modifiable UDTC risk factors 24 25 among participants. 26 27 28 29 INTRODUCTION 30 31 32 According to the WHO, more than 70% of the total deaths worldwide were related to 33 34 non-communicable diseases (NCDs) in 2016.1 Cancer is the second cause of NCDs, accounting 35

36 for 22% of total global deaths related to NCDs. Globally, in 2018, an estimated 572 034 http://bmjopen.bmj.com/ 37 38 individuals were diagnosed with oesophageal cancer (OC), and 1 033 701 individuals were 39 40 diagnosed with stomach cancer (SC), with approximately 50% of new cases occurring in China. 41 42 There are an estimated 508 585 cancer deaths of OC and 782 685 cancer deaths of SC in 2018, 43 2

44 accounting for 5.3% and 8.2% of cancer-cause deaths. Apparently, upper digestive tract cancer on September 27, 2021 by guest. Protected copyright. 45 46 (UDTC) (oesophagus, stomach) has become a significant morbidity and mortality source related 47 3 48 to cancer. According to the National Cancer Center, OC has been the sixth most common cancer 49 50 and the fourth most common cancer cause of cancer-related death. SC has been the second most 51 52 common cancer and the third most common cancer cause of cancer-related death. The incidence 53 54 rate of OC and SC was 17.87/100 000, and 29.31/100 0000, the mortality rate of them were 55 13.68/100 0000 and 21.16/100 0000, respectively, in China in 2015.3 Hence, UDTC has become a 56 57 major public health challenge in China, and the disease burden of it is also considerable.4 Due to 58 59 the population health-seeking behaviour and the diseases’ character, UDTC is mostly diagnosed at 60 3

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 35

1 2 3 a late stage, which is leading to a low survival rate. It was estimated that its five-year survival rate 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 was less than 20% if diagnosed at an advanced stage but is as high as 95% if detected at an earlier

7 5 6 7 8 stage. 9 10 Although the cause of UDTC is not clear, it is believed by researchers that the epidemic of 11 12 UDTC in China is attributed to the multiplicity of demographic factors, diet, lifestyle, family 13 14 health, environment, gastrointestinal history and genetic factors.8 It is well known that tobacco, 15 16 alcohol consumption, overweight or obesity, thermal irritation (drinking scalding liquids) and 17 18 insufficient intakeFor of vegetables peer and fruit, review consumption of pickled only and fried food are eight risk 19 20 factors that can be altered by a tangible action for UDTC control.9 10 11 12 13 21 22 23 Although parts of these risk factors have decreased because of a set of interventions 24 25 implemented by Chinese government,14 15 the others have increased and will continue to grow in 26 27 the next decades because of the rapid transition of urbanization, industrialization and ageing.16 17 28 29 Furthermore, several studies have indicated that these risk factors co-prevalence was common in 30 18 19 31 the population which would further increase the risk of UDTC. A comprehensive assessment 32 33 of the distribution and the status of UDTC risk factors co-prevalence is significant for cancer 34 35 prevention and control. Once we have such data, interventions can be planned and implemented

36 http://bmjopen.bmj.com/ efficiently to minimize these modifiable risk factors, thereby minimizing the health risks of 37 38 increasing UDTC-related mortality and morbidity. 39 40 41 The Yangzhong City of Jiangsu Province is one of the high-risk areas of UDTC, especially in 42 43 the rural area.8 In 2015, the incidence rate of OC was 69.2/100 000, the mortality rate of OC and

44 on September 27, 2021 by guest. Protected copyright. 45 SC was 70.24/100 0000 and 81.89/100 0000, respectively, in Yangzhong City, which is higher 46 47 compared with the average of the nation.20 21 Hence, Yangzhong City had been one of the project 48 49 sites of the Upper Digestive Tract Cancer Early Diagnosis and Treatment (UDTCEDAT) since the 50 51 2006.8 Many studies have estimated the risk factors for UDTC in different areas worldwide.9 10 11 52 53 12 13 The results reveal that risk factors for UDTC are widespread. The modifiable risk factors are 54 55 significantly crucial for the prevention and control of UDTC because these factors can be changed 56 57 by some healthy education or other interventions implemented by doctors and government, and 58 59 improved with the increase of personal health awareness. However, the evidence on the 60 4

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 35 BMJ Open

1 2 3 co-prevalence of these modifiable risk factors in high-risk areas is still limited. Thus, we aimed to 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 report the prevalence and co-prevalence of modifiable UDTC risk factors, and analyze the relevant 7 8 factors influencing modifiable UDTC risk factor co-prevalence among residents aged 40-69 years 9 in Yangzhong City which is a high-risk area of UDTC. 10 11 12 METHODS 13 14 15 Study population 16 17 18 For the present study,For we used peer secondary datareview collected from theonly screening of UDTC, focusing on 19 20 the early diagnosis and treatment of UDTC among high-risk populations (aged 40-69 years) in 21 22 Yangzhong City, China, from 2006 to 2017.9 22 We use the method of multistage stratified 23 24 cluster sampling to select the study sample. In the first stage, we stratified Yangzhong City 25 26 into six regions (Sanmao, Baqiao, Youfang, Xinglong, Xilai, and Xinba ) covering the whole of 27 28 Yangzhong. In the second stage, we randomly selected clusters of three regions (Baqiao, Youfang, 29 30 and Xinglong) by region distribution and economic level based on Yangzhong Yearbook data. In 31 the third stage, administration villages or neighbourhood communities in each chosen regions 32 33 were randomly selected with probability proportional to size. In the fourth stage, each resident 34 35 group or village group was selected from chosen administration villages or neighbourhood

36 http://bmjopen.bmj.com/ 37 communities. In the fifth stage, all man or woman eligible from each household in the sites 38 39 mentioned above were invited for cancer screening, unless they met the following exclusion 40 41 criteria: (1) history of UDTC or mental disorder; (2) contraindications for endoscopic 42 43 examinations and (3) inability to complete the whole interview or informed consent. Inclusion

44 on September 27, 2021 by guest. Protected copyright. 45 criteria for participants were as following: (1) aged 40-69 years; (2) permanent residents in 46 47 Yangzhong City and (3) willing to accept endoscopic examination. 48 49 50 Before the screening, we obtained written informed consent from all participants after informing 51 52 them about the backgrounds, objectives, procedures, benefits, confidentiality agreement of 53 54 personal information and possible consequences of the whole program. Then questionnaire-based 55 56 interview, physical examinations, laboratory tests were performed by professional investigators. 57 58 At last, the endoscopic examinations, pathological diagnosis and necessary therapy for 59 60 participants were conducted by well-trained doctors in People’s Hospital of Yangzhong City. The 5

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 35

1 2 3 screening procedure follows China's cancer screening and early diagnosis and treatment 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 22 6 technology program strictly. The data used in this study derived mainly from the questionnaire 7 8 and physical examinations. Finally, a total of 21 175 individuals were surveyed, with a response 9 rate of 60.9% (2 1175 / 34 743), 1 962 residents were excluded due to missing age, marital status, 10 11 or some other factors, leaving 19 213 participants available for the present survey. The sample 12 13 size accounted for about 17.3% of the total target population of Yangzhong City. We provided 14 15 health education about UDTC and the potential role of modifiable risk factors related to UDTC to 16 17 all eligible participants after collecting information relating to risk factors with the questionnaire. 18 For peer review only 19 Besides, we combined active and passive follow-ups to collect outcome information for 20 21 participants diagnosed with UDTC or precancerous lesions. We also performed a regular 22 23 re-examination for patients according to the diagnosis. 24 25 26 27 28 29 Questionnaire data collection 30 31 Before implementing data collection, training sessions organized by the expert group on 32 33 UDTCEDAT were provided for all staff. The aim of this study, the standard measurement 34 35 methods, how to perform questionnaires properly, and the concrete study procedure were included

36 http://bmjopen.bmj.com/ 37 in the training contents. At the end of the training sessions, all staff participated in the assessment 38 39 and proved to be qualified. 40 41 42 We use uniformly structured questionnaires to collect information through face-to-face 43

44 interviews. Each questionnaire took approximately 25 mins to complete. The questionnaire on September 27, 2021 by guest. Protected copyright. 45 46 information included demographic factors (gender, birthday, address, ID, marital status, and 47 48 household size), socioeconomic characteristic (education and annual family income), behavioural 49 50 factors (excessive alcohol consumption, current smoking), dietary habits (insufficient fresh 51 52 vegetables intake and insufficient fruit intake, consumption of pickled, fried and hot food), and 53 54 body mass index (BMI). 55 56 57 Physical examination 58 59 Physical examination included height and weight. Height and weight measurements were taken by 60 6

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 35 BMJ Open

1 2 3 height scale and digital weight, respectively, with the help of trained examiners based on a 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 standardized program. All subjects were asked to remove any footwear, hats, and heavy clothing 7 8 before height and weight were measured. Height was measured to the nearest 0.1 cm, while 9 weight was measured to the nearest 0.1 kg. BMI was calculated by dividing body weight (in 10 11 kilograms) by the square of height (in meters). 12 13 14 Assessment criteria 15 16 17 Definitions of UDTC modifiable risk factors and the co-prevalence of these risk factors 18 For peer review only 19 20 Eight modifiable UDTC risk factors were defined based on current national guidelines or 21 22 related references. Overweight/obesity was defined as BMI ≥24.0 kg/m2.23 24 25 Current smoking 23 24 was defined as self-reported having used any tobacco products, including cigarettes, cigars or 25 26 26 pipes daily continuously. Excessive alcohol consumption, insufficient fresh vegetables and fruit 27 27 28 intake were defined according to the Dietary Guidelines for Chinese residents (2016). 29 30 Accordingly, excessive alcohol consumption was defined as consumption of more than 25 gram 31 (for males) or 15 gram (for females) alcohol drinks per day after calculating pure alcohol based on 32 33 the type of alcohol they chose, insufficient vegetables intake as self-reported consumption of 34 35 vegetables less than 7 times per week and insufficient fruit intake as self-reported consumption of

36 http://bmjopen.bmj.com/ 37 fruit less than 7 times per week.27 Besides, self-reported consumption of pickled, fried or hot 38 39 food at least once a week were classified as “Yes” in dietary habits, respectively. 40 41 42 Based on considering the literature and the average number of risk factors in the research 43

44 population, co-prevalence of modifiable UDTC risk factors was defined as presenting at least four on September 27, 2021 by guest. Protected copyright. 45 46 related risk factors in one individual.24 28 47 48 49 Covariates 50 51 52 Covariates included in this study were demographic and socioeconomic information ascertained 53 by questionnaire, including age (40-44 years, 45-49 years, 50-54 years, 55-59 years, 60-64 years, 54 55 65-69 years), gender (male and female), marital status (single, currently married, 56 57 divorced/widowed/separated), educational status (no institutional education, primary school, 58 59 junior high school, senior high school and higher), household size (0-3, 4-5, ≥6) and annual family 60 7

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 35

1 2 3 income (tertiles: lower, middle and higher).29 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 Statistical analysis 7 8 9 Descriptive statistics were applied to describe the socio-demographic characteristics of the 10 11 sampled population. The difference in continuous variables was analyzed by student’s t-test, and 12 13 by χ² test to assess the differentials in the prevalence values among categorical variables. 14 15 Differences in men and women, the prevalence of each modifiable UDTC risk factor and the 16 17 distribution of modifiable UDTC risk factors co-prevalence (0, 1, 2, 3 and, ≥4) in 18 For peer review only 19 socio-demographic and other characteristics were described in the overall population, respectively. 20 21 Multiple logistic regression models were adopted to explore the association between relevant 22 23 characteristics and UDTC risk factors co-prevalence. Only the variables that we found statistically 24 25 significant at P <0.05 in the univariate analysis were included in the multiple logistic regression 26 27 models. The result of multiple logistic regression analyses was presented in terms of adjusted odds 28 29 ratios (AOR) and their respective 95% confidence intervals (CIs). All statistical analyses were 30 31 performed by SPSS software V.17.0. A two-sided P <0.05 was considered to be statistically 32 33 significant. 34 35 Participant involvement statement

36 http://bmjopen.bmj.com/ 37 38 No participants or public were included in the design phase of this study. No participants were 39 40 asked to advise on interpretation or writing up of results. Dissemination of the result of the 41 42 research to participants and relevant participants community was prohibited. All the participants 43

44 had the right to receive the result of health check if they wanted. on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 8

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 35 BMJ Open

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 9

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 35

1 2 3 Prevalence of modifiable UDTC risk factors 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 The prevalence of overweight or obesity, current smoking and excessive alcohol consumption in 7 8 this study was 45.3%, 24.1% and 16.2%, respectively. The prevalence of current smoking and 9 10 excessive alcohol consumption in men was significantly higher than in women (all P<0.001). In 11 12 addition, insufficient vegetables intake, insufficient fruit intake and the consumption of pickled, 13 14 fried and hot food in participants accounted for 66.1%, 94.5%, 68.1%, 36.0%, and 88.4%, 15 16 respectively. The prevalence of consumption of pickled and hot food was higher in women than in 17 18 men (all P<0.05)For (Table 2).peer As shown review in table 3, there wereonly significant differences in age, 19 20 marriage status, education level, and annual family income in the 8 UDTC risk factors (all 21 22 P<0.05). The prevalence of these 8 modifiable UDTC risk factors tends to be higher in single 23 24 participants, except for overweight or obesity (all P<0.001). Moreover, the prevalence of 25 26 excessive alcohol consumption, insufficient vegetables intake, insufficient fruit intake and the 27 28 consumption of pickled, fried and hot food varied significantly with the household size (all 29 30 P<0.001)(Table 3). 31 32 33 34 35

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 10

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 35 BMJ Open

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 Co-prevalence of modifiable UDTC risk factors 7 8 9 The prevalence of zero to eight modifiable UDTC risk factors participants had simultaneously in 10 11 the study (Overweight/obesity, current smoking, excessive alcohol consumption, insufficient 12 13 vegetables intake, insufficient fruit intake, the consumption of pickled, fried and hot food) was 14 15 0.1%, 1.3%, 6.4%, 22.6%, 23.2%, 22.4%, 17.5%, 5.5% and 1.0%, respectively (data shown partly 16 17 in Table 4). Among them, nearly all (99.9%) participants showed one or more modifiable UDTC 18 For peer review only risk factors, 98.6% showed two or more modifiable UDTC risk factors, 92.2% had three or more 19 20 modifiable UDTC risk factors and 69.7% presented at least four modifiable UDTC risk factors. 21 22 The average count of modifiable UDTC risk factors per participant in this study is 4.39. The 23 24 prevalence of co-prevalence of modifiable UDTC risk factors was higher in men than in women 25 26 (P<0.001). The prevalence of co-prevalence of modifiable UDTC risk factors was the highest in 27 28 the age group 50-54 and among single participants (all P<0.001). There was an increasing trend 29 30 towards modifiable UDTC risk factors co-prevalence with increasing education and annual family 31 32 income (all P<0.001). Moreover, modifiable UDTC risk factors co-prevalence was the highest 33 34 among those participants who had 0-3 family members (P<0.001) (Table 4). 35

36 http://bmjopen.bmj.com/ 37 The multivariable logistic regression analysis revealed that men were more likely to have 4 or 38 39 more modifiable UDTC risk factors compared with women (OR 2.302, 95% CI 2.145 to 2.471). 40 41 The prevalence of modifiable UDTC risk factors co-prevalence increased with age and then 42 43 decreased, peaking at 50-54 (OR 2.044, 95% CI 1.825 to 2.288). Participants who were married

44 on September 27, 2021 by guest. Protected copyright. 45 (OR 0.511, 95% CI 0.330 to 0.792) were less likely to have 4 or more modifiable UDTC risk 46 47 factors than those who were single. In addition, we divided the education into four groups, which 48 49 showed that increasing the level of education was a risk factor for modifiable UDTC risk factors 50 51 co-prevalence. Compared with participants who had a higher level of annual family income, those 52 who had middle (OR 0.218, 95% CI 0.197 to 0.241) and lower (OR 0.223, 95% CI 0.201 to 0.247) 53 54 level of annual family income were less likely to have 4 or more modifiable UDTC risk factors. 55 56 Modifiable UDTC risk factors co-prevalence were less common among participants who had more 57 58 than 6 family members than those who had less than 3 ones (OR 0.598, 95% CI 0.527 to 0.678) 59 60 (Table 5). 11

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 35

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 7 8 DISCUSSION 9 10 11 12 13 As far as we know, this is the first large population-based survey investigated the prevalence and 14 15 co-prevalence of eight modifiable UDTC risk factors and described the socio-demographic and 16 17 socio-economic factors associated with these among Yangzhong City residents aged 40-69 from 18 For peer review only 19 southeast China. The present study revealed that the prevalence and co-prevalence of overweight 20 21 or obesity, current smoking, excessive alcohol consumption, insufficient vegetables intake, 22 23 insufficient fruit intake and consumption of pickled, fried and hot food were high, which implied 24 25 the health risk of UDTC residents have in Yangzhong City. We found that the prevalence of 26 27 insufficient fruit intake and hot and pickled food consumption were the top three modifiable 28 29 UDTC risk factors in the population surveyed. Besides, 69.7% of the participants presented at 30 least four UDTC risk factors. 31 32 33 The prevalence of overweight or obesity (45.3%) in our study was higher than that observed 34 35 in Nanjing (35.6%),23 and nationwide population (42.0%),30 but our findings were closed to that

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 35 BMJ Open

7 36 19 8 (22.7%, 7.1%, and 10.9%), as well as in Nanbu (28.63%, 1.95%, and 6.11%), both of these 9 region mentioned above are high-risk areas of UDTC in China. 10 11 12 The co-prevalence of risk factors for chronic diseases is widespread.19 Several previous 13 14 studies reported the co-prevalence of chronic diseases in the Chinese population. For example, 15 16 among 49 247 Chinese aged 15 to 69 years from the 2007 China Chronic Disease and Risk Factor 17 18 Surveillance, the prevalenceFor peerof having zero, review one, two and at least only three chronic disease risk factors 19 20 were 9.1%, 33.9%, 32.4% and 24.6%, respectively.37 Also, other regional studies have examined 21 22 the co-prevalence of some specific chronic diseases in residents. Hong et al. reported that 30.1% 23 24 and 35.2% of the Nanjing population presented one and at least two cardiovascular diseases 25 26 (CVD) risk factors.23 Conversely, a much higher rate of CVD risk factors co-prevalence was 27 28 noticed by Ni et al. in Shenzhen City.24 29 30 31 In our present study, it was observed that 0.1%, 1.3%, 6.4%, 22.6% and 69.7% of participants 32 33 had zero, one, two, three and at least four modified UDTC risk factors, respectively, among 34 35 residents aged 40-69 years. The modified UDTC risk factors co-prevalence was prevalent

36 http://bmjopen.bmj.com/ considerably in Yangzhong City. Different estimates of the risk factors co-prevalence for UDTC 37 38 were found in the literature. He et al. 2019 showed that among residents aged 40-69 years in 39 40 UDTC high-risk areas, 33.08%, 35.99%, 16.76% and 11.93% of participants had one, two, three 41 42 and at least four oesophageal cancer risk factors, respectively.19 In another case-control study of 2 43

44 266 Chinese adults, 32.5% and 41.1% of the participants presented three and four or more risk on September 27, 2021 by guest. Protected copyright. 45 46 factors, respectively, for oesophageal or stomach cancer.18 Compared with these two studies 47 48 mentioned above,18 19 a much higher co-prevalence of risk factors was noticed in our present 49 50 study. The variations could change likely due to the difference in diagnostic criteria, the number 51 52 and kind of risk factors included in the research and the participants' age group. Overall, there are 53 54 some other national, and worldwide studies on the co-prevalence of some common or specific 55 56 chronic diseases risk factors. Still, the study on modifiable UDTC risk factors is limited. 57 58 59 The factors associated with modifiable UDTC risk factors co-prevalence included gender, 60 13

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 35

1 2 3 age, marriage status, education, annual family income and household size. We found the 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 prevalence of modifiable UDTC risk factors co-prevalence was lower in women compared with

7 19 23 37 8 men, which was consistent with findings from other settings. The possible reason could be 9 Chinese men are less aware of self-protect for chronic diseases and have worse health-seeking 10 11 behaviour, and may also attend more social occasions, tend to consume more tobacco/cigarette 12 13 high-salt, high-fat and high-calorie food compared with women.23 31 14 15 16 In addition, this study revealed that the prevalence of modifiable UDTC risk factors 17 18 co-prevalence wasFor increased peer with age, which review was consistent with only previous studies.19 23 24 37 Studies 19 20 showed that in Nanbu, China, the co-prevalence of oesophageal cancer risk factors increased with 21 22 age may attribute to the lower level of awareness, practice and willingness for health among the 23 24 elder.19 Meanwhile, it is also shown in table 5 that being over 65 protects against having more 25 26 than four risk factors. A possible reason for this difference is as follows: with the ageing of the 27 28 body and the deterioration of organ function, an elder individual possesses a higher risk of health 29 30 disorder and has a greater demand for medical care. As a result, this creates more opportunities to 31 32 get diagnosed with some health screening, including UDTC. Correspondingly, the elderly have 33 34 more chance to get a healthy education from physicians than the younger.38 Our study also showed 35 that single participants had more prevalence of modifiable UDTC risk factors co-prevalence 36 http://bmjopen.bmj.com/ 37 19 38 compared with participants who were currently married, following a previous study. A possible 39 40 explanation is that being a single older resident comes with its own economic and emotional

41 39 40 42 challenges, contributing to the unhealthy habit of lifestyle and diet. 43

44 Our study demonstrated that the level of socioeconomic status (SES, education, annual on September 27, 2021 by guest. Protected copyright. 45 46 family income) was positively associated with modifiable UDTC risk factors co-prevalence, 47 48 which was inconsistent with other reports.19 23 37 Residents with a higher level of SES are more 49 50 aware of control and prevention of chronic disease and have better health-seeking behaviour 51 52 compared with those with a lower level of SES.19 23 41 Moreover, the poor or lower education 53 54 participants may have relatively more inaccessibility and unaffordability to medical services.38 55 56 This paradox may be due to most of the participants enrolled in our study were from rural areas, 57 58 and their SES was generally low. However, it may also imply that the higher income may 59 60 contribute to unhealthy lifestyles,42 and knowledge alone may not be sufficient to change 14

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 35 BMJ Open

1 2 3 unhealthy lifestyles. Therefore, the level of education and income are two essential SES factors for 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 modifiable UDTC risk factors co-prevalence. 7 8 It is, however, important to note that participants who have more than 6 family members had 9 10 a lower prevalence of modifiable UDTC risk factors co-prevalence compared with those who have 11 12 less than 3 ones. Changes in household size are bound to affect the adjustment of the family diet. 13 14 As the household size increase, it is more likely to increase dietary diversity (e.g. fruits, vegetables 15 16 and milk) every day.43 Besides, the affection, information and economic support among family 17 18 members will alsoFor increase peersignificantly, reviewwhich can adjust and only correct the unhealthy lifestyle of 19 20 individuals.44 45 21 22 23 Our study explored the prevalence and influencing factors of modifiable UDTC risk factors 24 25 co-prevalence in the UDTC high-risk area, Yangzhong City, based on the community-based 26 27 project for UDTC screening with a large sample size. Additionally, the physical measurement and 28 29 the data collection implemented by trained interviewers strictly according to standard protocol and 30 31 instrument increase the validity of our results. In order to reduce the prevalence and co-prevalence 32 33 rate of UTDC risk factors, the screening teams should focus on individuals with co-prevalence of 34 35 risk factors in screening and improve their unhealthy lifestyles continually through a range of

36 http://bmjopen.bmj.com/ methods such as post-screening health education, personalized interventions, and disease 37 38 follow-up. The social impact of screening should be expanded to improve the compliance of 39 40 high-risk groups, thereby increasing the output of screening health benefits. Meanwhile, the 41 42 government should also help high-risk groups (especially the older and male groups) to improve 43

44 their health literacy and awareness of UTDC prevention through diversified education, motivation, on September 27, 2021 by guest. Protected copyright. 45 46 and publicity methods, such as health education, health talks, and mass media campaigns. By 47 48 guiding the culture of smoking, drinking, and other food culture, promote high-cultural groups to 49 50 transform their cognitive and economic advantages into advantages in UTDC prevention and 51 52 health care, and effectively change unhealthy habits. Besides, the government should focus on 53 54 single and residents with small household size in the high-risk groups in the process of health 55 56 education and the development of prevention strategies. The findings may also provide the 57 58 reference for departments in charge of the prevention and control of UDTC in Yangzhong City, 59 60 Jiangsu province, and relevant departments in other UDTC high-risk areas (e.g. Linzhou, 15

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 35

1 2 3 Feicheng, Yanting). 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 There were also several possible limitations to our study. Firstly, a cross-section study cannot 7 8 exam the causality or temporal relationship between the co-prevalence of modifiable UDTC risk 9 10 factors and its influencing factors. Secondly, the modifiable UDTC risk factors included in our 11 12 study were self-reported by participants, which may contribute to recall and reporting bias, except 13 14 BMI. Our study results were from Yangzhong City only, and cannot be generalized to the other 15 16 high-risk areas and the whole of southeast China. Additionally, the study response rate was 17 18 relatively low, particularlyFor amongpeer males, whichreview may affect the only results' representativeness. Finally, 19 20 our study only focused on the 8 modifiable UDTC risk factors, but there are far more than eight 21 22 risk factors for UDTC. Hence, further studies are needed. 23 24 25 Conclusion 26 27 28 In summary, this cross-sectional study shows that the prevalence and co-prevalence of modifiable 29 30 UDTC risk factors are high among participants in Yangzhong City. Our analyses indicate that 31 men, younger adults, single adults, and participants with higher SES or smaller household size are 32 33 susceptible to modifiable UDTC risk factors co-prevalence. Policies to prevent UDTC have 34 35 already been developed in the strategic plan and operational plan, however, the accuracy and

36 http://bmjopen.bmj.com/ 37 validity of implementing the undertaken policies are still insufficient. Consequently, extra 38 39 attention is required to pay on these high-risk groups during the progress of screening. Relative 40 41 departments also need to make effective public health programs targeting modifiable UDTC risk 42 43 factors that aim to decrease UDTC risk factors co-prevalence in high-risk groups from high-risk

44 on September 27, 2021 by guest. Protected copyright. 45 areas of UDTC. 46 47 48 Acknowledgements We would like to thank all participants in the present study. We are 49 50 also grateful to the Upper Digestive Tract Cancer Early Diagnosis and Treatment Program. 51 52 53 54 55 Contribution XF and RC are joint first authors. ZLH, DFQ, WQW, GQW, JYZ, JJW, GS 56 57 and XW contributed to the study design. ZLH, QZ, AWS, TQS performed the survey and 58 59 collected study data. XF and RC wrote the manuscript and ZLH, DFQ, WQW, JYZ, JJW, GS, 60 16

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 35 BMJ Open

36 http://bmjopen.bmj.com/ 37 38 39 Provenance and peer review Not commissioned; externally reviewed. 40 41 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 Data sharing statement No additional data are available. 46 47 48 49 50 Open access This is an Open Access article distributed in accordance with the Creative 51 52 Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to 53 54 distribute, remix, adapt, build upon this work non-commercially, and license their derivative 55 56 works on different terms, provided the original work is properly cited, and the use is 57 58 non-commercial. See:http://creativecommons.org/licenses/by-nc/4.0/ 59 60 17

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 35

1 2 3

44 treatment esophageal cancer in China (2014, Beijing). China Prac Med,2015,4:320-337.In on September 27, 2021 by guest. Protected copyright. 45 Chinese. 46 47 48 [8] Chen R, Ma S, Guan C, et al. The National Cohort of Esophageal Cancer-Prospective Cohort 49 Study of Esophageal Cancer and Precancerous Lesions based on High-Risk Population in 50 51 China (NCEC-HRP): study protocol. BMJ Open,2019,9(4):e027360. 52 53 [9] Salaspuro M. Interrelationship between alcohol, smoking, acetaldehyde and cancer. Novartis 54 Found Symp,2007,285:80-9. 55 56 57 [10] Somi MH, Mousavi SM, Naghashi S, et al. Is there any relationship between food habits in 58 the last two decades and gastric cancer in North-Western Iran?. Asian Pac J Cancer Prev. 59 60 2015,16(1):283-90. 18

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 35 BMJ Open

1 2 3 [11] Navarro-Silvera SA, Mayne ST, Risch HA, et al. Principal component analysis of dietary and

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 22 of 35

1 2 3 cross-sectional survey among Nanjing adults in China. BMJ Open,2018,8(6):e020530.

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 [24] Ni W, Weng RX, Yuan X, et al. Clustering of cardiovascular disease biological risk factors 7 among older adults in Shenzhen City, China: a cross-sectional study. BMJ 8 9 Open,2019,9(3):e024336. 10 11 [25] Department of disease control, ministry of health, PRC. Guidelines for the prevention and 12 13 control of overweight and obesity in Chinese adults. Beijing: people's medical publishing 14 house, 2006:2-4.In Chinese. 15 16 [26] Howitt C, Hambleton IR, Rose AM, et al. Social distribution of diabetes, hypertension and 17 18 related risk factorsFor in Barbados: peer a cross-sectional review study. BMJonly Open,2015,5(12):e008869. 19 20 [27] Chinese nutrition society. Dietary guidelines for Chinese residents (2016). Beijing: people's 21 22 medical publishing house, 2016In Chinese. 23 24 [28] Zaman MM, Bhuiyan MR, Karim MN, et al. Clustering of noncommunicable diseases risk 25 26 factors in Bangladeshi adults: An analysis of STEPS survey 2013. BMC Public 27 Health.2015,15:659. 28 29 [29] Xu F, Yin XM, Zhang M, et al. Family average income and body mass index above the 30 31 healthy weight range among urban and rural residents in regional Mainland China. Public 32 Health Nutr,2005,8(1):47-51. 33 34 35 [30] National Commission of Health Bureau of disease control and Prevention. Report on China

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 23 of 35 BMJ Open

1 2 3 factors in some cities of Hubei province . J of Pub health and Prev Med,2016,27(5):82-85.In

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 24 of 35

1 2 3

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 22

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Page 25 of 35 BMJ Open

BMJ Open Page 26 of 35

Page 27 of 35 BMJ Open

BMJ Open Page 28 of 35

Page 29 of 35 BMJ Open

BMJ Open Page 30 of 35

Page 31 of 35 BMJ Open

the consumption of pickled food, Fried food, the consumption of fried food; Hot food, the consumption of hot food on September 27, 2021 by guest. Protected copyright. 24 25 26 27 28 29 30 31 32 33 34 35 36 37 Table 4 The different number and co-prevalence of modifiable UDTC risk factors in residents by relevant characters in Yangzhong City adults aged 40-69 38 39 40 41 29 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

BMJ Open Page 32 of 35

1 2 3 4 5 Category None(0) single(1) Two(2) Three(3) Non-co-prevalence(≤3) Co-prevalence(≥4) χ² value P value 6 7 8 Total 14(0.1) 244(1.3) 1235(6.4) 4336(22.6) 5829(30.3) 13384(69.7) 9 10 Gender, n(%) 632.397 <0.001 11 12 men 2(0.0)For64(0.8) peer356(4.3) review1293(15.6) 1715(20.7) only 6553(79.3) 13 14 15 women 12(0.1) 180(1.6) 879(8.0) 3043(27.8) 4114(37.6) 6831(62.4) 16 http://bmjopen.bmj.com/ 17 Age group, years, n(%) 373.748 <0.001 18 19 40-44 1(0.0) 66(2.1) 193(6.1) 891(28.4) 1151(36.6) 1991(63.4) 20 21 22 45-49 0(0.0) 26(0.7) 165(4.2) 857(21.8) 1048(26.6) 2889(73.4) 23 on September 27, 2021 by guest. Protected copyright. 24 50-54 0(0.0) 15(.4) 112(3.0) 759(20.3) 886(23.7) 2857(76.3) 25 26 27 55-59 0(0.0) 48(1.3) 175(4.9) 724(20.3) 947(26.5) 2624(73.5) 28 29 60-64 0(0.0) 33(1.1) 273(8.8) 712(23.1) 1018(33.0) 2070(67.0) 30 31 65-69 13(0.8) 56(3.2) 317(18.3) 393(22.7) 779(45.0) 953(55.0) 32 33 34 Marrige, n(%) 67.346 <0.001 35 36 Single 0(0.0) 0(0.0) 9(4.4) 16(7.9) 25(12.3) 178(87.7) 37 38 39 40 41 30 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Page 33 of 35 BMJ Open

BMJ Open Page 34 of 35

1 2 3 4 5 ≥6 2(0.1) 11(0.6) 170(9.7) 376(21.5) 559(32.0) 1187(68.0) 6 7 8 9 10 11 12 For peer review only 13 14 15 Table 5 The multivariable logistic regression analysis of modifiable UDTC risk factors co-prevalence 16 http://bmjopen.bmj.com/ 17 Category Wald P value β SE OR 95% CI 18 19 Gender, n(%) 20 21 22 women - - - - 1.000 - 23 on September 27, 2021 by guest. Protected copyright. 24 men 533.13 <0.001 0.834 0.036 2.302 2.145 to 2.471 25 26 27 Age group, years, n(%) 28 29 40-44 - - - - 1.000 - 30 31 45-49 48.302 <0.001 0.381 0.055 1.464 1.315 to 1.631 32 33 34 50-54 153.543 <0.001 0.715 0.058 2.044 1.825 to 2.288 35 36 55-59 143.780 <0.001 0.710 0.059 2.033 1.811 to 2.283 37 38 39 40 41 32 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Page 35 of 35 BMJ Open

BMJ Open Page 36 of 35

Page 37 of 35 BMJ Open

1 2 3 4 5 6 7 8 9 10 11 12 For peer review only 13 14 15 16 http://bmjopen.bmj.com/ 17 18 19 20 21 22 23 24 on September 27, 2021 by guest. Protected copyright. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 35 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Article Type: Original research

Date Submitted by the 12-Mar-2021 Author:

Complete List of Authors: Feng, xiang; People's Hospital of Yangzhong City, Institute of tumour prevention and control Hua, Zhao-lai; People's Hospital of Yangzhong City, Institute of tumour prevention and control Zhou, Qin; People's Hospital of Yangzhong City, Institute of tumour prevention and control Shi, Ai-wu; People's Hospital of Yangzhong City, Institute of tumour prevention and control Song, Tong-qiu; People's Hospital of Yangzhong City, Institute of tumour prevention and control Qian, Dongfu; Nanjing Medical University, School of Health Policy &

Primary Subject Epidemiology Heading:

Secondary Subject Heading: Public health, Epidemiology, Oncology

EPIDEMIOLOGY, Endoscopy < GASTROENTEROLOGY, Risk management Keywords: < HEALTH SERVICES ADMINISTRATION & MANAGEMENT, Epidemiology < ONCOLOGY

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 27 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39

1 2 3

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 27 BMJ Open

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 Prevalence and co-prevalence of modifiable 6 7 risk factors for upper digestive tract cancer 8 9 10 among residents aged 40 to 69 years in 11 12 Yangzhong City, China: a cross-sectional 13 14 15 study 16 Xiang Feng,1 Zhaolai Hua,1 Qin Zhou1, Aiwu Shi,1 Tongqiu Song,1 Dongfu Qian,2 Ru Chen,3 17 3 3 4 5 6 6 18 GuiQi Wang, WenqiangFor Wei, peer Jinyi Zhou, review Jiejun Wang, Gang only Shao, Xi Wang 19 20 1 Institute of tumour prevention and control, People’s Hospital of Yangzhong City, Yangzhong, 21 22 China 23 2 School of Health Policy & Management, Nanjing Medical University, Nanjing, China 24 3 Cancer Registry Office, National Cancer Center/National Clinical Research Center for 25 Cancer/Cancer Hospital, Chinese Academy of Medical Science and Peking Union Medical 26 27 College, Beijing, China 28 4 Department of Non-communicable Disease Prevention, Jiangsu Provincial Center for Disease 29 Control and Prevention, Nanjing, China 30 31 5 Department of Oncology, Changzheng Hospital, Second Military Medical University, Shanghai, 32 China. 33 6 Department of Oncology, 903rd Hospital of PLA, Hangzhou, China. 34 35

36 Correspondence to Zhaolai Hua; [email protected]; Dr Wenqiang Wei; http://bmjopen.bmj.com/ 37 [email protected] 38 39 40 ABSTRACT 41 Objectives To describe the prevalence of modifiable risk factors for upper digestive tract cancer 42 (UDTC) and its co-prevalence, and investigate relevant influencing factors of modifiable UDTC 43 risk factors co-prevalence among residents aged 40-69 years in Yangzhong City, China. 44 on September 27, 2021 by guest. Protected copyright. 45 Design Cross-sectional study 46 Participants A total of 21 175 participants aged 40-69 years were enrolled in the study. 1 962 47 48 subjects were excluded due to missing age, marital status, or some other selected information. 49 Eventually, 19 213 participants were available for the present analysis. 50 Main outcomes measures Prevalence and co-prevalence of eight modifiable UDTC risk factors 51 (overweight or obesity, current smoking, excessive alcohol consumption, insufficient vegetables 52 53 intake, insufficient fruit intake, and the consumption of pickled, fried, and hot food) were 54 analyzed. 55 56 Results The prevalence of overweight/obesity, current smoking, excessive alcohol consumption, 57 58 insufficient vegetables intake, insufficient fruit intake, and the consumption of pickled, fried, and 59 60 hot food in this study was 45.3%, 24.1%, 16.2%, 66.1%, 94.5%, 68.1%, 36.0%, and 88.4%,

1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 27

1 2 3 respectively. Nearly all (99.9%) participants showed one or more UDTC risk factors, 98.6% of the 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 participants showed at least two risk factors, 92.2% of the participants had at least three risk 7 8 factors, and 69.7% of the participants had four or more risk factors. Multivariate logistic 9 regression analysis revealed that men, younger age, single, higher education, higher annual family 10 11 income and, smaller household size were more likely to present modifiable UDTC risk factors 12 13 co-prevalence. 14 15 Conclusions The prevalence and co-prevalence of modifiable UDTC risk factors are high among 16 17 participants in Yangzhong City. Extra attention must be paid to these groups who are susceptible 18 For peer review only 19 to risk factors co-prevalence during screening progress. Relative departments also need to make 20 21 significant public health programs that aim to decrease modifiable UDTC risk factors 22 23 co-prevalence among residents aged 40-69 years from high-risk areas of UDTC. 24 25 26 27 Strengths and limitations of this study 28 29 This is the first study examining the prevalence and co-prevalence of modifiable upper digestive 30 tract cancer (UDTC) risk factors and investigating relevant influencing factors in Yangzhong City, 31 with large and representative residents from southeast China. 32 33 34 Participants who volunteered to participate in our study are considered to be the high-risk group 35 for UDTC, which is significantly important for the prevention and control of UDTC in China.

36 http://bmjopen.bmj.com/ 37 38 A cross-section study cannot exam the causality or temporal relationship. 39 40 The modifiable UDTC risk factors included in our study were self-reported by participants, which 41 42 may contribute to recall and reporting bias, except BMI. 43

44 We only focus on the eight modifiable UDTC risk factors, which may underestimate the average on September 27, 2021 by guest. Protected copyright. 45 46 number of modifiable UDTC risk factors among participants. 47 48 INTRODUCTION 49 50 According to the WHO, more than 70% of the total deaths worldwide were related to 51 52 non-communicable diseases (NCDs) in 2016.1 Cancer is the second cause of NCDs, accounting 53 54 for 22% of total global deaths related to NCDs. Globally, in 2018, an estimated 572 034 55 56 individuals were diagnosed with oesophageal cancer (OC), and 1 033 701 individuals were 57 58 diagnosed with stomach cancer (SC), with approximately 50% of new cases occurring in China. 59 60 There are an estimated 508 585 cancer deaths of OC and 782 685 cancer deaths of SC in 2018,

2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 27 BMJ Open

1 2 3 accounting for 5.3% and 8.2% of cancer-cause deaths.2 Apparently, upper digestive tract cancer 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 (UDTC) (oesophagus, stomach) has become a significant morbidity and mortality source related

7 3 8 to cancer. According to the National Cancer Center, OC has been the sixth most common cancer 9 and the fourth most common cancer cause of cancer-related death. SC has been the second most 10 11 common cancer and the third most common cancer cause of cancer-related death. The incidence 12 13 rate of OC and SC was 17.87/100 000 and 29.31/100 000, the mortality rate of them was 14 15 13.68/100 000 and 21.16/100 000, respectively, in China in 2015.3 Hence, UDTC has become a 16 17 major public health challenge in China, and the disease burden of it is also considerable.4 Due to 18 For peer review only 19 the population health-seeking behaviour and the diseases’ character, UDTC is mostly diagnosed at 20 21 a late stage, which is leading to a low survival rate. It was estimated that its five-year survival rate 22 23 was less than 20% if diagnosed at an advanced stage but is as high as 95% if detected at an earlier 24 25 stage.5 6 7 26 27 Although the cause of UDTC is not clear, it is believed by researchers that the epidemic of 28 29 UDTC in China is attributed to the multiplicity of demographic factors, diet, lifestyle, family 30 31 health, environment, gastrointestinal history and genetic factors.8 It is well known that tobacco, 32 33 alcohol consumption, overweight or obesity, thermal irritation (drinking scalding liquids) and 34 35 insufficient intake of vegetables and fruit, consumption of pickled and fried food are eight risk

36 http://bmjopen.bmj.com/ 37 factors that can be altered by a tangible action for UDTC control.9 10 11 12 13 38 39 Although parts of these risk factors have decreased because of a set of interventions 40 14 15 41 implemented by the Chinese government, the others have increased and will continue to grow 42 16 43 in the next decades because of the rapid transition of urbanization, industrialization and ageing.

44 on September 27, 2021 by guest. Protected copyright. 17 45 Furthermore, several studies have indicated that these risk factors co-prevalence was common in

46 18 19 47 the population which would further increase the risk of UDTC. A comprehensive assessment 48 of the distribution and the status of UDTC risk factors co-prevalence is significant for cancer 49 50 prevention and control. Once we have such data, interventions can be planned and implemented 51 52 efficiently to minimize these modifiable risk factors, thereby minimizing the health risks of 53 54 increasing UDTC-related morbidity and mortality. 55 56 The Yangzhong City of Jiangsu Province is one of the high-risk areas of UDTC, especially in 57 58 rural areas.8 In 2015, the incidence rate of OC was 69.2/100 000, the mortality rate of OC and SC 59 60 was 70.24/100 0000 and 81.89/100 0000, respectively, in Yangzhong City, which is higher

3 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 27

1 2 3 compared with the average of the nation.20 21 Hence, Yangzhong City had been one of the project 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 sites of the Upper Digestive Tract Cancer Early Diagnosis and Treatment (UDTCEDAT) since the

7 8 9 10 11 8 2006. Many studies have estimated the risk factors for UDTC in different areas worldwide. 9 12 13 The results reveal that risk factors for UDTC are widespread. The modifiable risk factors are 10 11 significantly crucial for the prevention and control of UDTC because these factors can be changed 12 13 by some healthy education or other interventions implemented by doctors and government and 14 15 improved with the increase of personal health awareness. However, the evidence on the 16 17 co-prevalence of these modifiable risk factors in high-risk areas is still limited. Thus, we aimed to 18 For peer review only 19 report the prevalence and co-prevalence of modifiable UDTC risk factors and analyze the relevant 20 21 factors influencing modifiable UDTC risk factor co-prevalence among residents aged 40-69 years 22 23 in Yangzhong City, which is a high-risk area of UDTC. 24 25 METHODS 26 27 Study population 28 29 For the present study, we used secondary data collected from the screening of UDTC, focusing on 30 31 the early diagnosis and treatment of UDTC among high-risk populations (aged 40-69 years) in 32 33 Yangzhong City, China, from 2006 to 2017.9 22 We use the method of multistage stratified 34 35 cluster sampling to select the study sample. In the first stage, we stratified Yangzhong City

36 http://bmjopen.bmj.com/ 37 into six regions (Sanmao, Baqiao, Youfang, Xinglong, Xilai, and Xinba ), covering the whole 38 39 Yangzhong. In the second stage, we randomly selected clusters of three regions (Baqiao, Youfang, 40 41 and Xinglong) by region distribution and economic level based on Yangzhong Yearbook data. In 42 43 the third stage, administration villages or neighbourhood communities in each chosen regions

44 on September 27, 2021 by guest. Protected copyright. 45 were randomly selected with probability proportional to size. In the fourth stage, each resident 46 47 group or village group was selected from chosen administration villages or neighbourhood 48 communities. In the fifth stage, all man or woman eligible from each household in the sites 49 50 mentioned above were invited for cancer screening, unless they met the following exclusion 51 52 criteria: (1) history of UDTC or mental disorder; (2) contraindications for endoscopic 53 54 examinations and (3) inability to complete the whole interview or informed consent. Inclusion 55 56 criteria for participants were as following: (1) aged 40-69 years; (2) permanent residents in 57 58 Yangzhong City and (3) willing to accept endoscopic examination. 59 60 Before the screening, we obtained written informed consent from all participants after informing

4 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 27 BMJ Open

1 2 3 them about the backgrounds, objectives, procedures, benefits, confidentiality agreement of 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 personal information and possible consequences of the whole program. Then questionnaire-based 7 8 interview, physical examinations, laboratory tests were performed by professional investigators. 9 At last, the endoscopic examinations, pathological diagnosis and necessary therapy for 10 11 participants were conducted by well-trained doctors in People’s Hospital of Yangzhong City. The 12 13 screening procedure follows China's cancer screening and early diagnosis and treatment 14 15 technology program strictly.22 The data used in this study derived mainly from the questionnaire 16 17 and physical examinations. Finally, a total of 21 175 individuals were surveyed, with a response 18 For peer review only 19 rate of 60.9% (2 1175 / 34 743), 1 962 residents were excluded due to missing age, marital status, 20 21 or some other factors, leaving 19 213 participants available for the present survey. The sample 22 23 size accounted for about 17.3% of the total target population of Yangzhong City. We provided 24 25 health education about UDTC and the potential role of modifiable risk factors related to UDTC to 26 27 all eligible participants after collecting information relating to risk factors with the questionnaire. 28 29 Besides, we combined active and passive follow-ups to collect outcome information for 30 31 participants diagnosed with UDTC or precancerous lesions. We also performed a regular 32 33 re-examination for patients according to the diagnosis. 34 35

36 http://bmjopen.bmj.com/ 37 Questionnaire data collection 38 39 Before implementing data collection, training sessions organized by the expert group on 40 41 UDTCEDAT were provided for all staff. The aim of this study, the standard measurement 42 43 methods, how to perform questionnaires properly, and the concrete study procedure were included

44 on September 27, 2021 by guest. Protected copyright. 45 in the training contents. At the end of the training sessions, all staff participated in the assessment 46 47 and proved to be qualified. 48 We use uniformly structured questionnaires to collect information through face-to-face 49 50 interviews. Each questionnaire took approximately 25 mins to complete. The questionnaire 51 52 information included demographic factors (gender, birthday, address, ID, marital status, and 53 54 household size), socioeconomic characteristic (education and annual family income), behavioural 55 56 factors (excessive alcohol consumption, current smoking), dietary habits (insufficient fresh 57 58 vegetables intake and insufficient fruit intake, consumption of pickled, fried and hot food), and 59 60 body mass index (BMI).

5 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 27

1 2 3 Physical examination 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 Physical examination included height and weight. Height and weight measurements were taken by 7 8 height scale and digital weight, respectively, with the help of trained examiners based on a 9 standardized program. All subjects were asked to remove any footwear, hats, and heavy clothing 10 11 before height and weight were measured. Height was measured to the nearest 0.1 cm, while 12 13 weight was measured to the nearest 0.1 kg. BMI was calculated by dividing body weight (in 14 15 kilograms) by the square of height (in meters). 16 17 Assessment criteria 18 For peer review only 19 Definitions of UDTC modifiable risk factors and the co-prevalence of these risk factors 20 21 Eight modifiable UDTC risk factors were defined based on current national guidelines or 22 23 related references. Overweight/obesity was defined as BMI ≥24.0 kg/m2.23 24 25 Current smoking 24 25 was defined as self-reported having used any tobacco products, including cigarettes, cigars or 26 27 pipes daily continuously.26 Excessive alcohol consumption, insufficient fresh vegetables and fruit 28 29 intake were defined according to the Dietary Guidelines for Chinese residents (2016).27 30 31 Accordingly, excessive alcohol consumption was defined as consumption of more than 25 gram 32 33 (for males) or 15 gram (for females) alcohol drinks per day after calculating pure alcohol based on 34 35 the type of alcohol they chose, insufficient vegetables intake as self-reported consumption of

36 http://bmjopen.bmj.com/ 37 vegetables less than 7 times per week and insufficient fruit intake as self-reported consumption of 38 39 fruit less than 7 times per week.27 Besides, self-reported consumption of pickled, fried or hot 40 41 food at least once a week were classified as “Yes” in dietary habits, respectively. 42 43 Based on considering the literature and the average number of risk factors in the research

44 on September 27, 2021 by guest. Protected copyright. 45 population, co-prevalence of modifiable UDTC risk factors was defined as presenting at least four

46 24 28 47 related risk factors in one individual. 48 Covariates 49 50 Covariates included in this study were demographic and socioeconomic information ascertained 51 52 by questionnaire, including age (40-44 years, 45-49 years, 50-54 years, 55-59 years, 60-64 years, 53 54 65-69 years), gender (male and female), marital status (single, currently married, 55 56 divorced/widowed/separated), educational status (no institutional education, primary school, 57 58 junior high school, senior high school and higher), household size (0-3, 4-5, ≥6) and annual family 59 60 income (tertiles: lower, middle and higher).29

6 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 27 BMJ Open

1 2 3 Statistical analysis 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 Descriptive statistics were applied to describe the socio-demographic characteristics of the 7 8 sampled population. The difference in continuous variables was analyzed by student’s t-test and 9 by χ² test to assess the differentials in the prevalence values among categorical variables. 10 11 Differences in men and women, the prevalence of each modifiable UDTC risk factor and the 12 13 distribution of modifiable UDTC risk factors co-prevalence (0, 1, 2, 3 and ≥4) in 14 15 socio-demographic and other characteristics were described in the overall population, respectively. 16 17 Multiple logistic regression models were adopted to explore the association between relevant 18 For peer review only 19 characteristics and UDTC risk factors co-prevalence. Only the variables that we found statistically 20 21 significant at P <0.05 in the univariate analysis were included in the multiple logistic regression 22 23 models. The result of multiple logistic regression analyses was presented in terms of adjusted odds 24 25 ratios (AOR) and their respective 95% confidence intervals (CIs). All statistical analyses were 26 27 performed by SPSS software V.17.0. A two-sided P <0.05 was considered to be statistically 28 29 significant. 30 31 Patient and Public Involvement 32 33 No participants or public were included in the design phase of this study. No participants were 34 35 asked to advise on interpretation or writing up of results. Dissemination of the result of the

36 http://bmjopen.bmj.com/ 37 research to participants and relevant participants community was prohibited. All the participants 38 39 had the right to receive the result of health check if they wanted. 40 41 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

7 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 27

1 2 3 RESULTS 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 Socio-demographic and other characteristics of participants 7 8 The description of socio-demographic and other characteristics of 19 213 participants are 9 presented in Table 1. Of the participants (40-69 years, mean age 53.2±7.8 years) surveyed, 57.0% 10 11 were women, with a mean age of 52.9±7.8 years. More than 95.0% (man 95.6%, woman 94.8%) 12 13 of participants were married, nearly half (44.3%) (man 41.6%, woman 46.4%) had an education 14 15 level of primary school, 44.7% (man 44.6%, woman 44.7%) had a lower level of annual family 16 17 income, and the mean of household size and BMI were 4.0±1.4 (man 4.0±1.4, woman 4.0±1.4) 18 For peer review only 19 and 23.8±3.0 kg/m2 (man 23.9±3.0, woman 23.7±3.0), respectively. The differences between 20 21 men and women in age, marriage, education, household size and BMI were significant (all 22 23 P<0.01) (Table 1). In addition, of the cancer cases diagnosed in this study, the majority of the 24 25 oesophagus was squamous (54/57), while adenocarcinoma predominated in gastric (29/33) and 26 27 cardia (35/36) cancers. 28 29 30 31 32 33 34 35

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

8 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 27 BMJ Open

1 2 3 Prevalence of modifiable UDTC risk factors 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 The prevalence of overweight or obesity, current smoking and excessive alcohol consumption in 7 8 this study was 45.3%, 24.1% and 16.2%, respectively. The prevalence of current smoking and 9 excessive alcohol consumption in men was significantly higher than in women (all P<0.001). In 10 11 addition, insufficient vegetables intake, insufficient fruit intake and the consumption of pickled, 12 13 fried and hot food in participants accounted for 66.1%, 94.5%, 68.1%, 36.0%, and 88.4%, 14 15 respectively. The prevalence of consumption of pickled and hot food was higher in women than in 16 17 men (all P<0.05) (Table 2). As shown in table 3, there were significant differences in age, 18 For peer review only 19 marriage status, education level, and annual family income in the eight UDTC risk factors (all 20 21 P<0.05). The prevalence of these eight modifiable UDTC risk factors tends to be higher in single 22 23 participants, except for overweight or obesity (all P<0.001). Moreover, the prevalence of 24 25 excessive alcohol consumption, insufficient vegetables intake, insufficient fruit intake and the 26 27 consumption of pickled, fried and hot food varied significantly with the household size (all 28 29 P<0.001)(Table 3). 30 31 32 33 34 35

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

9 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 27

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 Co-prevalence of modifiable UDTC risk factors 6 7 The prevalence of zero to eight modifiable UDTC risk factors participants had simultaneously in 8 9 the study (Overweight/obesity, current smoking, excessive alcohol consumption, insufficient 10 11 vegetables intake, insufficient fruit intake, the consumption of pickled, fried and hot food) was 12 13 0.1%, 1.3%, 6.4%, 22.6%, 23.2%, 22.4%, 17.5%, 5.5% and 1.0%, respectively (data shown partly 14 15 in Table 4). Among them, nearly all (99.9%) participants showed one or more modifiable UDTC 16 17 risk factors, 98.6% showed two or more modifiable UDTC risk factors, 92.2% had three or more 18 For peer review only 19 modifiable UDTC risk factors and 69.7% presented at least four modifiable UDTC risk factors. 20 21 The average count of modifiable UDTC risk factors per participant in this study is 4.39. The 22 prevalence of co-prevalence of modifiable UDTC risk factors was higher in men than in women 23 24 (P<0.001). The prevalence of co-prevalence of modifiable UDTC risk factors was the highest in 25 26 the age group 50-54 and among single participants (all P<0.001). There was an increasing trend 27 28 towards modifiable UDTC risk factors co-prevalence with increasing education and annual family 29 30 income (all P<0.001). Moreover, modifiable UDTC risk factors co-prevalence was the highest 31 32 among those participants who had 0-3 family members (P<0.001) (Table 4). 33 34 The multivariable logistic regression analysis revealed that men were more likely to have 35

36 four or more modifiable UDTC risk factors compared with women (OR 2.302, 95% CI 2.145 to http://bmjopen.bmj.com/ 37 38 2.471). The prevalence of modifiable UDTC risk factors co-prevalence increased with age and 39 40 then decreased, peaking at 50-54 (OR 2.044, 95% CI 1.825 to 2.288). Participants who were 41 42 married (OR 0.511, 95% CI 0.330 to 0.792) were less likely to have four or more modifiable 43

44 UDTC risk factors than those who were single. In addition, we divided the education into four on September 27, 2021 by guest. Protected copyright. 45 46 groups, which showed that increasing the level of education was a risk factor for modifiable 47 48 UDTC risk factors co-prevalence. Compared with participants who had a higher level of annual 49 50 family income, those who had middle (OR 0.218, 95% CI 0.197 to 0.241) and lower (OR 0.223, 51 52 95% CI 0.201 to 0.247) level of annual family income were less likely to have four or more 53 54 modifiable UDTC risk factors. Modifiable UDTC risk factors co-prevalence were less common 55 56 among participants who had more than six family members than those who had less than three 57 58 ones (OR 0.598, 95% CI 0.527 to 0.678) (Table 5). 59 60

10 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 27 BMJ Open

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 DISCUSSION 6 7 8 As far as we know, this is the first large population-based survey investigated the prevalence and 9 co-prevalence of eight modifiable UDTC risk factors and described the socio-demographic and 10 11 socio-economic factors associated with these among Yangzhong City residents aged 40-69 from 12 13 southeast China. The present study revealed that the prevalence and co-prevalence of overweight 14 15 or obesity, current smoking, excessive alcohol consumption, insufficient vegetables intake, 16 17 insufficient fruit intake and consumption of pickled, fried and hot food were high, which implied 18 For peer review only 19 the health risk of UDTC residents have in Yangzhong City. We found that the prevalence of 20 21 insufficient fruit intake and hot and pickled food consumption were the top three modifiable 22 23 UDTC risk factors in the population surveyed. Besides, 69.7% of the participants presented at 24 25 least four UDTC risk factors. 26 27 The prevalence of overweight or obesity (45.3%) in our study was higher than that observed 28 29 in Nanjing (35.6%),23 and nationwide population (42.0%),30 but our findings were closed to that 30 31 observed in some other regional.24 31 The prevalence of current smoking in this population 32 33 (24.1%) was much higher than that in a cross-sectional study in Shenzhen (10.5%)24 and Barbados 34 35 (9.2%),32 which was consistent with a survey in Nanjing (24.5%).23 However, the rate was not as

36 http://bmjopen.bmj.com/ 37 high as reported (28.1%) in the China national nutrition and chronic disease survey (2015).30 Our 38 39 findings showed a high prevalence of excessive alcohol consumption in the Yangzhong population 40 30 41 (16.2%) relative to the national average of 11.1% in men and 2.0% in women. The rate of 42 43 excessive alcohol consumption we found was similar to Barbados, Nanbu and the prospective

44 on September 27, 2021 by guest. Protected copyright. 45 study of China Kadoorie Biobank, where excessive alcohol consumption rates were around

46 32 19 33 47 14.5%, 16.7% and 14.9%, respectively. We found higher levels of insufficient intake of 48 vegetables (66.1%) in this population than those observed in the Tanzania34 and Hubei Province35 49 50 where the insufficient intake of vegetables or fruit is 55.8% and 29.7%, respectively, while the 51 52 levels of insufficient intake of fruit (94.5%) in our study were also much higher than that observed 53 54 in the region mentioned above.34 35 Moreover, the proportion of the Yangzhong population had 55 56 dietary habits of consumption of pickled, fried and hot food were greater than the levels in Huaian 57 58 (22.7%, 7.1%, and 10.9%),36 as well as in Nanbu (28.63%, 1.95%, and 6.11%),19 both of these 59 60 region mentioned above are high-risk areas of UDTC in China.

11 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 27

1 2 3 The co-prevalence of risk factors for chronic diseases is widespread.19 Several previous 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 studies reported the co-prevalence of chronic diseases in the Chinese population. For example, 7 8 among 49 247 Chinese aged 15 to 69 years from the 2007 China Chronic Disease and Risk Factor 9 Surveillance, the prevalence of having zero, one, two, and at least three chronic disease risk 10 11 factors were 9.1%, 33.9%, 32.4% and 24.6%, respectively.37 Also, other regional studies have 12 13 examined the co-prevalence of some specific chronic diseases in residents. Hong et al. reported 14 15 that 30.1% and 35.2% of the Nanjing population presented one and at least two cardiovascular 16 17 diseases (CVD) risk factors.23 Conversely, a much higher rate of CVD risk factors co-prevalence 18 For peer review only 19 was noticed by Ni et al. in Shenzhen City.24 20 21 In our present study, it was observed that 0.1%, 1.3%, 6.4%, 22.6% and 69.7% of participants 22 23 had zero, one, two, three and at least four modified UDTC risk factors, respectively, among 24 25 residents aged 40-69 years. The modified UDTC risk factors co-prevalence was prevalent 26 27 considerably in Yangzhong City. Different estimates of the risk factors co-prevalence for UDTC 28 29 were found in the literature. He et al. 2019 showed that among residents aged 40-69 years in 30 31 UDTC high-risk areas, 33.08%, 35.99%, 16.76% and 11.93% of participants had one, two, three 32 33 and at least four oesophageal cancer risk factors, respectively.19 In another case-control study of 2 34 35 266 Chinese adults, 32.5% and 41.1% of the participants presented three and four or more risk

36 http://bmjopen.bmj.com/ 37 factors, respectively, for oesophageal or stomach cancer.18 Compared with these two studies 38 39 mentioned above,18 19 a much higher co-prevalence of risk factors was noticed in our present 40 41 study. The variations could change likely due to the difference in diagnostic criteria, the number 42 43 and kind of risk factors included in the research and the participants' age group. Overall, there are

44 on September 27, 2021 by guest. Protected copyright. 45 some other national, and worldwide studies on the co-prevalence of some common or specific 46 47 chronic diseases risk factors. Still, the study on modifiable UDTC risk factors is limited. 48 The factors associated with modifiable UDTC risk factors co-prevalence included gender, 49 50 age, marriage status, education, annual family income and household size. We found the 51 52 prevalence of modifiable UDTC risk factors co-prevalence was lower in women compared with 53 54 men, which was consistent with findings from other settings.19 23 37 The possible reason could be 55 56 Chinese men are less aware of self-protect for chronic diseases and have worse health-seeking 57 58 behaviour, and may also attend more social occasions, tend to consume more tobacco/cigarette 59 60 high-salt, high-fat and high-calorie food compared with women.23 31

12 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 27 BMJ Open

1 2 3 In addition, this study revealed that the prevalence of modifiable UDTC risk factors 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 19 23 24 37 6 co-prevalence was increased with age, which was consistent with previous studies. Studies 7 8 showed that in Nanbu, China, the co-prevalence of oesophageal cancer risk factors increased with 9 age may attribute to the lower level of awareness, practice and willingness for health among the 10 11 elder.19 Meanwhile, it is also shown in table 5 that being over 65 protects against having more 12 13 than four risk factors. A possible reason for this difference is as follows: with the ageing of the 14 15 body and the deterioration of organ function, an elder individual possesses a higher risk of health 16 17 disorder and has a greater demand for medical care. As a result, this creates more opportunities to 18 For peer review only 19 get diagnosed with some health screening, including UDTC. Correspondingly, the elderly have 20 21 more chance to get a healthy education from physicians than the younger.38 Our study also showed 22 23 that single participants had more prevalence of modifiable UDTC risk factors co-prevalence 24 25 compared with participants who were currently married, following a previous study.19 A possible 26 27 explanation is that being a single older resident comes with its own economic and emotional 28 29 challenges, contributing to the unhealthy habit of lifestyle and diet.39 40 30 31 Our study demonstrated that the level of socioeconomic status (SES, education, annual 32 33 family income) was positively associated with modifiable UDTC risk factors co-prevalence, 34 35 which was inconsistent with other reports.19 23 37 Residents with a higher level of SES are more

36 http://bmjopen.bmj.com/ 37 aware of control and prevention of chronic disease and have better health-seeking behaviour 38 39 compared with those with a lower level of SES.19 23 41 Moreover, the poor or lower education 40 38 41 participants may have relatively more inaccessibility and unaffordability to medical services. 42 43 This paradox may be due to most of the participants enrolled in our study were from rural areas,

44 on September 27, 2021 by guest. Protected copyright. 45 and their SES was generally low. However, it may also imply that the higher income may

46 42 47 contribute to unhealthy lifestyles, and knowledge alone may not be sufficient to change 48 unhealthy lifestyles. Therefore, the level of education and income are two essential SES factors for 49 50 modifiable UDTC risk factors co-prevalence. 51 52 It is, however, important to note that participants who have more than six family members 53 54 had a lower prevalence of modifiable UDTC risk factors co-prevalence compared with those who 55 56 have less than three ones. Changes in household size are bound to affect the adjustment of the 57 58 family diet. As the household size increase, it is more likely to increase dietary diversity (e.g. 59 60 fruits, vegetables and milk) every day.43 Besides, the affection, information and economic support

13 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 27

1 2 3 among family members will also increase significantly, which can adjust and correct the unhealthy 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 44 45 6 lifestyle of individuals. 7 8 Our study explored the prevalence and influencing factors of modifiable UDTC risk factors 9 co-prevalence in the UDTC high-risk area, Yangzhong City, based on the community-based 10 11 project for UDTC screening with a large sample size. Additionally, the physical measurement and 12 13 the data collection implemented by trained interviewers strictly according to standard protocol and 14 15 instrument increase the validity of our results. In order to reduce the prevalence and co-prevalence 16 17 rate of UTDC risk factors, the screening teams should focus on individuals with co-prevalence of 18 For peer review only 19 risk factors in screening and improve their unhealthy lifestyles continually through a range of 20 21 methods such as post-screening health education, personalized interventions, and disease 22 23 follow-up. The social impact of screening should be expanded to improve the compliance of 24 25 high-risk groups, thereby increasing the output of screening health benefits. Meanwhile, the 26 27 government should also help high-risk groups (especially the older and male groups) to improve 28 29 their health literacy and awareness of UTDC prevention through diversified education, motivation, 30 31 and publicity methods, such as health education, health talks, and mass media campaigns. By 32 33 guiding the culture of smoking, drinking, and other food culture, promote high-cultural groups to 34 35 transform their cognitive and economic advantages into advantages in UTDC prevention and

36 http://bmjopen.bmj.com/ 37 health care, and effectively change unhealthy habits. Besides, the government should focus on 38 39 single and residents with small household size in the high-risk groups in the process of health 40 41 education and the development of prevention strategies. The findings may also provide the 42 43 reference for departments in charge of the prevention and control of UDTC in Yangzhong City,

44 on September 27, 2021 by guest. Protected copyright. 45 Jiangsu province, and relevant departments in other UDTC high-risk areas (e.g. Linzhou, 46 47 Feicheng, Yanting). 48 There were also several possible limitations to our study. Firstly, a cross-section study cannot 49 50 exam the causality or temporal relationship between the co-prevalence of modifiable UDTC risk 51 52 factors and its influencing factors. Secondly, the modifiable UDTC risk factors included in our 53 54 study were self-reported by participants, which may contribute to recall and reporting bias, except 55 56 BMI. Our study results were from Yangzhong City only, and cannot be generalized to the other 57 58 high-risk areas and the whole of southeast China. Additionally, the study response rate was 59 60 relatively low, particularly among males, which may affect the results' representativeness. Finally,

14 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 27 BMJ Open

1 2 3 our study only focused on the eight modifiable UDTC risk factors, but there are far more than 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 eight risk factors for UDTC. Hence, further studies are needed. 7 8 Conclusion 9 In summary, this cross-sectional study shows that the prevalence and co-prevalence of modifiable 10 11 UDTC risk factors are high among participants in Yangzhong City. Our analyses indicate that 12 13 men, younger adults, single adults, and participants with higher SES or smaller household size are 14 15 susceptible to modifiable UDTC risk factors co-prevalence. Policies to prevent UDTC have 16 17 already been developed in the strategic plan and operational plan, however, the accuracy and 18 For peer review only 19 validity of implementing the undertaken policies are still insufficient. Consequently, extra 20 21 attention is required to pay to these high-risk groups during the progress of screening. Relative 22 23 departments also need to make effective public health programs targeting modifiable UDTC risk 24 25 factors that aim to decrease UDTC risk factors co-prevalence in high-risk groups from high-risk 26 27 areas of UDTC. 28 29 Acknowledgements We would like to thank all participants in the present study. We are 30 31 also grateful to the Upper Digestive Tract Cancer Early Diagnosis and Treatment Program. 32 33 34 35 Contribution XF and RC are joint first authors. ZLH, DFQ, WQW, GQW, JYZ, JJW, GS

36 http://bmjopen.bmj.com/ 37 and XW contributed to the study design. ZLH, QZ, AWS, TQS performed the survey and 38 39 collected study data. XF and RC wrote the manuscript, and ZLH, DFQ, WQW, JYZ, JJW, GS, 40 41 XW were responsible for manuscript revision. 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 Funding This study was supported by the National key research and development program 46 47 (Grant NO. 2016YFC0901400; 2016YFC1302800); National Natural Science Foundation of 48 China (Grant NO. 81974493); and National Key Technology R&D Program (Grant NO. 49 50 2006BAI02A05). 51 52 53 54 Competing interests None declared. 55 56 57 58 Patient consent Obtained. 59 60

15 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 27

1 2 3 Ethics approval This study was approved by the academic and ethical committee of the 4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 6 cancer hospital of Chinese academy of medical sciences. 7 8 9 Provenance and peer review Not commissioned; externally reviewed. 10 11 12 13 Data sharing statement Data are available upon reasonable request. 14 15 16 17 Open access This is an Open Access article distributed in accordance with the Creative 18 For peer review only 19 Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to 20 21 distribute, remix, adapt, build upon this work non-commercially, and license their derivative 22 23 works on different terms, provided the original work is properly cited, and the use is 24 25 non-commercial. See:http://creativecommons.org/licenses/by-nc/4.0/ 26 27 28 29 30 31 32 33 34 REFERENCES 35 [1] World Health Organization. World health statistics 2018: monitoring health for the SDGs,

36 http://bmjopen.bmj.com/ sustainable development goals. Geneva: World Health Organization, 37 38 2018.https://www.who.int/gho/publications/world_health_statistics/2018/en/. 39 [2] Bray F, Ferlay J, Soerjomataram I, et al. Global cancer statistics 2018: GLOBOCAN 40 estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J 41 42 Clin 2018;68:394-424. 43 [3] Zheng RS, Sun KX, Zhang SW, et al. Report of cancer epidemiology in China, 2015. Chin J

44 Oncol 2019;41:19-28.In Chinese. on September 27, 2021 by guest. Protected copyright. 45 46 [4] Chen WQ, Zheng RS, Baade PD, et al. Cancer statistics in China, 2015. CA Cancer J Clin 47 2016;66:115-32. 48 [5] Ajani JA, Bentrem DJ, Besh S, et al. Gastric cancer, version 2.2013: featured updates to the 49 NCCN Guidelines. J Natl Compr Canc Netw 2013;11:531-46. 50 51 [6] Ciocirlan M, Lapalus MG, Hervieu V, et al. Endoscopic mucosal resection for squamous 52 premalignant and early malignant lesions of the esophagus. Endoscopy 2007;39:24-9. 53 [7] Ma D, Yang F, Liao Z, et al. Expert opinion on early screening and endoscopic diagnosis and 54 55 treatment esophageal cancer in China (2014, Beijing). China Prac Med 2015;4:320-37.In 56 Chinese. 57 [8] Chen R, Ma S, Guan C, et al. The National Cohort of Esophageal Cancer-Prospective Cohort 58 59 Study of Esophageal Cancer and Precancerous Lesions based on High-Risk Population in 60 China (NCEC-HRP): study protocol. BMJ Open 2019;9:e027360.

16 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 27 BMJ Open

1 2 3 [9] Salaspuro M. Interrelationship between alcohol, smoking, acetaldehyde and cancer. Novartis

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 Found Symp 2007;285:80-9. 6 [10] Somi MH, Mousavi SM, Naghashi S, et al. Is there any relationship between food habits in 7 the last two decades and gastric cancer in North-Western Iran? Asian Pac J Cancer Prev 8 2015;16:283-90. 9 10 [11] Navarro-Silvera SA, Mayne ST, Risch HA, et al. Principal component analysis of dietary and 11 lifestyle patterns in relation to risk of subtypes of esophageal and gastric cancer. Ann 12 Epidemiol 2011;21:543-50. 13 14 [12] Andrici J, Eslick GD. Hot Food and Beverage Consumption and the Risk of Esophageal 15 Cancer: A Meta-Analysis. Am J Prev Med 2015;49:952–60. 16 [13] Gupta B, Kumar N, Johnson NW. Relationship of Lifetime Exposure to Tobacco, Alcohol 17 18 and Second For Hand Tobacco peer Smoke review with Upper aero-digestive only tract cancers in India: a 19 Case-Control Study with a Life-Course Perspective. Asian Pac J Cancer Prev 20 2017;18:347-56. 21 [14] Chen WQ, ZUO TT. Initial effect achievement of battles on upper digestive tract cancer in 22 23 China. Chin J Prev Med 2017; 51:378-80.In Chinese. 24 [15] Luan DC, Li SJ, Li H, et al. Change trends in health behaviors among residents in Liaoning 25 province,1991-2006. Chin J Public Health 2013;29:1509-11.In Chinese. 26 27 [16] McCormack VA, Boffetta P. Today's lifestyles, tomorrow's cancers: trends in lifestyle risk 28 factors for cancer in low- and middle-income countries. Ann Oncol 2011 22:2349–57. 29 [17] Li FX, Robson PJ, Chen Y, et al. Prevalence, trend, and sociodemographic association of five 30 31 modifiable lifestyle risk factors for cancer in Alberta and Canada. Cancer Causes Control 32 2009;20:395-407. 33 [18] Gu XP, Wang YC, Zhi HK, et al. Risk factors of esophageal and stomach cancer and their 34 ： 35 clustering in Dafeng municipality a case-control study. Chin J Public Health 2016;

36 32:1406-09.In Chinese. http://bmjopen.bmj.com/ 37 [19] He Q, Jing YH, Huang HR, et al. Prevalence and clustering of esophageal cancer-related risk 38 factors among rural residents in Nanbu County, Sichuan Province. Chin H Cancer Prev Treat 39 40 2019 26:1675-80.In Chinese. 41 [20] Tong HY, Zhang MM, Zhang HY. Analysis on the epidemiology trend and disease burden of 42 esophagus cancer in Yangzhong from 2004 to 2015. Modern Prev Med 43 2016;43:3665-68,3687.In Chinese. 44 on September 27, 2021 by guest. Protected copyright. 45 [21] Tong HY, Zhang MM, Sun LP, et al. Quantitative study on death caused by main chronic 46 diseases in Yangzhong city. Jiangsu J Prev Med 2017;28:502-4,508.In Chinese. 47 [22] DONG ZW. China's cancer screening and early diagnosis and treatment technology 48 49 program. Beijing: People's medical publishing house 2009.In Chinese. 50 [23] Hong X, Ye Q, He J, et al. Prevalence and clustering of cardiovascular risk factors: a 51 cross-sectional survey among Nanjing adults in China. BMJ Open 2018;8:e020530. 52 53 [24] Ni W, Weng RX, Yuan X, et al. Clustering of cardiovascular disease biological risk factors 54 among older adults in Shenzhen City, China: a cross-sectional study. BMJ Open 55 2019;9:e024336. 56 57 [25] Department of disease control, ministry of health, PRC. Guidelines for the prevention and 58 control of overweight and obesity in Chinese adults. Beijing: people's medical publishing 59 house, 2006:2-4.In Chinese. 60

17 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 27

1 2 3 [26] Howitt C, Hambleton IR, Rose AM, et al. Social distribution of diabetes, hypertension and

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 related risk factors in Barbados: a cross-sectional study. BMJ Open 2015;5:e008869. 6 [27] Chinese nutrition society. Dietary guidelines for Chinese residents (2016). Beijing: people's 7 medical publishing house, 2016.In Chinese. 8 [28] Cureau FV, Duarte P, dos Santos DL, et al. Clustering of risk factors for noncommunicable 9 10 diseases in Brazilian adolescents: prevalence and correlates. J Phys Act Health 2014; 11 11:942-9. 12 [29] Xu F, Yin XM, Zhang M, et al. Family average income and body mass index above the 13 14 healthy weight range among urban and rural residents in regional Mainland China. Public 15 Health Nutr 2005;8:47-51. 16 [30] National Commission of Health Bureau of disease control and Prevention. Report on China 17 18 national nutritionFor and peerchronic disease review survey (2015). Beijing: only People’s Medical Publishing 19 House 2015;11:33-50.In Chinese. 20 [31] Wang R, Zhang P, Gao C, et al. Prevalence of overweight and obesity and some associated 21 factors among adult residents of northeast China: a cross-sectional study. BMJ Open 22 23 2016;6:e010828. 24 [32] Howitt C, Hambleton IR, Rose AMC, et al. Social distribution of diabetes, hypertension and 25 related risk factors in Barbados: a cross-sectional study. BMJ Open 2015;5:e008869. 26 27 [33] Millwood IY, Walters RG, Mei XW, et al. Conventional and genetic evidence on alcohol and 28 vascular disease aetiology: a prospective study of 500 000 men and women in China. Lancet 29 2019;393:1831-42. 30 31 [34] Msambichaka B, Eze IC, Abdul R, et al. Insufficient Fruit and Vegetable Intake in a Low- 32 and Middle-Income Setting: A Population-Based Survey in Semi-Urban Tanzania. Nutrients 33 2018;10:222. 34 [35] Guo YL, Tan XD, Liu XZ, et al. Fruit and vegetable intake of adults and its influencing 35

36 factors in some cities of Hubei province . J of Pub health and Prev Med 2016;27:82-5.In http://bmjopen.bmj.com/ 37 Chinese. 38 [36] Wen JB, Sun ZM, Miao DD, et al. Influencing factors about early cancer of upper-digestive 39 40 tract among high-risk population in Huai’an City of Jiangsu Province. China cancer 41 2019;28:749-56.In Chinese. 42 [37] Li Y, Zhang M, Jiang Y, Wu F. Co-variations and clustering of chronic disease behavioral 43 risk factors in China: China Chronic Disease and Risk Factor Surveillance, 2007. PLoS One 44 on September 27, 2021 by guest. Protected copyright. 45 2012;7:e33881. 46 [38] Ahmed S, Tariqujjaman M, Rahman MA, et al. Inequalities in the prevalence of undiagnosed 47 hypertension among Bangladeshi adults: evidence from a nationwide survey. Int J Equity 48 49 Health. 2019,18:33. 50 [39] Floud S, Balkwill A, Canoy D, et al. Marital status and ischemic heart disease incidence and 51 mortality in women: a large prospective study. BMC Med 2014;12:42. 52 53 [40] Cao Z, Wang R, Cheng Y, et al. Adherence to a healthy lifestyle counteracts the negative 54 effects of risk factors on all-cause mortality in the oldest-old. Aging (Albany NY) 55 2019;11:7605-19. 56 57 [41] Prom-Wormley EC, Clifford JS, Bourdon JL, et al. Developing community-based health 58 education strategies with family history: Assessing the association between community 59 resident family history and interest in health education. Soc Sci Med 2021;271:112160. 60

18 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 27 BMJ Open

1 2 3 [42] Yang F, Qian D, Liu X, et al. Socioeconomic disparities in prevalence, awareness, treatment,

4 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from 5 and control of hypertension over the life course in China. Int J Equity Health 2017;16:100. 6 [43] Workicho A, Belachew T, Feyissa GT, et al. Household dietary diversity and Animal Source 7 Food consumption in Ethiopia: evidence from the 2011 Welfare Monitoring Survey. BMC 8 Public Health 2016;16:1192. 9 10 [44] Bot SD, Mackenbach JD, Nijpels G, et al. Association between Social Network 11 Characteristics and Lifestyle Behaviours in Adults at Risk of Diabetes and Cardiovascular 12 Disease. PLoS One 2016;11:e0165041. 13 14 [45] Verheijden MW, Bakx JC, van Weel C, et al. Role of social support in lifestyle-focused 15 weight management interventions. Eur J Clin Nutr 2005;59 Suppl 1:S179-S186. 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 27, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

19 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

BMJ Open Page 22 of 27

1 2 3 4 5 Table legends 6 7 Table 1 Socio-demographic and socio-economic characteristics of participants in Yangzhong City, China 8 9 Category Total(n=19213) Men(n=8268) Women(n=10945) t/ χ² value P value 10 11 Number, n(%) 19213(100.0) 8268(43.0) 10945(57.0) 12 For peer review only 13 Age, years, mean(SD) 53.2±7.8 53.5±7.8 52.9±7.8 5.61 <0.001 14 Age group, n(%) 33.29 <0.001 15

40-44 3142(16.4) 1270(15.4) 1872(17.1) http://bmjopen.bmj.com/ 16 17 45-49 3937(20.5) 1642(19.9) 2295(21.0) 18 50-54 3743(19.5) 1560(18.9) 2183.0(19.9) 19 55-59 3571(18.6) 1582(19.1) 1989(18.2) 20 21 60-64 3088(16.1) 1416(17.1) 1672(15.3) 22 65-69 1732(9.0) 798(9.7) 934(8.5) 23 24 Marrige, n(%) on September 27, 2021 by guest. Protected copyright. 93.41 <0.001 25 Single 203(1.1) 137(1.7) 66(0.6) 26 Currently married 18285(95.2) 7907(95.6) 10378(94.8) 27 28 Divorced, widowed or separated 725(3.8) 224(2.7) 501(4.6) 29 Education, n(%) 886.879 <0.001 30 No institutional education 1594(8.3) 228(2.8) 1366(12.5) 31 32 Primary school 8510(44.3) 3436(41.6) 5074(46.4) 33 Junior high school 7591(39.5) 3647(44.1) 3944(36.0) 34 Senior high school and higher 1518(7.9) 957(11.6) 561(5.1) 35 36 Annual family income, n(%) 0.04 0.981 37 lower 8585(44.7) 3689(44.6) 4896(44.7) 38 39 40 41 42 20 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Page 23 of 27 BMJ Open

1 2 3 4 5 Middle 5420(28.2) 2338(28.3) 3082(28.2) 6 Higher 5208(27.1) 2241(27.1) 2967(27.1) 7 8 Household size, n, mean(SD) 4.0±1.4 4.0±1.4 4.0±1.4 -3.17 0.002 9 BMI,kg/m2, mean(SD) 23.8±3.0 23.9±3.0 23.7±3.0 3.47 0.001 10 11 12 For peer review only 13 14 Table 2 Prevalence of modifiable UDTC risk factors in Yangzhong City adults aged 40-69 15 16 http://bmjopen.bmj.com/ 17 Factor Total (n=19213) Men (n=8268) Women (n=10945) χ² value P value 18 19 BMI, n(%) 2.62 0.106 20 21 Overweight or obesity 8695(45.3) 3797(45.9) 4898(44.8) 22 Normal weight or underweight 10518(54.7) 4471(54.1) 6047(55.2) 23

Current smoking, n(%) on September 27, 2021 by guest. Protected copyright. 6093.25 <0.001 24 25 No 14589(75.9) 3988(48.2) 10601(96.9) 26 Yes 4624(24.1) 4280(51.8) 344(3.1) 27 Excessive alcohol consumption, n(%) 3197.31 <0.001 28 29 No 16109(83.8) 5504(66.6) 10605(96.9) 30 Yes 3104(16.2) 2764(33.4) 340(3.1) 31 insufficient vegetables intake, n(%) 0.32 0.572 32 33 No 6519(33.9) 2787(33.7) 3732(34.1) 34 Yes 12694(66.1) 5481(66.3) 7213(65.9) 35 36 insufficient fruit intake, n(%) 0.10 0.749 37 No 1055(5.5) 459(5.6) 596(5.4) 38 39 40 41 42 21 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

BMJ Open Page 24 of 27

1 2 3 4 5 Yes 18158(94.5) 7809(94.4) 10349(94.6) 6 Pickled food consumption, n(%) 8.38 <0.05 7 8 No 6138(31.9) 2734(33.1) 3404(31.1) 9 Yes 13075(68.1) 5534(66.9) 7541(68.9) 10 Fried food consumption, n(%) 1.00 0.318 11 12 No For peer12293(64.0) review5323(64.4) only6970(63.7) 13 Yes 6920(36.0) 2945(35.6) 3975(36.3) 14 Hot food consumption, n(%) 27.10 0.001 15 < 16 No 2221(11.6) 1070(12.9) 1151(10.5) http://bmjopen.bmj.com/ 17 Yes 16992(88.4) 7198(87.1) 9794(89.5) 18 19 20 21 22 23 Table 3 Prevalence of modifiable UDTC risk factors by relevant characters in Yangzhong City adults aged 40-69 24 Category Overweight or obesity Smoking Drinking Vegetables Fruit Pickled on September 27, 2021 by guest. Protected copyright. food Fried food Hot food 25 Age group,years, n(%) 26 27 40-44 1492(47.5) 533(17.0) 348(11.1) 1888(60.1) 2902(92.4) 2033(64.7) 782(24.9) 2971(94.6) 28 45-49 1953(49.6) 827(21.0) 540(13.7) 2648(67.3) 3735(94.9) 2594(65.9) 1451(36.9) 3894(98.9) 29 30 50-54 1729(46.2) 1068(28.5) 723(19.3) 2506(67.0) 3538(94.5) 2669(71.3) 1325(35.4) 3723(99.5) 31 55-59 1607(45.0) 987(27.6) 695(19.5) 2314(64.8) 3380(94.7) 2366(66.3) 1312(36.7) 3464(97.0) 32 60-64 1271(41.2) 796(25.8) 537(17.4) 2131(69.0) 2949(95.5) 2158(69.9) 1285(41.6) 2343(75.9) 33 34 65-69 643(37.1) 413(23.8) 261(15.1) 1207(69.7) 1654(95.5) 1255(72.5) 765(44.2) 597(34.5) 35 P value <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 36 Marrige, n(%) 37 38 Single 77(37.9) 73(36.0) 43(21.2) 177(87.2) 200(98.5) 169(83.3) 143(70.4) 186(91.6) 39 40 41 42 22 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Page 25 of 27 BMJ Open

1 2 3 4 5 Currently married 8334(45.6) 4439(24.3) 2979(16.3) 12087(66.1) 17253(94.4) 12424(67.9) 6518(35.6) 16304(89.2) 6 Divorced, widowed or separated 284(39.2) 112(15.4) 82(11.3) 430(59.3) 705(97.2) 482(66.5) 259(35.7) 502(69.2) 7 8 P value <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 9 Education, n(%) 10 No institutional education 692(43.4) 75(4.7) 76(4.8) 781(49.0) 1479(92.8) 1286(80.7) 675(42.3) 1198(75.2) 11 12 Primary school For3699(43.5) peer2030(23.9) review1328(15.6) 5898(69.3) only8149(95.8) 4785(56.2) 2205(25.9) 7194(84.5) 13 Junior high school 3521(46.4) 2045(26.9) 1373(18.1) 4878(64.3) 7087(93.4) 5819(76.7) 3146(41.4) 7187(94.7) 14 Senior high school and higher 783(51.6) 474(31.2) 327(21.5) 1137(74.9) 1443(95.1) 1185(78.1) 894(58.9) 1413(93.1) 15 16 P value <0.001 <0.001 <0.001 <0.001 <0.001 http://bmjopen.bmj.com/ <0.001 <0.001 <0.001 17 Annual family income, n(%) 18 19 lower 3711(43.2) 2128(24.8) 1551(18.1) 4844(56.4) 8234(95.9) 5535(64.5) 1300(15.1) 7168(83.5) 20 Middle 2458(45.4) 1303(24.0) 857(15.8) 3566(65.8) 4953(91.4) 2923(53.9) 1521(28.1) 4876(90.0) 21 Higher 2526(48.5) 1193(22.9) 696(13.4) 4284(82.3) 4971(95.4) 4617(88.7) 4099(78.7) 4948(95.0) 22 23 P value <0.001 <0.05 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 24 Household size, n(%) on September 27, 2021 by guest. Protected copyright. 25 0-3 3583(45.6) 1916(24.4) 1381(17.6) 5137(65.4) 7384(94.1) 5512(70.2) 2989(38.1) 7051(89.8) 26 27 4-5 4285(44.6) 2285(23.8) 1428(14.8) 6574(68.4) 9159(95.2) 6216(64.6) 3153(32.8) 8467(88.0) 28 ≥6 827(47.4) 423(24.2) 295(16.9) 983(56.3) 1615(92.5) 1347(77.1) 778(44.6) 1474(84.4) 29 P value 0.064 0.601 0.001 0.001 0.001 0.001 0.001 0.001 30 < < < < < < 31 Notes: Smoking, current smoking; Drinking, excessive alcohol consumption; Vegetables, insufficient vegetables intake, Fruit, insufficient fruit intake; Pickled food, 32 the consumption of pickled food, Fried food, the consumption of fried food; Hot food, the consumption of hot food 33 34 35 36 37 38 39 40 41 42 23 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

BMJ Open Page 26 of 27

1 2 3 4 5 6 Table 4 The different number and co-prevalence of modifiable UDTC risk factors in residents by relevant characters in Yangzhong City adults aged 40-69 7 8 Category None(0) single(1) Two(2) Three(3) Non-co-prevalence(≤3) Co-prevalence(≥4) χ² value P value 9 Total 14(0.1) 244(1.3) 1235(6.4) 4336(22.6) 5829(30.3) 13384(69.7) 10 Gender, n(%) 632.397 <0.001 11 12 men 2(0.0)For64(0.8) peer356(4.3) review1293(15.6) 1715(20.7) only 6553(79.3) 13 women 12(0.1) 180(1.6) 879(8.0) 3043(27.8) 4114(37.6) 6831(62.4) 14 Age group, years, n(%) 373.748 <0.001 15 16 40-44 1(0.0) 66(2.1) 193(6.1) 891(28.4) 1151(36.6) 1991(63.4)http://bmjopen.bmj.com/ 17 45-49 0(0.0) 26(0.7) 165(4.2) 857(21.8) 1048(26.6) 2889(73.4) 18 19 50-54 0(0.0) 15(.4) 112(3.0) 759(20.3) 886(23.7) 2857(76.3) 20 55-59 0(0.0) 48(1.3) 175(4.9) 724(20.3) 947(26.5) 2624(73.5) 21 60-64 0(0.0) 33(1.1) 273(8.8) 712(23.1) 1018(33.0) 2070(67.0) 22 23 65-69 13(0.8) 56(3.2) 317(18.3) 393(22.7) 779(45.0) 953(55.0) 24 Marrige, n(%) on September 27, 2021 by guest. Protected copyright. 67.346 <0.001 25 Single 0(0.0) 0(0.0) 9(4.4) 16(7.9) 25(12.3) 178(87.7) 26 27 Currently married 12(0.1) 201(1.1) 1130(6.2) 4167(22.8) 5510(30.1) 12775(69.9) 28 Divorced, widowed or separated 2(0.3) 43(5.9) 96(13.2) 153(21.1) 294(40.6) 431(59.4) 29 Education, n(%) 417.766 <0.001 30 31 No institutional education 4(0.3) 56(3.5) 241(15.1) 379(23.8) 680(42.7) 914(57.3) 32 Primary school 8(0.1) 71(0.8) 503(5.9) 2396(28.2) 2978(35.0) 5532(65.0) 33 34 Junior high school 2(0.0) 107(1.4) 428(5.6) 1367(18.0) 1904(25.1) 5687(74.9) 35 Senior high school and higher 0(0.0) 10(0.7) 63(4.2) 194(12.8) 267(17.6) 1251(82.4) 36 Annual family income, n(%) 1078.75 <0.001 37 38 lower 6(0.1) 102(1.2) 588(6.8) 2500(29.1) 3196(37.2) 5389(62.8) 39 40 41 42 24 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Page 27 of 27 BMJ Open

1 2 3 4 5 Middle 5(0.1) 105(1.9) 470(8.7) 1403(25.9) 1983(36.6) 3437(63.4) 6 Higher 3(0.1) 37(0.7) 177(3.4) 433(8.3) 650(12.5) 4558(87.5) 7 8 Family member, n(%) 17.913 <0.001 9 0-3 5(0.1) 165(2.1) 438(5.6) 1642(20.9) 2250(28.7) 5600(71.3) 10 4-5 7(0.1) 68(0.7) 627(6.5) 2318(24.1) 3020(31.4) 6597(68.6) 11 12 ≥6 2(0.1)For11(0.6) peer170(9.7) review376(21.5) 559(32.0) only 1187(68.0) 13 14 15 16 http://bmjopen.bmj.com/ 17 Table 5 The multivariable logistic regression analysis of modifiable UDTC risk factors co-prevalence 18 19 Category Wald P value β SE OR 95% CI 20 Gender, n(%) 21 women - - - - 1.000 - 22 23 men 533.13 <0.001 0.834 0.036 2.302 2.145 to 2.471 on September 27, 2021 by guest. Protected copyright. 24 Age group, years, n(%) 25 40-44 - - - - 1.000 - 26 27 45-49 48.302 <0.001 0.381 0.055 1.464 1.315 to 1.631 28 50-54 153.543 <0.001 0.715 0.058 2.044 1.825 to 2.288 29 30 55-59 143.780 <0.001 0.710 0.059 2.033 1.811 to 2.283 31 60-64 32.963 <0.001 0.351 0.061 1.421 1.260 to 1.602 32 65-69 10.685 <0.001 -0.229 0.070 0.796 0.694 to 0.913 33 34 Marrige, n(%) 35 Single - - - - 1.000 - 36 Currently married 9.054 0.003 -0.671 0.223 0.511 0.330 to 0.792 37 38 Divorced, widowed or separated 10.813 0.001 -0.777 0.236 0.460 0.289 to 0.731 39 40 41 42 25 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

BMJ Open Page 28 of 27

1 2 3 4 5 Education, n(%) 6 No institutional education - - - - 1.000 - 7 8 Primary school 17.448 <0.001 0.256 0.061 1.291 1.145 to 1.456 9 Junior high school 51.436 <0.001 0.485 0.068 1.624 1.423 to 1.855 10 Senior high school and higher 40.562 <0.001 0.604 0.095 1.829 1.519 to 2.202 11 12 Annual family income, n(%) For peer review only 13 Higher - - - - 1.000 - 14 Middle 0.001 15 874.464 < -1.524 0.052 0.218 0.197 to 0.241 16 lower 799.154 <0.001 -1.502 0.053 http://bmjopen.bmj.com/ 0.223 0.201 to 0.247 17 Household size, n(%) 18 19 0-3 - - - - 1.000 - 20 4-5 82.359 0.429 -0.334 0.037 0.716 0.666 to 0.770 21 ≥6 64.364 <0.001 -0.514 0.064 0.598 0.527 to 0.678 22 23 Constant 64.447 <0.001 1.878 0.234 - - 24 on September 27, 2021 by guest. Protected copyright. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 26 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-042006 on 7 April 2021. Downloaded from

Page 29 of 27 BMJ Open

1 2 3 4 5 6 7 8 9 10 11 12 For peer review only 13 14 15 16 http://bmjopen.bmj.com/ 17 18 19 20 21 22 23 24 on September 27, 2021 by guest. Protected copyright. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 27 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60