Stochastic Learning Model for ASD Diagnosis Using EHR Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current diagnostic methods for Autism Spectrum Disorders (ASD) and Gestational Diabetes Mellitus (GDM) lack reliable biomarkers, leading to delayed interventions and increased costs, with existing tools requiring extensive data and laboratory tests, limiting their applicability and accuracy.
Innovation Solution
A computer-based method utilizing stochastic learning algorithms and Hidden Markov Models to predict ASD and GDM diagnoses from unprocessed raw data, including diagnostic codes, without the need for specific blood work or laboratory results, enabling early intervention and reducing diagnostic burdens.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional diagnostic methods using blood work and laboratory tests are used for ASD and GDM, then measurement precision may be improved, but device complexity and ease of operation deteriorate due to extensive data requirements and procedural burden
Solution Approach 1:
The patent extracts and removes the need for complex blood work and laboratory tests from the diagnostic process. Instead, it uses only unprocessed raw data from electronic health records such as diagnostic codes, patient demographics, and historical medical information. This extraction principle eliminates the burden of invasive procedures while maintaining diagnostic capability through computational analysis of existing data.
Solution Approach 2:
The patent replaces the mechanical and laboratory-based diagnostic system with a computational system using stochastic learning algorithms and Hidden Markov Models. Instead of physical blood tests and manual analysis, the system uses machine learning to analyze electronic health record data, substituting mechanical procedures with automated computational processing that reduces operational burden.
2Measurement precision
If comprehensive blood work and laboratory tests are performed for accurate diagnosis, then measurement precision improves, but loss of time increases due to repeated testing and delayed results
Solution Approach 1:
The patent performs preliminary diagnostic analysis using data that already exists in electronic health records before clinical appointments or laboratory tests are conducted. By analyzing unprocessed raw data such as diagnostic codes and historical medical information in advance, the system provides preliminary risk assessments that guide subsequent clinical decisions, eliminating the need for repeated testing and reducing overall diagnostic time.
Solution Approach 2:
The patent enables continuous diagnostic monitoring by repeatedly analyzing updated electronic health record data over time. Instead of discrete, time-consuming laboratory tests, the system continuously processes available data to update diagnostic risk assessments, maintaining measurement precision while eliminating gaps and delays between testing intervals.
3Measurement precision
If specific blood work and laboratory test results are required for diagnosis, then measurement precision improves, but device complexity and data requirements increase, limiting applicability
Solution Approach 1:
The patent creates a universal diagnostic system that can diagnose multiple conditions including Autism Spectrum Disorder and Gestational Diabetes Mellitus using the same computational framework and data sources. The system processes unprocessed raw data from electronic health records for various diagnoses without requiring condition-specific laboratory tests, making the diagnostic tool universally applicable across different medical conditions and patient populations.
Solution Approach 2:
The patent uses copies of existing data from electronic health records rather than requiring original laboratory test results. By analyzing diagnostic codes, historical medical information, and patient demographics that already exist in digital form, the system creates diagnostic assessments without needing physical blood samples or laboratory infrastructure, reducing data requirement burden while maintaining diagnostic reliability.
Data Source
AI summary
A diagnosis prediction (DP) computing device (102) receives training datasets from a health records server (108A), an insurance claims server (108B), and other third party servers (108C). DP computing device builds a model based on the training datasets and stores the model on a database (106) via a database server (104). Using the model and a stochastic learning algorithm, a risk estimator (110) determines a prediction of a disease or disorder diagnosis of a patient to a client device (112). The prediction is based on data gathered pertaining to the patient including unprocessed raw data comprising records of diagnostic codes generated during past medical encounters from an insurance claims database.