Appropriately characterised death rates among people living with HIV/AIDS prior to and following enrolment in HIV care is a useful surveillance indicator to prioritise public health action with respect to testing, referral and treatment procedures.

Many patients who are diagnosed with HIV infection do not link to care, while among those who do many are lost to follow-up; both these factors must be accounted for to produce unbiased estimates of mortality. Methods to address such biases require updating and further testing in a range of populations.

What does MeSH aim to do in this area?

MeSH_30_30Develop test and apply simplified tools to enhance the capacity of programs to estimate mortality among individuals who are diagnosed with HIV infection, enroled in HIV-care and/or starting anti-retroviral therapy (ART), by improved accounting for association between death and loss to follow up
MeSH_30_30Use available data from sources such as sample vital registration, verbal autopsy and community-based HIV surveillance applied to national and local HIV prevalence and service coverage data to obtain estimates of overall trends in AIDS deaths.

What is MeSH working on and achieved so far?

MeSH_30_30Developed a model allowing mortality estimates to be better adjusted to account for loss to follow-up. The model was revised based on feedback from the UNAIDS Reference Group on Estimates, Modelling and Projections and the MeSH Mortality Working Group and presented to the UNAIDS reference group meeting in May 2016. It will be included in an updated version of the UNAIDS Spectrum software. A systematic review of published studies that traced patients lost to follow up to ascertain their vital and treatment status and a paper describing the methods to adjust for Loss to Follow-Up have been published. This work was carried out in partnership with the IeDEA consortium.
MeSH_30_30Derived improved estimates of residual survival based on longitudinal CD4 count trajectories and data from care cohorts. A report based on this work was presented to the UNAIDS HIV modelling Reference Group meeting in October 2016. The methods are currently under review for incorporation in the Spectrum software.
MeSH_30_30Improved estimates of the mortality burden attributable to HIV have been calculated in the general population by conducting non-parametric survival analyses using South African data to estimate gains in the population-wide life expectancy at age 15 years since the introduction of ART. The analyses were published in a research paper in 2016. Further tailor-made analyses that were included in the 2016 World AIDS Day Report.
MeSH_30_30Carrying out patient tracing studies in sites where clinical records can be linked to demographic surveillance data to ascertain the vital status of individuals lost to follow-up from treatment programmes and better understand levels of mortality amongst this group.
MeSH_30_30Better understand patterns and trends in HIV-attributable mortality in African populations by developing novel analytical tools and apply them to population-based data sets.