Improving cohort coverage estimation using a data triangulation framework: the Swiss HIV Cohort Study example
Definitely a must-read before the summer break! The study by Duran et al. presents a novel data triangulation framework to estimate the coverage and representativeness of the Swiss HIV Cohort Study (SHCS) by integrating multiple independent data sources, including the Swiss national HIV/AIDS surveillance system, national antiretroviral therapy (ART) sales data, and previously published cohort coverage estimates. The aim was to provide a more robust assessment of how well the SHCS reflects the diagnosed HIV population in Switzerland and to identify demographic groups that may be underrepresented.
Using longitudinal data from 1985 to 2023, the authors estimated that the SHCS captures, on average, 62.4% of all diagnosed people living with HIV in Switzerland, 74.0% of individuals with AIDS, and 64.9% of people receiving ART. These estimates were consistent with previous publications and demonstrate that the SHCS remains one of the most comprehensive national HIV cohorts worldwide despite being a voluntary research cohort.
The study found that cohort coverage increased steadily during the early years of the HIV epidemic, reaching approximately 90% around 1999, but has declined over the past decade, falling to approximately 42% for newly diagnosed individuals in 2023. Sensitivity analyses suggested that this decline is unlikely to be explained by delayed enrolment and more likely reflects evolving models of HIV care, with increasing numbers of clinically stable individuals being managed outside SHCS-affiliated centres, together with demographic changes, migration, and psychosocial barriers affecting engagement in cohort research.
Although the SHCS continues to be broadly representative of the Swiss HIV epidemic, the analysis identified several demographic imbalances. Women, older adults, individuals with heterosexually acquired HIV infection, and people who inject psychoactive substances were consistently underrepresented, whereas men, younger adults, men who have sex with men, and individuals receiving complex or salvage ART regimens were relatively overrepresented. Regional differences in cohort coverage were also observed, with lower coverage in more decentralised parts of Switzerland.
Comparison with nationwide ART sales demonstrated a remarkably stable cohort coverage of approximately two-thirds across commonly prescribed ART regimens, supporting the validity of the triangulation approach. The authors argue that combining surveillance, treatment, and cohort data provides a practical and reproducible framework for monitoring cohort coverage in settings where direct linkage between databases is not feasible because of privacy regulations.
While the SHCS continues to represent the majority of diagnosed people living with HIV in Switzerland, targeted efforts will be required to improve recruitment and retention of underrepresented populations and to maintain the cohort’s relevance as HIV care continues to evolve. Congratulations, Jessy, for this paper.