Artificial Intelligence Causal Inference of Community Clinic HIV Viral-Load Monitoring Spacing Under Unmeasured Dried Blood Spot Degradation: A Healthcare Doubly Robust Virological Failure Evaluation

Dried blood spot (DBS) specimens used for HIV viral-load monitoring in community clinics degrade under ambient heat and humidity, and the degradation is unmeasured in routine practice. Because degradation is plausibly related both to monitoring interval and to virological failure, naive comparisons of monitoring spacing are confounded. The framework combines an artificial-intelligence-assisted degradation proxy estimated from specimen metadata with an augmented inverse-probability-weighted estimator that is consistent if either the treatment model or the outcome model is correctly specified. A replication-based power analysis across 2000 design cells, with between-cluster SD 6.0, residual SD 4.5 and intraclass correlation 0.64, indicates that 56 clusters are required to detect a 5-percentage-point effect at 80% power. The analysis is methodological: no observational data are used or claimed. The contribution is a design-level account of what can and cannot be identified about monitoring spacing when specimen integrity is unobserved, together with the cluster counts that a BACI evaluation would need.