Chapter 11: Monitoring and evaluation
Consolidated ARV guidelines, June 2013
11.4 Other monitoring considerations
Programmes are increasingly moving beyond coverage indicators to focus on critical outcomes, such as viral load suppression and immune reconstitution, and on the broader impact of HIV treatment, including HIV-related mortality and HIV incidence. However, programmes also need to measure potential unintended outcomes, such as HIV drug resistance and ARV-related toxicities. Periodic evaluations and implementation research are also central to reviewing programmes.
11.4.1 HIV drug resistance
WHO recommends the use of early warning indicators to help identify deficits in programme performance that favour the emergence of HIV drug resistance (Box 11.1). WHO also recommends that countries undertake surveillance of HIV drug resistance and provides specific guidance on how to do the surveys required.
11.4.2 Sentinel surveillance for ARV toxicity monitoring
Surveillance of the toxicity of ARV drugs is essential to identify and address preventable adverse events. Various approaches have been developed to monitor the toxicity of ARV drugs, including targeted and systematic surveillance reporting on specific types of toxicity and serious adverse events caused by a specific drug in targeted populations, and the pregnancy exposure registry following a cohort of pregnant women exposed to ART, including birth defect surveillance. WHO technical guidance on implementing toxicity monitoring at sentinel sites will become available in 2013.
11.4.3 Evaluation, including impact and programme performance, and implementation research
Routine monitoring should be complemented by systematic evaluations and programme reviews to assess the performance and effects of HIV programmes, either comprehensively or with respect to specific priority areas. Social science and implementation research are important to assess perceptions and values of service recipients and communities along with barriers, facilitators and experiences in delivering and receiving services.
Impact indicators, such as incidence, morbidity and mortality, are often difficult to measure. Guidance on the use of assays for recent infection to estimate HIV incidence at the population level has been recently developed (3), and guidance on monitoring mortality, including the cause of death, will be available in 2013. A short guide summarizing five methods to measure the impact of programmes for PMTCT (4) is already available, and detailed guidance that can be adapted to implement each method will become available in 2013.
Mathematical modelling is often undertaken to project various scenarios for programme planning and evaluating impact. Ensuring the availability of robust data is especially important when estimating the prevention impact of ARV drugs at the population level, as multiple sources of information and uncertainty come into play. Specific data collection efforts and models for particular contexts may provide more accurate estimates.