Scale-up of HIV index testing in an urban population: Experiences and achievements from Nairobi County, Kenya

Objective: To describe the implementation strategies of the index testing program across Nairobi County in Kenya, assess outcomes along the HIV index testing cascade (acceptance, elicitation ratio, HIV positivity and linkage to treatment), and assess annual changes along the HIV index testing cascade during the first 2 years of implementation. Methods: Retrospective analysis of programmatic aggregate data collected from October 2017 to September 2019 after the roll-out of index testing services in 48 health facilities in Nairobi County. Proportions and ratios were calculated for acceptance, elicitation ratio, testing uptake and HIV positivity. We compared these outcomes between years using a chi-squared test, Fisher's exact test or Wilcoxon sign test, and we assessed trends using the Mann-Kendall test. Results: Testing among eligible partners increased from 42.4% (1471/3470) to 74.9% (6114/8159) in the general population, and the positivity yield remained high across both years (25.2% in year 1 and 24.1% in year 2). Index testing positivity yield remained significantly higher than other testing modalities (24.3% vs. 1.3%, p < 0.001). The contribution of index testing services to the total number of HIV-positive individuals identified increased from 7.5% in the first year to 28.6% in the second year (p < 0.001). More men were tested, but the positivity yield was higher among women (30.0%) and those aged 50 years or older (32.4%). Testing eligible partners in key populations (KPs) decreased from 52.4% (183/349) to 40.7% (109/268) (p = 0.674); however, the HIV positivity yield increased from 8.6% to 23.9% (p < 0.001) by the second year of implementation. The HIV positivity yield from index testing remained higher than other testing modalities (14% vs. 0.9%, p < 0.001) for KPs. Conclusion: Index testing was well-accepted and effective in identifying individuals living with HIV in a Kenyan urban setting across both general populations and KPs. Ongoing adaptations to the strategies deployed as part of index testing services helped improve most of the outcomes along the index testing cascade.