AI Health

Spring 2021

Course Information

Course Date and Time: Thursday 12:00-3:00PM
Location: Web based
Instructor: Ying Ding
Office Hour: Thursday 10:30-11:30AM, or by appointment

Course Description

Recently, the U.S. healthcare industry has surpassed manufacturing and retail to become the largest employer in the country, with every 1 out of 8 Americans working in this sector. Policies and incentives have been established to promote IT in health to improve care and delivery. In this course, we will explore the major components of health IT systems, ranging from data semantics (ICD10), data interoperability (FHIR), diagnosis code (SNOMED CT), to workflow in clinical decision support systems. After establishing the good understanding of the fundamentals of health IT systems, we will dive deep into how AI innovations (e.g., machine learning, deep learning, computer vision) are transforming our healthcare system by introducing new concepts of mobile health, AI diagnosis, AI medicine, smart device, and intelligent delivery. This course will offer hands-on tutorials based on the real-world Electronic Health Record (EHR) data from MIMIC III (https://mimic.physionet.org/) released by MIT Critical Data. MIMIC-III (Medical Information Mart for Intensive Care III) contains de-identified health information from over forty thousand patients who stayed in critical care units of the Beth Israel Deaconess Medical Center between 2001 and 2012. These tutorials aim to enhance data search and analytics skills by providing practices related to database search, natural language processing, data visualization, machine learning, and deep learning. In this course, we will enhance the group learning experience and learning by doing, therefore, there will be many class activities. This course is designed for everyone, so no tech or programming background is required or desired.

Course Objectives

After attending this course, you should be able to achieve the following goals:

  • Be aware of current healthcare initiatives to deliver quality care
  • Understand the basic technologies of health IT systems including data semantics, data interoperability, workflow, and diagnosis
  • Be familiar with electronic health record systems (EHR systems)
  • Gain the overview of AI innovations in healthcare
  • Master practical skills of data search and analytics including database search, natural language processing, data visualization, machine learning, and deep learning

Recommended Textbooks

Data and Software

  • MIMIC III and MIMIC IV: contains de-identified data from over 40,000 patients who were admitted to Beth Israel Deaconess Medical Center in Boston, Massachusetts from 2001 to 2012 (Data and access).
  • Python Coding using Google Colab: https://colab.research.google.com/
  • Tableau: https://www.tableau.com/academic/students


No Lecture Class Activity Lab/Tutorial Notes
L1 Introduction knowing each other Intro to MIMIC dataset, form expert group Get acces to MIMIC III
L2 Evidence-based Care, i2b2 and OMOP paper discussion, compare common data models MIT ethic course, Group project client presentation MIT ethic course due, and get access to MIMIC III data, form a group for your group project
L3 EMR Semantics: ICD10 Paper discussion T1: MIMIC-Visual: Tableau I Form a group for self-learning tutorial
L4 EMR Semantics: SNOMED CT I paper, case, expert group T1: MIMIC-Visual: Tableau II Send the group project summary to TA
L5 EMR Semantics II: SNOMED CT II Paper Discussion T2: MIMIC-SQL MIMIC Tableau Gallery Due
L6 EMR Semantics: LONIC Focused group discussion T2: MIMIC-SQL, one query hackathon Group project
L7 EMR Semantics: RxNorm Focused group discussion, teach us one medical taxonomy T3: MIMIC-NLP: spacy and sciSpacy MIMI SQL Due
L8 Clinical Decision Support System Focused group discussion T3: MIMIC-NLP: sciSpacy and one NLP hackathon (bluebert, clinical bert) Group project
L9 EMR data share: FHIR Focused group discussion T4: MIMIC-ML LOS I Select papers to present in next two weeks
L10 AI health: ML/DL I (team paper presentation) Focused group discussion T4: MIMIC-ML LOS II MIMIC NLP Due
L11 AI health: ML/DL I (team paper presentation) focused group discussion T5: MIMIC-ML: Readmission I Group project
L12 AI health: imaging Focused group discussion T5: MIMIC-ML Readmission II (One ML hackathon) Self-learning tutorial due
L13 AI in drug discovery Focused group discussion Group project Group project
L14 AI health: Wrap up Chapter presentation of AI in healthcare group project MIMIC ML due
L15 Final Presentation Final Presentation Best student tutorial award (vote by the students) Group project
L16 Final Report due Demo/Poster session Final Group porject due, AI health data challenge Group project due

Homework and Grading

Individual Assignments (45%):

  • MIT Ethic Course (5%): pass the course
  • MIMIC Gallery (10%): Develop the visual gallery using MIMIC data using Tableau (Tableau file, video to show details with the goal that other students can follow and rebuild your gallery)
  • MIMIC SQL (10%): Develop 10 SQL queries using MIMIC data (powerpoint slides showing SQL queries, meaning and screenshots for results, video)
  • MIMIC NLP (10%): generate a NLP tutorial using MIMIC data (powerpoint slides, code, video)
  • MIMIC ML tutorial (10%): generate a ML tutorial using MIMIC data (powerpoint slides, code, video)

Group Assignments (45%):

  • Self-learning tutorial using MIMIC data (15%)
  • Group Project (30%): Build evidence-based care apps/tools
    • joining AI Health Data Challenge
    • using MIMIC datasets
    • data+design
    • a 3-5 page report (aiming for a workshop paper), powerpoint, demo, code, and video (assuming that you will teach it to other students), presentation

Class presentation, participation, and final presentation (10%)

Grading Guideline

Your written, web-based, and oral work will be evaluated according to four criteria; it must:

  • Be clearly written, marked up, and/or presented, and checked for spelling and grammar;
  • Demonstrate a degree of insight into the concepts, issues, and trends in both the areas you investigate in the assignments and in the course content;
  • Demonstrate a degree of originality in your reviews, analyses and projects; and
  • Display familiarity with the appropriate literature.

To receive a passing grade in this course, you must turn in all of the assignments and the term project and complete all the presentations. You cannot pass this course without doing all of the assigned work (which includes the final presentation), however, turning in all of the work is not a guarantee that you will pass the course.

Borderline grades will be decided (up or down) on the basis of class contributions and participation throughout the semester.

A 4.0 Outstanding achievement. Student performance demonstrates full command of the course materials and evinces a high level of originality and/or creativity that far surpasses course expectations.
A- 3.7 Excellent achievement. Student performance demonstrates thorough knowledge of the course materials and exceeds course expectations by completing all requirements in a superior manner.
B+ 3.3 Very good work. Student performance demonstrates above-average comprehension of the course materials and exceeds course expectations on all tasks as defined in the course syllabus.
B 3.0 Student performance meets designated course expectations and demonstrates understanding of the course materials at an acceptable level.
B- 2.7 Marginal work. Student performance demonstrates incomplete understanding of course materials.
C+ 2.3 Unsatisfactory work. Student performance demonstrates incomplete and inadequate understanding of course materials.
C 2.0
C- 1.7 Unacceptable work. Coursework performed at this level will not count toward the MLS or MIS degree. For the course to count toward the degree, the student must repeat the course with a passing grade.
D+ 1.3
D 1.0
D- 0.7  
F 0.0 Failing. Student may continue in program only with permission of the Dean.