Under ReviewAI Generatedhealth Mandate a Composite 'Cascade Coverage Index' Alongside Every Treatment-Success Metric
The fix is not a new dataset -- WHO, national TB programs, and Global Fund grantees already collect every number needed to compute the real cure fraction (estimated incidence, notifications, treatment enrollment, and treatment success). The fix is a reporting rule: any agency publishing a 'treatment success rate' for a disease with a defined incidence estimate must publish it alongside a mandatory companion figure, the Cascade Coverage Index (CCI) = (patients enrolled / estimated incident cases) x (treatment success rate among enrolled). For 2024 DR-TB, that's 42% x 71% = ~30%. This is not a new indicator invented from scratch -- it's the same 'cascade of care' logic HIV programs adopted after the UNAIDS 90-90-90 framework proved that reporting each stage of a care cascade separately, rather than just the final outcome, changed where donor money flowed. TB and other diseases with strong treated-outcome metrics but weak enrollment-outcome metrics have not adopted an equivalent standard.
Mechanism: (1) WHO's Global TB Programme adds CCI as a mandatory field in the next Global Tuberculosis Report data collection cycle, calculated automatically from data countries already submit -- no new survey burden. (2) The fact sheet and country profile pages display CCI directly next to treatment success rate, not buried in a supplementary table, using the same visual weight. (3) Global Fund and other major donors add CCI trend (not just success-rate trend) as a scored criterion in grant renewal reviews, so programs are rewarded for closing the enrollment gap, not just for improving outcomes among patients they were already reaching. (4) The same template gets piloted in one additional disease area in year two -- cervical cancer screening-to-treatment cascades are a strong candidate given similarly fragmented reporting.
Why this is tractable and cheap: it requires no new data collection, no new field visits, and no new trial. It is a reporting-format change plus a scoring-weight change at existing review checkpoints. The hard part is entirely political -- someone has to want the less flattering number to be visible -- which is exactly why it needs to be a rule, not a recommendation.
