Objectives: To explore models to improve HIV testing, linkage to care and treatment among men who have sex with men (MSM) in cooperation with community-based organizations (CBOs) in China.
Methods: We introduced a new model for HIV testing services targeting MSM in six cities in 2013.These models introduced provision of rapid HIV testing by CBO staff and streamlined processes for HIV screening, confirmation of initial reactive screening results, and linkage to care among diagnosed people. We monitored attrition along each step of the continuum of care from screening to treatment and compared program performance between 2012 and 2013. According to the providers of two rapid tests (HIV screening), four different services delivery models were examined in 2013: Model A = first screen at CDC, second at CDC (Model A = CDC+CDC), Model B = first and second screens at CBOs (Model B = CBO+CBO), Model C = first screen at CBO, second at Hospital (Model C = CBO+Hosp), and Model D = first screen at CBO, second at CDC (Model D = CBO+CDC). Logistic regressions were performed to assess advantages of different screening models of case finding and case management.
Results: Compared to 2012, the number of HIV screening tests performed for MSM increased 35.8% in 2013 (72,577 in 2013 vs. 53,455 in 2012). We observed a 5.6% increase in proportion of cases screened reactive receiving HIV confirmatory tests (93.9% in 2013 vs. 89.2% in 2012, χ2 = 48.52, p<0.001) and 65% reduction in loss to CD4 cell count tests (15% in 2013 vs. 43% in 2012, χ2 = 628.85, p<0.001). Regarding linkage to care and treatment, the 2013 pilot showed that the Model D had the highest rate of loss between screening reactive and confirmatory test among the four models, with 18.1% fewer receiving a second screening test and a further 5.9% loss among those receiving HIV confirmatory tests. The Model B and the Model C showed lower losses (0.8% and 1.3%) for newly diagnosed HIV positives receiving CD4 cell count tests, and higher rates of HIV positives referred to designated ART hospitals (88.0% and 93.3%) than the Model A and Model D (4.6% and 5.7% for CD4 cell count test, and 68.9% and 64.4% for referring to designated ART hospitals). The proportion of cases where the screening test was reactive that were commenced on ART was highest in Model C; 52.8% of cases commenced on ART compared to 38.9%, 34.2% and 21.1% in Models A, B and D respectively. Using Model A as a reference group, the multivariate logistic regression results also showed the advantages of Models B, C and D, which increased CD4 cell count test, referral to designated ART hospitals and initiation of ART, when controlling for program city and other factors.
Conclusions: This study has demonstrated that involvement of CBOs in HIV rapid testing provision, streamlining testing and care procedures and early hospital case management can improve testing, linkage to, and retention in care and treatment among MSM in China.