Knowledge is the only asset in Medical Affairs that compounds

HCP knowledge assessment means recording what a healthcare professional actually understands about a product and a scientific topic, and how that understanding changes over time. Almost every medical affairs function generates this information continuously and retains very little of it.

A field medical team learns an enormous amount every week. Which arguments land with which clinicians. Where the evidence base is thin. Which objections recur across a specialty. Which HCP has read the subgroup analysis and which has not.

Almost none of it is retained in a form the organisation can use, which is a strange outcome for a function whose entire purpose is scientific exchange. It also sits inside the measurement gap more generally — you cannot compound what you never recorded.

What happens when an MSL changes territory?

The honest answer is that the relationship restarts.

The incoming MSL inherits a contact list, a history of logged interactions, and whatever free text the previous person had time to write. What they do not inherit is the useful part: that this haematologist is unconvinced by the primary endpoint but interested in the real-world evidence, that she asked twice about a specific subgroup and never received a satisfactory answer, that the last conversation ended with an open evidence request nobody followed up.

That information existed. It was generated in the room, held in one person's memory, and lost at handover. The new MSL will spend two or three interactions rediscovering it, during which the HCP is answering questions she has already answered — which is, from her side, a reason to take fewer meetings.

Territory changes are routine. So is turnover. Which means a function whose value depends on accumulated understanding is systematically discarding that understanding on a rolling basis.

Why doesn't the CRM already solve this?

Because a CRM is an activity record, and activity is not the thing that needs recording.

A CRM answers: did a meeting happen, when, with whom, about which product. Those are the questions it was built for and it answers them well. What it does not have is a field for what the HCP understood at the end of the conversation, because understanding is not an event and cannot be logged as one.

The usual workaround is free text. An MSL writes a paragraph after the meeting, and that paragraph is the only place the substance lives. Free text fails for three reasons: it is written under time pressure at the end of a long day, it is unsearchable at scale, and it is not comparable between people or across time. Two MSLs describing the same level of HCP understanding will write completely different paragraphs.

So the information is not missing because nobody recorded it. It is missing because it was recorded in a form that cannot be aggregated, compared or carried forward.

What would a knowledge position actually record?

Not a single score for an HCP. That is the first mistake, and it collapses information rather than preserving it.

A knowledge position is held per product and per communication topic, because those are genuinely independent. A clinician may understand a mechanism of action thoroughly and have engaged very little with the safety data. Averaging those into one number destroys the only detail that would have been useful in planning the next conversation.

It records what was discussed and what was understood, which are different. Coverage says a topic was raised. A knowledge position says something about where the HCP got to.

It records change over time, not a static label. The value is in the direction of travel — an HCP who has moved substantially across three conversations is telling you something different from one who has not moved at all, even if both are currently at the same point.

And it is built from what the HCP said, not from what the MSL assumed. That is why insight yield matters here too: the same conversation that produces an insight also produces the evidence for updating a knowledge position.

Why knowledge is not the same as advocacy

This distinction is worth defending, because collapsing it is how a medical measure quietly becomes a commercial one.

An HCP can understand a product completely — the mechanism, the trial design, the safety profile, the subgroup data — and reasonably conclude that it is not appropriate for the patients she treats. That is a successful scientific outcome. She is informed, and her decision is evidence-based.

If a knowledge model treats prescribing or advocacy as the endpoint of understanding, then that outcome registers as a failure, and the function's measurement framework has quietly been redefined around commercial objectives. Everything Medical Affairs derives its standing from depends on not doing that.

So knowledge, position on adoption, and advocacy are three separate things. They correlate loosely. They should be recorded separately, and a well-informed sceptic should not look like a problem in the data.

What compounding changes over three years

At one interaction, a knowledge position is a data point of limited value. The MSL was in the room and remembers it.

At six interactions across eighteen months, it is a picture: which topics have moved, which have stalled, which questions recur, where this clinician's remaining uncertainty actually sits. Preparation for the seventh conversation starts from a materially better place than preparation for the second.

At the level of a therapeutic area across three years, it becomes something the organisation could not obtain any other way. Which scientific arguments hold up under challenge from specialists. Where the evidence base is consistently thin across a specialty. Which objections are regional and which are universal. That is direct input into evidence generation planning, and it comes from conversations the company was already having.

It is also, unlike almost everything else in the function, cumulative. Activity resets every quarter. Knowledge does not.

One consequence is worth being clear-eyed about. An asset that takes three years to build cannot be quickly rebuilt somewhere else, which makes portability a commercial question rather than a technical one. It is worth establishing at the start of a relationship — not the end — what happens to your data if you leave.

RocketMSL builds a knowledge position per HCP, per product and per topic, from the conversations your team is already having. See how knowledge compounds in the platform →