Artificial Intelligence Causal Effect Estimation of Community Antibiotic Stewardship Prescribing Under Variable Primary Care Consultation Coding Completeness: A Target Trial Emulation Evaluation of Healthcare Resistance Containment

Community antibiotic stewardship programmes are evaluated almost exclusively through routinely coded primary care consultation records, yet the completeness of those codes varies systematically across clinics and is rarely modelled as a source of bias. the present analysis treats the problem as a design question rather than a data-cleaning problem. It formalises the target trial for stewardship prescribing under incomplete consultation coding, derives the bias of the naive before-after estimator, and characterises the operating characteristics of a before-after-control-impact (BACI) estimator that exploits non-stewardship clinics as a comparison series. The evaluation is a design and power analysis, not an observational study: no patient records were analysed. Under a declared variance structure with between-cluster standard deviation 6.0, residual standard deviation 4.5 and intraclass correlation 0.64, 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. Fifty-four clusters are required for 80% power at a 5 percentage point effect. Coding completeness should enter the design stage as a stratification variable, and that the BACI contrast is the minimum defensible identification strategy for stewardship claims drawn from administrative primary care data.