Comparative Effectiveness HAP 823

George Mason University Department of Health Administration & Policy 

HAP823

 

Course Description

An advanced statistics course in analysis of massive data involving thousands of variables, using linear and logistic regression. The course focuses on use of (1) ridge regression and (2) propensity scores in analysis of data within electronic health records.  Topics include (1) counterfactual framework and assumptions, (2) data balancing, (3) matching or weighting, and (4) sensitivity analysis.  Slides► History►

Course Objectives

Upon completion of the course, one will be able to:

  1. Demonstrate mastery of assumptions of logistic and linear regression 
  2. Demonstrate mastery of counterfactual framework and assumptions in propensity matching or weighting
  3. Decide which regression model is applicable in various situations in health services management, policy and research.
  4. Analyze data using ridge logistic or linear regressions
  5. Reduce dimensionality of the data using correlation analysis
  6. Select among various methods of matching and weighting data
  7. Analyze project data using propensity-matched linear or logistic regressions.
  8. Interpret statistical outputs for ridge or propensity matched regression models.
  9. Present results, after and before data balancing, to policy makers.

Required Textbooks

  • No textbook is required.  Required reading is posted to the web.  Home

Pedagogy

Each class will consist of four parts:

  1. Learn one: instructor provides didactic lectures
    • Brief lectures of 10-15 minutes per module
    • Extensive online resources
    • Interactive lectures requiring active student participation
    •  Q&A on Twitter
  2. Do one: practical laboratory work within each lecture. 
    • Complete assignments during class time
    • Work in groups
    • Work on real datasets
  3. Teach one:  Students teach the topic they have learned to one another. 
    • All assignments are done by pair of students.  A different team member should be selected for each assignment.
    • Students are expected to present a narrated presentation in at least one of the lectures
  4. Evaluation
    • Students evaluate each lecture, not just at end of the course
    • Students work in teams but are evaluated by their own independent (unconfounded) performance across teams

Q&A on Twitter

You are invited to ask and receive answers on Twitter. Please follow the instructor at @DrAlemi.  Questions asked are answered so all students can receive the answer.  Answers typically include videos posted to twitter. You can also text 703 893 3799 for more private communication.  Emails are not answered in a timely fashion, please use Twitter for faster responses.  Each tweet should contain the relevant hash tag available in the lecture web site.  General tweets should use hash tag #effectiveness and directed to @DrAlemi. Get account►

Teach One Assignment

We rely on a method typically used in training of medical residents: "Learn one, do one, teach one."  Each student is expected to not only learn the concepts in the course, and do the assignments, but also teach a portion of the course. This active participation in teaching helps students learn the concepts in the course in more depth.  The best way to learn a topic is to teach it.  Students are expected to teach by preparing a brief video. Students select which topic they wish to teach.  They can teach about any aspect within the topic.  Typically students teach how to do the assignment in the week's topic. Following tools are needed for preparing an online project presentation:

  1. A microphone is necessary to narrate your slides.  Please do not rely on built in microphones for portable computers.
  2. You can capture screen shots and insert it into your slide presentation using Command and Print Screen keys.  MAC Users►
  3. Narrate your slides or use other video making software.  Narrate► Free Camstudio►   Camtasia►  SWF► IMovie►
  4. Convert the narrated slides to a video format that can be uploaded to the web.   
  5. Upload your narrated slides Author Stream► You-Tube►
    • Share your narrated slides publicly so all students in the class can view it.  Tweet the URL to using the hash tag for the related lecture.
    • Put within the description the following statement:  "This presentation was prepared as part of the HAP 823 course on Comparative Effectiveness  taught by Farrokh Alemi, Ph.D. at George Mason University  Department of Health Administration and Policy."  Add a sentence about yourself to the description.  Keep in mind that the work you are uploading will remain on the web for years to come and will help shape your career.  Maintain a professional attitude and presentation. 
    • If you wish to be exempted from sharing your work publicly, make a request to the instructor with your justification and alternative plans.
  6. Email your URL to all the students in the course and the instructor. 

Effective video presentations follow these rules:

  1. Explain the topic from different perspectives.   
  2. Make sure your presentation is accurate.
  3. Explain each step in the process and do not jump over what may seem trivial.
  4. Have a consistent style (the same font, capitalization policy, color, and size through out the video.
  5. Make one, and only one, point per slide. 
  6. Use little text, just enough so the reader can follow your narrated comments.
  7. Make sure all texts on slides are readable, even after compression to fit the video requirements.  A U-tube video reduces Power Point slides to 1/4 of their original size.  Phone views have additional reductions.
  8. Make sure the narration is clear and continuous.  Occasional hiccups are ok, keep them; they add color.  Set compression levels high enough to be understood in a noisy room. 
  9. Be brief.  Do not exceed 10 minutes. If you need more time, make multiple videos. 

Semester Long Project

In lieu of exam, you are to analyze publicly available data, using data balancing, and prepare a written paper, ready for submission, to a journal.  No team submission is allowed but students are expected and encouraged to help each other.  Request for data should be made no later than first day of class.  Read details of semester long project and timetable for producing portions of the project.  Read►

Students' Evaluation

  1. Teach One Assignment (20%): The Teach One assignment must be completed one week ahead of the related lecture.  Missing this deadline reduces the Teach One grade by 20%.  The URL for the Teach One assignment must be emailed to all students including the instructor. 
  2. Weekly Assignments (40%): You are required to complete all assignments one week after lecture day; assignments for lectures over two weeks are due one day after 2nd week's lecture. 
  3. Semester Long Project (40%): Assignments are due on first day, 4th week, 6th week, 10th week and 13th week of class
  4. Exams (0%):  There are no exams in this course. 

University grading policies are followed.

A                 4.00              94-100%
A-                3.67              90-93%
B+               3.33              87-89%
B                 3.00              83-86%
B-                2.67              80-82%
C                 2.00              70-79%
F                                        69% and below

Topical Outline

Please note that course syllabus may change at anytime prior to the date of the lecture.  Do not print ahead of schedule.  Check this page regularly for updates.  Assignments are due within 7 days of date of topic, unless otherwise noted. Late assignments are accepted.  Late assignments will receive 20% less grade.  All assignments must be completed with an additional student.  No two students should work on more than one project together.

  1. Review of linear and logistic regression
  2. Review independence
  3. Problems of observational data
  4. Counterfactual framework
  5. Propensity scores procedures
  6. Inverse propensity score weighting
  7. Stratified covariate balancing
  8. Analysis of cost effectiveness of interventions
  9. Analysis of impact of intervention on hospitalization rates 
  10. Analysis of prognosis using thousands of morbidities
  11. Use of network analysis in understanding context of text
  12. Final project presentation

Honor code

“To promote a stronger sense of mutual responsibility, respect, trust, and fairness among all members of the George Mason University community and with the desire for greater academic and personal achievement, we, the student members of the university community, have set forth this honor code: Student members of the George Mason University community pledge not to cheat, plagiarize, steal, or lie in matters related to academic work” (George Mason University Catalog, 2006-2007, p. 31).    

Individuals with Disabilities

George Mason University is committed to complying with the Rehabilitation Act of 1973 and the Americans with Disabilities Act of 1990 by providing reasonable accommodations for disabled applicants for admission, students, applicants for employment, employees, and visitors. Applicants for admission and students requiring specific accommodations for a disability should contact the Disability Resource Center at 703-993-2474, or the Equity Office at 703-993-8730. Applicants for employment and employees should contact Human Resources at 703-993-2600 or the Equity Office. Students are responsible for providing appropriate documentation and requesting reasonable accommodation in a timely manner (George Mason University Catalog).  

Working E-Mail & Tweeter Accounts

All communications are made by email and tweeter.  Students must be able to receive emails and to regularly, at least daily, review emails.  Do not allow the mailbox to become full.  Tweeter account is used for quick responses.  Students can also text the instructor at 703 893 3799


This page is part of the course on Comparative Effectiveness by Farrokh Alemi, Ph.D. Home► Email►