Two Pills Podcast: Increase Your Confidence (interval) Teaching Biostatistics!


Oh, biostatistics. A subject that strikes fear into anyone studying for an exam that contains them, someone presenting a journal club, or even analyzing your own data for research. Today, I am going to be describing a systematic approach to biostatistics that may help you in teaching the content and help your students with application.

Healthcare professionals are required to continuously update their knowledge; therefore, our students need the skills for life-long learning, as well as an appreciation for the scientific method. Biostatistics is the “basic science” of quantitative evaluation of evidence and students will need to require evidence for methods of: prevention, diagnosis, and therapy/management in the treatment of medical conditions. Students need to know how to interpret diagnostic procedures and apply them to individual patients. Students need to develop the skills to read the medical literature with confidence in their ability to evaluate the validity of articles.

Often, students are taught biostatistics in a lecture-based format. When I was taught biostats in professional school, I think I had last seen statistics in high school during AP statistics. As we’ll discuss, repetition is key for understanding and applying biostatistics. After they initially learn about biostatistics, their first presentations on statistical analysis may be in the high-pressure environment of a journal club. I think we have all seen the spectrum of confidence that students have when presenting statistics in a journal club.

I first became interested in augmenting my teaching of biostatistics in an interprofessional setting. I was working with a medical residency and they wanted to increase the structure of their journal club/biostatistics curriculum. The milestones that I attempted to address were:


PBLI -1: Locates, appraises, and assimilates evidence from scientific studies related to the patients’ health problems

Level 1: Describes basic concepts in clinical epidemiology, biostatistics, and clinical reasoning Categorizes the design of a research study

Level 2: Identifies pros and cons of various study designs, associated types of bias, and patient-centered outcomes Formulates a searchable question from a clinical question Evaluates evidence-based point-of-care resources

Level 3: Applies a set of critical appraisal criteria to different types of research, including synopses of original research findings, systematic reviews and meta-analyses, and clinical practice guidelines Critically evaluates information from others, including colleagues, experts, and pharmaceutical representatives, as well as patient-delivered information

Level 4: Incorporates principles of evidence-based care and information mastery into clinical practice

Level 5: Independently teaches and assesses evidence based medicine and information mastery techniques

The program had a journal club each month that was led by a resident. In addition to presenting the article, residents were also assigned a biostatistics term or two that they needed to include in their discussion. For example, January was types of bias. March was absolute risk reduction and relative risk reduction. Other topics included Type I and Type II error, incidence and prevalence, sensitivity/specificity, odds and hazard ratios, and predictive values. There was over a year’s worth of topics by reviewing one to two topics each month. By designing this structure, the faculty of the residency could ensure that residents would learn about all of these biostatistics terms while in the program.

Four years ago, I decided to implement a more structured approach to students learning about statistics while on rotations with me. I call it Stats Tuesday (similar to Fat Tuesday). Each Tuesday afternoon, we briefly discuss assigned statistics terms. Prior to our discussion, students split up the list of terms for that session. They are to be prepared with two things for each term. First, they are to provide a plain English definition of the terms, (not (a+b)/(c+d) formulas). Second, they are to present a real example of the term from a clinical trial.

I have found that this method leads to a great deal of discussion and clarification about the terms. Students often seem to be confused by power and about non-inferiority trials. As a small group, we are able to work through their questions and confusion about these terms. In addition to the review, it also seems that Stats Tuesday has increased their confidence in Journal Clubs. They are more equipped to interpret confidence intervals and to evaluate if a certain statistical test was appropriate.

For my rotation, the number of statistical terms increases each week. They begin with types of data. Next, we discuss alpha, beta, types of error, and confidence intervals. We then dive into ratios, risk reduction, and NNT/NNH. Finally, we review t-tests, ANOVA, Chi-square, and regressions.

Statistics are challenging, but they seem to become easier with repetition and application. It’s great to see the lightbulb go on for a student who has found these concepts overwhelming in the past. Hopefully this framework is helpful to you. There will always be a wild and wacky statistical test in a random journal article, but these types of reviews should help students tackle the most common biostatistics terminology.

Resources:

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5339583/

https://ep.bmj.com/content/105/4/236