The Inuence Factor of Health Seeking Behavior of the Caregivers for under-Five Sick Children in Southern Xinjiang, China

Background: Child mortality can be reduced by 20% with the caregivers’ appropriate health seeking behaviors. Therefore, in order to reduce child mortality in remote areas, we examined the Health Seeking Behavior (HSB) of caregivers of sick children under ve years old in Moyu County, Southern Xinjiang, which is the remotest and poorest areas in China, and the factors inuencing HSB of caregivers. Methods: We analyzed the factors inuencing the health seeking behavior of 846 caregivers of sick children with a questionnaire survey in Chinese and Uighur by a cross-sectional study. Binary logistic regression analysis was used to predict the inuencing factors on HSB of caregivers. Results: After the children’s illness was detected, 89.94% caregivers took their children to go to the doctor. 44.91% caregivers took their children to the doctor in the public medical institutions, while 40.07% caregivers chosen the private medical institutions. Compared with the private medical institutions, public medical institutions were likely to be chosen when caregivers consider the treatment cost or the children with cough, runny nose and stuffy nose, but considering children’s age (0.5times), reputation of doctor (0.6times), adequate drug resources (0.5times), the traditional habits on HSB (0.5times), and time to get to hospital (0.2times), the public medical institutions weren’t be chosen. Conclusions: The HSB of the caregivers for sick children in Southern Xinjiang isn’t optimistic. The doctors’ competencies, the quantity of medicines, the distance, the traditional habits inuencing caregivers to choose hospitals, which is also the most signicant thing to promote the health of rural children in Xinjiang and northwest China.


Background
Although the mortality has more than halved yet since 1990 across the globe, 5.6 million of children die every year before reaching their fth birthday. The mortality rate of children in economically backward in a remote area of the countryside is higher than that in other areas. However, most of these death in economically backward are preventable 1,2 , such as diarrhea, pneumonia, all of which play an important role in the cause of death 3,4 . Accessible health facilities, effective policy support and quali ed medical staffs endeavor to lower child mortality while health seeking behavior (HSB) for sick children is also important that caregivers can recognize the symptoms of those preventable diseases and seek appropriate care for children in time 5 . According to the report, children mortality can be signi cantly reduced with the caregivers' appropriate HSB 6,7 . Therefore, more attention should be paid to the HSB of caregivers of children in economically backward.
HSB of caregivers represent individual take any action to seek an appropriate remedy for whom they take care of, once they are aware of the child's health problem or illness [8][9][10] . In developing counties, only 20% caregivers can identify symptoms of pneumonia 10 . Child mortality in developing countries, particularly in poor areas, is much higher than in developed countries and regions. China and India, Nigeria, Pakistan, Ethiopia, the Democratic Republic of the Congo are account for half of the global under-ve deaths in 2017 11 . Therefore, to reduce child mortality, HSB of caregivers in poor areas should be paid more attention.
Xinjiang Uygur Autonomous Region is the remotest and poorest region in northwest of China. The under-5-year-old mortality rate in Xinjiang is about 2-3 times higher than the average mortality rate in China, at a rate of 25 deaths per 1,000 live births 7 . Moyu County is a national poverty county in China and the second most populous region in Xinjiang, Owing to Moyu County's remoteness, harsh environment, inconvenient tra c, medical and health resources shortage, child mortality in Moyu County is signi cantly higher than the average level in Xinjiang. In order to reduce child mortality, Integrated management of child-hood illness (IMCI) was introduced into Moyu County, Xinjiang in 2014. The doctors in public medical institutions were trained for children's disease according to standard treatment and scienti c mediation cycle by IMCI when the doctors in private medical institutions without training.
However, public medical institutions haven't been enough used and chosen by the caregivers. Therefore, the purpose of this study is to investigate the factors in uencing of the health behaviors of caregivers for children under ve years old in southern Xinjiang, and to provide basis for developing strategies, which can take full advantage of medical services so that sick child can receive timely treatment.

Study design and participants
To know about HSB of the caregivers and its in uencing factors, we conducted a questionnaire survey on caregivers of children under 5 years old in Moyu through a cross-sectional study. According to the economic development and the distance between the town and the county, we selected 8 townships from 17 townships in Moyu County by convenience sampling, and then randomly selected 4 villages in each selected township. All the caregivers in the selected villages were included in this survey. The caregivers were surveyed the latest health seeking behavior for under-5-year-old. If more than one child is under-5- year-old in a family, the HSB for the youngest child will be chosen. The inclusion criteria for the study were those caregivers of children who were sick within a year, without mental illness, infectious disease or serious organic diseases.
Considering the background of the Moyu County, based on Axel Kroeger's and Webair's research about HSB 3,12 , the preliminary questionnaire was developed after the discussion with the professors in Peking University, including three parts with subject characteristics, disorder characteristics, service characteristics, a total of 20 items. Then the content and expression of the preliminary questionnaire's items were discussed in two meetings with Xinjiang Health Project Group of Save the Children and adjusted through pre-survey. The nal version of the questionnaire consisted of 23 items, including 12 subject characteristics, 5 disorder characteristics and 6 service characteristics. Since the mostly native language in the southwestern of Xinjiang is Uighur, the items of the questionnaire was translated into Uighur and then translated back into Chinese with the content validity index (CVI) of 0.75 and test-retest reliability of 0.78 after four weeks.
Before the survey, the investigators explained the purpose and content of the survey and the bene ts and risks for caregivers. The caregivers volunteered to participate and can withdraw from the investigation at any stage of the investigation for no reason. After the caregivers fully understood the study, an informed consent is required to be signed for those who agreed to participate in the study. The questionnaire was conducted anonymously, and the privacy and con dentiality principles of the caregivers were always respected during the investigation, and all the information was guaranteed to be used only in the research.
HSB of caregivers is a dependent variable and was con rmed by the items, "Which medical institution did you prefer to bring your child to?" Responses included "village clinic, township hospital, county people's hospital, county maternal and child care service center, county uigur medical hospital, village private clinic, township private clinic, county private clinic". Public medical institutions included village clinic, township hospital, county people's hospital, county maternal and child care service center, county uigur medical hospital. Private medical institutions included village private clinic, Township private clinic and county private clinic.

Study variables
Disorder characteristics are symptoms of children' diseases and detected by caregivers. They consisted of 5 items, including the following: How long has the child been sick last time? Do you remember the speci c situation of the child's illness? What kind of uncomfortable symptoms did the child have? Responses include "coughing with breathing di cult or tachypnea, diarrhea, fever, nasal obstruction and runny nose, other(s) (Please elaborate)". What did you think about the symptoms and what was your plan when you realize your child is feeling uncomfortable last time? Responses includes "urgent or not, to see a doctor or not". If doctor told you that you should take your child to pay a return visit, could you do that?
Service characteristics are the information about medical institution and factors in uencing the choice of caregiver, including the following: What are the main reasons you choose this institution? Responses includes "expenditure, distance, medical level, adequate medication, injection and intravenous infusion, doctor' attitude, medical habits, time spent on the road, the decision of decision-makers in family, simple process, no deposit, no need pay all expenditure at once, other(s) (Please specify)". Please prioritize the factors which affect your choice of medical institutions. The distance from your home to the selective hospital is. How long does it take to go to the medical institution you choose?
Subject characteristics consist of 10 items about caregiver (including the following: gender, age, nation, education, occupation, village name, annual family income, number of members of family, relation between the caregiver and the child and the person who made the decision to seek medicine service when the child was sick last time) and 2 items about child (including the child's gender, age).

Ethics approval and informed consent
The caregivers volunteered to participate in the study. Each potential respondent was properly told of the study objectives, the related aspects to the study. This study protocol and procedure were approved by the Ethics Committee of Peking University Health Science Center (Ethical review approval number: IRB00001052-18001).

Statistical Analysis
Data was entered using Epidata and analyzed using SPSS 22.0. Firstly, all characters are shown in table  1 and table 2. Secondly, all variables are categorized, and their frequency is compared between the public medical institutions and the private medical institutions by chi-square tests in table 3 and table 4. Thirdly, we examined the association between variables and HSB in Table 5 with logistic regression analysis. P<0.05 was considered statistically signi cant. Per capita age and annual family income are considered as continuous variables.

Results
A total of 945 caregivers of sick children under 5 years old were recruited and 846 completed the questionnaire with response rate of 89.52%. In this survey, everyone is Uygur. Demographic data of the caregivers shows that the caregivers of sick children are mainly their mothers, accounting for 95.5%. Their ages are concentrated in 20-40 years old. The oldest was 50 and the youngest was 19. They generally have low education, either primary education or no education. Caregivers' occupations are mainly farmers, accounting for more than 90%. The demographic characteristics of the caregivers are detailed in Table 1.
After the child's illness was detected, 89.94% of the caregivers took their child to see a doctor, and the selected mainly hospitals were township hospital and private medical institutions. As shown in Table 2, country-level hospital and others are least chosen. Therefore, township hospitals and village clinics were merged into public medical institutions.
The characteristics compared between public medical institutions and private medical institutions are shown in table 3 and table 4

Discussion
In some developing countries, when child becomes ill, caregivers don't seek for medical care urgently, often leading to child deaths 13 . In Moyu, the medical or public health services are the poorest. The mortality rate of children under-5-year-old in Moyu is signi cantly higher than that in Xinjiang province, which is also 2-3 times higher than that in China 14 . Therefore, in order to reduce mortality rate, it is essential to pay attention to the health-seeking behavior of parents in remote areas. The researches involving the caregivers' health-seeking behavior for sick children mainly concentrate on the children with single disease such as diarrhea, visual impaired, or nodding syndrome [15][16][17] . The Niger study found that about one-third of caregivers wouldn't take their children to see a doctor in rural Niger when their children have acute diarrhea 18 . However, the major causes of death among children under-5-year-old include diarrhea, pneumonia, fever, etc. Therefore, our study focuses on the caregivers' health-seeking behavior for children under-5-year-old suffering from different diseases in remote rural areas, medical institutions chosen by caregivers and the factors that in uence the caregivers' seeking medical services.
An Iranian study found that about 70% of caregivers seek health care when their children have an acute diarrhea 19 . The study in western China showed that the rate of seek health care for diarrhea is 67.2% 20 . Another research based on vision health care program for left-behind children in Shanxi and Gansu showed that even if Symptoms are detected, health care services will be underutilized without a subsidy 20 . Our study found that about 89.94% of caregivers would take their children to see a doctor when their children got sick. It is higher than other research. On one hand, it indicate that with the development of the society, the caregivers in remote areas have gradually realized the importance of taking their children to see a doctor in time when symptoms of diseases were detected. On the other hand, in the process of recruit, some caregivers told us that their children didn't have any diseases in the past years. Therefore, it is referred that some caregivers may not have detected the signs of diseases. Propaganda and education about disease symptoms need to be strengthened.
In Moyu County, the differences between doctors in public medical institutions and private medical institutions doctors are as follows. First of all, the doctors in public medical institutions can treat children according to standard treatment and scienti c mediation cycle. Second, the cost can be reimbursed in a certain proportion. However, the cost of private medical institutions cannot be reimbursed. But doctors in private medical institutions tend to treat the children according to their parents' requirements, including the treatment cycle. For example, private medical institutions charge by day when public medical institutions charge by a course of treatments. If intravenous infusion takes 3 days and Caregivers chosen to private medical institutions for treatment on the rst day, they won't proper to continue to treatment next day when they felt relieved. Therefore, the public medical institutions are recommended by Moyu County. However, in our study, nearly half of caregivers chose to take their children to private medical institutions. So, we analyzed the in uencing factors of HSB.
The caregivers' health-seeking behavior is in uenced by various factors, such as the caregiver's age, education, and the symptoms of children's disease detected, as well as the current status of medical institutions, such as distance, cost, etc 15 . Our nding show that caregivers are likely to choose public medical institutions when children with cough, runny nose and stuffy nose. However, there is no signi cant different in caregivers choosing medical institutions for other diseases. In other words, there is no obvious relationship between caregivers' choice and the types of children's diseases, unless the disease is common. Caregivers are also likely to choose public medical institutions, when considering treatment cost. However, they are less likely to choose them when considering children's age, doctor's reputation, insu cient drugs, the traditional habits of HSB, time to hospital. Analyzing the factors in uencing caregivers' choice, we found that caregivers in Moyu County had more faith in the doctors in private medical institutions, which may be due to the traditional habits, adequate drug resources, easy to access.
Research about HSB in northwest of China is few. Gao's research shows that if caregivers lived farther, caregivers preferred to choose the lower level care for their children under 36 months with diarrhea in rural western China 20 . In addition to child's age, the education of caregivers also in uenced the care-seeking behaviors. In our research, only child's age is found to in uence the choice. we didn't nd caregivers' education has an in uence on health-seeking behavior, which is inconsistent with the results of relevant studies 5 . The reason may be that the overall caregivers' education is very low in Moyu. Most of caregivers only have primary education or no education, so the in uence of education level is not obvious.

Conclusion
Our study found that many factors may have impact on HSB of caregivers for children under-5-year-old in southern Xinjiang. Considering the cost of treatment can encourage caregivers to choose the doctor in public medical institutions. Considering the amounts of drugs, time to hospital, doctor' reputation, the traditional habit of HSB can prevent caregiver to choose the doctor in public medical institutions. More attention needs to be paid to the traditional habits of HSB and the status of insu cient drugs in public medical institutions, which is also the key to reduce children's death rate in the rural areas of Xinjiang and northwest China.

Declarations
Ethics approval and consent to participate The caregivers volunteered to participate in the study. Each potential respondent was properly told of the study objectives, the related aspects to the study. This study protocol and procedure were approved by the Ethics Committee of Peking University Health Science Center (Ethical review approval number: IRB00001052-18001).

Not applicable
Availability of data and materials The datasets used and/or analyzed during the current research are available from the corresponding author on request.

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

Funding
This study was supported by Save the Children (number CTXJ-2018-0148). We thank Save the Children China Programme Beijing O ce and Xinjiang O ce.
Authors' contributions LR, XZ, and HL conceived and designed the study. LR, FL acquired the data. LR, YD, YY, KZ, ZW, LZ, TZ, JH did the data analysis and data interpretation. LR, YD, YY, XZ and HL wrote, reviewed, and revised the manuscript.