Feasibility Study Of Voluntary Counseling And HIV Testing Based On Community

ObjectiveDepartment of Health and department in charge of family planning carry out Health AIDS Voluntary Counseling and Testing (VCT) services jointly. Explore the feasibility based on community organization and propose joint preliminary pattern. Carry on fitting data by using BP and GA-BP neural network on the MATLAB software. Compare their effects of fitting and determine the finally fitting model. In order to provide a basis for further intervention studies, write a sensitivity analysis program to study the extent of the impact of various factors on Health AIDS VCT services based on the MATLAB.MethodThis article uses a cross-sectional study combined qualitative and quantitative. This paper, taking the 60 management and technician personnel in department in charge of family planning and Department of Health from community of Ruijin, Xujiahui, Jinwan as object, know their cognitions and attitudes to HIV / AIDS / VCT, know the factors of carrying on the VCT project, with the methods of interview in order to Analyze and find out positive factor, perceived barriers and counter measure to carrying on the VCT project.437 health service personnel from 3 communities were interviewed to collect information with structured questionnaire through cluster sampling. Analyze their knowledge of AIDS and VCT with Chi-square test, ANOVA and non-parametric test.Calculate total scores of the AIDS knowledge and the VCT knowledge. Write programs based on BP, GA-BP neural network of MATLAB. Using these programs to realize data fit in this article. MSE, EPOCH, R2, adjusted R2 were determined as network performance evaluation. Compare network performance, select number of the neurons in the hidden layer, establish data fitting model through Chi-square test, ANOVA and non-parametric test. Develop the extent of the impact of various factors on carrying on Health AIDS VCT services jointly through sensitivity analysis based on the MATLAB.ResultsQualitative interview results suggest that the three advantages of network of family planning, support of the leaders and legal law are existed between Health departments and Family Planning departments, which are being in the community management, to cooperate the VCT services. The insufficient of community technical personnel VCT knowledge, lack of investment and training of leaders are important disadvantages. Assessment, more investment, incorporation VCT in the daily work and policy support will be the solutions. The first three advantages of community based VCT services are appropriate cooperation, wide clients and convenience community service in technical .The first three disadvantages are insufficient professional knowledge, lack of instability technical staff and worse conditions. Still more than 50% community staff comments the project. And there are eight parts can be conducted the project, which are VCT materials, pre-testing consultation, condom promotion, HIV screening, partner notification, post-testing consultation, SRH consultation and HIV(+)/AIDS followed..80% of staff took part in education services of voluntary counseling and testing based community. Qualitative interview results show that the affections of community service VCT are four aspects: the social environment (whether the discrimination in society, society can understand the acceptance), whether the results confidential, Service accessibility, quality of service.The results show that the mean scores of AIDS and VCT knowledge are 18.48±3.39 and 10.09±1.76 respectively of the 437 investigation objects. 30% of the junior staff thinks of the daily contact such as "sharing a razor", "sharing the toilet", "mosquito bites" and "courtesy kissing", etc. could spread HIV. Only 49.2% correctly recognized AIDS as statutory National B infectious diseases. Less than 5% community technologies could correctly answer the probability of infecting HIV by"sharing syringe with HIV (+) person", "having anal sex with HIV (+) person", ""having vaginal intercourse with HIV (+) person ", "maternal-neonatal transmission". More than 70% know the infected probability of accepting HIV (+) blood. About 60% investigations answered correctly of "window period". The rate of answering typical symptoms such as "Long-term low-heat," "ulcer," "lymph nodes" is more than 80%. But the others were less than 40%. More than 70% of the investigators consider VCT as HIV testing. Half of the interviewers do not know VCT is free. Less than 1/3 of them know CDC can make free HIV testing. 60% of people think "HIV (+) / AIDS should not get married".85% of them the target think servicers should refuse surgery and checking ring of HIV (+) / AIDS.GA-BP neural network is more robust than BP neural network whatever the hidden layer neurons is 5, 10, 15, 20(P <0.05). With the hidden layer neurons increase, GA-BP neural network reached the rate of the intended target and iterative steps within 1000(P<0.001). Wilcoxon test showed that: the adjust R2, R2 of GA-BP neural network were greater than the BP neural network. Kruskal-Wallis test results showed that: R2 reached the Maximum of GA-BP neural network and BP neural network at 15 of hidden layer neurons. GA-BP neural network fitting data by determined the number of hidden layer neurons was 15.Quantitative results shows gender, educational level, professional, VCT knowledge scores of affected attitude of communities technical staff to the project (P <0.05). Multi-Logistic regression results showed that women, university educated community technical staff is more easily agree with IPPF, health joint VCT services. GA-BP neural network data fitting results show that community-based technical personnel titles, with an average monthly income, AIDS knowledge scores, VCT knowledge scores, job department, unit level, working hours, level of education on whether they agree with project, health joint VCT services to a greater impact (sensitivity values are above 0.60). ConclusionCommunity based VCT service conducted by health departments and family planning departments. In order to make sure the project can progress well, more training for the community staff needed. Genetic algorithm can optimum the better weights and thresholds for BP neural network to overcome the slow convergence and local maximum. Sensitivity analysis by neural network can show the order of affections of input variables.