#50 – This Very Variability: Supervisor Practice Differences and Clinical Learning

Episode host: Jason R. Frank

Dr. Jason R. Frank, portrait.
Jason R. Frank
Photo: Erik Cronberg.

What do learners think when their supervisors do things differently? This paper provides an answer and introduces Variability Theory.

Episode Article

Mithoowani, S., Khattak, S., Lieberman, S., Tseng, E. K., Zeller, M. P., & van Merriënboer, J. (2024). Learning From Clinical Supervisor Practice Variability: Exploring Medical Resident and Fellow Experiences and Interpretations. Academic medicine: journal of the Association of American Medical Colleges, 99(3), 310–316.

Episode Notes


Early in the patient safety movement, practice variation between clinicians was identified as something to be addressed. When I was a trainee, we would whisper about how our teachers did things differently. Often we attributed a reason to these variations that related to some characteristic of our supervisors, such as “He is risk averse, so orders more tests”. Now I AM that supervisor, with my own strongly held

practice variations and quite a few foibles of my own. How does all this supervisor availability impact clinical learning?


Mithoowani et al from McMaster University set out to explore this phenomenon. They said “…This study explores how medical residents and fellows experience and interpret intersupervisor clinical practice variability and how these variations influence learning.”


Clinical practice variability was defined as “2 or more clinicians making different treatment decisions despite encountering a similar case…”.

The authors took a constructivist stance in this work and conducted semistructured interviews of senior trainees (PGY3+) internal medicine, hematology, and thrombosis during the COVID epidemic. The authors justify this population because of their shared experience working in thrombosis, an area with extensive evidence and rapid patient decisions. Participants were recruited purposively (that’s for Jon). All participants got $15Cdn (enough for a coffee). The first author, who knew and worked with these trainees, conducted all the interviews.

The authors said they used a constructivist grounded theory approach with concurrent and iterative data collection and analysis. Member checking was “optional”. They used variation theory to guide interview design and analysis. The authors provide a reflexivity statement. They continued until sufficiency was reached.

Variation theory

Variation theory, also known as “variation theory of learning,” is a pedagogical framework that aims to enhance learning by systematically varying aspects of the learning environment, teaching methods, or tasks. It originates from the work of Swedish educational researcher Ference Marton and his colleagues in the 1970s.

The core idea of variation theory is that learning occurs through discerning variation and invariance in the studied phenomenon. Learners must experience 3 patterns of variation—contrast, generalization, and fusion—to advance their understanding. Marton et al (2015) describe this as “powerful ways of acting” derived from “powerful ways of seeing”


The authors interviewed 17 trainees. All participants experienced intersupervisor clinical practice variability.

Where Variability Comes From:

  • Learners highlighted areas where the patient presented with a rare condition for which there was little evidence to guide practice.
  • Patient handover was a key area where one clinician may change plans already made.
  • The learners attributed variability to supervisor characteristics, including training, risk tolerance, research expertise, scope of practice, and willingness to deviate from medical evidence. Interpretation of and awareness of evidence were other areas of variability
  • Other participants in care decisions influenced variability: patient preferences, goals of care, priorities, families, and trainee perspectives.

Impact on Learning

  • Cues from supervisors helped trainees recognize key features of a case that they might have overlooked (contrast).
  • Knowing key features helped trainees recognize other patients that principles might apply to (generalization).
  • Trainees with more experience noted that they later realized that decision-making is more nuanced (fusion), helping them fine-tune their decision-making


The authors found that senior trainees’ experiences of clinical practice variation aligned with Variation theory.

The authors recommend that clinical learning design can be enhanced by incorporating these findings by:

  1. Ensuring that trainees are exposed to variability
  2. Ensuring trainees have the learning skills to explore variability
  3. Ensuring supervisors make their reasoning more explicit


This is another example of applying a theory of educational psychology as a framework for an HPE paper.

There are several expertise theories that have been explored in meded/HPE that are relevant to this paper, e.g. Dreyfuss & Dreyfuss.


Marton, F. (2015). Necessary conditions of learning. Routledge.


Related posts


A QR code linked to page on how to subscribe to PAPERs Podcast


The four members of the host team standing close together looking into the camera and smiling

Photo: Erik Cronberg

Read a presentation of our fabulous hosts