Case-based reasoning system for development of tablet formulations
Chemical Product Design and Engineering (CPD&E)
Chemical Product Design & Engineering - Poster (CPD&E - P)
Keywords: Tablet formulation, Case-based reasoning system
Tablets, a compressed solid dosage form in one unit, are by far the most popular medicinal preparations since they serve benefits of good compliance for the patients and constant quality and good stability for the manufacturers. The design of tablet formulations involves identifying not only an active ingredient providing curing effect, but other necessary excipients, i.e., fillers, binders, lubricants, disintegrants, which improve the quality of the tablets. Therefore, the properties of tablet are a complex of the properties of the active ingredient and excipients itself and the interaction among each components. Occasionally, tablet formulations have dealt with the criteria conflicted where high strength is required together with rapid disintegration. The design of a new tablet involves identifying not only appropriate excipeints but also their quantities in order to balance the properties of the drug, serving for the robust manufacturing and the desired dose of drug absorbed by the patients.
A case-based reasoning system, CBR, is a problem solving technique which imitates a human reasoning and tries to make a decision based on earlier experiences. CBR is beneficial when the problems are rather complicate in which the exact model can not be derived. The most advantage of this system is that a variety of experiences representing a rule, a constraint, some general law or advice, or recording past events of both good and bad results are acceptable. In this study, a CBR for tablet formulations was created by the data from tablet formulations of a number of active ingredients. The application of CBR involved:
1) a database of tablet formulations; tablet composition, amount of active ingredient and excipients, tablet properties (diameter, hardness, disintegration time, drug dissolution),
2) knowledge base for adaptation concluded in the formulation repository; chemical and physical properties of each drug and excipient,
3) a set of rules and functions for measuring similarity; criteria for desired tablet formulations (diameter, hardness, disintegration time, drug dissolution),
4) rules for adaptation implemented in data analysis unit; specific properties following Pharmacopeia’s criteria.
The system made tablet formulation exemplars containing case attributed values that identify the problem type and distinguish one problem type from another. The case attributes that identify the problem type are used as indices in retrieval. Then the exemplars retrieved by the indices are further filtered by applying a similarity to assess their closeness to the problem. This similarity function measures the degree of similarity between two different cases, or the distance between cases. Every case can be characterized by a set of its features in the N-dimensional or multi-dimensional objects and thereby calculated weighted dimension of case’s features. After feature weighting, the correlation among features is analyzed and give the set of features which most affect on the result. Finally, by applying the genetic algorithm, the most frequently occurring excipient in the retrieved exemplars is selected for the desired formulation.
In this study CBR is applied for tablet formulations of the drugs in common, such as drotaverine, glibenclmide and paracetamol. Results show that CBR proposes appropriate ingredients for the desired tablet formulation with an optimal quantities within the limits of the tablet manufacturing process and Pharmacopeia specifications.
Presented Wednesday 19, 13:30 to 15:00, in session Chemical Product Design & Engineering - Poster (CPD&E - P).
