Socrates Software Solutions, LLC, an Artificial Intelligence (AI) focused startup, seeks funding for the development and marketing of Med Sage, a Cloud Based AI Clinical Decision Support System for Ambulatory Care and Outpatient Care. Med Sage will be built with a patent pending set of concepts, algorithms, and processes that overcome the computing and business viability challenges to developing AI based clinical decision support applications that can be used on a broad scale by ambulatory and outpatient care providers across the U.S. and across the world by simply logging into the cloud-based interface. The processes and algorithms needed for development and operationalization of Med sage has not only been worked out and outlined (e.g. much of it described in the provisional patent application) but a functional initial prototype of Med Sage has already been developed and is available.
An initial prototype of Med Sage is already available for exploration. Screen shots of the prototype has been provided below. To obtain weblink to the prototype please contact Hussain Yusuf at [email protected] Note: Please access and explore the prototype application accessible through the link above only for the purpose of considering funding the proposal that is the context of this email. Please only share the link on this page with other persons only for that same purpose.
What Med Sage Will Enable:
- An approved user (e.g. clinician) to log into the user interface
- Input patient information including Patient history, Exam findings, Lab findings, Imaging findings, Procedure findings, Social/Family history.
- Input the patient information by selecting patient findings from drop down lists that are automatically brought up based on partial typing in of text in the patient fields. These patient information fields will be organized in multiple tabs.
- Input patient information by dragging and dropping a PDF file or XML file containing patient information [such as a pdf file generated by the clinicians EMR software].
- Submit the patient information to generate a Digital Report that will include -
- Differential Diagnoses - a list of 10-25 Dx that scored the highest based on the patient inputs; these will be ordered by their scores
- For each of the diagnoses in the Digital Report, there will be:
- Clinical Guidance Information
- Lists of most common Lab tests & results, Imaging & results, Dx procedures & findings, Therapeutic procedures, Medications, and Comorbidities seen among patients with the diagnosis in a nationally representative sample to ambulatory care patients.
- Publication links to high metric (e.g. high altmetric) publications associate with the diagnosis.
- If the user chooses, she/he can also directly input diagnoses (instead of generating diagnoses through inputting patient findings) into an application tab with fields for such and then submit these diagnoses to generate a Digital Report that has the same outputs as when generated by inputting patient findings.
The Opportunities for AI Based Clinical Decision Support
- Increasing body of medical knowledge can pose a challenge to clinicians in ensuring that all necessary aspects are taken into account during patient management
- However, advances in medical data, informatics, and AI methodologies now enable the development of computerized systems and services to support clinicians in patient management.
The Challenges to Developing AI Based Clinical Decision Support - Disease occurrence and progression is a complex process influenced by multiple interacting factors – thus, difficult to appropriately capture in computer programs and algorithms
- Requires large volumes patient data that meet specific criteria related to types of information captured, data completeness, data standardization, and data representativeness
- Requires large computing power and processing times
- All of the above reduce business viability
Innovations in Med Sage
- Computation-intensive steps occur during application development and maintenance >> Reducing computation time and costs during individual instances of use of the application.
- Innovative algorithm for cumulative weighted scoring for deriving diagnoses that accounts for:
- Strength of association between abnormal findings and the disease
- Strength of association between normal findings and ruling out the diseases
- How specific is the abnormal finding to the ds
- How common is the disease
- Comparative weighting of positive and negative associations and other predictors.
- Uses large pools of EMR data to develop Nationally Representative Sample of Out-Patient Visits.
- Regression and other algorithms to generate vast repository of positive and negative predictive values between abnormal and normal clinical findings and ICD-10-CM based Diagnoses.
- Clinical support material related to important management aspects are based on representative EMR data pool.
Advantages of Med Sage vs Other Potential Systems
- Development of diagnosis and management algorithms based on medical text is very resource intensive and not well aligned with large volumes of codified EMR based data that is available. >>> Med Sage, instead, uses real-world EMR patient data to achieve its objectives.
- Multiple Regression based approaches can be too restrictive to specific circumstances (if explanatory leaning) or insufficient to derive diagnoses (if predictive leaning) and models require large computing resources and time to run – thus decreasing practical viability. >>> Med Sage, uses Regression modeling only in application development and maintenance phases and avoids it during actual use by the user.
- Neural network, decision tree, and random forest based approaches can be accurate in deriving diagnoses. However, the high computing intensity and time required decrease their practicality. >>> Med Sage, instead uses an innovation based on predictive and rule out scores and weights of patient findings vis-a-vis potential diagnoses.
The Business Case
Revenue will be generated through per use fees and/or subscription fees,
Near Term Outlook
- Each year ~800,000,000 Ambulatory Care Visits occur in the U.S.
- If Med Sage was used in 1/1,000 visits, this would result in >> 800,000 instances of use every year
- There are ~300,000 Primary Care Providers in the U.S.
- If 1% subscribed to Med Sage, this would result in ~3,000 subscribers
- Globally there is much larger potential for use by clinical care providers
Long-term Outlook
- Med Sage will be expanded into clinical decision support for other medical specialties
- Further developments of AI capabilities of Med sage will enable narrowing potential Diagnoses and recommending more specific management plans
- Special Note: The Core Basis of Med sage is the vast pool of predictive values and rule out values within binary sets of all ICD-10-CM diseases diagnoses and all coded clinical findings. This will be generated during development of Med Sage. This will be resource that can potentiate many other developments.
Major Resource Needs for Med Sage - Salaries for development manager and applications programmers
- Large pool/s of de-identified EMR data
- Subscriptions to large cloud-based data storage and analytic platform
- Clinical consultants for identification, curation, and/or development of clinical guidance documents.
- Marketing
Broadscale AI based clinical decision support is an inevitability. The question is: who will be the pioneers?
Thank you.
Have a good day.
Hussain Yusuf, M.D., M.P.H.
Socrates Software Solutions, LLC
Email: [email protected]
Ready to Ask For Funding for your company?
Post a Funding Request