Protocols change between editions.

Why a row says “stopped naming” and never “stopped doing”

If a procedure is in the earlier edition and not the later one, all we know is that the later book does not name it. We do not know the agency stopped doing it. A protocol book guides care; it is not a list of everything a crew may do. The procedure may have moved to a separate policy, a state rule, a training bulletin or a medical director’s memo, and crews may still perform it every day.

So no row here says an agency stopped doing anything, and if you find one, it is a bug. A row appears only when both editions were read successfully. A book we could not read produces no rows, because an unread document would otherwise look like an agency that dropped everything.

What we deliberately do not publish

We can tell you that 5 authorization changes and 7 certification changes were detected between editions, and we will not draw them as a trend. They did not hold up: the readings pointed both ways in similar numbers, and none was solid enough to publish. Our reading of the books is reliable when comparing one region to another, but not when comparing one agency to its own past. Until that is fixed, a trend drawn from these would show the limits of our method, not a change in EMS.

The half that does hold — who decides, compared across regions at one moment — is on the authorization page.

How far this reaches

11 agencies have a published change here (AL, IL, MA, MD, MI, NE, OH, PA, VA), out of roughly 26,883 licensed EMS agencies nationally. This is not a national picture of how EMS is changing and cannot be read as one: an agency is here only if it publishes dated editions, keeps them readable, and published at least two we could hold side by side. What it is good for is showing that the changes are real, specific, and findable.

Who decides: authorization across regions · The protocol census · Methodology