Ethic approval
This study was undertaken with ethic approval from the Peking University biomedical ethics committee, Duke University Institutional Review Board for Clinical Investigations and School of Population Health Research Ethics Committee at the University of Queensland.
Study sample
Study subjects were recruited from five different Chinese medical schools, Duke University in U.S., and the University of Queensland in Australia. All those schools were chosen for convenience based on established collaboration within the network of the Center of Excellence at the George Institute. Chinese students were majoring in clinical medicine, enrolled in 2nd, 4th and 7th year at Peking University Health Science Center (PUHSC) in Beijing, School of Medicine, Xi’an Jiaotong University (XJTU) in Xi’an, China Medical University (CMU) in Shenyang, Ningxia Medical University (NMU) in Ningxia, and Changzhi Medical College (CMC) in Changzhi. American students were majoring in clinical medicine, enrolled in the Doctor of Medicine (MD) or MD/PhD (MSTP) program at Duke University, School of Medicine in Durham, North Carolina, U.S.. Australian students were majoring in clinical medicine, enrolled in 1st, 2nd, 3rd or 4thyear at the School of Medicine, the University of Queensland.
In terms of eligibility criteria, the participants had to have been enrolled before October 2012. This was because by the time the survey was administered, students enrolled by October 2012 would be in their 2nd or greater year at their school. Students were considered eligible if they were enrolled in their 2nd, 4th, or 7th year. These cohorts were selected to represent students from pre-clinical, new-clinical, and post-clinical populations from five different Chinese medical universities. The three cohorts of students began schooling in 2006, 2009 and 2011.
Based on a sample size calculation from the pilot study, the target sample size was 150 students from each grade, or 450 students in total, from the five medical universities in China. This was based on an estimate that 55 % students were considered to have positive attitudes towards preventive counselling from the pilot study (α = 0.05, absolute permissible error = 5 %, relative error = 10 % response rate = 70 %). The sample size was calculated as: N = U(1-α) 2/ Relative error2. We included all medical students from Duke University and from the University of Queensland. The total number of enrollments between October 2012 and November 2012 was 425 students from Duke University, 500 students from the University of Queensland, and 1568 students from the five Chinese medical schools. The non-response rates in the Duke University, University of Queensland, and five Chinese medical universities were 54.4, 50.5, and 79.2 %, respectively.
Survey tool and distribution
This study examined the prevalence of healthy lifestyles among medical students and its association with attitudes towards preventive counseling of NCDs using a 32-item, self-reported online survey that took medical students approximately 15 min to complete.
The web-based survey consisted of four sections: 1. general health status; 2. personal health behaviors including physical activity, diet, mental health, smoking, alcohol consumption, health goals and health care seeking behaviors; 3. health care opinions and attitudes; 4. personal information. Many of the questions were adapted from validated sources, including the U.S. Centers for Disease Control (CDC) Behavioral Risk Factor Surveillance System (BRFSS) [13], CDC National Health Interview Survey [14], NHIS Tobacco Questions: 1997–2010 [15], and International Physical Activity Questionnaire (IPAQ) [16]. The inclusion of questions in the survey was based on consultations with the research team at the George Institute in China, a leader in the field of medical students’ lifestyles. The authors also referred to the study “Healthy Doctor = Healthy Patient”, by Professor Erica Frank from the University of British Columbia [17]. There were also questions that have been developed by the research team.
Before the formal implementation of the survey, we conducted a paper-based pilot study at Peking University Health Science Center in April 2012 to assess the acceptance of the questionnaire and made modifications as needed. Also to assess the feasibility of online survey and the response rate, an online pilot survey was administered to students at Peking University Health Science Center in September 2012 and further modifications were made as needed.
The final questionnaire was distributed in Chinese to Chinese medical students and in English to American and Australian medical students. Care was taken to ensure the consistence between the Chinese version and English version of the questionnaire. If the source of the questions were from validated English questionnaires, the questions were translated to Chinese and back translation was done to ensure the consistence of the meaning. And vice versa, if the source of questions were in Chinese, they were translated to English and back translation was done. Students were sent an email with a link to the questionnaire and were instructed to complete it within 2 weeks. In an additional effort to maximize the response rate, each medical school administrator followed up emails by sending four reminder emails, 4 weeks apart, between October 2012 and November 2012. Students who completed the survey received monetary compensation for their time and effort.
The standard of healthy lifestyles (Exposure Variables)
The four domains of exposure variables were diet, exercise, smoking, and alcohol consumption. Those complying with dietary recommendations in the U.S. and Australia consumed more than five servings of vegetables and/or fruits per day [18]. In comparison, participants in China who reported consuming at least 200 grams of fruits and 300 grams of vegetables per day were considered to be in compliance with dietary recommendations [19, 20]. Participants who reported at least 150 min or more of moderate-to-vigorous physical activity per week were considered to be in compliance with exercise recommendations [21]. Participants were considered smokers if they reported having smoked at least 100 cigarettes in their lifetime and were currently still smoking cigarettes [22]. Alcohol consumption was divided into binge drinking and heavy drinking. Heavy drinkers were those who consumed, on average, at least one drink per day for women and at least two drinks per day for men in month [23]. Binge drinkers were those who reported drinking more than four drinks for women and more than five drinks for men in one sitting at any time during the past month. Participants who met dietary and exercise recommendations and who reported being non-smokers [15], non-heavy drinkers, and non-binge drinkers, were considered to have an overall healthy lifestyle.
Outcome measures
The primary outcome measure of this study was the association between the healthy lifestyles and the attitudes towards preventive counselling.
Statistical analysis
Univariate logistic regression analyses were performed to obtain an odds ratio (OR), 95 % confidence interval (95%CI) and P-value, to assess the association between healthy behaviors and attitudes towards preventive counseling among medical students at Duke University, the University of Queensland, and five medical schools in China. In addition, data was pooled according to country in order to compare findings from the U.S, Australia and China.
Differences in health behaviors and attitudes towards preventive counseling, due to level of education and level of professional development, may exist among students from different cohorts. Therefore, data was stratified by year to avoid effect modification. Prevalence estimates were obtained for all participants on self-reported body mass index, healthy behaviors (i.e., diet, exercise, smoking, and alcohol), and attitudes towards preventative counseling for each of these domains.
Multivariate logistic regression analyses were adjusted for gender to evaluate the association between healthy behaviors and attitudes towards preventative counseling between males and females. The main exposure variables (independent variables) were physical activity, balanced diet, weight control, stress management, smoking behavior and alcohol consumption. The main outcome variables (dependent variables) were students’ attitudes toward preventative counseling, which included basic knowledge of chronic diseases, stratified by year of training, country, and adjusted for gender; perceived adequacy of training; school’s promotion of each healthy habit; and so on. Collinearity between the domain variables was assessed by simple tabulation and any collinearity problems were noted in the tables. All analyses were performed using SPSS17.0 [24].