Teaching Statistics to Health Professions Students

In development for my new course, "Using Statistics in Medicine", for UCDavis CPE's Post-Baccalaureate Health Professions program, I am focusing on two primary goals:

1. Helping students who will eventually use research to inform their clinical or pharmacy practice learn to use this information in the best ways possible, and 

2. Exposing these students to hands-on applications of statistics using simple, modifiable python code examples via Google Colab Notebooks, to show them how statistics are implemented today in the real world, using medical data examples.

Statistical literacy and an understanding of probability are both very important for everyone in the age of large language models, and an understanding of statistics in the context of research and research computing are essential to the education of anyone who is moving into the health professions today.  

In the near future, doctors, nurses and pharmacists will spend even more time interacting with data and data derivatives, including statistical summaries, insights / predictions, and prescriptive probabilities, than ever before.  Will they gain the strong, foundational understanding of statistics fundamentals that they will need in order to interact wisely with new decision support technologies, or make informed decisions about which research findings to integrate into their evidence-based clinical practice? 

Only if those of us teaching medical statistics expose them to modern methods and take the time to demonstrate the use and proper interpretation of statistical techniques in both clinical and research contexts.

The Research Context

Many methods may be used to describe, evaluate and predict the relationships that exist between people’s health and their environments, disease states, treatments and interventions.

Medical research falls into many categories, including Clinical Research, Health Outcomes Research, Health Policy Research, Quality Improvement  and Epidemiological Research.

Epidemiology: the study and analysis of the determinants of health and disease in a population

Important uses:

• Improve the value of treatments and interventions (higher quality, lower cost)

• Improve the evidentiary base of medicine (RCT is the gold standard)

• Improve safety and quality, tailor interventions, and mitigate risks

Clinical Research should be a structured, documented process that investigates facts & theories and explores connections!

It ought to be systemic in its methods of examining conditions and outcomes, and a process that generates evidence for decision making.

Clinical research is always subject to examination for:

• Reliability: reproducible, dependable measurement

• Validity: actually measures what it is intended to measure

Health Professionals Engage with Research Daily

• Strive for evidence-based practice <- provide the best quality care based on sound scientific data / best evidence available

• Learn to deploy continuous critical inquiry as part of the clinical decision-making process

• Engage with, examine, and choose when and how to use clinical research, daily

To use research ethically and well, all health professionals must understand:

Methods: how research can or should be conducted, including study design and statistical method application

Evaluation: ways the outcomes of research can be evaluated, and the pros and cons of each type of evaluation method

Error & Bias: mistakes that can be made in study design, measurement, evaluation, and mistakes that are baked into the data we use

Here are a few examples of critiques of research that will be discussed and illustrated in the course.  These are basic critiques, and once these have been illustrated and discussed, more advanced critiques of methods and design will be introduced.

1. Were the researchers transparent about methods and evaluation?

2. Do they offer an explanation for their choice of either?

3. Do they discuss potential bias or errors in their discussion or limitations sections?

4. Is the data they used available, and if not, do they have a good reason?

5. Can you reach out to the contact person with questions and get a response?

Google Colab Notebook Examples

To demystify statistical programming and create an interactive, hands-on understanding of statistics for health professions students, I am creating a series of examples for each unit of my course, using Google Colab Notebooks.  Each notebook focuses on a different aspect of statistics or probability, and most make use of real-world health data.

Here are some examples of the notebooks I have created for the course so far:

More notebooks and walkthroughs coming soon! 🔜