Biography
Dr. Tushar Garg is a Technical Officer in the Innovations & Grants team at the Stop TB Partnership. He works on the Start4All project, evaluating diagnostic algorithms and novel TB diagnostics in 7 countries, and on the TB REACH initiative, which is focused on improving programmatic aspects of tuberculosis case detection and treatment outcomes. Previously, he was a postdoctoral fellow in infectious disease epidemiology at Johns Hopkins University and has managed TB programs in rural India. He holds an MPH from Johns Hopkins University and an MBBS from Sawai Man Singh Medical College, India.
Key Impacts
A Bayesian cost-effectiveness model of TB screening and diagnosis in seven high TB burden low- and middle-income countries
The scale-up of more cost-effective algorithms could lead to increased TB case detection through broader coverage.
Source: Conference 2024
Differentiated TB abnormality thresholds for artificial intelligence interpretation of chest X-rays
While differentiated abnormality threshold scores did not show improvement in both sensitivity and specificity compared to the 0.5 threshold, different thresholds were required to achieve 90% sensitivity for both sex and age groups highlighting the need to adjust thresholds based on context, target population for screening, and testing capacity.
Source: Conference 2024
Testing more for less: Evidence on pooling specimens for molecular testing at healthcare facilities and by active case finding in communities across seven high-burden countries
Pooled testing using Xpert Ultra has high performance and lower cost per person tested, meeting WHO accuracy thresholds in screening test target product profiles. Our results suggest that scale-up of pooling can lead to cost savings and expanded access to RMD in high burden settings.
Source: Conference 2024
Testing more for less: evidence on pooling specimens for molecular testing at healthcare facilities and by active case finding in communities across seven high-burden countries
Impact details available upon request.
Source: Conference 2024