Two decades and roughly 200,000 calls is enough time to watch a good QA habit stop working, not because anyone stopped caring about it, but because it was never built to survive growth in the first place. That's the lesson behind this post: most of what looks like "good QA" at a small agency is really a workaround for limited review capacity, and workarounds don't scale.
Sampling isn't a QA strategy. It's a capacity limit wearing a strategy's clothes
A QA officer who can personally read every serious call at a 5,000-call agency cannot personally read every serious call at a 40,000-call agency. Something has to give, and historically that something has been coverage: review 5 or 10 percent, hope the sample is representative, and treat "nothing got flagged" as reassurance rather than what it usually is, an artifact of not having looked. The absolute number of unreviewed charts grows every year right alongside the agency, even while the reviewed percentage stays flat.
Three habits that felt like good QA and were actually scale workarounds
Reviewing the same handful of chart types every meeting. When review time is scarce, it gets spent on the calls that feel highest-risk, cardiac arrests, pediatric traumas, controlled substance administrations. That's not wrong, but it quietly trains a QA program to only ever look where it already expects to find something, which is exactly how documentation drift in routine, lower-acuity calls goes unnoticed for years.
Treating "not flagged" as "reviewed and fine." Those are different claims. One means a human or a system looked and found nothing. The other means nobody looked. Sample-based QA collapses the two together, and it's an easy distinction to lose sight of once a program has run the same way for a while.
Measuring QA activity instead of QA coverage. Meetings held, charts discussed, corrective actions issued, all real work, but none of it answers the question that actually matters: what fraction of what your crews document each month never gets seen by anyone in a QA role.
What actually scales
NHTSA's guidance on EMS quality improvement has made a version of this point since before most current EMS software existed: QI has to be a routine, structural part of how an EMS system operates, not an occasional audit layered on top of it (A Leadership Guide to Quality Improvement for Emergency Medical Services Systems, NHTSA). The more recent EMS Agenda 2050 vision pushes the same idea further, describing an EMS system that is evidence-based and data-driven as a baseline expectation, not an aspiration (EMS Agenda 2050, NHTSA Office of EMS). Neither document was written with software in mind, but both describe the same requirement: review has to keep pace with volume, structurally, not through more heroic effort from whoever is doing QA that week.
In practice that means three things. Full-volume review, because a QA program's coverage shouldn't depend on how many hours a QA officer has this month, as we wrote about in The End of the 10% Blind Spot. A deterministic, traceable rule set doing the primary pass, so the review is reproducible and defensible regardless of chart volume, the same reasoning behind keeping AI a second layer rather than the decision-maker, covered in Black Box vs. Glass Box: Explainable AI for EMS QA. And review that isn't tied to a single ePCR's schema, so growth, a mutual-aid arrangement, or an eventual charting-platform change doesn't force a QA program to start over, which is the same case we made in Vendor-Neutral EMS QA Software.
Why we built Integritas EMSQA around this
We built Integritas EMSQA the way we did because we'd already watched sampling fail at scale from the inside. Full-volume, NEMSIS-native review with a deterministic rules engine as the primary pass, AI layered on top rather than replacing it, on whatever ePCR an agency already runs. Not because full-volume review is a nice-to-have feature, but because it's the only version of QA that doesn't quietly get worse every year an agency grows.
The EMS quality improvement program that scales is the one built for growth from day one
The habits that feel like good QA at a small agency, sampling, informal review, activity over coverage, aren't wrong choices. They're the choices available when review capacity is capped by how much one person can read. The way out isn't asking that person to read faster. It's building a QA program that doesn't depend on how much any one person can personally review.
Request a demo and we'll walk through what full-volume review actually looks like on the ePCR your agency already runs.
Integritas EMSQA is NEMSIS v3.5.1 "Receive and Process" certified and works with any NEMSIS-compliant ePCR.