TY  - JOUR
T1  - Reconstructing the Causal Dose–Exposure–Response Chain under Adherence Variation, Disease Progression, Concomitant Therapy, and Measurement Error
A1  - Mateo Alvarez
A1  - Sofia Herrera
A1  - Diego Cruz
A1  - Andres Castro
JF  - International Journal of Pharmaceutical Research and Allied Sciences
JO  - Int J Pharm Res Allied Sci
SN  - 2277-3657
Y1  - 2025
VL  - 14
IS  - 3
DO  - 10.51847/ld93VGQWjn
SP  - 103
EP  - 112
N2  - Observed exposure–response associations are often interpreted as evidence that changing dose or exposure will alter clinical outcome. Such interpretation may fail when prescribed dose differs from administered dose, adherence changes over time, disease severity modifies pharmacokinetic disposition, prior response influences subsequent treatment, concomitant medicines alter exposure or pharmacodynamic state, or measurement and sampling processes determine which observations enter the analysis. This Original Causal PK/PD Architecture Article proposes the Causal Dose–Exposure–Response Reconstruction Architecture, an original conceptual synthesis that separates treatment intention, realized administration, absorbed drug input, latent systemic exposure, measured concentration, pharmacodynamic state, latent biological response, and observed outcome. The architecture represents disease progression, time-varying physiology, clinical dose modification, and concomitant therapy as temporally ordered processes whose causal roles depend on the estimand and treatment history. It also distinguishes measurement error, informative sampling, censoring, and missingness from biological variability. The contribution provides a causal graph, construct definitions, testable propositions, failure modes, and staged identification, estimation, and validation requirements. Its central proposition is that no single exposure metric, adherence measure, pharmacokinetic model, pharmacodynamic model, causal estimator, or digital method can independently reconstruct the chain. Validation must separately address dose-input accuracy, structural assumptions, latent-state recovery, observation models, external transportability, causal identification, and decision consequences. The architecture may support study design, assumption auditing, model criticism, and validation planning, but it does not constitute an empirically validated causal model, clinical dosing recommendation, regulatory standard, digital twin, or deployment-ready decision system.
UR  - https://ijpras.com/article/reconstructing-the-causal-doseexposureresponse-chain-under-adherence-variation-disease-progressio-9ckkckazzmdudkw
ER  -