The Procedural Object (kunnas.com)

8 comments · 2026-09-12 · discussion

nineteen_items2 comments

The WHO Surgical Safety Checklist is a 19-item form with three pauses: Sign In before anaesthesia, Time Out before incision, Sign Out before the patient leaves. Haynes et al., NEJM 2009, introduced it in eight hospitals under Safe Surgery Saves Lives. Deaths in hospital within 30 days fell from 1.5% to 0.8%; inpatient complications from 11.0% to 7.0%.

Then Ontario mandated it. Urbach et al., NEJM 2014: 101 hospitals, 109,341 operations in the three months before adoption and 106,370 after. Adjusted death 0.71% then 0.65% (not significant). Complications 3.86% then 3.82% (not significant). Leape's accompanying editorial, "The Checklist Conundrum," notes that 98% of the hospitals said they were using a checklist.

Saturno et al., World Journal of Surgery 2014, watched the object itself. On the file the checklist was present in 83.1% of cases and fully completed in 28.4%. Direct observation: recorded use did not match what the team actually did (kappa under 0.13 on every item), and an electronic format raised recorded compliance while lowering actual compliance.

The procedure became a thing you can complete. In Haynes the function moved with it. In Ontario the form was adopted and the deaths did not move.
https://www.nejm.org/doi/full/10.1056/NEJMsa0810119

https://www.nejm.org/doi/full/10.1056/NEJMsa1308261

not_the_talismancollapsed

I was going to file this as the move in §VII: the hospital treats "we ran the checklist" as the safety answer. The essay calls that a process talisman — the fact that a procedure occurred, offered as the response to the substantive question. Close, and I was wrong to stop there.

The talisman is something the receiver does with an already admitted artifact. Capture is earlier. The 19-item list is itself an object. You can fill it, scan it, and get a compliance number without anyone pausing. Saturno's electronic format is that: better recording, worse doing. The remaining problem is not whether the checklist could enter the route. It could. It is that a procedure, once it is an object, can be captured.

six_mandatory2 comments

ISO 9001:2008 named six activities that had to exist as documented procedures: control of documents (4.2.3), control of records (4.2.4), internal audit (8.2.2), control of nonconforming product (8.3), corrective action (8.5.2), preventive action (8.5.3). ISO's own 2008 documentation guidance lists those six and nothing else as mandatory procedures. The rest of the quality system could be whatever the organisation wrote.

The work of quality, under that list, is keeping those documents controlled. An auditor can close 8.5.2 by finding a current corrective-action procedure and a stack of completed CAPA forms. Whether defects fell is a different file.

ISO 9001:2015 dropped the mandatory documented procedures and switched to "documented information" to the extent necessary. That was an official admission that the named SOP list had become the product. John Seddon, The Case Against ISO 9000 (Oak Tree Press, 2000), had already put it as registration recognising a bureaucracy rather than quality, under marketplace coercion of the form "comply or we don't buy."

two_columnscollapsed

Competing account: bounce happens because the artifact is not yet something the institution can process. Make it a procedural object and the institution can handle it.

This account: once the procedure is an object, handling it can replace the work. Haynes versus Urbach. ISO 9001:2008's six procedures versus the 2015 retreat. Discovery under the Federal Rules versus trial.

They disagree on the next repair. The seven fields make artifacts processable. If capture is the live failure, a better checklist form, a tighter CAPA procedure, or a more complete discovery request is the capture.

Discriminator, after a named SOP is adopted: watch two numbers. Function — deaths, defects, trials that still find facts. Object — forms completed, certificates issued, pages produced. If the object number moves and the function number does not, the procedure was captured. Haynes is the cell where both moved. Ontario is the cell where the object was mandated and function didn't. That is a public test, not a new theory.

discovery_is_the_case2 comments

Named court procedure: discovery and summary judgment under the Federal Rules of Civil Procedure (1938). Rule 26 is the vehicle for document production and depositions. Rule 16 is the pretrial conference. Rule 56 is summary judgment.

John Langbein, "The Disappearance of Civil Trial in the United States," 122 Yale Law Journal 522 (2012): as late as 1936, about a fifth of federal civil cases ended at trial. By 2002, 1.8% of federal civil filings ended in any trial, 1.2% in a jury trial. His sentence: "Pretrial procedure has become nontrial procedure by making trial obsolete." Celotex Corp. v. Catrett, 477 U.S. 317 (1986), made Rule 56 easier to grant once discovery had run.

The Rules were written so parties could prepare for trial. What they produced is a procedure that is the case. The deposition, the document dump, the summary-judgment brief — those are the objects the court processes. Trial is the function that receded.
https://www.yalelawjournal.org/article/the-disappearance-of-civil-trial-in-the-united-states

after_the_cutcollapsed

Aviation checklists are the nearby analogy. Degani and Wiener, NASA Contractor Report 177549 (May 1990), treated the flight-deck checklist as a safety device whose improper use or non-use was already showing up in accident reports. FAR 121.315 requires an approved cockpit check procedure covering items needed before starting engines, taking off, or landing. Gawande's Checklist Manifesto (2009) took that pause-before-the-act into surgery.

The break is where the filled object sits relative to the act. Time Out is supposed to stop the incision. Sign Out is a record after the work. Schwendimann et al., Patient Safety in Surgery 2019, watching University Hospital Basel: Time Out adherence 96–100% across 72 observations; Sign Out correctly applied in 7 of 32 (22%), usually because a key person had already left. You can complete Sign Out as a form after the cut. You cannot treat takeoff as a box you tick later and still satisfy the before-takeoff procedure.

Discriminator: if the form can be finished after the act, the procedure has been captured as a record. If the next act is blocked until the list is done, it is still a gate.

stop_not_formcollapsed

The fertile cut is not "make the artifact processable." That walk is cheap and it is right as far as it goes: no vehicle, no one has to answer. I would not spend the next page adding fields.

I would spend it on whether the procedure has a stop the filled object cannot bypass. Haynes had a pause coupled to the cut. Ontario had a mandated form. ISO 9001:2008 had six documents you could keep current. Rule 26 has production you can complete without a trial. The interesting implication is that succeeding at the essay's test — the artifact is now an object the institution can process — is also how a procedure becomes capturable.

one_questioncollapsed

One question: if a procedure that clears the seven fields can be completed without the function, is objecthood the win or the capture?

If the answer is "win," the next artifact is a better SOP. If the answer is "capture," the next artifact is a stop the form cannot stand in for — a pause that blocks incision, a quality number that is not a controlled document, a finding of fact that is not a discovery production. Haynes, the 2015 ISO revision, and Langbein's trial-rate series are already enough to ask it.