Dynamic Healthcare Message Customization via Predictive Patient Profiling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for addressing non-compliance and non-adherence in healthcare, such as questionnaires and biophysical tests, are time-consuming and do not provide an accurate estimation of adherence, leading to reduced treatment efficacy and increased healthcare costs.
Innovation Solution
A computer-implemented method that uses a trained mathematical model to predict patient profiles, generating dynamically customized messages for both patients and healthcare practitioners, reducing the number of questions needed and improving response reliability by providing real-time feedback and personalized interventions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional questionnaires and biophysical tests are used to assess non-compliance, then comprehensive data can be collected, but the process becomes time-consuming and reduces patient engagement
Solution Approach 1:
The system performs preliminary assessment by analyzing electronic health record data before administering the full questionnaire. This preliminary action identifies patients at risk of non-adherence early in the process, allowing clinicians to focus detailed assessment only on those who need it, thereby reducing overall assessment time while maintaining accuracy.
Solution Approach 2:
The system extracts and analyzes specific high-value data elements from electronic health records (such as demographic information, medical history, and treatment data) to create a risk profile. This extraction of critical information separates the essential assessment components from the full questionnaire, enabling faster initial screening.
2Measurement precision
If comprehensive questionnaires are administered to all patients, then complete adherence data can be obtained, but patient engagement and response reliability decrease
Solution Approach 1:
The system applies different assessment intensities to different patient groups based on their risk profiles. High-risk patients receive comprehensive assessment while low-risk patients receive streamlined assessment. This localized quality approach ensures that each patient receives the appropriate level of scrutiny, maintaining response reliability for those who need it while preserving engagement for others.
Solution Approach 2:
Instead of administering full questionnaires to all patients, the system uses partial assessment through EHR data analysis for initial screening. This partial action is sufficient to identify most at-risk patients, and full questionnaires are administered only when necessary, thereby maintaining engagement while obtaining adequate adherence data.
3Measurement precision
If manual assessment of non-compliance is performed, then detailed individual evaluation is possible, but productivity and scalability are reduced
Solution Approach 1:
The system replaces manual mechanical assessment with automated computational analysis of EHR data. Machine learning algorithms process patient data to generate adherence risk scores, eliminating the need for manual review of each patient record while maintaining or improving assessment accuracy. This substitution dramatically increases productivity and scalability.
Solution Approach 2:
The system enables self-assessment through automated EHR data collection and analysis. The assessment process serves itself by automatically gathering relevant data from existing electronic records without requiring manual intervention, thereby increasing throughput while maintaining individualized assessment quality.
Data Source
AI summary
A computer-implemented method for providing a dynamically customized subject message and a dynamically customized health practitioner message to each subject and health practitioner in a healthcare facility, respectively, wherein each health practitioner is assigned to a plurality of subjects, the method comprising for each subject receiving subject data of the subject, receiving responses of the subject to one or more first questions of a questionnaire, generating an input dataset from the subject data and the received responses, applying a trained mathematical model on the generated input dataset to predict a profile of the subject, determining that a predetermined configuration requirement has been met, receiving healthcare facility data of the healthcare facility, and determining performance information of at least one of the plurality of subjects based on the healthcare facility data and on the subject data and predicted profile.