Health Education Services Utilization and Its Determinants Among Migrants : a Cross-sectional Study in Urban-rural Connection Districts of Beijing, China

Background: Domestic migration poses a challenge for China as migrants have little access to preventive healthcare services and are vulnerable to certain risks and diseases. This research sought to unveil and explore the determinant factors associated with health education utilization as a key aspect in basic public health services for migrants in Beijing, China. Methods: A sample of 863 inter-provincial migrants, 18 years old and above, was selected by three-stage stratied cluster sampling method in urban-rural fringe areas of Beijing during 2016 to 2017. Face-to-face structured interviews were conducted in the questionnaire survey. The effects of the explanatory variables on health education utilization from predisposing, enabling, health behaviors and need variables were used to demonstrate by Anderson health service utilization model. Results: The study revealed that 61.6% migrants desired to receive health education, while only 53.8% of them received in the past year. There were differences in the utilization and needs of health education among the migrants in different ages and genders. Many migrants desired to gain access to various types of health education information from the internet. Chi-square independence test lists such major determinant factors in migrants whole health education as age, "Hukou" registration system, marital status, education level, long-term residence plan in Beijing, 1 or more children in Beijing, employment status, housing source, average daily working time, exercises, health knowledge, smoking, self-rated health. The binary logistic regression indicates that the migrants with younger age, high education level, one or more children in Beijing, exercises and good self-rated health were more likely to receive whole health education. The results also show that average daily working time of enabling variables and exercise of health behavior variables were the strong and consistent determinants of three types of health education utilization, including communicable, non-communicable and occupational diseases. Conclusion: Gaps exist between the needs and utilization in health education and more attention should be given to the migrants with heavy workload and low education level. Feasible policies and measures, such as multiple health information channels, should be vigorously implemented to ensure equitable and easy access to health education for migrants.

control, and treatment for general public and for the target group, particularly in the marginalized and migrant populations [16][17][18]. It is also likely to be instrumental in effectively addressing growing health care costs and in preventing or mitigating the negative effects of migration on health systems and society. While previous study found that compared with relatively high use of medical care, preventive care was used less frequently among migrants [19].
However, current contents and traditional face-to-face education of health information ignore the actual utilization situation and fail to meet the needs of the general public with an increasing sense of health [20]. In order to implement successful policies to address social and health inequalities among the migrant populations, policy makers need to understand what barriers migrants face, and also need to identify and answer their health needs.
As the political, economic and cultural center of China, Beijing attracts tens of thousands of migrants from all over the country every year [21].
Systematic research on health education utilization behaviors and in uence factors of migrants is far from su cient. This research results are crucial to disease prevention and health promotion for the migrants in China. We aim to (1) evaluate differences between utilization behaviors and needs of health education; (2) put forth the potential major determinant factors of health education utilization behaviors in different sociodemographic, health behavior and health outcome setting for migrants in Beijing by using the simpli ed Anderson health service utilization model. The comparisons and inferences could help us gure out the obstacles in migrants health education, and take targeted intervention measures to improve health literacy, to control disease and to promote health status.

Analytic Framework
Andersen health service utilization model, a well-validated theoretical framework, can predict determinants of health services utilization, take into consideration both individual and societal determinants [22]. In the model, health education utilization is determined by three dynamics: predisposing, enabling, and need variances (PEN). Social demographic characteristics including sex, age, race, etc, can be divided into predisposing factors, which increase one's needs for health education services. For example, a person with strong belief in health education services for effective disease prevention is more likely to seek health education. Enabling factors are individual, family and community resources support, and can facilitate or impede the use of health education services. Need factors represent both actual and self-perceived needs for health education services. In our study, health behavior variances served as a key dynamics parameter and were integrated to evaluate the predictions of a new model. Also we used a feedback loop to illustrate the relationship between health education behaviors (seek health education, and not seek health education) and other aspects. (See gure 1) Health education seeking behaviors (seek health service or not) is determined by four dynamics: predisposing (demographic and social structure), enabling (individual and family resources), health behavior variances (health promotion behaviors and health hazard behaviors), and need variances (chronic diseases and self-evaluation general health).

Ethics Statement
The study was undertaken as a part of Beijing Philosophy and Social Science Planning Project, a population-based cross-sectional survey on risk factors of health status for migrants. It was approved by the Ethical Committee of Capital Medical University, Beijing, China. Data were collected from a cross-sectional survey in urban-rural fringe areas of Beijing in 2016. The parents or guardians are main decision-makers in public health services for children, thus those children under 18 were not included in this study. Written informed consent was obtained from each participant involved in this study. All participants' information will be anonymized and kept con dential.

Data acquisition and study population
A eldwork survey of Public Health Service Utilization of Migrants Population in Beijing Urban-Rural Fringe Areas was performed from June 2016 to January 2017. All respondents were at least 18 years old, including inter-provincial migrants residing or working in the sampling regions (for no less than six months).
Exclusion criteria consisted of the following: migrants who were not able to respond, those with mental health issues, and tourists in Beijing. The migrants dwell mainly in 5 (Chaoyang, Haidian, Fengtai, Daxing, Changping districts) out of 16 districts in Beijing. Five districts were divided into two types of region based on the number of migrants, including the region with more than 1 million migrants (Chaoyang, Haidian districts) and the region with 0.5 to 1 million migrants (Fengtai, Daxing, Changping districts). A sample of 1,000 migrants was chosen from two of the ve districts in Beijing by using strati ed three-stage cluster random sampling, as follows: Stage One, one district was chosen from the rst region (Haidian district), and the other district was chosen from the second region (Fengtai district). Stage Two, two streets were chosen respectively from each sampled districts according to the population size and social economic status. Stage Three, the total number of 1,000 migrants was recruited and investigated from four streets, with Haidian and Fengtai each 500 respectively. Data were collected via face-to-face interview and a total of 863 respondents were analyzed after excluding the data missing information on any of variables in the research. These samples can be considered representative of migrants population in Beijing as they presented similar distribution of age and gender status compared with those in the larger population, according to the report on China's migrant population development in 2015.
The structured questionnaire includes four parts of variances as follows: predisposing factors (sociodemographic characteristics), enabling factors (individual/family resources), need factors (general health status), health behavior factors (health promotion behaviors and health hazard behaviors) and health education seeking behavior (receiving health education or not).
In this survey, utilization and needs of health education were measured by responses of three questions: The health education scheme of Basic Public Health Service Program includes ve main forms, including receipt of "materials of health education", "health education bulletin board", "public health consultation", "health knowledge lecture", and "individualized health education" by population from CHSCs.
-Binary: 1 for received any form of health education at least once, 0 for did not receive.
2. What was/were the type/types of health education have you received?
The types of health education mainly consisted of "occupational disease prevention and therapy", "child healthcare", "antenatal, prenatal and postpartum healthcare", "communicable disease prevention and therapy", "non-communicable disease prevention and therapy", "adolescent healthcare", "menopause healthcare" and "aged healthcare", etc. Respondents should answer the question according to their utilization of health education.
3. What is/are the type/types of health education that you want to receive in the future?
The types of health education were same as in the question two. Respondents should answer the question according to their needs.
Quality-assurance measures for this survey include questionnaire evaluation, training investigators, and eldwork supervision to monitor the survey procedure.
It was reviewed, edited, and validated by experts from health administration and community health service institutions (CHSIs) prior to implementation. A trial survey covering 50 persons was implemented during June 6 to 11, 2016 for better understanding the questionnaire and the eldwork procedures. Double entry and validation were adopted for all data using EpiData software (Version 3.1, EpiData Association, Odense, Denmark). Discrepancies, compared and analyzed from the two databases, would be clari ed by reviewing the original data source.

Statistical analysis
A person was the unit of seeking health education at least once in the past year, rather than total number of times health education received. Chisquare independence test was used to analyze the differences in the categorical variables. Odds Ratios (ORs) and their 95% con dence intervals (CIs) were calculated using binary logistic regression analysis. In the rst step, descriptive statistics and chi-square independence test were calculated, strati ed by the categories of health education utilization (received and did not receive health education). In the second step, multivariate logistic regression analysis was performed to predict the potential major determinant factors in migrants health education utilization, and in possible confounding variables control. The full model consisted of all predisposing, enabling, health-related behaviors and the need factors were entered into the model. The full list of independent variables is summarized in the Table 1. All statistical analyses were performed using IBM Statistical Package for Social Science version 20.0 (SPSS Inc., Chicago IL, US) and all tests are two sided. The signi cance level for all analyses was set at P<0.05.

Utilization of health education in the past year
In this study, 61.6% migrants desired to receive health education, while only 53.8% of them received it. Sample characteristics were presented in Table 2. A total of 863 inter-provincial migrants (males 383; females 480) in Beijing above 18 years old were investigated. The age of migrants varies from 18 to 76 years old, average age 36.6±11.2 years old.
Descriptive statistics and chi-square independence test were used to describe the information and to analyze the in uence factors of health education utilization by sociodemographic factors. Chi-square independence test showed that age, "Hukou" registration system, marital status, education level, plan to reside for a long time in Beijing, have at least one child in Beijing, employment status, housing source, average daily working time, do exercises, health knowledge, smoking, self-evaluation general health status are the major determinants affecting migrants to receive health education. (See Table 2) Gender, "Hukou" registration system, education level, plan to reside for a long time in Beijing, have at least one child in Beijing, employment status, housing source, average daily working time, do exercises, acquire health knowledge are the major determinants affecting migrants to receive communicable disease health education. "Hukou" registration system, housing source, average daily working time, do exercises are the major determinants affecting migrants to receive non-communicable disease health education. Age, "Hukou" registration system, marital status, education level, plan to reside for a long time in Beijing, have at least one child in Beijing, employment status, average daily working time, do exercises, acquire health knowledge, smoking, are the major determinants affecting migrants to receive occupational disease health education. (See Table 3) Table 4 shows that the top ve types of health education received by male migrants were "occupational disease prevention and therapy" (19.1%), "child healthcare" (14.9%), "antenatal, prenatal and postpartum healthcare" (10.2%), "communicable disease prevention and therapy" (9.9%), and "non-communicable disease prevention and therapy" (9.1%). However, the top ve types of health education that male migrants wanted to receive were "non-communicable disease prevention and therapy" (58.2%), "communicable disease prevention and therapy" (46.7%), "child healthcare" (44.9%), "aged healthcare"(42.0%) and "occupational disease prevention and therapy" (36.8%). The top ve types of health education received by female migrants were "antenatal, prenatal and postpartum healthcare" (18.8%), "child healthcare" (16.5%), "communicable disease prevention and therapy" (14.4%), and "non-communicable disease prevention and therapy" (12.7%), "occupational disease prevention and therapy" (11.9%). However, the top ve types of health education that female migrants wanted to receive were "non-communicable disease prevention and therapy" (63.5%), "aged healthcare"(51.7%), "communicable disease prevention and therapy" (50.6%), "child healthcare" (48.3%) and "adolescent healthcare" (40.2%). Table 5 shows that the types of health education received by migrants in different age groups. "Occupational disease prevention and therapy" (30.9%), "communicable disease prevention and therapy" (15.5%), and "adolescent healthcare" (18.2%) were received by migrants in the age group of 18 to 24 years old more than other groups. "Antenatal, prenatal and postpartum healthcare" (24.4%) and "child healthcare" (22.8%) were received by migrants in the age group of 25 to 34 years old more than other groups. "Non-communicable disease prevention and therapy" (15.5%), "aged healthcare" (15.5%) and "menopause healthcare" (10.3%) were received by migrants in the age group of 55 years old and above more than other groups. Table 6 shows that the types of health education desired to receive by migrants in different age groups. "Occupational disease prevention and therapy" (54.5%), "adolescent healthcare" (54.5%), and "antenatal, prenatal and postpartum healthcare" (50.9%) were desired to receive by migrants in the age group of 18 to 24 years old more than other groups. "Non-communicable disease prevention and therapy" (37.6%), "aged healthcare" (81.0%) were wanted to receive by migrants in the age group of 55 years old and above more than other groups. "Child healthcare" (59.9%), "communicable disease prevention and therapy" (53.8%), and "menopause healthcare" (45.2%) were wanted to receive by migrants respectively in the age groups of 25 to 34 years old, 35 to 44 years old, and 45 to 54 years old more than other groups.
Additionally, the top three pathways through which migrants want to acquire health information were television broadcasting (62.8%), internet (webpage and WeChat) (58.2%), and professionals and health managers (37.4%).

Multivariate Logistic Regression Model
The ndings indicate that all of the full models were able to distinguish between migrants with receiving health education and those without receiving health education, and all predictors were statistically signi cant at the 0.000 level (Model , χ2=186.467, P=0.000; Model , χ2=49.367, P=0.000; Model , χ2=39.895, P=0.000; Model , χ2=90.941, P=0.000). In the model summary, Model explained between 19.6 per cent (Cox and Snell R square) and 26.3 per cent (Nagelkerke R square) of the variance in health education utilization as a whole. Model explained between 5.6 per cent (Cox and Snell R square) and 10.7 per cent (Nagelkerke R square) of the variance in communicable disease health education utilization as a whole. Model explained between 4.6 per cent (Cox and Snell R square) and 9.1 per cent (Nagelkerke R square) of the variance in non-communicable disease health education utilization of migrants as a whole. Model explained between 10.1 per cent (Cox and Snell R square) and 17.7 per cent (Nagelkerke R square) of the variance in occupational disease health education utilization as a whole. (See table 7 Table 9 predicts the determinants of communicable disease, non-communicable disease, and occupational disease health education utilization for migrants (Model , Model , and Model ) by multivariate logistic regression. Model shows that the chances of receiving communicable disease health education decreased 71.0% in average daily working time more than 8 hours (OR=0.290), compared with migrants with average daily working time less than 8 hours.
Additionally, migrants who do exercises (OR=2.204), acquire health knowledge (OR=1.954) were more likely to receive communicable disease health education. Model indicates that divorced or widowed migrants were at 4.448-times (OR=4.448) higher chance of receiving non-communicable disease health education, compared with unmarried migrants. The chances of receiving non-communicable disease health education decreased 74.6% in average daily working time more than 8 hours (OR=0.254), compared with migrants with average daily working time less than 8 hours. Additionally, migrants who do exercises (OR=2.436), were more likely to receive non-communicable disease health education. Model indicates that the chances of receiving occupational disease health education decreased 58.3% in married (OR=0.417), and 68.5% in average daily working time more than 8 hours (OR=0.315), compared with unmarried migrants and average daily working time less than 8 hours respectively. Migrants who have formal work (OR=2.001), do exercises (OR=1.827), were more likely to receive occupational disease health education.

Discussion
This study attempted to describe the differences between the needs and utilization of health education, and assess the major determinants associated with the health education utilization for migrants in urban-rural fringe areas of Beijing, to better facilitate their health education utilization.

Utilization and needs of health education
Previous researches indicated that the advantage of "healthy migrant effect" ( rst-generation migrants are often healthier with lower overall morbidity and mortality than local-born populations) will diminish dramatically, particularly in middle age [23,24], together with demanding work schedules, poor working and living environment, insu cient health literacy, and negative attitudes toward the health preventive behaviors. Our research revealed that though many migrants were aware of signi cance of health education, and expressed a desire to gain access to health information for enhancing their well-being, yet low utilization rate lingered and only 34.5% migrants received health education in the past year. It was self-evident that most of the migrants investigated were middle-aged, and they had age-appropriate health education needs, such as antenatal, prenatal and postpartum healthcare, aged healthcare, and noncommunicable disease prevention and therapy. Furthermore, with the evolution of migration model, migrants should take the responsibilities for caring for their child(ren) and parent(s), thus they have relatively high needs of child, adolescent, and aged healthcare. It also seems strange that there were low rates of occupational disease health education utilization and needs actually, which are consistent with the previous study in Xi'an that more than 50% migrants have not received occupational safety and health protection training [25]. Furthermore, there might be a reason to explain the low needs of occupational disease health education among migrants. Different from acute occupational diseases (occupational allergic contact dermatitis, occupational poisoning) that always occur after a relatively brief exposure, the common chronic occupational diseases (pneumoconiosis, musculoskeletal disorders, psychological stress at work, occupational tumors) which occupy the majority of occupational diseases, only occur after prolonged exposure to relevant hazards [26,27]. Migrants with low health literacy, unstable job, and limited knowledge of occupational hazards, would not pay enough attention to occupational diseases, even if chronic occupational diseases having occurred.

Determination factors of whole health education utilization
The rate of health education utilization was higher in the groups of 18 to 24 and 25 to 34 years old than that in other three age groups, particularly in the rst age group. The result was similar to a previous China-based study that migrants in the group of 25 to 34 years old have higher rate of health education utilization than other age groups, but the rate of migrants in age group of 18 to 24 years old was opposite [28]. There might be two factors for the differences. Firstly, the second-generation migrants with higher education level and relatively stable working condition, were born after 1980, had better health literacy and could acquire reliable health information from various sources, compared with the rst-generation migrants [29]. Meanwhile, this research also indicated that migrants with high education level and regular exercises had a higher likelihood of receiving health education. The prevalence of health literacy was related to health knowledge, health decisions, health behaviors and health outcomes of the population [30][31][32]. On the contrary, low education level, accompanied by low literacy and health awareness, pose di culties and barriers in understanding complex health-related information, health practices and outcomes [33]. Secondly, selection bias would in uence the results due to insu cient sample size of migrants in age group of 18 to 24 years old. In the future, migrants in this age group should be studied in terms of their health education utilization behaviors and in uencing factors as a unique group.
Determination factors of three types of health education utilization Gaps exist between the needs and utilization of three types of health education for general migrants population investigated, including communicable disease, non-communicable disease and occupational disease. Our study focuses on the determinations factors of three types of health education utilization.
For migrants, we observed that both average daily working time of enabling variables and do exercises of health behavior variables contributed signi cantly to the variances in three types of health education utilization. Migrants who worked more than 8 hours daily and not to do exercises were less likely to use three types of health education than migrants with working time less than 8 hours and do exercises. Put another way, the heavy workloads and poor health awareness for migrants reduce the opportunity to receive health information. Normally, people who do not exercise regularly are lacking of health awareness to access health education. Additionally, consistent with the previous research [28], migrants with formal work are more likely to receive occupational disease health education than those with informal work. Worldwide, the patterns of employment for migrants are similar, developing countries or developed ones. Migrants are overwhelmingly employed in 3Ds jobs (dirty, dangerous and degrading), covering the service sector, production, construction and maintenance, transportation, which have more health-related risks compared with other industries [34][35][36]. In accordance with state regulations, employers have the responsibility to provide regular training, to educate their employees about occupational hazards, and to require them to strictly abide by safety rules. Nevertheless, migrants working in small-and medium-sized enterprises are at higher risk, due to the de ciency of occupational disease health education and supervision, compared with those in large-sized enterprises [37]. Meanwhile, the informal and temporary working status, long working time and situational stress cut down the needs of acquiring health information, especially on occupational disease for migrants.

Device of new health education tools
Advances of information technology witness that smartphones and internet have become an integral part of our lives, and are widely used in health information research. The report of "Internet adoption, social media usage, and smartphone ownership rates in 37 countries across the world in 2017" from Pew research center revealed that the rates of internet use, smartphone ownership and social media use were 71%, 68% and 60% respectively in China [38]. Different from traditional text-based health education tools, including brochures, lea ets, newspapers, web-based social media tools offer more convenient and effective methods of delivering health information [39]. Take the WeChat application as an example, as a free instant messaging application for smartphones, it plays an important part in modern lifestyles. WeChat can serve as a bridge between information technology and frequent multimedia messages to provide health support and management through the communication and transmission of voices, texts, pictures and videos over great distances [40]. Information related to the prevention and treatment of various diseases can be acquired and requested at any time via such applications [41,42]. The increasing number of smartphones in China provides a mobile platform for delivering health education. As in our study, 58.2% migrants want to acquire health information via the internet. Therefore, mHealth interventions strategies have enormous potentials as an educational tool for behavioral change to further control the spread of epidemics, development of the chronic disease for migrants.

Limitations
The study has several limitations. Firstly, cross-sectional survey cannot be determined the time-effect and causality accurately compared with the cohort study, and cannot evaluate the effects of health education compared with the intervention study. Secondly, recall biases on self-report might underestimate the information on health education utilization. Thirdly, health education utilization behaviors were measured as a dichotomous variable (the received or did not receive of health education), rather than measured by the intensity of received health education. Finally, although the questionnaire was designed according to the previous theories and experiences, several signi cant potential determination factors may not be considered in the model, such as community resource factors. Future research is needed to explain the dynamic and cyclical causal relationships of Anderson's health service utilization model by identifying more variables.

Conclusion
The ndings of the survey contribute to our understanding of the health education utilization, determination factors, and needs of health education among Chinese migrants. There were certain gaps between the needs and utilization in different types of health education. Compared with the rst-generation migrants, second-generation migrants had higher rate of health education utilization. Additionally, average daily working time of enabling variables and do exercises of health behavior variables in the Anderson health service utilization model, was a dominant predictor of three types diseases of health education utilization, including CDs, NCDs and occupational diseases. Many migrants desired for more health education information from internet. In the next step, we should focus on the health education utilization for migrants with heavy workload and low education level.
The ndings of this research might be useful for establishing basic public health service network. It also suggests that policy makers should take feasible policies and measures to overcome obstacles and to break down barriers for migrants, including fully implementing health education intervention strategies and policies, providing multiple health information channels, ensuring easy and equitable access to health education. In the future, more comprehensive studies should be carried out to evaluate the e ciency of health intervention strategies to improve the acquisition and utilization of basic public health services for migrants.
Abbreviations SARS: Severe Acute Respiratory Syndrome CD: communicable disease NCD: chronic non-communicable disease RMB: Renminbi CHSIs: community health service institutions PEN: predisposing, enabling, and need variances Declarations Ethics approval and consent to participate It was approved by the Ethical Committee of Capital Medical University, Beijing, China. Data were obtained from a 2016 cross-sectional survey in urban-rural connection districts of Beijing. Written informed consent was obtained from each participant involved in this study. For participants under the age of eighteen, written informed consents were obtained from their guardians. All participants' information was kept con dential and tracked anonymously with an identi cation number only.

Consent for publication
Not applicable.

Availability of data and materials
The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
The study was undertaken as a part of Beijing Philosophy and Social Science Planning Project (15SHC035), which is a population-based cross sectional survey on risk factors of health status for migrants. Apart from inputs on overall progress of this research, as funding body, the funders had no role in the design, collection, analysis, interpretation of data, or writing of the manuscript related to this study.
Authors' contributions SS drafted the manuscript and acquired data. SS and HRZ acquired, analyzed, and interpreted the data. JYXX and XLC analyzed the data. YLZ and MRW revised the manuscript. JD conceived the study design and revised the manuscript. All authors read and approved the nal manuscript.     Model : Binary logistic regression analysis of predictors of communicable disease health education utilization of migrants in the past month.
Model : Binary logistic regression analysis of predictors of non-communicable disease health education utilization of migrants in the past month.
Model : Binary logistic regression analysis of predictors of occupational disease health education utilization of migrants in the past month. Model I: Multivariate logistic regression analysis of predictors of health education receipt by migrants.