There is a mechanism at work in senior decision forums that most advisory professionals have experienced without being able to name precisely. Decision-makers cannot directly evaluate complex reasoning under the conditions of a real executive meeting, so they do not try. Instead they construct an estimate of reasoning quality from whatever signals are available to them, and they allocate weight on the basis of that estimate. The reasoning itself is not what gets weighed. What gets weighed is what the reasoning looks like to someone trying to assess it quickly, under pressure, with incomplete information and finite cognitive bandwidth.

For Medical Affairs professionals in pharmaceutical organisations, this mechanism produces distortions that follow a consistent pattern across different companies, different forums, different decision-makers. The pattern is not coincidental. It arises from four structural features of the Medical Affairs role that each interact with the compression mechanism in a specific way, producing a specific kind of distortion. Understanding what those distortions are, and why they are predictable, is the precondition for doing anything useful about them.

That is what this essay addresses.

Counterfactual value compresses to silence

Most of what Medical Affairs does well leaves no visible trace. The regulatory risk that was identified early and avoided. The data package that held up under scruti ny because its limitations were characterised accurately before submission. The commercial overreach that was caught before physician engagement was compromised. These are real contributions, often consequential ones. They are also constituted by things that did not happen.

Inferential compression works on what is observable. A decision-maker building an estimate of an advisory professional’s reliability is drawing on the record of their contributions. When the most consequential contributions are not in that record, they do not register as positive evidence. They register as absence of evidence.

That distinction matters. Absence of evidence does not compress to a negative signal. It compresses to silence, and silence in a mechanism trying to build an estimate of who can be relied upon is not neutral. It leaves that estimate unrevised in a positive direction. Over time, the Medical Affairs professional whose best work is systematically invisible accumulates a record that does not reflect what they have actually contributed. The compression mechanism has not made an error. It has been working with an incomplete data set, and incomplete data sets produce estimates that drift from reality in predictable ways.

Advisory authority compresses to uncertain attribution

Medical Affairs recommends. It does not decide. Every contribution Medical Affairs makes reaches its effect, if it has any, by passing through someone else’s judgment and then their action. The causal chain runs from the advisory reasoning to the decision-maker’s cognition to their decision to the outcome. Each step is a handoff, and this happens at multiple levels simultaneously: within a single recommendation, across a program of work, and over the course of a relationship between a Medical Affairs leader and the executive team they advise.

When a decision-maker is trying to reconstruct who was responsible for a good outcome, they follow the most accessible causal path. The person who made the call is the proximate cause, visible and unambiguous. The advisor whose reasoning shaped the call is further back and requires more inferential work to reach. Under cognitive constraint, that work often does not get done, and the outcome accrues to the decision-maker.

This happens even when the advisory reasoning was decisive, when the recommendation was genuinely the thing that made the difference. The compression mechanism is not misreading the situation. It is correctly identifying the proximate cause and stopping there because full reconstruction of the causal chain is not available in the time.

Combined with the first feature, this creates a specific double bind. When the contribution prevents a bad outcome, the prevention is invisible. When the contribution shapes a good outcome, the attribution goes elsewhere. Medical Affairs is systematically under-represented in its own record from both ends, not through any individual’s failure to acknowledge contribution, but through the structure of how attribution under cognitive constraint actually works.

Calibrated uncertainty compresses to apparent fragility

Medical Affairs professionals are expected to represent uncertainty accurately. That is not a communication preference. It is part of our training, a regulatory requirement and a scientific obligation. Understating uncertainty, suppressing confidence intervals, presenting a cleaner picture than the evidence warrants, these are not available options. The obligation is to say what the evidence actually shows, including where it is limited, where the conclusions are conditional, and what would change the recommendation.

The compression mechanism uses confidence as one of its proxies for reasoning quality. The template of what authoritative expert judgment looks like is clear in its direction and settled in its conclusions. Contributions that are appropriately qualified do not fit that template. Under compression, qualification reads as fragility, as the signal of someone who is not sure rather than someone who is being precise about what they know and what they do not.

The Medical Affairs professional who says the evidence supports a course of action in this population under these conditions, but that there are reasons to be cautious about extending the conclusion to a different context, is doing the job correctly. In a senior executive forum under time pressure, that sentence arrives as a qualified recommendation, and the qualification is the proxy the compression mechanism has available. It reads as hesitation. The Medical Affairs professional who simplifies, who presents a clean recommendation without the epistemic scaffolding, produces a contribution that compresses more cleanly. They appear more certain. They get more weight. The accuracy of the representation has worked against the person who produced it.

There is no resolution to this within the terms of the standard advice. Presenting more confidently, if it means suppressing genuine uncertainty, is not available as a strategy. Medical Affairs cannot solve this by abandoning the thing that makes its advice worth having in the first place. The distortion is structural, which is precisely why it persists regardless of how capable the individuals involved are.

Multi-domain integration compresses to surface signals

The most sophisticated thing Medical Affairs does is hold constraints across domains simultaneously. Scientific evidence quality, regulatory pathway implications, commercial access dynamics, operational feasibility, ethical obligations. These do not operate on the same time horizons, they do not resolve in the same direction, and they require different kinds of expertise to fully evaluate. Integrating them into a defensible directional recommendation is the work.

In a senior executive forum, no single observer holds the full constraint architecture. The regulatory lead can fully evaluate the regulatory dimension. The commercial lead can fully evaluate the commercial one. Neither can fully evaluate how the scientific evidence quality bears on the regulatory pathway, or how that bears on the commercial access question, or how the ethical considerations interact with all three. The integration is exactly what is invisible to the individual domain experts who make up the room.

When the reasoning architecture cannot be reconstructed, the compression mechanism defaults to evaluating what it can actually assess. The confidence of the presentation. The structural coherence of the framing. The apparent certainty of the conclusion. These are the proxies available when the integration cannot be recovered, and they carry the entire evaluative burden that full reconstruction would otherwise bear. The more genuinely integrative the reasoning, the harder it is to compress accurately, and the more heavily the mechanism falls back on surface signals that may bear little relationship to the quality of what produced them.

Why the distortions interact

These four distortions do not sit beside each other. They compound.

Counterfactual value and advisory authority create an attribution problem that is more severe than either alone. Medical Affairs, whose most important contributions are invisible, is also the function whose causal role is mediated. Two independent routes to positive attribution are blocked simultaneously, and the compression mechanism is trying to build an estimate from a record that is incomplete from both ends at once.

Calibrated uncertainty and multi-domain integration create what the model calls a legibility problem, a gap between how sophisticated the reasoning actually is and how much of it can be reconstructed by the people evaluating it. The Medical Affairs professional who most needs to represent complexity accurately is also working in a role whose complexity is hardest to reconstruct from the outside. When genuine uncertainty in a cross-domain recommendation is represented accurately, it arrives as qualification in a forum where no single observer can follow the full integration that produced it. The more accurate the representation, the more it reads as doubt, because the architecture that would allow a different reading is not available to those in the room.

Across counterfactual value, advisory authority, calibrated uncertainty, and multi-domain integration, the same underlying problem recurs. The compression mechanism is working on whatever is available, and for Medical Affairs that set is systematically smaller than for most other advisory functions. Together these features leave the mechanism with the surface signals of the delivery as its primary input, regardless of what the reasoning underneath actually is.

That is not a random outcome. It is the predictable consequence of a specific mechanism operating on a specific set of structural features, which is why it follows the same pattern across different organisations, different forums, different decision-makers. The conditions are the same. The mechanism is the same. The distortions are the same.

The title of this essay could be thought of as a communication failure, but what it actually describes is a structural one. The signal Medical Affairs sends, careful, qualified, integrated, multi-constrained, is real and it is accurate. The signal the compression mechanism reads is something else: confidence, certainty, directional simplicity. The gap between the two is not closed by improving the communication. It is closed by changing what is available to the mechanism in the first place. That is a different problem, and it has a different answer.