MedaSync
MedaSync, a forward-thinking healthcare technology company, specializes in creating AI-driven software solutions that streamline MDS (Minimum Data Set) management and improve revenue integrity across all-payer systems for Skilled Nursing Facilities. From April 2019 to June 2022, I collaborated with MedaSync part-time to build and enhance their serverless infrastructure and data processing pipelines.
My role involved developing an AWS serverless infrastructure to process unstructured medical data efficiently, integrating NLP capabilities, and creating daily customer intake reports. The goal was to harness the power of AI and automation to reduce manual workflows, ease the burden on overworked staff, and optimize financial performance for Skilled Nursing Facilities.
Key Projects and Contributions
AWS Serverless Lambda Pipeline for Medical Code Extraction
Developed a comprehensive AWS serverless Lambda pipeline to transform raw intake documents for patients at nursing homes and extract Medicare/Medicaid codes. This project aimed to automate and streamline the processing of unstructured medical data, reducing manual effort and improving data accuracy.
Technologies and Applications:
- OCR Tool: Google Tesseract for converting raw PDFs to text.
- NLP Integration: AWS Comprehend Medical for extracting medical information.
- Serverless Infrastructure: AWS Lambda for processing data.
- Data Storage: AWS DynamoDB for storing processed data.
- Data Transmission: SFTP delivers daily customer intake reports.
Project Phases and Processes:
OCR Processing:
- Received raw PDF documents and used Google Tesseract to perform Optical Character Recognition (OCR) to convert the documents into text format.
NLP with AWS Comprehend Medical:
- Sent the text to AWS Comprehend Medical to extract medical data.
- Evaluated the extracted data to a confidence interval of 70% to ensure business acceptability.
Code Extraction and Transformation:
- Extracted specific types of medical codes from the Comprehend Medical output.
- Transformed the extracted data into standard codes.
- Further transformed the standard codes into Medicare/Medicaid codes.
Data Storage and Reporting:
- Posted the processed information into AWS DynamoDB based on the client ID from the original request.
- Created AWS Lambda functions to generate daily customer intake reports, extracting data from DynamoDB and SQL databases and transmitting them via SFTP in CSV format.
Outcomes
This pipeline significantly improved the efficiency and accuracy of processing unstructured medical data. Automating the extraction and transformation of Medicare/Medicaid codes reduced manual workflows, minimized errors, and ensured consistency in coding. The serverless infrastructure allowed MedaSync to process large volumes of data in real time, enhancing its capability to provide actionable insights and improve revenue integrity for skilled nursing facilities.
Python Flask/React Application for Prescription Search
Developed a proof of concept (POC) Python Flask/React application for fuzzy searching prescription names. The application consumed a dictionary of prescriptions provided by the client. It offered a simple interface for users to search for specific medications/drugs and retrieve the appropriate Medicare/Medicaid code.
Technologies and Applications:
- Backend: Python Flask
- Frontend: React
- Search Functionality: Fuzzy search on prescription names
Outcomes
This POC application enhanced the user experience by providing a fast and efficient way to search for prescription names and retrieve the associated Medicare/Medicaid codes. The successful implementation of fuzzy search algorithms ensured that users could find relevant information even with partial or misspelled inputs, significantly improving the accuracy and usability of the application.
Impact and Practices
Effective AI and Data Processing Integration
MedaSync significantly benefited from the integration of AI and advanced data processing techniques, which streamlined its workflows and improved overall efficiency. By leveraging AWS services and NLP, MedaSync could handle vast amounts of unstructured data, transforming it into valuable insights that enhanced revenue integrity for Skilled Nursing Facilities.
Robust Compliance and Data Security
Ensuring compliance and data security was a cornerstone of this engagement. By utilizing AWS's robust security features and following best practices for data handling, MedaSync maintained the highest data privacy and security standards, reinforcing its reputation as a trusted healthcare technology provider.
Outcomes and Reflections
The collaboration with MedaSync underscored the importance of combining technical excellence with AI integration and automated data processing in the healthcare sector. MedaSync was well-positioned to deliver innovative and secure healthcare solutions by improving data processing capabilities and ensuring robust compliance. This engagement highlighted the critical role of AI, serverless infrastructure, and continuous improvement in achieving technological excellence in healthcare.
Conclusion
Partnering with MedaSync was a rewarding experience showcasing AI's transformative power and serverless infrastructure in healthcare technology. Developing an AWS serverless Lambda pipeline enabled efficient processing of unstructured medical data, while AWS Comprehend Medical integration provided valuable clinical insights. These contributions were instrumental in improving revenue integrity and operational efficiency for Skilled Nursing Facilities. The success of these initiatives aligns with Dev3loper.ai's mission to leverage cutting-edge technology and best practices to drive transformative change across diverse industries.
Get In Touch
We'd love to hear from you! Whether you have a question about our services, need a consultation, or just want to connect, our team is here to help. Reach out to us through the form, or contact us directly via social media.
