Artificial Intelligence Causal Effect Estimation of Oral PrEP Discontinuation Under Variable Pharmacy Dispensing Record Linkage: A Target Trial Emulation Evaluation of Healthcare HIV Prevention Equity

Pharmacy dispensing records are the primary source for measuring oral pre-exposure prophylaxis (PrEP) discontinuation, yet their linkage to clinical and demographic data is variable and often incomplete. the present analysis evaluates whether a target trial emulation framework, combined with machine learning-based probabilistic record linkage, can recover unbiased causal estimates of PrEP discontinuation under realistic linkage error. Because no observational data were available, the study characterises the operating characteristics of a before-after-control-impact (BACI) design under a declared variance structure (between-cluster SD = 6.0, residual SD = 4.5, intraclass correlation = 0.640). Power was estimated across 2000 replications per design cell for cluster counts of 8, 16, 24 and 32 and effect sizes of 3, 5 and 8 percentage points.54 clusters are required to detect a 5-percentage-point effect at 80% power. Minimum detectable effect curves and coefficient plots illustrate the precision achievable at each design point. The contribution is a formal framework that treats record linkage error as a design parameter rather than a nuisance, and specifies the conditions under which causal conclusions about PrEP discontinuation are defensible. The framework has implications for equity monitoring in HIV prevention programmes where dispensing data are fragmented across pharmacy systems.