Reviewing surgical quality when you cannot be in the room

01

Define the review question first

A request to review the session is too broad. It invites a tour of whatever evidence happens to exist and produces comments nobody can compare across months. So the question comes first. Are we checking adherence to a standard? Hunting the cause of a recurring variation? Confirming readiness for more independent work? Testing whether the last round of feedback was applied?

Each question needs different evidence. Without one, more photos and more forms only add noise.

02

Build an evidence chain, not an evidence pile

The evidence should tell the sequence of the work, from planning through execution to documentation and outcome, in an order the reviewer does not have to guess at. Which item came first and which version is final should never be a puzzle.

A chain needs consistent labels, known capture points and enough context to interpret what is visible. It also has to respect privacy. Include only what the judgement requires, keep it in approved storage, and strip identifiers that add nothing to the decision.

03

Ask for context without inviting excuses

I ask for a consistent context record: what was expected, what changed, when it became known, what was decided and which constraint shaped the decision. The clinical reviewer can then distinguish an explained variation from a gap requiring further investigation.

The frame keeps the review fair without letting the standard move after the event.

Context does not erase a gap. It classifies one. Knowledge, judgement, process design, resource, or documentation, and the classification decides who owns the fix.

04

Separate observation, interpretation and action

The fastest way to lose a surgeon's trust from a distance is to jump from a partial image to a verdict about competence. So the feedback is built in three layers, written in that order. The observable fact. The interpretation, with its confidence stated. The action. When someone pushes back, we can find which layer the disagreement actually lives in.

That is usually the difference between an argument and a correction.

Say what was seen, what it may mean, and what happens next. The moment those blur into one verdict, the conversation becomes about the verdict.

05

Calibrate the reviewers

A shared checklist does not produce consistency. People read the same scale differently, weigh evidence differently, and hold different thresholds for escalation. Calibration is slower and it works: the same material reviewed independently, judgements compared, the differences argued out in the open. The goal is not agreement for its own sake. It is shared meaning about what each standard looks like when it appears in evidence.

Keep the agreed examples. After a few cycles they become the real reference, and they are worth more than the scoring label they sit under.

06

Design feedback that can travel back into work

A long review report proves effort and changes nothing. What travels is a short list of prioritised actions, a reason attached to each, and a named point of re-check. I separate four kinds of gap because they need four kinds of owner: a risk needing immediate control, a skill needing practice, a process needing redesign, and a record needing cleaner evidence.

Hand all four to the same person and three of them will wait.

The loop closes when the next session's evidence tests whether the feedback was applied. If it does not, the review was a letter, not a control.

07

Know what remote review cannot prove

Some judgements need the room: direct observation, physical context, an angle nobody captured. A responsible reviewer says so plainly instead of stretching thin evidence to reach a conclusion. Absence of evidence is not evidence of failure, and a polished submission is not proof that the unobserved work was strong.

Record the limit, request what is missing, and when the question cannot be answered remotely, get on the plane.