As a social scientist, one is trained always to view the world of data analysis through skeptical and critical eyes.  In the real world, that makes you skeptical of conclusions based on limited studies that have not weathered competing hypotheses and alternative conclusions.  It is with those eyes that I read Rich Rifkin's column in the Davis Enterprise.

Without a lot of the work of Mr. Rifkin, the issue of employee compensation would be something just a few people on the council would know anything about.  So, I do not want to go too far in criticizing Mr. Rifkin.  But what Mr. Rifkin is guilty of is going too far to show that he has an open mind and that he is willing to change his views based on new data and information.  That's an admirable trait, but one must always be cognizant of drawing too broad a conclusion from a limited inquiry and small sets of data.

---

The question Mr. Rifkin asks is if it is necessary for the fire department to show up with two engines followed by the EMR ambulance every time someone calls for a medical emergency.

Writes Mr. Rifkin, "In February of 2008, I wrote a column critical of the city's emergency medical response policy. 'What confounds me,' I wrote, 'is why it's necessary to always have so much overkill? Why does a person suffering chest pains need one ambulance and its crew plus an entire company of firefighters and all of the fire trucks and gear? Couldn't an ambulance do that job by itself?'"

"Then a month ago, I learned that the city keeps records which show exactly when an ambulance is dispatched, when a fire truck is dispatched and when each arrives. That, I was sure, would prove that we didn't need to send out the fire trucks so often," he continues.  "But I was wrong."

"The records sample I requested were from July 2010. There were 283 medical calls that month. For the first 15 days of July, I noted the call time, the DFD company which was dispatched, the arrival of AMR, the arrival of DFD, who was on scene first and the neighborhood of each call," he writes.  "I stopped making notes at half a month because I was exhausted and the pattern was obvious: 72 percent of the time, the Davis Fire Department arrives first; the other 28 percent it's AMR."

He goes further to say,  "The fire trucks not only arrive first, they arrive three or more minutes before the ambulance on 28 percent of all calls. Three extra minutes is huge if someone has had a heart attack or a stroke. The ambulance never beat the fire trucks on scene by a significant amount of time," he summarizes.

He then issues his conclusion from half a month of data he randomly selected:  "The July numbers disprove the notion that the fire trucks are chasing ambulances. We need firefighters on scene to care for patients at least until the ambulance arrives," he writes and then concludes, "Unless we have more ambulances stationed in Davis - there are supposed to be two, but a single ambulance responds almost every time - we have no choice but to keep sending out the fire trucks.  Lives depend on the DFD's speedy response time."

The problem here is that he has not done much of an analysis.  He simply notes an observation and draws a very broad conclusion from a very limited set of data.

Several people wrote to me yesterday about their surprise with Mr. Rifkin's column and what they saw as an overly-broad conclusion.

In my view there are several steps that Mr. Rifkin never took in this inquiry.

The first question is really about the data he is looking at.  He finds that ambulances 72% of the time arrive after the firefighters.  And 28% of the time they arrive three or more minutes after the fire trucks.  But he never bothers to try to find out if that is not in fact by design.  He simply assumes from limited data that the ambulance arrives late by some systemic factor without exploring what that factor might be.

One alternative hypothesis for that is that the ambulance arrives after the fire truck because they are not, in fact, first responders to the accident or injury scene. They do need to show up first because the paramedics in the fire station are trained to be the ones that respond and treat the individuals on the scene, and that the role of the ambulance personnel is not to treat injuries, it is to transport the victims/patienst to the hospital after their condition is stabilized by the firefighters.

That is a big unexplored question in Mr. Rifkin's column.  He assumes that the fire trucks arrive first because somehow they are inherently faster or better located than ambulances.  And in part, that may be due to the different functions each perform.

A second point that Mr. Rifkin uncharacteristically leaves unexplored is the core assumption of his test, that we necessarily need three units to respond to each emergency.  His test only looks at time of arrival, i.e. speed of the response.  It does not look at the weight of the response and the need for four individuals in two vehicles to show up to each scene, along with an ambulance.

Given the fact that the vast majority of calls for service for the fire department are medical, not fires, why do we need to send all personnel on every call?  The fire department will tell you that the reason they need it is for simultaneous calls and to make sure that they have all equipment on the scene, in case of the need to respond to a fire after an injury.  However, that is a high degree of waste for a minority of responses, and it seems possible that they could produce some alternative to be able to deploy and rendezvous a four-person team to a fire, even if both vehicles do not start out at the same location or if they have to mix and match to deploy from one event to another.

But Mr. Rifkin, in his column, never explores the possibility that, even given the speed of response, we could create a different system for deployment that keeps the element of speed, but eliminates the use of too much personnel and equipment.

Firefighters will tell you there are times, even in an injury situation, that they need more than two individuals.  I do not dispute or dismiss that possibility.  That means there are two possible alternatives.  One is that they deploy a two-person team to a basic accident scene, but can have an alternative under certain circumstances to deploy a four-person team, if needed.  The other alternative is to train police officers, who also end up on the scene and often beat the firefighters to an incident, to assist in medical attention.

That is also a possible fix for the four-person entry team.  We justify having four-person firefighter teams largely because of the OSHA regulations that require two persons in - two persons out when there needs to be an entry into a burning building.  The problem is that the department in Davis may only have three or four such entries a year, and the rest of the time they may be overstaffed.  One solution might be, in a pinch, before the next unit arrives, to have a trained police officer serve as the fourth person.

Many cities are now going to multi-use emergency services where police and fire services are actually interchangeable, and the individuals shift between law enforcement and firefighting duties.

The bottom line with all of this is that Mr. Rifkin seems willing to discard his hypothesis far too soon, based on a very small amount of data and a very broad reading of that data.

A more thorough analysis may show there are alternatives that are cheaper and more efficient that still deliver a good, speedy, and quality emergency medical response to Davis residents.

---David M. Greenwald reporting