Medical History Generation via Medication-Based Expert System
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing computer-based medical history taking systems are inefficient due to the need for patients to answer numerous irrelevant questions, leading to errors from misunderstanding, forgetfulness, and fatigue, with no mechanism to identify the most important questions upfront, resulting in an error rate of 3%-7%.
Innovation Solution
A patient-driven system utilizing an expert system and machine learning to generate a medical history based on the patient's medications, which prioritizes questions using a two-tiered drop-down list and sequential questioning to ensure accuracy and relevance, leveraging historical data to refine answers and reduce errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a predetermined set of questions is used in computerized medical history systems, then all pertinent questions are included and structure is maintained, but patients must answer too many irrelevant questions leading to errors from misunderstanding, forgetfulness, and fatigue
Solution Approach 1:
The questionnaire is segmented into multiple levels: a brief initial screening phase that captures key information quickly, followed by optional detailed sections. Patients can complete the essential brief version in minimal time, while those needing more comprehensive assessment can proceed to additional questions, thus reducing time loss for most patients while maintaining reliability through the structured brief assessment.
Solution Approach 2:
The system implements a brief version that captures the essential medical history information needed for most cases, representing a partial action that suffices for routine assessments. This brief version reduces the burden of answering all possible questions while maintaining sufficient accuracy for common medical evaluations.
2Adaptability or versatility
If branch-chain logic is used to hone in on specific patient complaints, then question relevance is improved, but the system still requires patients to answer numerous questions before reaching relevant topics
Solution Approach 1:
The system performs preliminary action by presenting a brief screening questionnaire first that quickly identifies the patient's chief complaints and relevant symptom categories. Based on these initial responses, the system then adaptively presents only the specific detailed questions relevant to the patient's condition, avoiding the need to navigate through unrelated questions and reducing overall complexity perception.
Solution Approach 2:
The questionnaire structure is made dynamic rather than static. The system adapts the question sequence and content based on patient responses in real-time, presenting different question paths for different patient presentations. This dynamic adaptation improves patient-specific relevance while managing complexity through intelligent routing rather than rigid hierarchical structures.
3Productivity
If patients self-report medical information in computer interviews, then data collection efficiency is improved, but error rates increase to 3%-7% due to patient misunderstanding, forgetfulness, and carelessness
Solution Approach 1:
The system implements feedback mechanisms including confirmation questions that ask patients to verify their responses, consistency checks that compare answers across different questions, and validation rules that flag potentially erroneous responses for review. This feedback loop maintains high data collection efficiency while reducing errors by immediately identifying and correcting misunderstandings or inconsistencies.
Solution Approach 2:
The system prepares patients beforehand by providing clear instructions, examples, and context for each question section. It also includes built-in validation and guidance that cushion against potential errors by preventing common mistakes before they occur, such as offering response options in multiple formats (text, checkboxes, dropdowns) to accommodate different patient comprehension levels and reduce forgetfulness-related errors.
Data Source
AI summary
A system and method for generating a medical history that is determined based on the patient's medication list. Medications are taken for specific indications, i.e., disease and/or symptom, and the system includes a knowledge base of all known medications and associated indications. Preferably, an expert system allows the patient, nurse or other user, to enter all or part of the name of the patient's medications, and creates a list of probable medications by determining which of the known medications the patient most likely takes. Based on a user selected probable medication, the system creates a list of probable medical indications associated with the selected medication. Based on a user selected probable indication, the system generates a medical history for the patient. The probable medications and probable medical indications are determined preferably based on the patient's demographic data, historical data for other patients, and responses to follow-up questions generated by the system.