This study was approved by the Ethics Committee of Tongji Medical College, Huazhong University of Science and Technology (Number: 2018IECS175). All participants were informed about the goal of this study and informed consent was obtained from electronic signatures of all participants. Data were collected anonymously. All methods were carried out in accordance with the Declaration of Helsinki.
Study design and sample size
A cross-sectional survey was designed and conducted among Chinese antibiotic users from July 1, 2018 to September 30, 2018, based on the China’s largest online survey platform (Questionnaire Star, https://www.wjx.cn). Questionnaire Star is a professional service platform for electronic questionnaire design and data collection, which has been widely used by researchers. This study focused on following three aspects of inappropriate use behaviors: SMA, SSA, and NAAT. The NAAT consisted of four types of non-adherence behaviors including missing antibiotics, increasing antibiotic dosage, decreasing antibiotic dosage, and discontinuing antibiotics. To ensure the stability and reliability of the survey results, the sample size should be as large as possible. In our study, the maximum sample size was calculated with the minimum prevalence of these target behaviors. We used the formula:\(n={z}^{2}p(1-p)/{d}^{2}\), where n is the sample size, p is the prevalence of the research target, z is the normal deviation (1.96) and d is the margin of error (d = 0.1p). When determining parameter p, we took the prevalence of target behaviors among publics in the past 6 months as the criterion according to previous studies. Among them, the prevalence of non-adherence behavior with proactively increasing antibiotic dosage was the lowest with 8.3%–22.0% [17, 19]. Therefore, we took the median value and set the confidence interval (CI) at 95% to calculate the sample size of 2152. Considering the possibility of invalid surveys (about 15%), at least 2532 participants should be surveyed in each region.
According to the China Health Statistical Yearbook (2018) [20], the population who participated in the online questionnaire were divided into three regions: eastern, central and western China. To ensure the representativeness of the survey samples in each region, the administrator would close the network link and end the questionnaire survey when the number of survey samples in each region reached 2532 participants.
Measurements
To reflect the prevalence of inappropriate use behaviors of antibiotics among Chinese antibiotic users, we selected three main outcomes: the percentage of SMA, the percentage of SSA, and the percentage of NAAT. In this study, SMA was defined as answering “yes” to the question: “During the past 6 months, did you self-administer antibiotics without a doctor’s prescription?” SSA was defined as answering “yes” to the question: “During the past 6 months, did you self-storage antibiotics at home for future use?” NAAT was measured according to four types of non-adherence behaviors that were developed from previous studies [21]: (1) Did you ever miss medication during the course of antibiotic treatment? (2) Did you ever increase the dosage by yourself during the course of antibiotic treatment? (3) Did you ever reduce the dosage by yourself during the course of antibiotic treatment? (4) Once symptoms disappear, do you immediately discontinue the course of antibiotics? Each item measures a type of specific non-adherence behavior, and dichotomous responses (Yes/No) are captured. The item is scored with 0 point if the response is “Yes” and scored with 1 point if the answer is “No”. The total scores range from 0 to 4, with the higher the score, the higher the antibiotic user’s adherence. The total scores ≤ 2 are considered as “non-adherence”, and the total scores > 2 are considered as “adherence.”
To identify the relevant factors of inappropriate use behaviors of SMA, SSA, and NAAT, we also collected the demographic and sociological characteristics, including gender, age, place of residence (urban/rural), educational level (junior high school or below/ high school/junior college or above), self-perceived economic status (good/ average/poor), self-perceived health status (good/ average/poor), and antibiotic knowledge. Antibiotic knowledge is assessed by asking 10 questions in three sections: the knowledge regarding antibiotic role, antibiotic use and antibiotic resistance. These 10 questions are estimated using a scoring scheme, with score of 1 for a correct response and 0 for an incorrect response. The total scores of knowledge range between 0 and 10. We divided the total scores of antibiotic knowledge into three levels as following: (1) high level: 8–10 scores; (2) medium level: 3–7 scores; (3) low level: 0–2 scores. In the questionnaire, Cronbach’s α coefficients of non-adherence behavior scale and antibiotic knowledge scale were 0.71 and 0.84 (> 0.70), respectively. Confirmatory factor analyses showed that the standardized factor loading of each item of non-adherence behavior scale ranged from 0.52 to 0.57 (> 0.40), and these values of the antibiotic knowledge scale were all above 0.70. The results of these tests indicated that the questionnaire had good reliability and validity.
Data collection
We designed a questionnaire and generated a valid QR code of this questionnaire on the online survey platform. Then, we recruited 50 graduate and undergraduate students from the School of Public Health at Huazhong University of Science and Technology as investigators in July 2018. Four of them refused the request to distribute questionnaires, with a participation rate of 92%. After the unified training, the investigators would get the QR code of questionnaire. During the summer vacation, all sent out electronic questionnaires through their personal social software such as “WeChat”. The respondents scanned the QR code on their mobile phones to fill in the questionnaire. The data administrator checked the questionnaire and sorted out the data through the background every week. When each region (eastern, central and western China) reached the minimum required survey sample of 2532, the administrator closed the network link and finished the survey.
Before the survey, we conducted a questionnaire pilot test to ensure the smooth implementation of all steps including the respondents’ login on the platform, questionnaire filling and submission. we set the following inclusion criteria that the respondents filling in the questionnaire must meet: (1) adults aged 18 years or above; (2) living in the area for more than 6 months; (3) normal cognitive and understanding abilities; (4) aware of the term “antibiotic” and having used antibiotics. To ensure that only one questionnaire per respondent can be filled out, we limited one IP address per submission on the online survey platform and required to log in with a personal mobile phone number (in China, the mobile phone number is real-name system) before filling in the questionnaire. During the survey, to improve the enthusiasm of respondents, we paid 3 Yuan in cash to each person who completed the questionnaire as an economic incentive. To ensure that respondents must read and answer the questionnaire carefully, we set three quality inspection questions at different locations, namely: “Where is the capital of China?”, “What is 7 minus 2?”, and “What is 1 plus 3?” If the answer to any of these three questions was incorrect, the questionnaire would be marked as invalid. After the respondents completed the questionnaires, all the questionnaires were automatically input into a data file, and two researchers independently collated and verified the data to ensure the reliability of the survey data.
During the process of the survey, 21,874 respondents visited the questionnaire link, and 17,062 respondents actually completed the questionnaire. Since this was an online survey, the true response rate could not be determined here. Among all the questionnaires, 1280 questionnaires were excluded because of unqualified quality inspection, and 256 questionnaires were also excluded because of respondents under the age of 18. Finally, 15,526 valid questionnaires were obtained.
Statistical analysis
Descriptive analysis of the demographic and sociological characteristics, antibiotic knowledge, and various inappropriate use behaviors of the participants was conducted. The classified variables were expressed as frequency and percentage, and the continuous variables were expressed as mean and standard deviation (SD). To examine the possible impact of the controversial Q5 and Q9 from the antibiotic knowledge scale (Table 2) on the study results [22, 23], a sensitivity analysis was performed by deleting these two items. When these two items were removed, the total scores for antibiotic knowledge became between 0 and 8. The total scores of antibiotic knowledge were re-divided into three levels as following: (1) high level: 7–8 scores; (2) medium level: 3–6 scores; (3) low level: 0–2 scores. Furthermore, binary logistic regression model was used to explore the factors associated with the inappropriate use behaviors of SMA, SSA, and NAAT among antibiotic users. Adjusted odds ratio (aOR) and 95% CI for each variable were given. All analyses were performed using SAS 9.4 (SAS Inc., Cary, NC). Statistical tests were double-tailed; a P value < 0.05 was considered to be significant.