Picture a late-breaking session ending on a Tuesday afternoon. By Wednesday morning an interventional cardiologist somewhere in Asia Pacific has seen the headline slide on social media, and by Thursday she is asking the rep in the cath lab what it means for her calcified lesions. The rep cannot, and should not, discuss data that has not been through medical review. The question is how short the gap between approval and a field that actually holds the approved message can be.
That is the problem a MedTech vascular field force in Asia Pacific asked us to work on. The program ran from May 2024 to July 2026. Across seven congress challenges over 26 months, each run in a window of about three days, the program recorded 190,000 answers, found 24,183 knowledge gaps and verified 49% of them closed. The headline is not the average. Pooled across all seven waves, closure sat between 46% and 52% in each of the five main content areas (other content blocks closed 43%), and the same three-day format ran congress after congress, which is what turned a one-off event into a repeatable operating rhythm.
In interventional cardiology, the argument changes at every congress
Pharma teams tend to plan around a label and a handful of major trial read-outs spread over years. Vascular device teams live on a faster clock. TCT, EuroPCR and ESC each bring late-breaking trials, new imaging and physiology data and fresh comparisons between device strategies. Physicians come back from those meetings with questions, and the rep who can discuss the new data credibly, within days, earns the next case.
The client's field force covered coronary and peripheral vascular devices across the region: guidewires and balloons, vessel preparation, drug-eluting stents, vessel closure, intravascular imaging and physiology. The roles were mixed: sales reps, clinical specialists who spend much of their day in the cath lab, and product managers who build the regional story.
The traditional response to a big congress is a debrief webinar a week or two later, followed by updated slides once medical-legal review has cleared them. That review is a prerequisite, not an obstacle: nothing should reach the field before it. The avoidable delay is what often comes after approval, when the approved message waits for the next meeting slot. And nobody knows who absorbed the debrief. The webinar recording shows attendance. It does not show whether a clinical specialist can explain what a non-inferiority result means for her physician's practice.
Timing matters for a second reason. Murre and Dros's 2015 replication of Ebbinghaus, a single-participant relearning experiment with nonsense syllables, illustrates forgetting under that specific experimental method. It does not tell us how fast congress evidence fades for a clinical specialist, but it argues for rehearsing the approved message soon after it is released, and more than once.
We built a congress challenge that runs in the days after the data lands
The design was simple to describe and demanding to run.
- Each challenge was built on the evidence that had just been presented at a congress, plus the parts of the portfolio that evidence supports.
- Questions came from approved material only. Once medical had approved the congress content, it was turned into questions quickly, and experts signed off the questions themselves before anything went live.
- Each challenge ran in a window of about three days, as a battle: one-minute challenges, five questions in 60 seconds, against real colleagues across the region, with notifications pulling people back in.
- Gap Fixing sent each person the explanation for their own wrong answers, and questions came back later to verify the fix.
We ran seven of these between May 2024 and July 2026, with 536 questions in play across them. Participation added up to 2,468 player places across the challenges, with 486 in the largest.
Two choices mattered most. The first was the short window. Three days is long enough for people across different time zones and cath lab schedules to play several rounds, and short enough to feel like an event. The second was pairing new evidence with the portfolio. A trial result on its own is trivia. The same result tied to device selection, vessel preparation or imaging becomes something a rep can use in the next case.
Half the gaps closed, and the five main content areas landed within six points
We count a gap as one person wrong on one question at first attempt. It is closed only when the same person answers the same question correctly later. Across the seven challenges, 24,183 gaps were found and 11,932 verified closed, 49%. The other 12,251 were flagged for follow-up coaching.
By content area:
- Congress data from TCT, EuroPCR and ESC (late-breaking trials): 7,238 gaps, 51% closed.
- Guidewires and balloons (design and selection): 5,394 gaps, 46% closed.
- Vessel preparation and stents (orbital atherectomy, drug-eluting stents, vessel closure): 4,229 gaps, 52% closed.
- Value proposition and competitors (portfolio positioning, comparisons): 3,837 gaps, 49% closed.
- Imaging and physiology (OCT, IVUS, FFR and INOCA): 2,562 gaps, 49% closed.
- Other content blocks: 923 gaps, 43% closed.
Congress data was the largest block, which is exactly what the format is for. It exists to move new evidence into the field while it still matters, and that is where the most first-attempt errors will naturally sit.
The more important number is the spread. Pooled across seven waves and more than two years, every main area closed between 46% and 52%. Be careful with what that shows: it is a spread across content areas, not across waves, so it does not establish that each congress produced the same result. What it does show is that no content area was left far behind over the life of the program, and that the same format could be run repeatedly. For a commercial leader, that is enough to plan congress readiness as a process instead of a scramble, and to ask for wave-by-wave closure as the next test.
A trial result on its own is trivia. The same result tied to device selection becomes something a rep can use in the next case.
The stubborn gaps were about statistics and selection, not facts
We tracked individual questions in one challenge focused on calcified lesions, with 345 players. Three of them show the difference between a gap that repetition fixes and one that needs a conversation.
Guidewire selection: a clinical reasoning gap at the table
One question asked for the three factors that drive guidewire choice in calcified lesions: how heavily the vessel is calcified, where the lesion sits anatomically and what crossing strategy the operator plans. In the challenge summary, 110 of 345 players could not state all three. The question-level record we can trace covers 84 of them, wrong at first attempt: 53 were verified as corrected and 31 were still wrong at the close. The remaining 26 do not appear in that trace, so we cannot say whether they corrected it.
This is not trivia. It is the reasoning a clinical specialist uses when a physician asks which wire to open with. Get one factor wrong and the recommendation that follows is weaker, whatever the rep knows about the product.
Non-inferiority: most knew it, the rest held on
Another question dealt with a late-breaking one-year comparison of a drug-eluting balloon strategy against drug-eluting stents. Most players knew the headline result, that the balloon strategy met non-inferiority. But of the 27 who were wrong at first attempt, only 6 corrected it. The wrong answers split between "inferior" and "failed non-inferiority".
The numbers are small, and we would not over-read them. The pattern, though, is one we see often: people who misread a statistical framing tend to keep misreading it. Non-inferiority is a design concept, not a fact to memorise, and repeating the answer does not teach the concept.
"Numerically higher, not significant": the overclaim risk
A third question asked about the lesion-level target lesion failure result from an observational study. The correct reading was that the rate was numerically higher but not statistically significant. At first attempt, 34 people said it was significantly lower, a much more flattering reading. By the close, 15 still said so. Of 45 wrong at first, 20 were verified as corrected and 25 were not.
This one worries us most, because it is the gap that creates compliance risk. A rep who overstates a non-significant result is not just wrong. He is making exactly the kind of claim medical-legal review exists to prevent, in front of a physician who has read the paper.
Five lessons from seven congress cycles
- Speed only matters if you measure it. The value of a three-day window after approval is not that it is fast. It is that it shows, within the week, who holds the approved message and who does not. One later right answer is a verified correction in the assessment, not proof of how someone discusses the data in the cath lab, and it does not rule out a lucky guess.
- Pair evidence with the portfolio. Congress data alone produced the largest block of gaps. Tying it to selection, preparation and imaging kept the questions close to real case decisions.
- Watch the statistical framing questions. Non-inferiority and non-significance were where wrong answers held on most stubbornly. These deserve a coaching slot, ideally led by a clinical specialist or medical colleague.
- Treat selection gaps as clinical reasoning. The guidewire question was about how to think, not what to recall. Use coaching with real case examples as the next intervention to test, followed by a retest.
- Repeatability is the business case. A format that ran seven times, with the five main content areas closing within six points of each other over the whole program, lets a leader budget and plan the next congress as routine work.
A fair objection: 49% is lower than the launch and kickoff cases
It is. In our case library, the 16 programs tied to a business moment closed 55% of their gaps against 41% for the other 84, an observed association, and some single launches have gone much higher. A congress program sitting at 49% can look like an underperformer.
We think that reading misses what makes congress content different. Every wave introduces evidence nobody in the field has seen before, and each window is only about three days. A launch program runs on one body of content for weeks and can keep coming back to it. Here, the content resets every few months. Closing about half of the gaps on genuinely new material, in three-day windows, over seven waves, is a harder task than closing more on material that stays put.
The fairer comparison is with what normally happens after a congress: a debrief nobody measures. Against that, knowing which half of the gaps closed, and exactly which 12,251 did not, is the useful part.
What we would do next
We would build a standing coaching track for the statistical framing gaps. Non-inferiority, non-significance, subgroup caution and observational versus randomised evidence come up at every congress. A short, recurring session with medical colleagues, fed by the open-gap list from each challenge, would address the concept once instead of the answer seven times.
We would carry open gaps forward between congresses. Selection questions like the guidewire one are not tied to a single meeting. Bringing the still-open ones back in the next challenge, alongside the new evidence, would show whether coaching moved them.
And we would give product managers the congress-by-congress readiness view by country. With seven waves of data, they can see which markets absorb new evidence fastest and plan launch support for new devices accordingly.
Where Atrivity fits
These challenges ran on Atrivity: battles in a synchronised window, one-minute challenges, Gap Fixing for each person's own mistakes and readiness by country and topic across all seven waves. The platform made the rhythm measurable. The editorial work, picking which new evidence mattered for the portfolio, was done jointly with the client's regional team. If your next big congress is a few months away, one question for your planning meeting: how will you know, by the Friday after, who in your field can explain the late-breaker?