Biologic Alternative Recommendation System Using NLP and Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional approaches for selecting alternatives to biologics do not consider patient information and identifying adverse drug reactions (ADRs), leading to unaffordable high-cost treatments and limited options for patients, especially in polypharmacy scenarios, resulting in reactive treatment decisions and potential rehospitalization.
Innovation Solution
A method and system that utilize hardware processors to extract metadata from e-prescriptions, perform text mining on biomedical literature, and apply relative scoring techniques to recommend low-cost biosimilars or interchangeable drugs, considering genetic information and ADRs, while identifying primary and secondary genes responsible for drug efficacy and interactions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If conventional approaches are used to select alternative biosimilars, then the selection process is simple, but patient information is not considered leading to unaffordable high-cost treatments
Solution Approach 1:
The system segments the alternative selection process into multiple components: extracting patient metadata (genetic information, previous diseases, current medications), identifying primary ADRs through text mining, evaluating reference biologic drugs, and generating a structured recommendation report. This segmentation allows comprehensive patient information analysis while maintaining process organization.
Solution Approach 2:
The system introduces an intermediary processing layer that includes metadata extraction modules, text mining engines, and recommendation generation components. This intermediary layer bridges the gap between simple biosimilar selection and comprehensive patient-specific analysis, enabling informed decision-making without overwhelming complexity.
2Reliability
If comprehensive patient information analysis is performed to identify ADRs and recommend alternatives, then treatment decisions are more informed, but the system complexity increases
Solution Approach 1:
The system performs preliminary actions by extracting and storing patient metadata (genetic information, previous diseases, current medications) before the actual recommendation process. It also pre-identifies primary ADRs through text mining of biomedical literature, so that when a recommendation is needed, the analysis is already partially complete, reducing real-time complexity.
Solution Approach 2:
The system implements feedback mechanisms by evaluating multiple reference biologic drugs against patient-specific criteria and selecting the optimum alternative based on comprehensive comparison. The structured recommendation report provides feedback on why certain alternatives are recommended, enabling verification and refinement of the decision-making process.
3Object-affected harmful factors
If low-cost biosimilars are recommended, then treatment affordability improves, but access to information about alternatives is not available with physicians at consultation time
Solution Approach 1:
The system performs preliminary information gathering and analysis by maintaining a database of reference biologic drugs, their ADRs, and alternative biosimilars. Patient metadata is extracted and analyzed in advance, so that when a physician needs a recommendation during consultation, the system can quickly generate informed alternatives without time-consuming research at the moment of decision-making.
4Adaptability or versatility
If patient-specific factors are considered for biologic alternatives, then treatment personalization improves, but the ability to identify secondary adverse reactions in polypharmacy scenarios is limited
Solution Approach 1:
The system segments the ADR identification process into primary ADRs (directly associated with the prescribed biologic) and secondary ADRs (resulting from interactions with other medications). This segmentation is achieved by analyzing patient metadata including current medications and using text mining to identify interaction-related adverse reactions, enabling comprehensive detection in polypharmacy scenarios.
Data Source
AI summary
High cost of biotherapy drug makes it unaffordable for patients to seek treatments. Further, access to information related to new low-cost alternatives like Biosimilars and Interchangeable may not be available with the physician at the time of consultation. Most of the conventional approaches aims to select an alternative biosimilar for a reference drug without considering patient's information. The present disclosure recommends a list of low-cost alternatives to high-cost reference drugs thereby enabling the physician to get timely and updated information on development of Biosimilars. The solution leverages Natural Language Processing (NLP) technology to extract known adverse events for a reference drug and a relative scoring based technique to identify and optimum alternative to prescribed biologics. The capability of the solution is further extended to identify secondary adverse events due to multiple drugs, thereby providing a clinical decision support system to help physicians take an informed decision.