The M&M Conference and the Question: What Did We Know When?

Wednesday, 3:30 p.m., the large conference room. The Morbiditäts- und Mortalitätskonferenz — the M&M conference — is discussing a postoperative bleed that was recognized late. The discussion is focused and fair — up to the question at which such discussions tip: was the drop in hemoglobin already present when the transfer to the general ward was decided? Two memories contradict each other. The record contains the lab value with the time of draw, but not when the result reached whom. The conference settles on "unclear" and moves to the next item. The moment in which the department could have learned something goes unused.

"What did we know when?" is the core question of every M&M conference. It separates the decision that was defensible with the knowledge of the time from the system problem that failed to bring the knowledge to where the decision was made. And it disciplines hindsight bias: whoever knows the outcome systematically overestimates what was recognizable beforehand. Without a reliable answer to this question, a conference negotiates memories — and memories always negotiate loyalties, too.

What is known about effective M&M conferences

The German Medical Association (Bundesärztekammer) published a methodological guide for morbidity and mortality conferences in 2016, systematizing structure, moderation and case selection [1]. The research on the subject is, honestly, limited: a 2023 systematic review found that standardized structures and organizational support go along with learning effects and system improvements — measured, however, mostly by surveys and self-assessments of the participants, hardly by objective patient outcomes [2]. A second systematic review, of team-based M&M conferences in the perioperative setting, names three recurring elements of effective formats: case selection and work-up before the conference, a standardized presentation format with moderation, and follow-up of the measures decided [3].

The first element deserves a second look. Work-up before the conference presupposes a record that can carry the reconstruction: a timeline of the decision points, the provenance of the central statements, the gaps declared as gaps. Direct study evidence that more complete records improve the outcomes of M&M conferences does not exist — the requirement is documented, the effect is not. But every M&M moderator knows the difference between a conference that starts from a worked-up timeline and one that has to improvise it in the room.

Just culture needs facts

The second precondition is cultural. M&M conferences live on participants speaking openly — and people speak openly when the setting puts system questions before questions of blame, with professional accountability intact. The German Coalition for Patient Safety (Aktionsbündnis Patientensicherheit) formulated these principles — freedom from sanctions, confidentiality, system orientation — early on for incident reporting systems [4]; for the M&M conference they apply in kind.

Here, culture and record touch. A conference that depends on memories almost inevitably personalizes: statement stands against statement, and whoever remembers better or speaks louder shapes the result. A conference that starts from an evidenced timeline can talk about systems instead of memories. The facts relieve the participants of the role of having to be their own witnesses — that is just culture as a working basis.

The tone is set by the leadership. The chief physician's first question decides whether the next forty minutes serve clarification or defense: "Who missed this?" closes mouths; "What should the system have offered differently?" opens them. Departments where the M&M conference counts as the best continuing education of the month usually have both — a leadership that models system questions, and cases worked up well enough that the discussion has facts to work against.

The record that can answer

What must a record be able to do for "What did we know when?" to become answerable? Three things, much of it convention before technology:

• Two timestamps per piece of information: when it came into being (draw, examination) and when it was available (arrival of the report). Where systems do not log acknowledgment, the documented consequence helps — the order, the note.
• Provenance per statement: own finding, external letter, patient's account. A "known" diagnosis without a source is a blind spot in the case work-up.
• Decision points recognizable as such: transfer, change of therapy, discharge — together with what was on hand at that moment.

In the aiomics platform, the second point is in production as architecture: every field of the prepared record knows its source, field by field, with a four-level confidence taxonomy — from "verified" through "single source" and "patient-reported" to "AI-extracted." For the case work-up before an M&M conference, this means the question of where a statement comes from and what supports it can be answered with a click. How this provenance works in detail is described in provenance tracking: every statement with its source (in German). It was built for admission, billing and payer audits; in the service of learning it shows the same value.

Four points for the next conference

1. Case work-up before the meeting, with a timeline of the decision points: what was on hand when, what arrived when — and where the record is silent, the gap is presented as a gap.
2. A standardized format and moderation, for instance along the lines of the German Medical Association's guide [1] — the form protects the openness.
3. System questions first: what prevented existing knowledge from reaching the decision?
4. Measures with owners and deadlines — and a conference that calls up its own resolutions again. Follow-up is the element that, in the reviews, separates effective conferences from ceremonial ones [3].

Why reporting systems see only a fraction of the events that concern a department, and what anonymity assurances have to carry, we discuss in CIRS: anonymity as architecture (in German). And if you would like to read regularly about error culture, documentation and AI in everyday hospital work: our weekly briefing Visite (German; English edition Grand Rounds is in preparation) covers exactly that — collegially, without the brochure tone.

Sources

1. Bundesärztekammer. Methodischer Leitfaden Morbiditäts- und Mortalitätskonferenzen (M&MK). Berlin; 2016. https://www.bundesaerztekammer.de/fileadmin/user_upload/_old-files/downloads/pdf-Ordner/QS/M_Mk.pdf
2. Systematic review of morbidity and mortality meeting standardization: does it lead to improved learning and system change? Queensland Health; 2023. https://psnet.ahrq.gov/issue/systematic-review-morbidity-and-mortality-meeting-standardization-does-it-lead-improved
3. Systematic review on team-based morbidity and mortality conferences in the perioperative setting. 2023. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10513145/
4. Aktionsbündnis Patientensicherheit e. V. Empfehlungen zur Einführung von Critical Incident Reporting Systemen (CIRS). https://www.aps-ev.de/