MiDATA
Need help with biostatistics or machine learning?
MiDATA can help, contact mi.data@utoronto.ca for more information
What to Do With All of MiDATA?
Bridging the gap between clinical expertise and the science of managing and analyzing medical imaging data is challenging. To provide direction for data management as well as the analysis and reporting of research findings, we have introduced a data science unit – MiDATA – offering users an environment geared towards a “soup to nuts” approach to medical imaging research methodology, statistics, and machine learning. The Department of Medical Imaging of the University of Toronto is one of the largest in North America with faculty, residents, and fellows based at nationally and internationally renowned hospitals conducting cutting-edge clinical research in the greater Toronto area. The challenge of any successful research and educational program is bridging the “know-do” gap. The goal of MiDATA is to facilitate impactful research through the efficient and creative use of a mentored learning environment.
Our program consists of three aspects: research education and mentorship, study design and analysis, and knowledge translation.
The three main areas of focus are:
• statistics for medical research.
• machine learning and statistics for medical image analysis
Professor. Pascal Tyrrell (he/him)
Associate Professor - Director, Data Science
Courses
Data Doctor: Monthly Stats & Research Clinic with MiDATA
First Monday of Each Month from 4 - 6 p.m.
Location: University of Toronto,
263 McCaul Street | 4th floor conference room | Toronto, ON M5T 1W7
We will also be making it hybrid. Click HERE for the zoom link.
Research is a cornerstone of both the Department of Medical Imaging and the Institute of Medical Science (IMS) at UofT, driving innovation and advancing healthcare. To support this, the MiDATA program hosts Data Doctor, a monthly drop-in session where residents, fellows, and faculty from Medical Imaging, along with IMS graduate students, can receive expert guidance on the statistical and methodological aspects of their research projects. These sessions are designed to provide personalized, project-focused advice in a collaborative environment, addressing challenges in study design, data analysis, and advanced statistical techniques.
All participants are encouraged to bring their inquiries, datasets, and research projects for personalized expert support. Led by Prof. Pascal Tyrrell and his MiDATA team, these sessions follow the practical, approachable teaching style of the “Biostats in a Nutshell” course, ensuring participants leave with concrete strategies to enhance the quality and rigor of their research.
AI in Medicine
Upcoming Courses and Resources
Artificial Intelligence (AI) is a rapidly developing field and has been revolutionizing the medical field to benefit medicine. This course uses recent literature, papers, and examples to look at the applications of AI tools to see how they can directly assist patients and medical professionals. This 0.25 FCE course (MSC 1114H) is primarily for students in a MSc, PhD, residency/ fellowship program that would like to gain insight on the emerging field. Some topics discussed include challenges in the medical field and how AI can help, applications of AI/ML, and future of AI in medicine.
Biostatistics in a Nutshell
Upcoming Courses and Resources
Introduction to statistics for medicine (biostatistics). This 0.25 FCE course (MSC1107H) is primarily for students in a MSc, PhD, or residency/ fellowship program that has a requirement of data analysis. Students learn about important fundamental statistical principles by studying the analysis of real data. These analyses introduce students to the following four outcome variables:
- Continuous: serum cholesterol, BMI and FVC
- Binary: death or cure yes/no
- Count: number of deaths, number of cures
- Survival time: time to death, time to cure
The course will include the statistics of diagnostic accuracy and agreement, two important tools in medical research. Students must complete an analysis of data collected during their program. Topics of discussion:
- Paired and unpaired proportions
- Paired and unpaired means
- Frequency data and diagnostic tests
- Statistical design
- Agreement
Resources
Our helpful resource website for both Biostatistics and Machine Learning is Mi-DATA.ca
Statistical Software - Open Source
- RStudio
- R tutorial to start from https://www.w3schools.com/r/
- EZR (menu driven interface for medical stats)