Abstract
OBJECTIVE: HIV cohorts are an important source of clinical data for informing public health policies and programmes. However, the generalizability of cohort findings to the wider population of people diagnosed with HIV in each country remains unclear. In this work, we assessed the representativeness of six large national HIV cohorts within Europe. DESIGN AND METHODS: Individual-level cohort data were provided from national cohorts in France, Germany, Greece, Italy, Spain and the United Kingdom. Analysis focused on new HIV diagnoses reported to The European Surveillance System (TESSy) during three time periods (2000-2004, 2005-2009 and 2010-2013), to allow for temporal changes. Cohort and TESSy records were matched and compared by age, sex, transmission mode, region of origin and CD4+ cell count at diagnosis. The probability of being included in each cohort given demographic characteristics was estimated and used to generate weights inversely proportional to the probability of being included. RESULTS: Participating cohorts were generally representative of the national HIV-diagnosed population submitted to TESSy. However, people who inject drugs, those born in a country other than that reporting the data, those with low CD4 cell counts at diagnosis, and those more than 55 years were generally underrepresented in the cohorts examined. CONCLUSION: These European cohorts capture a representative sample of the HIV-diagnosed populations in each country; however some groups may be underrepresented.