How Telvorn teaches clinical reasoning
A Telvorn case runs through five stages. Students work alone, compare their reasoning in small groups, learn from a facilitator-led debrief, reflect on their decisions and then answer again. The sequence is deliberate. Each stage is designed to support a different aspect of learning and clinical reasoning.
Stage 1: Making an initial judgement
Students begin by working through the case on their own. They decide what questions to ask, arrive at an initial judgement and rate their confidence before any discussion or teaching takes place.
The purpose of this stage is to require commitment before instruction. Rather than being shown the correct approach and then applying it, students must first reveal how they are thinking. The educational value comes from making an attempt, identifying gaps in understanding and creating a need for feedback.
Manu Kapur's work on productive failure found that learners who struggle with a problem before being taught it tend to understand the concept better afterwards. An attempt, even an unsuccessful one, helps reveal what the learner does and does not yet understand. Slamecka and Graf (1978) showed that people remember what they generate better than what they read, while Richland, Kornell and Kao (2009) found that unsuccessful attempts to answer a question before studying it still improved later learning.
The confidence rating serves two purposes. Butterfield and Metcalfe's work on the hypercorrection effect (2001) found that errors held with high confidence are more likely to be corrected once feedback arrives than errors held with low confidence. Dunlosky and Rawson (2012) showed that learners who misjudge what they know learn less because they stop studying what they believe they have mastered.
Telvorn records confidence alongside every decision so that beliefs can be compared with performance, a measure known as calibration.
Stage 2: Testing reasoning with others
Small groups see each member's decisions side by side and settle on one set of answers with reasons. Afterwards each student rates confidence again, privately.
The purpose of this stage is to test reasoning through discussion. Students are required to explain, defend and revise their thinking in response to alternative interpretations of the same case.
In a 2009 study in Science, Michelle Smith and colleagues found that students who discussed a question with a neighbour went on to answer a similar one better, even when neither had been right at the start. Discussion seems to improve understanding, not just spread it.
Chi and Wylie's ICAP framework (2014) offers an account. It ranks learning activities from passive to active to constructive to interactive and predicts that interactive activities, where learners build on one another's reasoning, lead to the deepest learning.
The second confidence rating completed in private shows whether the discussion changed a student's mind, or only her answer. Confidence that rises after a group has converged on a wrong answer is as informative as confidence that falls.
Stage 3: Learning from disagreement
The facilitator receives an anonymised summary that leads with where the cohort's reasoning diverged. The correct answers stay hidden until the debrief.
This stage allows teaching to respond to learner needs rather than assumptions. Instead of following a predetermined explanation, the debrief begins with where learners disagreed, struggled or showed uncertainty.
This follows just-in-time teaching, developed by Gregor Novak and colleagues in the 1990s, in which the instructor reads students' answers beforehand and builds the session around what they got wrong. Time goes where it is needed, not where the teacher assumed it would be. Black and Wiliam's 1998 review of formative assessment reached a similar conclusion: adjusting teaching to what students actually understand improves learning.
Learner responses are anonymised so that common misconceptions can be discussed openly without focusing on individual students.
Stage 4: Making sense of the experience
After the debrief, students answer four questions. How did that feel? Where did my answer differ from the evidence, and why? What principle will I take from this? What will I do differently next time?
The aim of this stage is to convert an experience into a lesson that can be used again. Reflection encourages learners to move beyond whether they were right or wrong and focus on how they arrived at their decisions.
Donald Schön argued in 1983 that professionals improve by reflecting on what they did, and David Kolb's learning cycle (1984) makes the same point: an experience teaches little until the learner draws a general lesson and tries it again.
The principle question performs a particularly important role. Chi and colleagues (1994) found that students prompted to explain a text to themselves as they read it understood it better than students who simply read it twice. Asking students to explain why their answer differed from the evidence, and to state the principle they will take away, calls for the same kind of self-explanation.
Stating a principle helps learners identify what may apply beyond the case in front of them.
Hattie and Timperley's 2007 review found that feedback telling learners where to go next can be among the most powerful, which is the job of the final question. The first question reflects Kluger and DeNisi's 1996 analysis, in which feedback made performance worse in over a third of cases, with the harm growing as feedback drew attention to the self and away from the task. Asking how the result felt acknowledges that reaction before asking the student to think past it.
Stage 5: Revisiting the decision
The case does not end with the first answer. Additional information becomes available and students must decide whether their original judgement still stands. They explain their reasoning, reassess their confidence and see the consequences of the decision they make.
Roediger and Karpicke (2006) found that students who took a recall test on a passage remembered more a week later than students who restudied it, although restudying did better on a test five minutes afterwards. Over the longer term, retrieval strengthens memory more than review does.
The research on reasoning gives a second reason. Graber, Franklin and Gordon (2005) found premature closure, the failure to keep considering alternatives after an initial diagnosis, to be the single most common cognitive cause in 100 cases of diagnostic error in internal medicine. Rather than moving on once an answer has been given, students are asked to reconsider it when new information becomes available.
Importantly, changing an answer is not the goal. Better reasoning is. Keeping a correct decision because the evidence still supports it can be just as appropriate as changing a wrong one. What matters is the quality of the explanation.
Showing the consequence draws on branched-outcome simulation in pharmacy education (Benedict and Schonder, 2011), where learners see what follows from the choice they made, not only whether it was right.
What ties the stages together
Across the five stages the confidence rating is taken three times: when the student first commits, after the group has discussed, and after the reassessment.
This is the thread that runs through the entire design. Telvorn is interested not only in whether learners are right, but whether they know when they are likely to be right and when they should be uncertain.
Set beside their accuracy, those three confidence ratings show whether learners are becoming better calibrated over time: trusting their judgement when it is justified and questioning it when it is not.
The five stages therefore work together as a single learning cycle. Learners commit to an answer, test it against others, correct it where the evidence disagrees, extract a general lesson and apply that lesson to a fresh decision.