"No adverse events reported" is a statement about reporting, not about harm

The sentence that should make us pause

Among this week's regulatory notices, one line deserves more attention than its Type II classification suggests. Health Canada lists a recall of Vitros A1c slides because certain generations may produce positively biased results. QuidelOrtho has confirmed 58 complaints and, as of 15 October 2026, no adverse events reported.

Read that quickly and it sounds reassuring. Read it as a pharmacovigilance professional and it should prompt a question: what would an adverse event even look like here? A biased laboratory result does not injure anyone directly. If harm occurs, it occurs downstream, through a clinical decision made on a wrong number. Nobody in that chain is prompted to file a report saying the slide caused it. The records do not say whether any such decisions were made, and that is the point: nor, probably, does anyone else yet.

"No adverse events reported" tells us how many reports arrived. It tells us little about how much harm occurred.

The pattern across this week's notices

The same gap between a hazard and a report runs through other items. The Boston Scientific OptiCross recall in Canada covers revised flushing guidance, reinforced air-management practices and forthcoming Instructions For Use updates. That is a use-related risk, one where the trigger is how a procedure is performed and not a defect a clinician would flag as a product failure. B. Braun's bloodline set recall stems from a missing solvent step that left some dialyzer connectors inadequately bonded in three lots, a manufacturing deviation invisible to the end user until something goes wrong. And Medline's exit from the hospital bed market followed reports of two deaths and 16 serious injuries linked to fire and entrapment hazards.

Meanwhile, regulators are working on the front end. The MHRA's #MedSafetyWeek runs from 2 to 8 November 2026 and comes with NHS login integration and a progressive web app for Yellow Card. Making reporting easier is an admission that reports do not flow on their own, and that the count of reports is partly a function of friction.

The strongest counter-argument

The fair objection is that this is over-reading. Fifty-eight complaints with no associated events is, on its face, a benign picture, and a PV system that treats every complaint as a latent safety case will drown in noise. The records give no denominator, no information on how many results were affected, and no clinical outcomes, so any claim of hidden harm would be speculation. Regulators classified this as Type II, not the most serious tier, and that judgement deserves respect.

I accept all of that. My argument is not that harm occurred. It is that a team cannot currently demonstrate it did not. Those are different positions, and only one is defensible in an inspection or a signal review. An absence of events is evidence only if someone would have reported an event had it happened.

What a mature PV system does with this

QPPVs and PV teams should ask of every product, medicine or device-adjacent, three questions. Who would notice harm? Through what pathway would it reach us? And does our intake design make that report likely or unlikely? For products whose output feeds a clinical decision, the pathway runs through the decision, and complaint and quality data become the nearest thing to a safety signal we will see.

That means complaint handling and PV cannot be separate archives. A cluster of complaints is a prompt to look for outcomes, not a reason to wait for them to arrive unprompted.