Amber K Gupta

personal photo of Amber K Gupta

Health Informatics

Date
By:Amber K Gupta

taught by Jon Duke

Syllabus

Course Page

In the CS6440 Introduction to Health Informatics course at Georgia Tech, I have learned a variety of key concepts that are essential for understanding and working in the field of health informatics. Here is a summary of the main concepts I covered:

1. Fundamentals of Health Informatics:

  • Introduction to Health Informatics: Overview of the field, its history, and its role in improving healthcare delivery.

  • Current State of Healthcare: Understanding the existing healthcare infrastructure, processes, and the role of technology in healthcare delivery.

  • Technical Challenges in Healthcare: Identifying and analyzing the technical issues faced by healthcare systems, such as interoperability, data security, and scalability.

  • Laws, Policy & Regulations: Learning about the legal and regulatory frameworks governing health informatics, including HIPAA, HITECH Act, and other relevant policies.

  • Software Development in Healthcare: Understanding the software development lifecycle specific to healthcare applications, including requirements gathering, design, implementation, testing, and maintenance.

  • Health Data Standards: Understanding standards like HL7, FHIR, ICD, SNOMED, and LOINC for interoperability and data exchange.

2. Electronic Health Records (EHRs):

  • EHR Systems: Structure, functionality, and implementation of EHR systems.

  • Data Standards: Familiarity with healthcare data standards such as OHDSI Atlas and OMOP CDM, which facilitate data interoperability and analytics.

  • Data Management: Best practices for capturing, storing, and retrieving health data within EHR systems.

  • Interoperability: Techniques and challenges in achieving interoperability between different EHR systems.

    • HL7 FHIR: Understanding and implementing the HL7 FHIR standard for health data exchange.

    • HAPI FHIR: Learning to use the HAPI FHIR framework to implement FHIR-based solutions.

    • SMART on FHIR: Developing applications that leverage the SMART on FHIR platform for integrating with EHR systems.

    • OMOP on FHIR: Combining the OMOP Common Data Model (CDM) with FHIR for advanced data integration and analysis.

    • CDS Hooks: Implementing clinical decision support systems using CDS Hooks.

    • DaVinci: Understanding and applying DaVinci project guidelines for payer-provider data exchange.

3. Clinical Decision Support Systems (CDSS):

  • CDSS Design: Principles and frameworks for designing and implementing clinical decision support systems.

  • Knowledge Representation: Methods for representing medical knowledge in CDSS, including rules-based systems and machine learning approaches.

4. Health Information Exchange (HIE):

  • HIE Models: Understanding different models of health information exchange, including centralized, federated, and hybrid approaches.

  • Data Sharing: Strategies and challenges in sharing health information across different organizations and systems.

5. Privacy and Security:

  • HIPAA Regulations: Understanding the Health Insurance Portability and Accountability Act (HIPAA) and its implications for data privacy and security.

  • Data Security: Techniques for securing health information, including encryption, access control, and audit trails.

6. Data Analytics in Healthcare:

  • Healthcare Analytics: Techniques for analyzing health data to derive insights and support decision-making.

  • Predictive Modeling: Applying statistical and machine learning methods to predict health outcomes and trends.

  • Big Data in Healthcare: Challenges and opportunities associated with the use of big data in healthcare.

  • Tools for Data Analysis and Visualization: Gaining proficiency in tools like R Studio, ClarityNLP, Leaf, and Excel for analyzing and visualizing healthcare data.

7. Patient-Centered Care:

  • Patient Engagement: Tools and strategies for engaging patients in their own care through technology.

  • mHealth and Telemedicine: Use of mobile health (mHealth) applications and telemedicine to support patient care and monitoring.

8. Health Informatics Research:

  • Research Methods: Approaches to conducting research in health informatics, including study design, data collection, and analysis.

  • Evaluation of Health IT: Methods for evaluating the impact and effectiveness of health information technology interventions.

9. Healthcare Delivery Systems:

  • Healthcare Systems: Structure and functioning of different healthcare delivery systems, including hospitals, primary care, and integrated care models.

  • Quality Improvement: Techniques for improving the quality and efficiency of healthcare delivery through informatics.

10. Ethical and Legal Issues:

  • Ethical Considerations: Ethical issues in health informatics, including patient consent, data ownership, and the digital divide.

  • Legal Frameworks: Understanding legal frameworks and policies that impact the use of information technology in healthcare.

These concepts have provided me a comprehensive understanding of how information technology can be leveraged to improve healthcare delivery, patient outcomes, and overall healthcare system efficiency.

In this course at Georgia Tech, I gained practical experience through various homework assignments and projects. Here's a summary of the practical experiences I acquired:

1. Working with Health Data Standards:

  • HL7 and FHIR Implementation: Practical tasks involving the use of HL7 (Health Level Seven) and FHIR (Fast Healthcare Interoperability Resources) standards for health data exchange and interoperability.

  • Coding with ICD, SNOMED, and LOINC: Assignments involving the use of coding systems like ICD (International Classification of Diseases), SNOMED (Systematized Nomenclature of Medicine), and LOINC (Logical Observation Identifiers Names and Codes) for standardizing health data.

  • Using State-of-the-Art Tools: Gaining practical experience with advanced health informatics tools and standards such as HL7 FHIR, SMART, CDS Hooks, OMOPonFHIR, OHDSI Atlas, and OMOP CDM.

  • Software Development: Developing and deploying health informatics solutions, with an emphasis on building scalable, secure, and interoperable systems.

2. Electronic Health Records (EHRs):

  • EHR System Design and Implementation: Projects involving the design and implementation of EHR systems, including data capture, storage, and retrieval.

  • Data Interoperability: Practical exercises focusing on achieving interoperability between different EHR systems, including data mapping and transformation tasks.

3. Clinical Decision Support Systems (CDSS):

  • CDSS Development: Creating simple clinical decision support tools using rules-based systems or integrating machine learning models to support clinical decision-making.

  • Knowledge Representation: Implementing methods for representing medical knowledge in CDSS, such as using ontologies and decision trees.

4. Health Information Exchange (HIE):

  • HIE Implementation: Projects involving the design and implementation of health information exchange systems, including the setup of centralized, federated, or hybrid HIE models.

  • Data Sharing Solutions: Developing and testing solutions for secure and efficient data sharing across different healthcare organizations.

  • Implementing FHIR-Based Solutions: Developing solutions that comply with the HL7 FHIR standard to ensure data interoperability between different healthcare systems.

  • SMART on FHIR Applications: Creating applications that integrate with EHR systems using the SMART on FHIR platform.

  • Advanced Data Integration: Using OMOP on FHIR to integrate and analyze diverse health datasets.

5. Privacy and Security:

  • HIPAA Compliance: Practical tasks ensuring compliance with HIPAA regulations, including implementing data security measures like encryption and access controls.

  • Security Audits: Conducting security audits and vulnerability assessments on health information systems to ensure data protection.

6. Healthcare Data Analytics:

  • Data Analysis Projects: Using statistical and machine learning techniques to analyze healthcare data for insights and decision support.

  • Predictive Modeling: Developing predictive models for various healthcare applications, such as disease prediction, patient readmission risk, and treatment outcomes.

  • Big Data Handling: Working with large healthcare datasets, employing techniques to manage, process, and analyze big data effectively.

7. Patient-Centered Care Technologies:

  • mHealth App Development: Designing and developing mobile health (mHealth) applications to support patient engagement and self-management.

  • Telemedicine Solutions: Implementing telemedicine platforms and conducting usability testing to enhance patient care and accessibility.

8. Research and Evaluation in Health Informatics:

  • Research Projects: Conducting research studies in health informatics, including designing studies, collecting and analyzing data, and interpreting results.

  • Health IT Evaluation: Evaluating the impact and effectiveness of health information technology solutions through case studies and real-world applications.

  • Programming in Health Informatics: Enhancing programming skills relevant to health informatics, including working with APIs, data transformation, and software development best practices.

  • Designing Healthcare Solutions: Developing the ability to design meaningful healthcare solutions that address specific clinical needs and challenges.

9. Quality Improvement Projects:

  • Process Improvement: Applying informatics tools to improve healthcare processes and outcomes, such as reducing wait times, enhancing care coordination, and improving patient safety.

  • Data-Driven Decision Making: Using data analytics to support quality improvement initiatives and evidence-based decision-making in healthcare settings.

10. Ethical and Legal Considerations:

  • Ethical Scenarios: Analyzing and addressing ethical issues in health informatics, such as patient consent, data privacy, and the digital divide.

  • Legal Frameworks: Applying legal knowledge to practical situations, ensuring that health informatics solutions comply with relevant laws and regulations.

These practical experiences have equipped me with the skills needed to design, implement, and evaluate health informatics solutions, as well as to handle real-world data and address challenges in the healthcare industry.

Here is summary of Labs and Practicum Project Completed part of Coursework:

Lab 1 - Electronic Health Records

Submission available on request

Objective: To introduce students to healthcare delivery workflows, focusing on Electronic Medical/Health Records (EMR/EHR), and to help them understand the U.S. healthcare landscape by reflecting on various incentives within the system and using the OpenEMR tool.

Quick Summary:

  • Exercise 1 - Reflecting on Various Wants:

    • Reflect on experiences and considerations as a patient, provider, and payer in the U.S. healthcare system.

    • Understand different needs and challenges in healthcare through brainstorming and subjective answers.

  • Exercise 2 - Electronic Health Records:

    • Deploy OpenEMR locally using Docker or access a provided server.

    • Gain hands-on experience with EHR systems by completing various tasks in OpenEMR.

Lab 2 - Introduction to FHIR Resources

Submission available on request

Objectives

  • Understand the structure of FHIR Resources, including handling list/array elements and FHIR data types.

  • Manipulate resources to perform simple tasks.

  • Accomplish paged bundle navigation.

Quick Summary

This lab focuses on exploring FHIR (Fast Healthcare Interoperability Resources) and HAPI FHIR (Healthcare Application Programming Interface). It is divided into two parts:

  1. Part I (Exercises 1-4): These exercises involve working with the FHIR Patient Resource, Observation Resource, and US Core Patient Profile. They emphasize understanding data structures using the HAPI FHIR library.

  2. Part II (Exercises 5-6): These exercises focus on Bundles, including navigation and understanding meta information as well as working with resources within the bundles.

All exercises are implemented using Java and submitted via Gradescope. The tools used include Java 1.8 and any IDE of choice (IntelliJ recommended). Resources provided include lab starter code, HAPI FHIR Primer, and FHIR specifications.

Lab 3 - FHIR Servers (OMOPonFHIR)

Submission available on request

Objective: The lab aims to introduce students to the integration of FHIR standards with the OMOP Common Data Model (CDM) using OMOPonFHIR. It focuses on deploying an OMOPonFHIR server, performing data mappings, and understanding how these technologies can improve data sharing in healthcare.

Quick Summary:

  • Deployment: Use Docker Compose to deploy an OMOPonFHIR server and a Postgres database.

  • Configuration: Download and load necessary vocabulary files (e.g., SNOMED, LOINC) from OHDSI Athena.

  • Testing: Post provided FHIR JSON files to the server and verify their correct mapping to the OMOP database.

  • Development Environment: Set up and manage dependencies using Python Poetry, and test the implementation using pytest.

  • Exercises: Answer questions and implement code to demonstrate understanding of data mappings and interactions between FHIR and OMOP CDM.

Tools Used:

  • Docker

  • Python Poetry

  • PgAdmin4 or DataGrip

  • OHDSI Athena

This lab provides hands-on experience in modernizing healthcare data interoperability through practical deployment and data mapping exercises.

Lab 4 - SMART on FHIR

Submission available on request

Objective: The lab aims to teach students how to deploy and develop a SMART on FHIR application that manages medications within an acute care hospital setting. It focuses on leveraging SMART on FHIR standards and APIs to build secure and interoperable healthcare applications.

Quick Summary:

  • Deployment: Use Angular to deploy a SMART on FHIR app locally, utilizing the SMART on FHIR sandbox server for simulation.

  • Configuration: Set up the development environment using Node.js and Angular CLI, and configure the SMART on FHIR sandbox launcher.

  • Development: Modify the provided TypeScript code to add features for searching, updating, and creating MedicationRequest resources using the SMART on FHIR Client.js library.

  • Exercises:

    1. Search Medication Requests: Implement functionality to list all MedicationRequest resources for the current patient.

    2. Update MedicationRequest: Implement functionality to update the status field of a MedicationRequest resource.

    3. Create MedicationRequest: Implement functionality to create a new MedicationRequest resource with specified fields.

    4. SMART on FHIR Scopes: Specify additional scopes required for the application based on SMART on FHIR guidelines.

Tools Used:

  • Node.js and Node Version Manager (NVM)

  • Angular and Angular CLI

  • SMART on FHIR sandbox and Client.js library

  • Provided starter code in TypeScript

This lab provides hands-on experience in developing a healthcare app using SMART on FHIR standards, enhancing students' understanding of FHIR APIs and secure app development in healthcare environments.

Lab 5 - Clinical Summarization with Generative AI

Submission available on request

Objective: To teach students how to use Generative AI to summarize clinical notes stored in FHIR (Fast Healthcare Interoperability Resources) DocumentReference resources, addressing the challenge of navigating voluminous clinical notes in Electronic Medical Records (EMRs).

Summary:

  • Clinical Summaries: Understand the importance of clinical summaries in EMRs, which include information like allergies, medications, social history, vital signs, encounters, family history, and immunizations.

  • Problem Statement: The challenge of quickly accessing relevant information from extensive clinical notes, especially in emergency situations.

  • Generative AI: Introduction to using Generative AI and Language Models to summarize clinical notes efficiently.

  • Hands-On Experience: Gain practical experience with the T5-Small Model for clinical summarization.

  • FHIR Ecosystem: Learn about the FHIR ecosystem, specifically how clinical notes are stored and accessed using DocumentReference resources.

  • Tools Used: Utilize Google Colab for executing provided notebooks, with instructions on setting up the environment, conducting experiments, and preparing submissions for autograding.

This lab provides foundational knowledge and practical skills in applying Generative AI to enhance the accessibility and usability of clinical notes in healthcare settings.

Lab 6 - Data Engineering / ML with FHIR

Submission available on request

Objective:

The lab aims to transform FHIR (Fast Healthcare Interoperability Resources) data into a format suitable for machine learning algorithms. The main task is to create a features dataset to predict if a patient will develop a stroke, using Python and the pandas package.

  • Utilize FHIR for feature engineering in machine learning.

  • Convert FHIR bulk export data into a tabular format for machine learning.

  • Understand FHIR data model and apply common data transformations.

Tools

  • Python (>=3.10)

  • pandas package

  • Docker (optional)

Steps

  1. Setup: Install Python, pandas, and optionally Docker.

  2. Data Transformation: Convert FHIR ndjson files to create new features.

  3. Feature Engineering:

    • Create unique patient IDs and calculate ages.

    • One-hot encode marital status.

    • Extract and process glucose and triglycerides data using specific LOINC codes.

    • Impute missing data and apply mean normalization.

  4. Label Creation: Create a binary indicator for stroke occurrence.

Practicum Project:

Objective:

The objective of the Practicum Project is to provide students with an opportunity to delve deeply into a specific area of Health Informatics. It allows them to apply their knowledge and skills to real-world problems, develop practical solutions, and gain hands-on experience in the healthcare technology ecosystem.

Summary:

The Practicum Project runs throughout the semester, becoming the primary focus in the latter half. Students, either individually or in teams of up to five, work on projects in various domains of healthcare technology such as system interoperability, data analytics, end user applications, security and privacy, public health, and social determinants of health. Each project is guided by a TA Mentor from the instructional team, who offers domain expertise and oversees the project's progress.

Students can either choose from instructor-developed topics or propose their own, with a requirement to scope their project to fit within an estimated 100 hours of work over the course. This includes planning, research, development, and regular check-ins with their TA Mentor to ensure the project's alignment with objectives and feasibility. The final deliverable is assessed based on the quality, complexity, and thoroughness of the work .

Submissions for Practicum Project:

Submitted code available on request

Project Proposal and Planning

Final Presentation

Academic paper based on project research work

Summary of Academic paper submitted:

Abstract: The paper discusses the development of a web application designed to help individuals track their alcohol and tobacco consumption and related health vitals. The app also performs population-level trend analysis using anonymized patient data to correlate substance use with health conditions.

Background: Excessive alcohol and tobacco use are significant public health issues, contributing to various serious health conditions. The motivation for the project is to provide a tool for self-management to reduce substance use and avoid negative health outcomes.

Problem Statement: The goal is to reduce excessive alcohol and tobacco consumption by allowing individuals to track their consumption and health vitals over time, and to inform them about increased risks of certain health conditions based on their profiles.

Solution: The team developed a web application using React for the frontend, Flask and MySQL for the backend, and Python for analytics. The application allows users to register, log their substance use and health vitals, and visualize trends. The team faced challenges in finding suitable data for alcohol consumption and decided to use tobacco smoking data from the FHIR database.

Methodology: The application uses a 3-tier architecture to separate the backend from the frontend, ensuring flexibility and scalability. The team performed data analysis to correlate tobacco use with health conditions, focusing on trends across different age groups and other demographics.

Features:

  • User Registration: Users can create profiles and log in.

  • Vitals Tracking: Users can input and track heart rate, respiratory rate, height, weight, and BMI.

  • Substance Consumption Tracking: Users can log their alcohol and tobacco use.

  • Dashboard: The app provides visualizations of health and consumption trends.

  • Population-Level Analysis: The app provides insights based on trend analysis of population data, projecting increased health risks for smokers compared to non-smokers.

Outcome and Further Work: The application shows promise in helping users manage their substance use and understand the associated health risks. Future improvements include obtaining more comprehensive data, refining trend analysis, and integrating real-time feedback and more detailed consumption tracking.

References: The paper cites various sources, including studies on alcohol and tobacco use, public health databases, and technical resources for the FHIR, LOINC, and SNOMED CT standards.

In conclusion, the paper describes a comprehensive effort to develop a health tracking application focused on alcohol and tobacco use, incorporating both individual tracking and population-level trend analysis to provide meaningful health insights.

Final Submission Document

Here is summary of final submission for a project "Patient-Facing Alcohol and Tobacco Smoking Consumption & Health Tracking App with Population-Level Trend Reporting" by Team Marvel.

Section I: Project Overview

  • Team Members and Roles:

    • Amadhya Anand: Full stack developer

    • Amber Gupta: Backend developer/Data Analysis/Lifescience Industry SME

    • Madisen Arurang: Backend developer

    • Sudhanshu Jaiswal: Backend/Infrastructure developer

    • Sushmitha Bedere: Front end developer

  • Mentor:

    • TA Mentor: Jaewoo Park

    • External Mentor: Not applicable

  • Project Task Status:

    • A detailed status report on project tasks, issues, and risks.

    • The project involved creating a technical framework, data design, and developing features for a health tracking app using ReactJS, Flask, and MySQL.

    • Major milestones include data modeling, prototype development, trend analysis, and final integration.

Section II: Project Artifacts

Section III: Project Presentation

Section IV: Project Documentation

Additional Details

  • Issues and Risks Tracker:

    • Document outlines the risks and issues encountered during the project, such as interdependencies causing rework, learning curves for Flask, and limited data volumes for trend analysis.

The document provides a comprehensive view of the project's objectives, progress, challenges, and outcomes.