Advanced analysis of formulation parameters governing encapsulation efficiency: drug delivery system

Area of Science:

• Pharmaceutical Sciences
• Computational Chemistry
• Systems Biology

Background:

• Encapsulation efficiency (%EE) is a key performance metric for niosomal drug delivery systems.
• Predictive modeling of %EE often lacks the causal insights needed for intervention-oriented reasoning.
• Existing methods struggle to differentiate predictive relevance from true causal influence in formulation development.

Purpose of the Study:

• To apply a causal discovery framework to understand the causal relationships between drug properties, formulation, process conditions, and %EE in niosomal systems.
• To reveal assumption-aware causal structures from literature-derived data.
• To move beyond predictive modeling towards intervention-focused understanding.

Main Methods:

• Utilized a causal discovery framework on 116 observational samples of thin-film hydration for niosomal formulations.
• Employed constraint-based (PC Algorithm), score-based (Greedy Equivalence Search - GES), and functional (Linear Non-Gaussian Acyclic Model - LiNGAM) causal discovery models.
• Selected input variables based on literature reporting, physicochemical relevance, and data availability.

Main Results:

• Encapsulation efficiency (%EE) was confirmed as a downstream outcome in the causal graphs.
• Intrinsic drug properties, particularly lipophilicity, were identified as upstream constraints rather than direct intervention levers.
• A significant divergence was observed between predictive feature importance and causal effect estimates, highlighting the limitations of purely predictive approaches.

Conclusions:

• Causal discovery offers a robust framework for understanding complex relationships in niosomal drug delivery, surpassing traditional predictive modeling.
• Intrinsic drug properties should be viewed as constraints guiding formulation, not as direct targets for manipulating %EE.
• The developed framework supports hypothesis generation, counterfactual analysis, and structured experimental design for optimizing niosomal formulations.