Artificial Intelligence Sparse-Data Estimation of Snakebite Antivenom Stock-Out Risk Under Variable District Pharmacy Dispensing Ledger Completeness: A Causal Transportability Evaluation of Healthcare Procurement Allocation

District pharmacy dispensing ledgers record antivenom issue irregularly, and the completeness of those records varies across districts in ways that are plausibly related to stock-out risk itself. the present analysis treats the problem as one of causal transportability rather than missing-data imputation: the target quantity is the effect of a procurement allocation rule on stock-out risk in a target district population, and the available evidence is a set of source districts whose ledger completeness differs systematically from the target. A formal framework is developed in which ledger completeness is represented as a selection mechanism on the observed sample, and identification of the target effect is shown to require an adjustment set that blocks the selection pathway. A sparse-data estimator is proposed that combines a doubly robust moment condition with a completeness-weighted transport formula, and its large-sample behaviour is characterised. Because no observational data are supplied, the paper reports the operating characteristics of the design under a declared variance structure (between-cluster SD = 6.0, residual SD = 4.5, intraclass correlation = 0.64), with power estimated across 2000 design cells. The analysis is a methodological and theoretical contribution: it specifies the conditions under which antivenom stock-out risk can be transported from ledger-complete to ledger-incomplete districts, and it identifies the completeness threshold below which transport fails.