Although the concept of the HIV treatment cascade has reached nearly ubiquitous acceptance in international HIV policy and research, methods for estimating it vary drastically. These variations become increasingly important as the focus of the HIV response shifts from emergency response to long-term outcomes and financial and organizational sustainability. We review the history of the cascade and the current literature and develop the first comprehensive typology of cascade scope and methods.
We define the cascade scope in terms of both breadth (range from first to final event) and depth (given breadth, number of cascade stages that analyzed). We distinguish cascade measurement according to four dimensions: denominator–denominator linkage (data used for cascade construction are linked at the individual level across stages); denominator–numerator linkage (data are linked at the individual level within each stage); single vs. multiple populations from which data sources are drawn; and longitudinal vs. cross-sectional design.
Everything else equal, we would prefer broader and deeper cascades, denominator–denominator linkage, denominator–numerator linkage, single population, and longitudinal data over their respective alternatives. Increased investments in population-based cohorts and data linkage are required to complement clinical cohorts for ‘broad’ longitudinal cascade analyses.
aHarvard T.H. Chan School of Public Health, Boston, Massachusetts, USA
bWellcome Trust Africa Centre for Health and Population Studies, University of KwaZulu-Natal, Durban, South Africa
cDivision of Infection and Immunity
dMRC Clinical Trials Unit, University College London, London, UK
Correspondence to Till Bärnighausen, MD, ScD, MSc, MSc, 665 Huntington Avenue, Building 1, room 1104, Boston, MA 02115, USA. Tel: +1 617 432 1232; e-mail: email@example.com