Modelling submitted for consideration by WHO in development of the 2015 ARV Guidelines revision

Further to the modelling analyses that the HIV Modelling Consortium undertook in support of the 2013 consolidated guidelines, the HIV Modelling Consortium were approached by Dr. Meg Doherty, a member of the HIV MC Steering Committee and Coordinator of Treatment and Care in the Department of HIV/AIDS at the WHO, to ask if modelling could once again be conducted for consideration in the development of the 2015 guidelines revision. Further to consultations with the World Health Organization, and our network of mathematical modellers, it was agreed to progress with the following work streams:


Work Stream 1: Priority setting for HIV care and treatment: a ‘population perspective’

Led by Jeffrey Eaton (HIV Modelling Consortium Secretariat)

Antiretroviral treatment (ART) has reduced mortality in countries affected by HIV and is thought to be a key component to the long-term control of HIV epidemics, but deaths due to HIV remain high. Guidelines and national strategic plans are now centred on how to improve and expand HIV care and treatment to further reduce HIV mortality and morbidity. Programmatic priorities should be informed by the weaknesses in the care that result in the greatest population health losses. Identifying the sources of failure —lack of diagnosis, linkage to care, timely ART initiation, care on ART or retention on ART — will enable strategic programmatic investment.

Work stream aims:

  • Synthesise empirical data about the engagement in care among those dying from HIV in current programmes.
  • Use mathematical modelling to examine how the sources of HIV mortality evolve as ART programmes expand and mature.
  • Compare the potential benefits from reducing losses at different stages of care, and how priorities may shift over time.

The methodological approach was to review and synthesise existing epidemiologic analysis and modelling. Modelling analyses were contributed by Eran Bendavid for Rwanda, Jack Olney for western Kenya, Andrew Phillips for Malawi, and Dan Klein & Anna Bershteyn for South Africa. In addition, Emma Slaymaker and George Reniers shared data analysis for the ALPHA Network Sites, as did Andrew Boulle for the Western Cape. 

Work stream 2: The costs and benefits of alternative approaches to monitoring patients on antiviral therapy

Led by Andrew Phillips and Paul Revill (HIV Modelling Consortium Key Partners)

This study uses modeling and cost-effectiveness analysis to compare the costs and health outcomes associated with alternative patient monitoring strategies in light of the most recent data on clinical effectiveness, behavioural responses to monitoring strategies and costs.  An individual-based stochastic simulation model (HIV Synthesis model) that captures the benefits of patient monitoring to the population (i.e. through reduced HIV transmission) as well as the patient was used to model a number of different monitoring strategies for a resource limited setting.

This work stream sought to contribute to the formulation of programmatic guidance by addressing the following questions:

  • For countries without access to viral load monitoring, what would be predicted to be the most cost-effective monitoring strategy?
  • For countries without access to viral load monitoring, what would be predicted to be the most cost-effective monitoring strategy?
  • For countries scaling up viral load monitoring, what can modelling tell us about the pros and cons of plasma vs DBS vs POC?
  • For countries with viral load monitoring in place:
    • Is there a schedule for monitoring stable patients that is likely to be more cost-effective than that which is currently recommended (i.e. at 6mo and then every 12mo)?
    • Is there a threshold for determining failure that is likely to be more cost-effective than that which is currently recommended (i.e. 1000 copies/mL)?

Efforts were made to reflect programmatic realities in sub-Saharan Africa. As such, results were refined at a workshop held in Harare, Zimbabwe in March 2015, which was attended by 42 delegates from a range of institutions (Ministries of Health/national laboratories; academic; international policy organizations and non-governmental organizations (NGOs)), including 19 directly involved in delivery of ART and patient monitoring in 5 countries (Zimbabwe, Malawi, Uganda, Kenya and South Africa). This workshop allowed for feedback on the modelling and cost-effectiveness work undertaken and for further discussion of issues faced by programmes in implementing patient monitoring. A report of the workshop proceedings (‘Implementation issues for monitoring people on ART in low-income settings in sub-Saharan Africa’) can be found in the meeting reports page of the HIV Modelling Consortium website.

A manuscript summarising the results of this work has been submitted to a peer-reviewed journal for publication and will be made available on the HIV Modelling Consortium website once published. Contact Ellen McRobie for more details