Study on influencing factors to HIV health services among MSM: based on Andersen Behavioral Model

: Background: We assessed the utilization of HIV health services and its influencing factors on consistent condom use, HIV testing and HIV counseling among men who have sex with men (MSM), so as to provide a theoretical basis for future infectious disease prevention and control strategies and health services policy formulation. Methods : This study is a cross-sectional study. From April 2013 to October 2014, a sample survey was conducted in southwest China including Chongqing, Sichuan, Xinjiang and Guangxi, and an anonymous self-administered questionnaire survey was conducted among MSM who met the requirements and were recruited. Based on Anderson Behavioral Model, the questionnaire divided the influencing factors into predisposing factor, enabling factor and need factor. There were 1727 valid questionnaires. SAS 9.4 was used for univariate analysis and multivariate Logistic regression analysis to explore the factors influencing the utilization of health services. AMOS Graphics 24.0 was used to construct the path diagram through path analysis to explore the effect among various factors. Results: In the survey of HIV health services, 9.96% of respondents consistently used condoms, 78.00% had HIV testing, and 60.63% had HIV counseling. Among the predisposing factor, the older respondents are, the easier it is to be tested for HIV(β=0.078, p ＜ 0.001). Among the enabling factor, urban registration is a factor that promotes HIV testing and counseling(β=0.064,p=0.003 and β=0.072, p=0.002). Among the need factor, HIV knowledge score is also a key point affecting testing and counseling(β=0.157, p ＜ 0.001 and β=0.184, p ＜ 0.001). The diagnosis of STD can promote respondents counseling(β=0.051, p=0.031). Depression is a contributing factor to consistent condom use(β=0.051,p=0.033), but negative to HIV counseling(β=-0.119, p ＜ 0.001). Conclusions: For these groups, MSM with high-risk characteristics should be identified as a priority in the future public health services. HIV knowledge should be promoted in health education, physical and mental health diagnosis and treatment should also be strengthened.

diseases [1], which is a threat to public health and quality of life. In recent years, the AIDS epidemic in China has grown rapidly, from 440,000 in 2011 to 950,000 in 2019 [2]. Among them, MSM is the main key population of HIV infection. Homosexual transmission has become the second most common HIV transmission route after heterosexual transmission, accounting for about a quarter of new HIV cases. Despite some success in HIV prevention efforts, the rate of HIV transmission among MSM population still remains high [3]. Therefore, it is very important to pay attention to the need and utilization of health services for this population.
Health services mainly refer to medical treatment, prevention, health care, rehabilitation and other activities provided to residents by the health system with certain health resources, including prevention and control of infectious diseases [4], disease screening [5] and disease surveillance [6]. Among them, HIV-related health services mainly include: HIV testing, HIV counseling, condom use and so on. As early as the beginning of the 21st century, existing countries recognized the importance of strengthening public health services to combat infectious diseases and conducted continuous research and improvement [7]. A Canadian study found that the private health services of HIV counselling and point-of-care(POC) testing based on urban communities were acceptable and feasible to a certain extent [8]. In addition, the Internet can be used as a potential point of access to health screening to address inequalities in health services, according to a US online survey [9].
And some Chinese scholars also have integrated the newly developed M-Health app with HIV health services to prevent disease transmission and reduce HIV infection rate.
In health services, the priority groups often focus on: adolescents [10], mothers and children [11], migrants [12], female sex workers [13] , etc. MSM, as a special population, has been largely ignored by health service projects [9]. In areas with laws and policies forbidden, some key population is more difficult to access them [10]. However, reasonable evaluation of acquisition and quality of health services can help reduce disparities and health care inequalities in MSM and other sexual minorities populations [9]. Future research should also give priority to a gender-specific organizational framework to understand and address the complex situation of limited MSM access to HIV health services.
Several studies have shown that social structure, psychological factors, stigma, homophobia, and policy issues can all affect MSM engagement and positivity in HIV health services [14]. There is also evidence that MSM have been rejected by family members, publicly humiliated and laughed by health care workers when they disclose their sexual orientation [15]. Because of the stigma and discrimination from health providers and neglect by health systems, it poses a significant challenge to HIV care and treatment for this population.
Therefore, based on Anderson Behavioral Model and path analysis, this paper takes consistent condom use, HIV testing and HIV counseling as health services utilization items, and discusses the utilization and influencing factors of health services for MSM. Priority should be given to the identification of high-risk characteristics of this population, and intervention or further research should be conducted to improve the utilization of HIV health services and reduce the incidence of new infections, and to provide more new ideas for future service supply strategies.

Participants and procedures
This study is a cross-sectional study. From April 2013 to October 2014, in four regions of southwest China, including Chongqing, Sichuan, Xinjiang and Guangxi, qualified subjects were first sought through advertisements on gay websites and QQ groups, and then more people were recruited through "snowball" sampling. Inclusion criteria include :(1) signing of informed consent. (2) Age≥18 years and ≤65. (3) Having sex at least once every two weeks on average. (4) Had at least one sexual partner in the previous month. The method of anonymous self-filling questionnaire was adopted, which was collected on the spot and checked for completeness and logic.

Definitions of HIV health services
HIV health services in this study include: consistent condom use, HIV testing and HIV counseling. Consistent condom use is defined as the use of condom every time including anal and oral sex. The project measured the number of times respondents had inserted sex and the number of times they used condoms during sex.
HIV testing refers to whether the respondents had been tested for HIV, and they chose "yes" or "no". HIV counselling refers to whether the respondents have had AIDS-related counselling, and they choose "yes" or "no".

Andersen Behavioral Model
The Anderson Behavioral Model chosen for this study was originally developed in the 20th century to describe the factors that influence the use of health services [16]. Has been used to guide the examination of predictors associated with a variety of health outcomes, such as drug use among people living with HIV [17].
The model has been widely used in hypertension disease management [9], women's mental health [18], breast cancer screening [19,20], etc. Taking Anderson Behavioral Model as the theoretical basis and incorporating it into the analysis of influencing factors can better explain and understand the influence of variables on outcome indicators, and can also better clarify the context of health services, so as to provide targeted suggestions for future policy making.
The model divided the influencing factors into: predisposing factor, enabling factor and need factor.
Predisposing factor refers to an individual's unchangeable nature, including demographic variables and social structural variables [21]. Age, nationality, degree of education, employment and sexual role are included into this factor. Enabling factor refers to that can promote or hinder the use of services, including personal resources and regional health services resources [21].Household registration, income, marriage, find sexual partners through the Internet and commercial sexual service are included into this factor. Need factor refers to the subjective understanding of the disease and clinical objective diagnosis results [21]. HIV knowledge score, diagnosis of sexually transmitted disease (STD), anxiety and depression are included into this factor.

HIV knowledge score
In the basic information, the HIV knowledge scale(Cronbach's alpha = 0.672) was made up of 13 questions based on a revision of the International AIDS Knowledge Survey General Scale [22]. The answers include "true", "false" and "don't know". Based on the answers, there is a 1 point for correct answers, and 0 points for incorrect or unknown answers. The higher score is, the more knowledge the respondents had about HIV.

Anxiety
Anxiety was measured by the self-rated anxiety scale [23], and related research showed that the scale had dependable reliability and validity [24](Cronbach's alpha = 0.86). The scale is composed of 20 items, and the selected score from 0 to 4 represents the frequency of personal experience of the respondents, among which 5 questions are scored in reverse. The final added score was adjusted by multiplying by 1.25, with a total score of > 50 indicating a possible clinical anxiety.

Depression
Depression was measured by the Center for Epidemiological Studies Depression Scale [25], used for epidemiological investigation to screen out patients with depression for further diagnosis(Cronbach's alpha = 0.87). The scale is composed of 20 items, and the selected score from 0 to 3 represents the frequency of personal experience of the respondents, among which 4 questions are scored in reverse. The final total score >16 indicates the possibility of clinical depression.

Statistical analysis
Firstly, descriptive analysis was conducted for each variable in the questionnaire, and then univariate analysis and multivariate Logistic regression analysis were conducted according to outcome variables: consistent condom use, HIV testing and HIV counseling. According to the results of multivariate analysis, a path diagram was constructed to explore the impact of each path. P<0.05 was considered statistically significant. In path analysis, the fitting index of the model is referred to:  2 /df ： 1~3, p>0.05, IFI>0.9, TLI>0.9, CFI>0.9, RMESEA>0.10. SAS 9.4 was used for univariate analysis and multivariate analysis, and AMOS Graphics 24.0 was used for path analysis.

Results
In this HIV health services survey, 1,914 questionnaires were collected in Chongqing, Sichuan, Xinjiang and Guangxi. The questionnaires that did not meet the requirements were excluded: 7 questionnaires did not complete HIV testing, 8 questionnaires did not complete HIV counseling, and 103 questionnaires did not complete condom use. 30 questionnaires were incomplete HIV knowledge score scale, 25 questionnaires were incomplete depression scale and 14 questionnaires were incomplete anxiety scale. A total of 1727 valid questionnaires met the requirements, with an effective recovery rate of 90.23%.
In this health services survey, 9.96% of respondents consistently used condoms, 78.00% had HIV testing, and 60.63% had HIV counseling. The results of univariate analysis are shown in the table. (Table 1) Among the respondents who continuously used condoms, the rate was statistically different for those who played different sexual roles (P=0.0022), and the "1" and "0" sexual roles were 13.16% and 7.14% respectively.
The continuous condom use rate of unmarried people was 9.10%, that of married people 9.64%, and that of divorced people 18.71%, with statistical differences (P=0.0144). The continuous condom use rate was 11.86% in depressed people and 8.77% in non-depressed people, and the rate was higher in depressed people (P=0.0365). There was no statistical difference in other factors.
Among the respondents who had been tested for HIV, there were statistically significant differences in different ages (P<0.0001), and the rates of 18-25 years old, 25-35 years old, and ≥35 years old were 48.84%, 74.57%, and 83.23% respectively. The HIV testing rate of respondents of Han nationality was lower than that of other nationalities (P=0.0017). There were statistically significant differences with different education levels, employment status, sexual roles and income, and urban areas was higher than that in rural areas (P< 0.0001). Respondents with higher HIV knowledge score and no depression had a higher testing rate (P<0.05).
Among respondents who had HIV counseling, the rates varied by age, employment, income, and sexual roles.
The HIV counseling rate of Han nationality was lower than that of other nationalities (P=0.0077), and urban areas was higher than that in rural areas (P<0.0001). Respondents with high HIV knowledge score and no depression and anxiety had higher HIV counseling rate (P<0.05). The results of multivariate Logistic regression analysis are shown in the table. (Table 2) Those in sexual roles who only did "1" were 2.229 times more likely to continue using condoms than those who only did "0" (1.157 to 4.292). The probability of continuous condom use was 0.660 times (0.473 to 0.921) of the respondents without depression among the predisposing factor, that is, the more depressed the respondents were, the more likely they were to continue to use condoms.  Based on path analysis, the fitting degree of the model is good ( 2 /df=2.960, p=0.001). Due to the large sample size (N=1727) in this study, P-value may be small, but other fitting indexes meet the requirements ( It can be seen that, among the predisposing factor, the older age is, the easier it is to be tested for HIV. Among the enabling factor, urban area is a factor that promotes HIV testing and counseling. Among the need factor, the higher HIV knowledge score and STD diagnosis were the factors that also promoted testing and counseling, while the respondents with depression were more likely to use condoms every time, but it was not easy to conduct testing and counseling. (Table 4,5) ( Figure 1)

Discussion
Andersen's model was adapted to identify factors associated with HIV health services by including sets of variables. The model helped uncover factors that may be ignored before, especially among MSM. Among the health services utilization projects examined in this study, the utilization of HIV counseling and HIV testing was good, but the status of consistent condom use was not optimistic, which is accordant with the description of Mengran [26]: a low level of intentions to use condoms consistently has been reported in Chinese MSM population.
According to the Anderson behavioral model, need factor reflects how people view their own health, subjective cognition of disease and clinical diagnosis for individual physical condition, is the most direct and important factor which influences health services utilization. Therefore, it is considered to be one of the powerful predictor in health services [27]. It can also be seen from this study that need factor is the main factor affecting HIV health services. Among them, HIV knowledge score, diagnosis of STD and depression of respondents had statistically significant effects.
In the 1727 valid questionnaires collected this time, the average HIV knowledge score was 9.31±2.58, among which, ≥11 accounted for 37% and <11 accounted for 63%, indicating that the degree of HIV knowledge was generally moderate. HIV knowledge is the main factor affecting HIV testing and HIV counseling, which is also consistent with the research ideas of Sofia [28]. In fact, as early as 10 years ago, scholars proposed the Information, Motivation and Behavioral Skills (IMB) Model, which has guiding significance in HIV risk reduction interventions [29]. The information in the IMB model mainly includes subjective information and objective information, of which objective information includes knowledge [30]. More and more studies have found that HIV knowledge plays a key role in the prevention and control of AIDS. The studies of Simukai and Doris [31,32] also indicated that HIV knowledge and attitude were associated with non-condom use and highrisk sexual behavior, so knowledge promotion and popularization based on media platforms were urgently needed. At the same time, Chilot [33] has pointed out that a very low comprehensive knowledge of HIV/AIDS among women of reproductive age in Ethiopia is one of the major reasons for the increase of HIV infections.
It has been confirmed in the literature that educational programs [34], sexual education and communication activities [35] can contribute to the improvement of knowledge. However, it remains to be discussed whether increased HIV knowledge will necessarily lead to improved behavior, because of the phenomenon of "knowledge-practice separation". According to the theory of Knowledge, Attitude and Practice (KAP) and the study of Min-Jin Peng [36], we can see that the phenomenon of "knowledge-practice separation" does exist, and at the same time, improving self-efficacy may help to solve this problem. Therefore, in the future, more studies are needed to evaluate the HIV knowledge of high-risk groups and explore relevant influencing factors, so as to solve the problem of "knowledge-practice separation" or "knowledge-attitude separation" in this group for reducing the incidence of new HIV infection.
According to the results of this study, individuals who are clinically diagnosed with STD are more likely to undergo HIV testing and HIV counseling. Conversely, individuals with depression are less likely to seek counseling and testing facing to general public, but more likely to use condoms every time they have sex. Our analysis shows that the impact of STDs on MSM is also significant. With undiagnosed or untreated STDS will also increase the risk of HIV transmission [37]. It makes sense, therefore, that if an individual is diagnosed with STD, they will be more likely to use HIV health services because of their fear of AIDS and perceived disease risk. likely as heterosexual men to experience some mental health disorders [38]. Due to the lack of supportive environment, the vulnerability of personal privacy information to disclosure in the implementation of HIV health services, as well as AIDS-related discrimination and stigma pose a threat to the quality of health services, and have a serious impact on patients' mental health and the quality of health services. Research by Rames [39] and Yuchen Mao [40] suggests that social stigma is common, and this stigma and social stress can lead to depression, reduced quality of life and negative treatment outcomes. We suggested that regulatory agencies should work to reduce stigma effects. Public health interventions should aim at increased access and effective utilization of services for both HIV/AIDS and mental health services [41].
In addition to the physical, psychological and cognitive factors of the individuals mentioned above, other factors will also affect HIV health services. The enabling factor in the Anderson Behavioral Model refers to the resources or means by which an individual has access to health services, usually involving individual and community resources such as health insurance, income, wealth, availability of services and urban classes.
User fees for healthcare services present a barrier to patients accessing healthcare and reduce detection of serious infectious diseases [42]. Because in China, HIV testing is free only in CDC, and testing in other institutions such as hospitals is charged, and is not covered by medical insurance. At the same time, the location of HIV testing is only under some settings, which makes it easier for MSM living in urban areas to access health services. It is therefore proposed to increase the number of health service points in rural areas in order to address the uneven distribution of resources between regions. Of course, there are other problems in HIV health services, such as limited data [43], unreasonable organization of health institutions [44], and high cost (time cost and human resources) for a health service [2]. The need to combine medical, sociological and psychological considerations also poses a great challenge to the development of health services. Especially for MSM, it needs to be more cautious and careful.

Conclusion
Based on Anderson Behavioral Model and path analysis, this paper studies the factors affecting HIV health services. Finally, we conclude that MSM population has a good utilization of health services, but the consistent condom use is not ideal. HIV knowledge is the main factor affecting HIV testing and HIV counseling, followed by age, household registration and other factors. At the same time, the physical and psychological state of the population will also affect the demand for and use of health services. Need factor is the main factor that determine the utilization of health services. The government and relevant departments should strengthen the popularization of disease knowledge and the diagnosis and treatment of individual physical and mental diseases. MSM population with high-risk characteristics should be identified as a priority in the future public health service delivery strategy. In view of these groups, the publicity and education of HIV knowledge should be strengthened, and the utilization of HIV health services should be improved. It is the focus of future research to provide new ideas for health services policy formulation by combining regional, economic, health resources, privacy, psychological problems and other factors. We hope that our study can encourage discussions of HIV health services， and set the stage for sharing and creating for service innovation。 Ethics approval and consent to participate: All procedures of this study were in accordance with the ethical approval granted by the Ethics Committee of Chongqing Medical University. Informed consent was obtained in writing from all individual participants included in the study, and all methods were carried out in accordance with relevant guidelines and regulations. The Ethics Committee of Chongqing Medical University has reviewed the proposed use of human subjects in the above-mentioned projects. It is recognized that the rights and the welfare of the subjects are adequately protected; the potential risks are outweighed by potential benefits. We approve the project implementation according to plan.

Consent for publication: Not applicable.
Availability of data and materials: The datasets generated and analysed during the current study are not publicly available due [REASON WHY DATA ARE NOT PUBLIC: Because our respondents are sensitive crowd, and related researches are in progress.] but are available from the corresponding author on reasonable request.